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 * On build targets where the multilib postfix is set in the build
34 configuration the libdir directory was changing based on whether
35 the lib directory with the multilib postfix exists on the system
36 or not. This unpredictable behavior was removed and eventual
37 multilib postfix is now always added to the default libdir. Use
38 `--libdir=lib` to override the libdir if adding the postfix is
43 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
44 function codes, this function can only cause problems for calling
49 * Add a configurable flag to output date formats as ISO 8601. Does not
50 change the default date format.
54 * Version of MSVC earlier than 1300 could get link warnings, which could
55 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
56 Support for this flag has been removed.
60 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
61 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
62 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
63 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
64 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
68 * The signatures of the functions to get and set options on SSL and
69 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
70 Some source code changes may be required.
74 * The public definitions of conf_method_st and conf_st have been
75 deprecated. They will be made opaque in a future release.
77 *Rich Salz and Tomáš Mráz*
79 * Client-initiated renegotiation is disabled by default. To allow it, use
80 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
81 flag, or the "ClientRenegotiation" config parameter as appropriate.
85 * Add "abspath" and "includedir" pragma's to config files, to prevent,
86 or modify relative pathname inclusion.
90 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
91 validated. Please consult the README-FIPS and
92 README-PROVIDERS files, as well as the migration guide.
94 *OpenSSL team members and many third party contributors*
96 * For the key types DH and DHX the allowed settable parameters are now different.
100 * The openssl commands that read keys, certificates, and CRLs now
101 automatically detect the PEM or DER format of the input files.
103 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
105 * Added enhanced PKCS#12 APIs which accept a library context.
109 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
113 * Added support for Kernel TLS (KTLS).
115 *Boris Pismenny, John Baldwin and Andrew Gallatin*
117 * Support for RFC 5746 secure renegotiation is now required by default for
118 SSL or TLS connections to succeed.
122 * The signature of the `copy` functional parameter of the
123 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
124 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
125 the signature of the `pub_decode` functional parameter of the
126 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
127 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
131 * The error return values from some control calls (ctrl) have changed.
135 * A public key check is now performed during EVP_PKEY_derive_set_peer().
139 * Many functions in the EVP_ namespace that are getters of values from
140 implementations or contexts were renamed to include get or get0 in their
141 names. Old names are provided as macro aliases for compatibility and
146 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
147 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
148 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
153 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
156 * The output from the command line applications may have minor
161 * The output from numerous "printing" may have minor changes.
165 * Windows thread synchronization uses read/write primitives (SRWLock) when
166 supported by the OS, otherwise CriticalSection continues to be used.
170 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
171 work on read only BIO source/sinks that do not support these functions.
172 This allows piping or redirection of a file BIO using stdin to be buffered
173 into memory. This is used internally in OSSL_DECODER_from_bio().
177 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
178 this function would return one of the values OSSL_STORE_INFO_NAME,
179 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
180 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
181 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
182 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
183 using this function should be amended to handle the changed return value.
187 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
188 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
189 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
190 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
191 contains more than one certificate identifier: This means that all
192 certificates referenced there MUST be part of the validation chain.
196 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
197 RC5, DESX and DES have been moved to the legacy provider.
201 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
202 RIPEMD-160 have been moved to the legacy provider.
206 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
211 * The deprecated functions EVP_PKEY_get0_RSA(),
212 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
213 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
214 well as the similarly named "get1" functions behave differently in
219 * A number of functions handling low-level keys or engines were deprecated
220 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
221 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
222 EVP_PKEY_get0_siphash().
226 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
227 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
228 will need to load the legacy crypto provider. This includes these PBE
229 algorithms which use this KDF:
230 - NID_pbeWithMD2AndDES_CBC
231 - NID_pbeWithMD5AndDES_CBC
232 - NID_pbeWithSHA1AndRC2_CBC
233 - NID_pbeWithMD2AndRC2_CBC
234 - NID_pbeWithMD5AndRC2_CBC
235 - NID_pbeWithSHA1AndDES_CBC
239 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
240 BIO_debug_callback() functions.
244 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
245 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
249 * The RAND_METHOD APIs have been deprecated.
253 * The SRP APIs have been deprecated.
257 * Add a compile time option to prevent the caching of provider fetched
258 algorithms. This is enabled by including the no-cached-fetch option
259 at configuration time.
263 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
264 count of PKCS12_DEFAULT_ITER.
266 *Tomáš Mráz and Sahana Prasad*
268 * The openssl speed command does not use low-level API calls anymore.
272 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
275 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
277 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
281 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
282 providers may supply their own group implementations (using either the "key
283 exchange" or the "key encapsulation" methods) which will automatically be
284 detected and used by libssl.
286 *Matt Caswell, Nicola Tuveri*
288 * The undocumented function X509_certificate_type() has been deprecated;
292 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
296 * Removed RSA padding mode for SSLv23 (which was only used for
297 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
298 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
303 * Deprecated the obsolete X9.31 RSA key generation related functions.
305 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
306 is not allowed to return a value > 1, this is no more taken as failure.
308 *Viktor Dukhovni and David von Oheimb*
310 * Deprecated the obsolete X9.31 RSA key generation related functions
311 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
312 BN_X931_generate_prime_ex().
316 * The default key generation method for the regular 2-prime RSA keys was
317 changed to the FIPS 186-4 B.3.6 method.
321 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
325 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
329 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
330 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
332 *Rich Salz, Richard Levitte, and David von Oheimb*
334 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
338 * Deprecated `OCSP_parse_url()`.
342 * Validation of SM2 keys has been separated from the validation of regular EC
347 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
348 switches: a validation failure triggers an early exit, returning a failure
349 exit status to the parent process.
353 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
354 to ignore unknown ciphers.
358 * The `-cipher-commands` and `-digest-commands` options
359 of the command line utility `list` have been deprecated.
360 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
364 * Added convenience functions for generating asymmetric key pairs:
365 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
366 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
370 * All of the low level EC_KEY functions have been deprecated.
372 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
374 * Deprecated all the libcrypto and libssl error string loading
379 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
380 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
385 * The `-crypt` option to the `passwd` command line tool has been removed.
389 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
394 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
398 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
399 EVP_PKEY_get1_tls_encodedpoint().
403 * The security callback, which can be customised by application code, supports
404 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
405 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
409 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
410 interface. Their functionality remains unchanged.
414 * Added new option for 'openssl list', '-providers', which will display the
415 list of loaded providers, their names, version and status. It optionally
416 displays their gettable parameters.
420 * Removed EVP_PKEY_set_alias_type().
424 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
425 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
429 * Changed all "STACK" functions to be macros instead of inline functions. Macro
430 parameters are still checked for type safety at compile time via helper
435 * Remove the RAND_DRBG API
437 *Paul Dale and Matthias St. Pierre*
439 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
440 as well as actual hostnames.
444 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
445 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
446 conversely, silently ignore DTLS protocol version bounds when configuring
447 TLS-based contexts. The commands can be repeated to set bounds of both
448 types. The same applies with the corresponding "min_protocol" and
449 "max_protocol" command-line switches, in case some application uses both TLS
452 SSL_CTX instances that are created for a fixed protocol version (e.g.
453 `TLSv1_server_method()`) also silently ignore version bounds. Previously
454 attempts to apply bounds to these protocol versions would result in an
455 error. Now only the "version-flexible" SSL_CTX instances are subject to
456 limits in configuration files in command-line options.
460 * Deprecated the `ENGINE` API. Engines should be replaced with providers
465 * Reworked the recorded ERR codes to make better space for system errors.
466 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
467 given code is a system error (true) or an OpenSSL error (false).
471 * Reworked the test perl framework to better allow parallel testing.
473 *Nicola Tuveri and David von Oheimb*
475 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
476 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
480 * 'Configure' has been changed to figure out the configuration target if
481 none is given on the command line. Consequently, the 'config' script is
482 now only a mere wrapper. All documentation is changed to only mention
485 *Rich Salz and Richard Levitte*
487 * Added a library context `OSSL_LIB_CTX` that applications as well as
488 other libraries can use to form a separate context within which
489 libcrypto operations are performed.
493 * Handshake now fails if Extended Master Secret extension is dropped
498 * Dropped interactive mode from the `openssl` program.
502 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
504 *David von Oheimb and Shane Lontis*
506 * Deprecated `EC_METHOD_get_field_type()`.
510 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
511 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
512 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
516 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
520 * Add CAdES-BES signature verification support, mostly derived
521 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
523 *Filipe Raimundo da Silva*
525 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
529 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
530 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
534 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
538 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
539 EC_KEY_precompute_mult().
543 * Deprecated EC_POINTs_mul().
547 * Removed FIPS_mode() and FIPS_mode_set().
551 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
555 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
556 EC_POINT_get_Jprojective_coordinates_GFp().
560 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
561 arrays to be more easily constructed via a series of utility functions.
562 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
563 the various push functions and finally convert to a passable OSSL_PARAM
564 array using OSSL_PARAM_BLD_to_param().
568 * The security strength of SHA1 and MD5 based signatures in TLS has been
573 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
574 contain a provider side internal key.
578 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
582 * Project text documents not yet having a proper file name extension
583 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
584 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
588 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
589 have been converted to Markdown with the goal to produce documents
590 which not only look pretty when viewed online in the browser, but
591 remain well readable inside a plain text editor.
593 To achieve this goal, a 'minimalistic' Markdown style has been applied
594 which avoids formatting elements that interfere too much with the
595 reading flow in the text file. For example, it
597 * avoids [ATX headings][] and uses [setext headings][] instead
598 (which works for `<h1>` and `<h2>` headings only).
599 * avoids [inline links][] and uses [reference links][] instead.
600 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
602 [ATX headings]: https://github.github.com/gfm/#atx-headings
603 [setext headings]: https://github.github.com/gfm/#setext-headings
604 [inline links]: https://github.github.com/gfm/#inline-link
605 [reference links]: https://github.github.com/gfm/#reference-link
606 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
607 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
609 *Matthias St. Pierre*
611 * The test suite is changed to preserve results of each test recipe.
612 A new directory test-runs/ with subdirectories named like the
613 test recipes are created in the build tree for this purpose.
617 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
618 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
619 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
621 *David von Oheimb, Martin Peylo*
623 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
624 It supports arbitrary request and response content types, GET redirection,
625 TLS, connections via HTTP(S) proxies, connections and exchange via
626 user-defined BIOs (allowing implicit connections), persistent connections,
627 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
628 The legacy OCSP-focused (and only partly documented) API
629 is retained for backward compatibility, while most of it is deprecated.
633 * Added `util/check-format.pl`, a tool for checking adherence to the
634 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
635 The checks performed are incomplete and yield some false positives.
636 Still the tool should be useful for detecting most typical glitches.
640 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
641 If domain name resolution yields multiple IP addresses all of them are tried
642 after `connect()` failures.
646 * All of the low level RSA functions have been deprecated.
650 * X509 certificates signed using SHA1 are no longer allowed at security
655 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
656 modified to use PKEY APIs. These commands are now in maintenance mode
657 and no new features will be added to them.
661 * The command line utility rsautl has been deprecated.
665 * The command line utilities genrsa and rsa have been modified to use PKEY
666 APIs. They now write PKCS#8 keys by default. These commands are now in
667 maintenance mode and no new features will be added to them.
671 * All of the low level DH functions have been deprecated.
673 *Paul Dale and Matt Caswell*
675 * All of the low level DSA functions have been deprecated.
679 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
680 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
684 * Deprecated low level ECDH and ECDSA functions.
688 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
692 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
693 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
694 a new formulation to include all the things it can be used for,
695 as well as words of caution.
699 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
703 * All of the low level HMAC functions have been deprecated.
705 *Paul Dale and David von Oheimb*
707 * Over two thousand fixes were made to the documentation, including:
708 - Common options (such as -rand/-writerand, TLS version control, etc)
709 were refactored and point to newly-enhanced descriptions in openssl.pod.
710 - Added style conformance for all options (with help from Richard Levitte),
711 documented all reported missing options, added a CI build to check
712 that all options are documented and that no unimplemented options
714 - Documented some internals, such as all use of environment variables.
715 - Addressed all internal broken L<> references.
719 * All of the low level CMAC functions have been deprecated.
723 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
724 functions have been deprecated.
726 *Paul Dale and David von Oheimb*
728 * Corrected the documentation of the return values from the `EVP_DigestSign*`
729 set of functions. The documentation mentioned negative values for some
730 errors, but this was never the case, so the mention of negative values
733 Code that followed the documentation and thereby check with something
734 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
738 * All of the low level cipher functions have been deprecated.
740 *Matt Caswell and Paul Dale*
742 * Removed include/openssl/opensslconf.h.in and replaced it with
743 include/openssl/configuration.h.in, which differs in not including
744 <openssl/macros.h>. A short header include/openssl/opensslconf.h
745 was added to include both.
747 This allows internal hacks where one might need to modify the set
748 of configured macros, for example this if deprecated symbols are
749 still supposed to be available internally:
751 #include <openssl/configuration.h>
753 #undef OPENSSL_NO_DEPRECATED
754 #define OPENSSL_SUPPRESS_DEPRECATED
756 #include <openssl/macros.h>
758 This should not be used by applications that use the exported
759 symbols, as that will lead to linking errors.
763 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
764 used in exponentiation with 512-bit moduli. No EC algorithms are
765 affected. Analysis suggests that attacks against 2-prime RSA1024,
766 3-prime RSA1536, and DSA1024 as a result of this defect would be very
767 difficult to perform and are not believed likely. Attacks against DH512
768 are considered just feasible. However, for an attack the target would
769 have to re-use the DH512 private key, which is not recommended anyway.
770 Also applications directly using the low-level API BN_mod_exp may be
771 affected if they use BN_FLG_CONSTTIME.
776 * Most memory-debug features have been deprecated, and the functionality
777 replaced with no-ops.
781 * Added documentation for the STACK API.
785 * Introduced a new method type and API, OSSL_ENCODER, to
786 represent generic encoders.
790 * Added a .pragma directive to the syntax of configuration files, to
791 allow varying behavior in a supported and predictable manner.
792 Currently added pragma:
796 This allows dollar signs to be a keyword character unless it's
797 followed by a opening brace or parenthesis. This is useful for
798 platforms where dollar signs are commonly used in names, such as
799 volume names and system directory names on VMS.
803 * Added functionality to create an EVP_PKEY from user data.
807 * Change the interpretation of the '--api' configuration option to
808 mean that this is a desired API compatibility level with no
809 further meaning. The previous interpretation, that this would
810 also mean to remove all deprecated symbols up to and including
811 the given version, no requires that 'no-deprecated' is also used
812 in the configuration.
814 When building applications, the desired API compatibility level
815 can be set with the OPENSSL_API_COMPAT macro like before. For
816 API compatibility version below 3.0, the old style numerical
817 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
818 For version 3.0 and on, the value is expected to be the decimal
819 value calculated from the major and minor version like this:
821 MAJOR * 10000 + MINOR * 100
825 -DOPENSSL_API_COMPAT=30000 For 3.0
826 -DOPENSSL_API_COMPAT=30200 For 3.2
828 To hide declarations that are deprecated up to and including the
829 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
830 given when building the application as well.
834 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
835 access to certificate and CRL stores via URIs and OSSL_STORE
838 This adds the following functions:
840 - X509_LOOKUP_store()
841 - X509_STORE_load_file()
842 - X509_STORE_load_path()
843 - X509_STORE_load_store()
844 - SSL_add_store_cert_subjects_to_stack()
845 - SSL_CTX_set_default_verify_store()
846 - SSL_CTX_load_verify_file()
847 - SSL_CTX_load_verify_dir()
848 - SSL_CTX_load_verify_store()
852 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
853 The presence of this system service is determined at run-time.
857 * Added functionality to create an EVP_PKEY context based on data
858 for methods from providers. This takes an algorithm name and a
859 property query string and simply stores them, with the intent
860 that any operation that uses this context will use those strings
861 to fetch the needed methods implicitly, thereby making the port
862 of application written for pre-3.0 OpenSSL easier.
866 * The undocumented function NCONF_WIN32() has been deprecated; for
867 conversion details see the HISTORY section of doc/man5/config.pod
871 * Introduced the new functions EVP_DigestSignInit_ex() and
872 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
873 EVP_DigestVerifyUpdate() have been converted to functions. See the man
874 pages for further details.
878 * Over two thousand fixes were made to the documentation, including:
879 adding missing command flags, better style conformance, documentation
882 *Rich Salz, Richard Levitte*
884 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
885 X25519, X448, Ed25519 and Ed448.
889 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
894 * Deprecated the public definition of `ERR_STATE` as well as the function
895 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
900 * Added ERR functionality to give callers access to the stored function
901 names that have replaced the older function code based functions.
903 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
904 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
905 ERR_peek_error_all() and ERR_peek_last_error_all().
907 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
908 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
909 ERR_func_error_string().
913 * Extended testing to be verbose for failing tests only. The make variables
914 VERBOSE_FAILURE or VF can be used to enable this:
916 $ make VF=1 test # Unix
917 $ mms /macro=(VF=1) test ! OpenVMS
918 $ nmake VF=1 test # Windows
922 * Added the `-copy_extensions` option to the `x509` command for use with
923 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
924 all extensions in the request are copied to the certificate or vice versa.
926 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
928 * Added the `-copy_extensions` option to the `req` command for use with
929 `-x509`. When given with the `copy` or `copyall` argument,
930 all extensions in the certification request are copied to the certificate.
934 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
935 they generate are by default RFC 5280 compliant in the following sense:
936 There is a subjectKeyIdentifier extension with a hash value of the public key
937 and for not self-signed certs there is an authorityKeyIdentifier extension
938 with a keyIdentifier field or issuer information identifying the signing key.
939 This is done unless some configuration overrides the new default behavior,
940 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
944 * Added several checks to `X509_verify_cert()` according to requirements in
945 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
946 (which may be done by using the CLI option `-x509_strict`):
947 * The basicConstraints of CA certificates must be marked critical.
948 * CA certificates must explicitly include the keyUsage extension.
949 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
950 * The issuer name of any certificate must not be empty.
951 * The subject name of CA certs, certs with keyUsage crlSign,
952 and certs without subjectAlternativeName must not be empty.
953 * If a subjectAlternativeName extension is given it must not be empty.
954 * The signatureAlgorithm field and the cert signature must be consistent.
955 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
956 must not be marked critical.
957 * The authorityKeyIdentifier must be given for X.509v3 certs
958 unless they are self-signed.
959 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
963 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
964 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
968 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
969 used even when parsing explicit parameters, when loading a encoded key
970 or calling `EC_GROUP_new_from_ecpkparameters()`/
971 `EC_GROUP_new_from_ecparameters()`.
972 This prevents bypass of security hardening and performance gains,
973 especially for curves with specialized EC_METHODs.
974 By default, if a key encoded with explicit parameters is loaded and later
975 encoded, the output is still encoded with explicit parameters, even if
976 internally a "named" EC_GROUP is used for computation.
980 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
981 this change, EC_GROUP_set_generator would accept order and/or cofactor as
982 NULL. After this change, only the cofactor parameter can be NULL. It also
983 does some minimal sanity checks on the passed order.
988 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
989 An attack is simple, if the first CMS_recipientInfo is valid but the
990 second CMS_recipientInfo is chosen ciphertext. If the second
991 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
992 encryption key will be replaced by garbage, and the message cannot be
993 decoded, but if the RSA decryption fails, the correct encryption key is
994 used and the recipient will not notice the attack.
995 As a work around for this potential attack the length of the decrypted
996 key must be equal to the cipher default key length, in case the
997 certifiate is not given and all recipientInfo are tried out.
998 The old behaviour can be re-enabled in the CMS code by setting the
999 CMS_DEBUG_DECRYPT flag.
1003 * Early start up entropy quality from the DEVRANDOM seed source has been
1004 improved for older Linux systems. The RAND subsystem will wait for
1005 /dev/random to be producing output before seeding from /dev/urandom.
1006 The seeded state is stored for future library initialisations using
1007 a system global shared memory segment. The shared memory identifier
1008 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1009 the desired value. The default identifier is 114.
1013 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1014 when primes for RSA keys are computed.
1015 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1016 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1017 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1018 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1019 This avoids possible fingerprinting of newly generated RSA modules.
1023 * Correct the extended master secret constant on EBCDIC systems. Without this
1024 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1025 negotiate EMS will fail. Unfortunately this also means that TLS connections
1026 between EBCDIC systems with this fix, and EBCDIC systems without this
1027 fix will fail if they negotiate EMS.
1031 * Changed the library initialisation so that the config file is now loaded
1032 by default. This was already the case for libssl. It now occurs for both
1033 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1034 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1038 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1039 where the former acts as a replacement for `ERR_put_error()`, and the
1040 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1041 `ERR_raise_data()` adds more flexibility by taking a format string and
1042 an arbitrary number of arguments following it, to be processed with
1047 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1048 to check if a named provider is loaded and available. When called, it
1049 will also activate all fallback providers if such are still present.
1053 * Enforce a minimum DH modulus size of 512 bits.
1057 * Changed DH parameters to generate the order q subgroup instead of 2q.
1058 Previously generated DH parameters are still accepted by DH_check
1059 but DH_generate_key works around that by clearing bit 0 of the
1060 private key for those. This avoids leaking bit 0 of the private key.
1064 * Significantly reduce secure memory usage by the randomness pools.
1068 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1073 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1074 algorithms. An implementation of a key exchange algorithm can be obtained
1075 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1076 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1077 the older EVP_PKEY_derive_init() function. See the man pages for the new
1078 functions for further details.
1082 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1086 * Removed the function names from error messages and deprecated the
1091 * Removed NextStep support and the macro OPENSSL_UNISTD
1095 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1096 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1097 Also removed "export var as function" capability; we do not export
1098 variables, only functions.
1102 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1103 an error and 1 indicating success. In previous versions of OpenSSL this
1104 was a void type. If a key was set longer than the maximum possible this
1109 * Support SM2 signing and verification schemes with X509 certificate.
1113 * Use SHA256 as the default digest for TS query in the `ts` app.
1117 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1121 * Default cipher lists/suites are now available via a function, the
1122 #defines are deprecated.
1126 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1127 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1128 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1132 * Join the directories crypto/x509 and crypto/x509v3
1136 * Added command 'openssl kdf' that uses the EVP_KDF API.
1140 * Added command 'openssl mac' that uses the EVP_MAC API.
1144 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1145 as default directories. Also added the command 'openssl info'
1146 for scripting purposes.
1150 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1155 * Add prediction resistance to the DRBG reseeding process.
1159 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1160 mandated by IEEE Std 1619-2018.
1164 * Added newline escaping functionality to a filename when using openssl dgst.
1165 This output format is to replicate the output format found in the `*sum`
1166 checksum programs. This aims to preserve backward compatibility.
1168 *Matt Eaton, Richard Levitte, and Paul Dale*
1170 * Removed the heartbeat message in DTLS feature, as it has very
1171 little usage and doesn't seem to fulfill a valuable purpose.
1172 The configuration option is now deprecated.
1176 * Changed the output of 'openssl {digestname} < file' to display the
1177 digest name in its output.
1181 * Added a new generic trace API which provides support for enabling
1182 instrumentation through trace output.
1184 *Richard Levitte & Matthias St. Pierre*
1186 * Added build tests for C++. These are generated files that only do one
1187 thing, to include one public OpenSSL head file each. This tests that
1188 the public header files can be usefully included in a C++ application.
1190 This test isn't enabled by default. It can be enabled with the option
1191 'enable-buildtest-c++'.
1195 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1199 * Add KMAC to EVP_MAC.
1203 * Added property based algorithm implementation selection framework to
1208 * Added SCA hardening for modular field inversion in EC_GROUP through
1209 a new dedicated field_inv() pointer in EC_METHOD.
1210 This also addresses a leakage affecting conversions from projective
1211 to affine coordinates.
1213 *Billy Bob Brumley, Nicola Tuveri*
1215 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1216 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1217 those algorithms that were already supported through the EVP_PKEY API
1218 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1219 and scrypt are now wrappers that call EVP_KDF.
1223 * Build devcrypto engine as a dynamic engine.
1225 *Eneas U de Queiroz*
1227 * Add keyed BLAKE2 to EVP_MAC.
1231 * Fix a bug in the computation of the endpoint-pair shared secret used
1232 by DTLS over SCTP. This breaks interoperability with older versions
1233 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1234 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1235 interoperability with such broken implementations. However, enabling
1236 this switch breaks interoperability with correct implementations.
1238 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1239 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1243 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1247 * Changed the license to the Apache License v2.0.
1251 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1253 - Major releases (indicated by incrementing the MAJOR release number)
1254 may introduce incompatible API/ABI changes.
1255 - Minor releases (indicated by incrementing the MINOR release number)
1256 may introduce new features but retain API/ABI compatibility.
1257 - Patch releases (indicated by incrementing the PATCH number)
1258 are intended for bug fixes and other improvements of existing
1259 features only (like improving performance or adding documentation)
1260 and retain API/ABI compatibility.
1264 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1268 * Remove the 'dist' target and add a tarball building script. The
1269 'dist' target has fallen out of use, and it shouldn't be
1270 necessary to configure just to create a source distribution.
1274 * Recreate the OS390-Unix config target. It no longer relies on a
1275 special script like it did for OpenSSL pre-1.1.0.
1279 * Instead of having the source directories listed in Configure, add
1280 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1285 * Add GMAC to EVP_MAC.
1289 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1293 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1294 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1295 to facilitate the continued use of MACs through raw private keys in
1296 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1300 * Deprecate ECDH_KDF_X9_62().
1304 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1305 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1306 are retained for backwards compatibility.
1310 * AES-XTS mode now enforces that its two keys are different to mitigate
1311 the attacked described in "Efficient Instantiations of Tweakable
1312 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1313 Details of this attack can be obtained from:
1314 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1318 * Rename the object files, i.e. give them other names than in previous
1319 versions. Their names now include the name of the final product, as
1320 well as its type mnemonic (bin, lib, shlib).
1324 * Added new option for 'openssl list', '-objects', which will display the
1325 list of built in objects, i.e. OIDs with names.
1329 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1330 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1335 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1336 improves application performance by removing data copies and providing
1337 applications with zero-copy system calls such as sendfile and splice.
1341 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1345 * `PKCS12_parse` now maintains the order of the parsed certificates
1346 when outputting them via `*ca` (rather than reversing it).
1350 * Deprecated pthread fork support methods.
1354 * Added support for FFDHE key exchange in TLS 1.3.
1361 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1363 * Fixed an SM2 Decryption Buffer Overflow.
1365 In order to decrypt SM2 encrypted data an application is expected to call the
1366 API function EVP_PKEY_decrypt(). Typically an application will call this
1367 function twice. The first time, on entry, the "out" parameter can be NULL and,
1368 on exit, the "outlen" parameter is populated with the buffer size required to
1369 hold the decrypted plaintext. The application can then allocate a sufficiently
1370 sized buffer and call EVP_PKEY_decrypt() again, but this time passing a non-NULL
1371 value for the "out" parameter.
1373 A bug in the implementation of the SM2 decryption code means that the
1374 calculation of the buffer size required to hold the plaintext returned by the
1375 first call to EVP_PKEY_decrypt() can be smaller than the actual size required by
1376 the second call. This can lead to a buffer overflow when EVP_PKEY_decrypt() is
1377 called by the application a second time with a buffer that is too small.
1379 A malicious attacker who is able present SM2 content for decryption to an
1380 application could cause attacker chosen data to overflow the buffer by up to a
1381 maximum of 62 bytes altering the contents of other data held after the
1382 buffer, possibly changing application behaviour or causing the application to
1383 crash. The location of the buffer is application dependent but is typically
1389 * Fixed various read buffer overruns processing ASN.1 strings
1391 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1392 structure which contains a buffer holding the string data and a field holding
1393 the buffer length. This contrasts with normal C strings which are repesented as
1394 a buffer for the string data which is terminated with a NUL (0) byte.
1396 Although not a strict requirement, ASN.1 strings that are parsed using OpenSSL's
1397 own "d2i" functions (and other similar parsing functions) as well as any string
1398 whose value has been set with the ASN1_STRING_set() function will additionally
1399 NUL terminate the byte array in the ASN1_STRING structure.
1401 However, it is possible for applications to directly construct valid ASN1_STRING
1402 structures which do not NUL terminate the byte array by directly setting the
1403 "data" and "length" fields in the ASN1_STRING array. This can also happen by
1404 using the ASN1_STRING_set0() function.
1406 Numerous OpenSSL functions that print ASN.1 data have been found to assume that
1407 the ASN1_STRING byte array will be NUL terminated, even though this is not
1408 guaranteed for strings that have been directly constructed. Where an application
1409 requests an ASN.1 structure to be printed, and where that ASN.1 structure
1410 contains ASN1_STRINGs that have been directly constructed by the application
1411 without NUL terminating the "data" field, then a read buffer overrun can occur.
1413 The same thing can also occur during name constraints processing of certificates
1414 (for example if a certificate has been directly constructed by the application
1415 instead of loading it via the OpenSSL parsing functions, and the certificate
1416 contains non NUL terminated ASN1_STRING structures). It can also occur in the
1417 X509_get1_email(), X509_REQ_get1_email() and X509_get1_ocsp() functions.
1419 If a malicious actor can cause an application to directly construct an
1420 ASN1_STRING and then process it through one of the affected OpenSSL functions
1421 then this issue could be hit. This might result in a crash (causing a Denial of
1422 Service attack). It could also result in the disclosure of private memory
1423 contents (such as private keys, or sensitive plaintext).
1428 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1430 * Fixed a problem with verifying a certificate chain when using the
1431 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1432 the certificates present in a certificate chain. It is not set by default.
1434 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1435 the chain that have explicitly encoded elliptic curve parameters was added
1436 as an additional strict check.
1438 An error in the implementation of this check meant that the result of a
1439 previous check to confirm that certificates in the chain are valid CA
1440 certificates was overwritten. This effectively bypasses the check
1441 that non-CA certificates must not be able to issue other certificates.
1443 If a "purpose" has been configured then there is a subsequent opportunity
1444 for checks that the certificate is a valid CA. All of the named "purpose"
1445 values implemented in libcrypto perform this check. Therefore, where
1446 a purpose is set the certificate chain will still be rejected even when the
1447 strict flag has been used. A purpose is set by default in libssl client and
1448 server certificate verification routines, but it can be overridden or
1449 removed by an application.
1451 In order to be affected, an application must explicitly set the
1452 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1453 for the certificate verification or, in the case of TLS client or server
1454 applications, override the default purpose.
1459 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1460 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1461 renegotiation ClientHello omits the signature_algorithms extension (where it
1462 was present in the initial ClientHello), but includes a
1463 signature_algorithms_cert extension then a NULL pointer dereference will
1464 result, leading to a crash and a denial of service attack.
1466 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1467 (which is the default configuration). OpenSSL TLS clients are not impacted by
1471 *Peter Kästle and Samuel Sapalski*
1473 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1475 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1476 create a unique hash value based on the issuer and serial number data
1477 contained within an X509 certificate. However it was failing to correctly
1478 handle any errors that may occur while parsing the issuer field (which might
1479 occur if the issuer field is maliciously constructed). This may subsequently
1480 result in a NULL pointer deref and a crash leading to a potential denial of
1486 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1487 padding mode to correctly check for rollback attacks. This is considered a
1488 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1493 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1494 functions. Previously they could overflow the output length argument in some
1495 cases where the input length is close to the maximum permissable length for
1496 an integer on the platform. In such cases the return value from the function
1497 call would be 1 (indicating success), but the output length value would be
1498 negative. This could cause applications to behave incorrectly or crash.
1503 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1504 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1505 could be exploited in a side channel attack to recover the password. Since
1506 the attack is local host only this is outside of the current OpenSSL
1507 threat model and therefore no CVE is assigned.
1509 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1514 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1516 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1517 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1518 If an attacker can control both items being compared then this could lead
1519 to a possible denial of service attack. OpenSSL itself uses the
1520 GENERAL_NAME_cmp function for two purposes:
1521 1) Comparing CRL distribution point names between an available CRL and a
1522 CRL distribution point embedded in an X509 certificate
1523 2) When verifying that a timestamp response token signer matches the
1524 timestamp authority name (exposed via the API functions
1525 TS_RESP_verify_response and TS_RESP_verify_token)
1530 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1532 * Certificates with explicit curve parameters are now disallowed in
1533 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1537 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1538 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1539 conversely, silently ignore DTLS protocol version bounds when configuring
1540 TLS-based contexts. The commands can be repeated to set bounds of both
1541 types. The same applies with the corresponding "min_protocol" and
1542 "max_protocol" command-line switches, in case some application uses both TLS
1545 SSL_CTX instances that are created for a fixed protocol version (e.g.
1546 TLSv1_server_method()) also silently ignore version bounds. Previously
1547 attempts to apply bounds to these protocol versions would result in an
1548 error. Now only the "version-flexible" SSL_CTX instances are subject to
1549 limits in configuration files in command-line options.
1553 * Handshake now fails if Extended Master Secret extension is dropped
1558 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1560 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1562 * Fixed segmentation fault in SSL_check_chain()
1563 Server or client applications that call the SSL_check_chain() function
1564 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1565 dereference as a result of incorrect handling of the
1566 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1567 or unrecognised signature algorithm is received from the peer. This could
1568 be exploited by a malicious peer in a Denial of Service attack.
1573 * Added AES consttime code for no-asm configurations
1574 an optional constant time support for AES was added
1575 when building openssl for no-asm.
1576 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1577 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1578 At this time this feature is by default disabled.
1579 It will be enabled by default in 3.0.
1583 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1585 * Revert the change of EOF detection while reading in libssl to avoid
1586 regressions in applications depending on the current way of reporting
1587 the EOF. As the existing method is not fully accurate the change to
1588 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1589 branch and will be present in the 3.0 release.
1593 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1594 when primes for RSA keys are computed.
1595 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1596 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1597 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1598 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1599 This avoids possible fingerprinting of newly generated RSA modules.
1603 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1605 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1606 while reading in libssl then we would report an error back to the
1607 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1608 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1609 therefore give a hint as to what went wrong.
1613 * Check that ed25519 and ed448 are allowed by the security level. Previously
1614 signature algorithms not using an MD were not being checked that they were
1615 allowed by the security level.
1619 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1620 was not quite right. The behaviour was not consistent between resumption
1621 and normal handshakes, and also not quite consistent with historical
1622 behaviour. The behaviour in various scenarios has been clarified and
1623 it has been updated to make it match historical behaviour as closely as
1628 * *[VMS only]* The header files that the VMS compilers include automatically,
1629 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1630 that the C++ compiler doesn't understand. This is a shortcoming in the
1631 compiler, but can be worked around with `__cplusplus` guards.
1633 C++ applications that use OpenSSL libraries must be compiled using the
1634 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1635 functions. Otherwise, only functions with symbols of less than 31
1636 characters can be used, as the linker will not be able to successfully
1637 resolve symbols with longer names.
1641 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1642 The presence of this system service is determined at run-time.
1646 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1651 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1653 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1654 number generator (RNG). This was intended to include protection in the
1655 event of a fork() system call in order to ensure that the parent and child
1656 processes did not share the same RNG state. However this protection was not
1657 being used in the default case.
1659 A partial mitigation for this issue is that the output from a high
1660 precision timer is mixed into the RNG state so the likelihood of a parent
1661 and child process sharing state is significantly reduced.
1663 If an application already calls OPENSSL_init_crypto() explicitly using
1664 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1667 *Matthias St. Pierre*
1669 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1670 used even when parsing explicit parameters, when loading a encoded key
1671 or calling `EC_GROUP_new_from_ecpkparameters()`/
1672 `EC_GROUP_new_from_ecparameters()`.
1673 This prevents bypass of security hardening and performance gains,
1674 especially for curves with specialized EC_METHODs.
1675 By default, if a key encoded with explicit parameters is loaded and later
1676 encoded, the output is still encoded with explicit parameters, even if
1677 internally a "named" EC_GROUP is used for computation.
1681 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1682 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1683 NULL. After this change, only the cofactor parameter can be NULL. It also
1684 does some minimal sanity checks on the passed order.
1689 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1690 An attack is simple, if the first CMS_recipientInfo is valid but the
1691 second CMS_recipientInfo is chosen ciphertext. If the second
1692 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1693 encryption key will be replaced by garbage, and the message cannot be
1694 decoded, but if the RSA decryption fails, the correct encryption key is
1695 used and the recipient will not notice the attack.
1696 As a work around for this potential attack the length of the decrypted
1697 key must be equal to the cipher default key length, in case the
1698 certifiate is not given and all recipientInfo are tried out.
1699 The old behaviour can be re-enabled in the CMS code by setting the
1700 CMS_DEBUG_DECRYPT flag.
1705 * Early start up entropy quality from the DEVRANDOM seed source has been
1706 improved for older Linux systems. The RAND subsystem will wait for
1707 /dev/random to be producing output before seeding from /dev/urandom.
1708 The seeded state is stored for future library initialisations using
1709 a system global shared memory segment. The shared memory identifier
1710 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1711 the desired value. The default identifier is 114.
1715 * Correct the extended master secret constant on EBCDIC systems. Without this
1716 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1717 negotiate EMS will fail. Unfortunately this also means that TLS connections
1718 between EBCDIC systems with this fix, and EBCDIC systems without this
1719 fix will fail if they negotiate EMS.
1723 * Use Windows installation paths in the mingw builds
1725 Mingw isn't a POSIX environment per se, which means that Windows
1726 paths should be used for installation.
1731 * Changed DH_check to accept parameters with order q and 2q subgroups.
1732 With order 2q subgroups the bit 0 of the private key is not secret
1733 but DH_generate_key works around that by clearing bit 0 of the
1734 private key for those. This avoids leaking bit 0 of the private key.
1738 * Significantly reduce secure memory usage by the randomness pools.
1742 * Revert the DEVRANDOM_WAIT feature for Linux systems
1744 The DEVRANDOM_WAIT feature added a select() call to wait for the
1745 /dev/random device to become readable before reading from the
1746 /dev/urandom device.
1748 It turned out that this change had negative side effects on
1749 performance which were not acceptable. After some discussion it
1750 was decided to revert this feature and leave it up to the OS
1751 resp. the platform maintainer to ensure a proper initialization
1752 during early boot time.
1754 *Matthias St. Pierre*
1756 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1758 * Add build tests for C++. These are generated files that only do one
1759 thing, to include one public OpenSSL head file each. This tests that
1760 the public header files can be usefully included in a C++ application.
1762 This test isn't enabled by default. It can be enabled with the option
1763 'enable-buildtest-c++'.
1767 * Enable SHA3 pre-hashing for ECDSA and DSA.
1771 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1772 This changes the size when using the `genpkey` command when no size is given.
1773 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1774 generation commands to use 2048 bits by default.
1778 * Reorganize the manual pages to consistently have RETURN VALUES,
1779 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1780 util/fix-doc-nits accordingly.
1782 *Paul Yang, Joshua Lock*
1784 * Add the missing accessor EVP_PKEY_get0_engine()
1788 * Have commands like `s_client` and `s_server` output the signature scheme
1789 along with other cipher suite parameters when debugging.
1793 * Make OPENSSL_config() error agnostic again.
1797 * Do the error handling in RSA decryption constant time.
1801 * Prevent over long nonces in ChaCha20-Poly1305.
1803 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1804 for every encryption operation. RFC 7539 specifies that the nonce value
1805 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1806 and front pads the nonce with 0 bytes if it is less than 12
1807 bytes. However it also incorrectly allows a nonce to be set of up to 16
1808 bytes. In this case only the last 12 bytes are significant and any
1809 additional leading bytes are ignored.
1811 It is a requirement of using this cipher that nonce values are
1812 unique. Messages encrypted using a reused nonce value are susceptible to
1813 serious confidentiality and integrity attacks. If an application changes
1814 the default nonce length to be longer than 12 bytes and then makes a
1815 change to the leading bytes of the nonce expecting the new value to be a
1816 new unique nonce then such an application could inadvertently encrypt
1817 messages with a reused nonce.
1819 Additionally the ignored bytes in a long nonce are not covered by the
1820 integrity guarantee of this cipher. Any application that relies on the
1821 integrity of these ignored leading bytes of a long nonce may be further
1822 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1823 is safe because no such use sets such a long nonce value. However user
1824 applications that use this cipher directly and set a non-default nonce
1825 length to be longer than 12 bytes may be vulnerable.
1827 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1833 * Add DEVRANDOM_WAIT feature for Linux systems
1835 On older Linux systems where the getrandom() system call is not available,
1836 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1837 Contrary to getrandom(), the /dev/urandom device will not block during
1838 early boot when the kernel CSPRNG has not been seeded yet.
1840 To mitigate this known weakness, use select() to wait for /dev/random to
1841 become readable before reading from /dev/urandom.
1843 * Ensure that SM2 only uses SM3 as digest algorithm
1847 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1849 * Change the info callback signals for the start and end of a post-handshake
1850 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1851 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1852 confused by this and assume that a TLSv1.2 renegotiation has started. This
1853 can break KeyUpdate handling. Instead we no longer signal the start and end
1854 of a post handshake message exchange (although the messages themselves are
1855 still signalled). This could break some applications that were expecting
1856 the old signals. However without this KeyUpdate is not usable for many
1861 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1863 * Timing vulnerability in DSA signature generation
1865 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1866 timing side channel attack. An attacker could use variations in the signing
1867 algorithm to recover the private key.
1869 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1874 * Timing vulnerability in ECDSA signature generation
1876 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1877 timing side channel attack. An attacker could use variations in the signing
1878 algorithm to recover the private key.
1880 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1885 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1886 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1887 of two gigabytes and the error handling improved.
1889 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1890 categorized as a normal bug, not a security issue, because the DRBG reseeds
1891 automatically and is fully functional even without additional randomness
1892 provided by the application.
1894 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1896 * Add a new ClientHello callback. Provides a callback interface that gives
1897 the application the ability to adjust the nascent SSL object at the
1898 earliest stage of ClientHello processing, immediately after extensions have
1899 been collected but before they have been processed. In particular, this
1900 callback can adjust the supported TLS versions in response to the contents
1905 * Add SM2 base algorithm support.
1909 * s390x assembly pack: add (improved) hardware-support for the following
1910 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1911 aes-cfb/cfb8, aes-ecb.
1915 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1916 parameter is no longer accepted, as it leads to a corrupt table. NULL
1917 pem_str is reserved for alias entries only.
1921 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1922 step for prime curves. The new implementation is based on formulae from
1923 differential addition-and-doubling in homogeneous projective coordinates
1924 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1925 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1926 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1927 to work in projective coordinates.
1929 *Billy Bob Brumley, Nicola Tuveri*
1931 * Change generating and checking of primes so that the error rate of not
1932 being prime depends on the intended use based on the size of the input.
1933 For larger primes this will result in more rounds of Miller-Rabin.
1934 The maximal error rate for primes with more than 1080 bits is lowered
1937 *Kurt Roeckx, Annie Yousar*
1939 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1943 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1944 moving between systems, and to avoid confusion when a Windows build is
1945 done with mingw vs with MSVC. For POSIX installs, there's still a
1946 symlink or copy named 'tsget' to avoid that confusion as well.
1950 * Revert blinding in ECDSA sign and instead make problematic addition
1951 length-invariant. Switch even to fixed-length Montgomery multiplication.
1955 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1956 step for binary curves. The new implementation is based on formulae from
1957 differential addition-and-doubling in mixed Lopez-Dahab projective
1958 coordinates, modified to independently blind the operands.
1960 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1962 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1963 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1964 EC_METHODs to implement their own specialized "ladder step", to take
1965 advantage of more favorable coordinate systems or more efficient
1966 differential addition-and-doubling algorithms.
1968 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1970 * Modified the random device based seed sources to keep the relevant
1971 file descriptors open rather than reopening them on each access.
1972 This allows such sources to operate in a chroot() jail without
1973 the associated device nodes being available. This behaviour can be
1974 controlled using RAND_keep_random_devices_open().
1978 * Numerous side-channel attack mitigations have been applied. This may have
1979 performance impacts for some algorithms for the benefit of improved
1980 security. Specific changes are noted in this change log by their respective
1985 * AIX shared library support overhaul. Switch to AIX "natural" way of
1986 handling shared libraries, which means collecting shared objects of
1987 different versions and bitnesses in one common archive. This allows to
1988 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1989 doesn't affect the way 3rd party applications are linked, only how
1990 multi-version installation is managed.
1994 * Make ec_group_do_inverse_ord() more robust and available to other
1995 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1996 mitigations are applied to the fallback BN_mod_inverse().
1997 When using this function rather than BN_mod_inverse() directly, new
1998 EC cryptosystem implementations are then safer-by-default.
2002 * Add coordinate blinding for EC_POINT and implement projective
2003 coordinate blinding for generic prime curves as a countermeasure to
2004 chosen point SCA attacks.
2006 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2008 * Add blinding to ECDSA and DSA signatures to protect against side channel
2009 attacks discovered by Keegan Ryan (NCC Group).
2013 * Enforce checking in the `pkeyutl` command to ensure that the input
2014 length does not exceed the maximum supported digest length when performing
2015 a sign, verify or verifyrecover operation.
2019 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2020 I/O in combination with something like select() or poll() will hang. This
2021 can be turned off again using SSL_CTX_clear_mode().
2022 Many applications do not properly handle non-application data records, and
2023 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2024 around the problems in those applications, but can also break some.
2025 It's recommended to read the manpages about SSL_read(), SSL_write(),
2026 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2027 SSL_CTX_set_read_ahead() again.
2031 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2032 now allow empty (zero character) pass phrases.
2036 * Apply blinding to binary field modular inversion and remove patent
2037 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2041 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2042 binary and prime elliptic curves.
2046 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2047 constant time fixed point multiplication.
2051 * Revise elliptic curve scalar multiplication with timing attack
2052 defenses: ec_wNAF_mul redirects to a constant time implementation
2053 when computing fixed point and variable point multiplication (which
2054 in OpenSSL are mostly used with secret scalars in keygen, sign,
2055 ECDH derive operations).
2056 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2059 * Updated CONTRIBUTING
2063 * Updated DRBG / RAND to request nonce and additional low entropy
2064 randomness from the system.
2066 *Matthias St. Pierre*
2068 * Updated 'openssl rehash' to use OpenSSL consistent default.
2072 * Moved the load of the ssl_conf module to libcrypto, which helps
2073 loading engines that libssl uses before libssl is initialised.
2077 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2081 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2083 *Ingo Schwarze, Rich Salz*
2085 * Added output of accepting IP address and port for 'openssl s_server'
2089 * Added a new API for TLSv1.3 ciphersuites:
2090 SSL_CTX_set_ciphersuites()
2091 SSL_set_ciphersuites()
2095 * Memory allocation failures consistently add an error to the error
2100 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2101 in libcrypto when run as setuid/setgid.
2105 * Load any config file by default when libssl is used.
2109 * Added new public header file <openssl/rand_drbg.h> and documentation
2110 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2112 *Matthias St. Pierre*
2114 * QNX support removed (cannot find contributors to get their approval
2115 for the license change).
2119 * TLSv1.3 replay protection for early data has been implemented. See the
2120 SSL_read_early_data() man page for further details.
2124 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2125 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2126 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2127 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2128 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2129 configuration has been separated out. See the ciphers man page or the
2130 SSL_CTX_set_ciphersuites() man page for more information.
2134 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2135 in responder mode now supports the new "-multi" option, which
2136 spawns the specified number of child processes to handle OCSP
2137 requests. The "-timeout" option now also limits the OCSP
2138 responder's patience to wait to receive the full client request
2139 on a newly accepted connection. Child processes are respawned
2140 as needed, and the CA index file is automatically reloaded
2141 when changed. This makes it possible to run the "ocsp" responder
2142 as a long-running service, making the OpenSSL CA somewhat more
2143 feature-complete. In this mode, most diagnostic messages logged
2144 after entering the event loop are logged via syslog(3) rather than
2149 * Added support for X448 and Ed448. Heavily based on original work by
2154 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2155 objects loaded. This adds the functions OSSL_STORE_expect() and
2156 OSSL_STORE_find() as well as needed tools to construct searches and
2157 get the search data out of them.
2161 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2162 version of OpenSSL should review their configuration settings to ensure
2163 that they are still appropriate for TLSv1.3. For further information see:
2164 <https://wiki.openssl.org/index.php/TLS1.3>
2168 * Grand redesign of the OpenSSL random generator
2170 The default RAND method now utilizes an AES-CTR DRBG according to
2171 NIST standard SP 800-90Ar1. The new random generator is essentially
2172 a port of the default random generator from the OpenSSL FIPS 2.0
2173 object module. It is a hybrid deterministic random bit generator
2174 using an AES-CTR bit stream and which seeds and reseeds itself
2175 automatically using trusted system entropy sources.
2177 Some of its new features are:
2178 - Support for multiple DRBG instances with seed chaining.
2179 - The default RAND method makes use of a DRBG.
2180 - There is a public and private DRBG instance.
2181 - The DRBG instances are fork-safe.
2182 - Keep all global DRBG instances on the secure heap if it is enabled.
2183 - The public and private DRBG instance are per thread for lock free
2186 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2188 * Changed Configure so it only says what it does and doesn't dump
2189 so much data. Instead, ./configdata.pm should be used as a script
2190 to display all sorts of configuration data.
2194 * Added processing of "make variables" to Configure.
2198 * Added SHA512/224 and SHA512/256 algorithm support.
2202 * The last traces of Netware support, first removed in 1.1.0, have
2207 * Get rid of Makefile.shared, and in the process, make the processing
2208 of certain files (rc.obj, or the .def/.map/.opt files produced from
2209 the ordinal files) more visible and hopefully easier to trace and
2210 debug (or make silent).
2214 * Make it possible to have environment variable assignments as
2215 arguments to config / Configure.
2219 * Add multi-prime RSA (RFC 8017) support.
2223 * Add SM3 implemented according to GB/T 32905-2016
2224 *Jack Lloyd <jack.lloyd@ribose.com>,*
2225 *Ronald Tse <ronald.tse@ribose.com>,*
2226 *Erick Borsboom <erick.borsboom@ribose.com>*
2228 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2229 as documented in RFC6066.
2230 Based on a patch from Tomasz Moń
2232 *Filipe Raimundo da Silva*
2234 * Add SM4 implemented according to GB/T 32907-2016.
2235 *Jack Lloyd <jack.lloyd@ribose.com>,*
2236 *Ronald Tse <ronald.tse@ribose.com>,*
2237 *Erick Borsboom <erick.borsboom@ribose.com>*
2239 * Reimplement -newreq-nodes and ERR_error_string_n; the
2240 original author does not agree with the license change.
2244 * Add ARIA AEAD TLS support.
2248 * Some macro definitions to support VS6 have been removed. Visual
2249 Studio 6 has not worked since 1.1.0
2253 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2254 without clearing the errors.
2258 * Add "atfork" functions. If building on a system that without
2259 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2260 requirements. The RAND facility now uses/requires this.
2268 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2269 not possible to disable entirely. However, it's still possible to
2270 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2273 To disable, configure with 'no-ui-console'. 'no-ui' is still
2274 possible to use as an alias. Check at compile time with the
2275 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2276 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2280 * Add a STORE module, which implements a uniform and URI based reader of
2281 stores that can contain keys, certificates, CRLs and numerous other
2282 objects. The main API is loosely based on a few stdio functions,
2283 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2284 OSSL_STORE_error and OSSL_STORE_close.
2285 The implementation uses backends called "loaders" to implement arbitrary
2286 URI schemes. There is one built in "loader" for the 'file' scheme.
2290 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2291 then adjusted to work on FreeBSD 8.4 as well.
2292 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2293 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2297 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2298 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2299 error code calls like this:
2301 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2303 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2304 that can be encoded in C. For the foreseeable future, this will only
2307 *Richard Levitte and Tim Hudson*
2309 * Removed BSD cryptodev engine.
2313 * Add a build target 'build_all_generated', to build all generated files
2314 and only that. This can be used to prepare everything that requires
2315 things like perl for a system that lacks perl and then move everything
2316 to that system and do the rest of the build there.
2320 * In the UI interface, make it possible to duplicate the user data. This
2321 can be used by engines that need to retain the data for a longer time
2322 than just the call where this user data is passed.
2326 * Ignore the '-named_curve auto' value for compatibility of applications
2329 *Tomáš Mráz <tmraz@fedoraproject.org>*
2331 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2332 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2333 alerts across multiple records (some of which could be empty). In practice
2334 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2335 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2336 support this at all. Supporting it adds significant complexity to the
2337 record layer, and its removal is unlikely to cause interoperability
2342 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2343 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2344 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2349 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2350 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2352 *Richard Levitte, Andy Polyakov*
2354 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2359 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2360 platform rather than 'mingw'.
2364 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2365 success if they are asked to add an object which already exists
2366 in the store. This change cascades to other functions which load
2367 certificates and CRLs.
2371 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2372 facilitate stack unwinding even from assembly subroutines.
2376 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2377 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2381 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2382 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2383 which is the minimum version we support.
2387 * Certificate time validation (X509_cmp_time) enforces stricter
2388 compliance with RFC 5280. Fractional seconds and timezone offsets
2389 are no longer allowed.
2393 * Add support for ARIA
2397 * s_client will now send the Server Name Indication (SNI) extension by
2398 default unless the new "-noservername" option is used. The server name is
2399 based on the host provided to the "-connect" option unless overridden by
2400 using "-servername".
2404 * Add support for SipHash
2408 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2409 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2410 prevent issues where no progress is being made and the peer continually
2411 sends unrecognised record types, using up resources processing them.
2415 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2416 using the algorithm defined in
2417 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2421 * Heartbeat support has been removed; the ABI is changed for now.
2423 *Richard Levitte, Rich Salz*
2425 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2429 * The RSA "null" method, which was partially supported to avoid patent
2430 issues, has been replaced to always returns NULL.
2437 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2439 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2440 used even when parsing explicit parameters, when loading a encoded key
2441 or calling `EC_GROUP_new_from_ecpkparameters()`/
2442 `EC_GROUP_new_from_ecparameters()`.
2443 This prevents bypass of security hardening and performance gains,
2444 especially for curves with specialized EC_METHODs.
2445 By default, if a key encoded with explicit parameters is loaded and later
2446 encoded, the output is still encoded with explicit parameters, even if
2447 internally a "named" EC_GROUP is used for computation.
2451 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2452 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2453 NULL. After this change, only the cofactor parameter can be NULL. It also
2454 does some minimal sanity checks on the passed order.
2459 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2460 An attack is simple, if the first CMS_recipientInfo is valid but the
2461 second CMS_recipientInfo is chosen ciphertext. If the second
2462 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2463 encryption key will be replaced by garbage, and the message cannot be
2464 decoded, but if the RSA decryption fails, the correct encryption key is
2465 used and the recipient will not notice the attack.
2466 As a work around for this potential attack the length of the decrypted
2467 key must be equal to the cipher default key length, in case the
2468 certifiate is not given and all recipientInfo are tried out.
2469 The old behaviour can be re-enabled in the CMS code by setting the
2470 CMS_DEBUG_DECRYPT flag.
2475 * Use Windows installation paths in the mingw builds
2477 Mingw isn't a POSIX environment per se, which means that Windows
2478 paths should be used for installation.
2483 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2485 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2486 This changes the size when using the `genpkey` command when no size is given.
2487 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2488 generation commands to use 2048 bits by default.
2492 * Prevent over long nonces in ChaCha20-Poly1305.
2494 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2495 for every encryption operation. RFC 7539 specifies that the nonce value
2496 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2497 and front pads the nonce with 0 bytes if it is less than 12
2498 bytes. However it also incorrectly allows a nonce to be set of up to 16
2499 bytes. In this case only the last 12 bytes are significant and any
2500 additional leading bytes are ignored.
2502 It is a requirement of using this cipher that nonce values are
2503 unique. Messages encrypted using a reused nonce value are susceptible to
2504 serious confidentiality and integrity attacks. If an application changes
2505 the default nonce length to be longer than 12 bytes and then makes a
2506 change to the leading bytes of the nonce expecting the new value to be a
2507 new unique nonce then such an application could inadvertently encrypt
2508 messages with a reused nonce.
2510 Additionally the ignored bytes in a long nonce are not covered by the
2511 integrity guarantee of this cipher. Any application that relies on the
2512 integrity of these ignored leading bytes of a long nonce may be further
2513 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2514 is safe because no such use sets such a long nonce value. However user
2515 applications that use this cipher directly and set a non-default nonce
2516 length to be longer than 12 bytes may be vulnerable.
2518 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2524 * Added SCA hardening for modular field inversion in EC_GROUP through
2525 a new dedicated field_inv() pointer in EC_METHOD.
2526 This also addresses a leakage affecting conversions from projective
2527 to affine coordinates.
2529 *Billy Bob Brumley, Nicola Tuveri*
2531 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2532 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2536 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2540 * Remove the 'dist' target and add a tarball building script. The
2541 'dist' target has fallen out of use, and it shouldn't be
2542 necessary to configure just to create a source distribution.
2546 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2548 * Timing vulnerability in DSA signature generation
2550 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2551 timing side channel attack. An attacker could use variations in the signing
2552 algorithm to recover the private key.
2554 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2559 * Timing vulnerability in ECDSA signature generation
2561 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2562 timing side channel attack. An attacker could use variations in the signing
2563 algorithm to recover the private key.
2565 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2570 * Add coordinate blinding for EC_POINT and implement projective
2571 coordinate blinding for generic prime curves as a countermeasure to
2572 chosen point SCA attacks.
2574 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2576 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2578 * Client DoS due to large DH parameter
2580 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2581 malicious server can send a very large prime value to the client. This will
2582 cause the client to spend an unreasonably long period of time generating a
2583 key for this prime resulting in a hang until the client has finished. This
2584 could be exploited in a Denial Of Service attack.
2586 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2591 * Cache timing vulnerability in RSA Key Generation
2593 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2594 a cache timing side channel attack. An attacker with sufficient access to
2595 mount cache timing attacks during the RSA key generation process could
2596 recover the private key.
2598 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2599 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2604 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2605 parameter is no longer accepted, as it leads to a corrupt table. NULL
2606 pem_str is reserved for alias entries only.
2610 * Revert blinding in ECDSA sign and instead make problematic addition
2611 length-invariant. Switch even to fixed-length Montgomery multiplication.
2615 * Change generating and checking of primes so that the error rate of not
2616 being prime depends on the intended use based on the size of the input.
2617 For larger primes this will result in more rounds of Miller-Rabin.
2618 The maximal error rate for primes with more than 1080 bits is lowered
2621 *Kurt Roeckx, Annie Yousar*
2623 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2627 * Add blinding to ECDSA and DSA signatures to protect against side channel
2628 attacks discovered by Keegan Ryan (NCC Group).
2632 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2633 now allow empty (zero character) pass phrases.
2637 * Certificate time validation (X509_cmp_time) enforces stricter
2638 compliance with RFC 5280. Fractional seconds and timezone offsets
2639 are no longer allowed.
2643 * Fixed a text canonicalisation bug in CMS
2645 Where a CMS detached signature is used with text content the text goes
2646 through a canonicalisation process first prior to signing or verifying a
2647 signature. This process strips trailing space at the end of lines, converts
2648 line terminators to CRLF and removes additional trailing line terminators
2649 at the end of a file. A bug in the canonicalisation process meant that
2650 some characters, such as form-feed, were incorrectly treated as whitespace
2651 and removed. This is contrary to the specification (RFC5485). This fix
2652 could mean that detached text data signed with an earlier version of
2653 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2654 signed with a fixed OpenSSL may fail to verify with an earlier version of
2655 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2656 and use the "-binary" flag (for the "cms" command line application) or set
2657 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2661 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2663 * Constructed ASN.1 types with a recursive definition could exceed the stack
2665 Constructed ASN.1 types with a recursive definition (such as can be found
2666 in PKCS7) could eventually exceed the stack given malicious input with
2667 excessive recursion. This could result in a Denial Of Service attack. There
2668 are no such structures used within SSL/TLS that come from untrusted sources
2669 so this is considered safe.
2671 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2677 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2679 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2680 effectively reduced to only comparing the least significant bit of each
2681 byte. This allows an attacker to forge messages that would be considered as
2682 authenticated in an amount of tries lower than that guaranteed by the
2683 security claims of the scheme. The module can only be compiled by the
2684 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2686 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2692 * Add a build target 'build_all_generated', to build all generated files
2693 and only that. This can be used to prepare everything that requires
2694 things like perl for a system that lacks perl and then move everything
2695 to that system and do the rest of the build there.
2699 * Backport SSL_OP_NO_RENGOTIATION
2701 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2702 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2703 changes this is no longer possible in 1.1.0. Therefore the new
2704 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2705 1.1.0 to provide equivalent functionality.
2707 Note that if an application built against 1.1.0h headers (or above) is run
2708 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2709 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2713 * Removed the OS390-Unix config target. It relied on a script that doesn't
2718 * rsaz_1024_mul_avx2 overflow bug on x86_64
2720 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2721 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2722 Analysis suggests that attacks against RSA and DSA as a result of this
2723 defect would be very difficult to perform and are not believed likely.
2724 Attacks against DH1024 are considered just feasible, because most of the
2725 work necessary to deduce information about a private key may be performed
2726 offline. The amount of resources required for such an attack would be
2727 significant. However, for an attack on TLS to be meaningful, the server
2728 would have to share the DH1024 private key among multiple clients, which is
2729 no longer an option since CVE-2016-0701.
2731 This only affects processors that support the AVX2 but not ADX extensions
2732 like Intel Haswell (4th generation).
2734 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2735 was originally found via the OSS-Fuzz project.
2740 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2742 * bn_sqrx8x_internal carry bug on x86_64
2744 There is a carry propagating bug in the x86_64 Montgomery squaring
2745 procedure. No EC algorithms are affected. Analysis suggests that attacks
2746 against RSA and DSA as a result of this defect would be very difficult to
2747 perform and are not believed likely. Attacks against DH are considered just
2748 feasible (although very difficult) because most of the work necessary to
2749 deduce information about a private key may be performed offline. The amount
2750 of resources required for such an attack would be very significant and
2751 likely only accessible to a limited number of attackers. An attacker would
2752 additionally need online access to an unpatched system using the target
2753 private key in a scenario with persistent DH parameters and a private
2754 key that is shared between multiple clients.
2756 This only affects processors that support the BMI1, BMI2 and ADX extensions
2757 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2759 This issue was reported to OpenSSL by the OSS-Fuzz project.
2764 * Malformed X.509 IPAddressFamily could cause OOB read
2766 If an X.509 certificate has a malformed IPAddressFamily extension,
2767 OpenSSL could do a one-byte buffer overread. The most likely result
2768 would be an erroneous display of the certificate in text format.
2770 This issue was reported to OpenSSL by the OSS-Fuzz project.
2775 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2777 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2778 platform rather than 'mingw'.
2782 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2783 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2784 which is the minimum version we support.
2788 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2790 * Encrypt-Then-Mac renegotiation crash
2792 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2793 negotiated where it was not in the original handshake (or vice-versa) then
2794 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2795 and servers are affected.
2797 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2802 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2804 * Truncated packet could crash via OOB read
2806 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2807 cipher is being used, then a truncated packet can cause that host to
2808 perform an out-of-bounds read, usually resulting in a crash.
2810 This issue was reported to OpenSSL by Robert Święcki of Google.
2815 * Bad (EC)DHE parameters cause a client crash
2817 If a malicious server supplies bad parameters for a DHE or ECDHE key
2818 exchange then this can result in the client attempting to dereference a
2819 NULL pointer leading to a client crash. This could be exploited in a Denial
2822 This issue was reported to OpenSSL by Guido Vranken.
2827 * BN_mod_exp may produce incorrect results on x86_64
2829 There is a carry propagating bug in the x86_64 Montgomery squaring
2830 procedure. No EC algorithms are affected. Analysis suggests that attacks
2831 against RSA and DSA as a result of this defect would be very difficult to
2832 perform and are not believed likely. Attacks against DH are considered just
2833 feasible (although very difficult) because most of the work necessary to
2834 deduce information about a private key may be performed offline. The amount
2835 of resources required for such an attack would be very significant and
2836 likely only accessible to a limited number of attackers. An attacker would
2837 additionally need online access to an unpatched system using the target
2838 private key in a scenario with persistent DH parameters and a private
2839 key that is shared between multiple clients. For example this can occur by
2840 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2841 similar to CVE-2015-3193 but must be treated as a separate problem.
2843 This issue was reported to OpenSSL by the OSS-Fuzz project.
2848 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2850 * ChaCha20/Poly1305 heap-buffer-overflow
2852 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2853 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2854 crash. This issue is not considered to be exploitable beyond a DoS.
2856 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2861 * CMS Null dereference
2863 Applications parsing invalid CMS structures can crash with a NULL pointer
2864 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2865 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2866 structure callback if an attempt is made to free certain invalid encodings.
2867 Only CHOICE structures using a callback which do not handle NULL value are
2870 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2875 * Montgomery multiplication may produce incorrect results
2877 There is a carry propagating bug in the Broadwell-specific Montgomery
2878 multiplication procedure that handles input lengths divisible by, but
2879 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2880 and DH private keys are impossible. This is because the subroutine in
2881 question is not used in operations with the private key itself and an input
2882 of the attacker's direct choice. Otherwise the bug can manifest itself as
2883 transient authentication and key negotiation failures or reproducible
2884 erroneous outcome of public-key operations with specially crafted input.
2885 Among EC algorithms only Brainpool P-512 curves are affected and one
2886 presumably can attack ECDH key negotiation. Impact was not analyzed in
2887 detail, because pre-requisites for attack are considered unlikely. Namely
2888 multiple clients have to choose the curve in question and the server has to
2889 share the private key among them, neither of which is default behaviour.
2890 Even then only clients that chose the curve will be affected.
2892 This issue was publicly reported as transient failures and was not
2893 initially recognized as a security issue. Thanks to Richard Morgan for
2894 providing reproducible case.
2899 * Removed automatic addition of RPATH in shared libraries and executables,
2900 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2904 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2906 * Fix Use After Free for large message sizes
2908 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2909 message larger than approx 16k is received then the underlying buffer to
2910 store the incoming message is reallocated and moved. Unfortunately a
2911 dangling pointer to the old location is left which results in an attempt to
2912 write to the previously freed location. This is likely to result in a
2913 crash, however it could potentially lead to execution of arbitrary code.
2915 This issue only affects OpenSSL 1.1.0a.
2917 This issue was reported to OpenSSL by Robert Święcki.
2922 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2924 * OCSP Status Request extension unbounded memory growth
2926 A malicious client can send an excessively large OCSP Status Request
2927 extension. If that client continually requests renegotiation, sending a
2928 large OCSP Status Request extension each time, then there will be unbounded
2929 memory growth on the server. This will eventually lead to a Denial Of
2930 Service attack through memory exhaustion. Servers with a default
2931 configuration are vulnerable even if they do not support OCSP. Builds using
2932 the "no-ocsp" build time option are not affected.
2934 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2939 * SSL_peek() hang on empty record
2941 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2942 sends an empty record. This could be exploited by a malicious peer in a
2943 Denial Of Service attack.
2945 This issue was reported to OpenSSL by Alex Gaynor.
2950 * Excessive allocation of memory in tls_get_message_header() and
2951 dtls1_preprocess_fragment()
2953 A (D)TLS message includes 3 bytes for its length in the header for the
2954 message. This would allow for messages up to 16Mb in length. Messages of
2955 this length are excessive and OpenSSL includes a check to ensure that a
2956 peer is sending reasonably sized messages in order to avoid too much memory
2957 being consumed to service a connection. A flaw in the logic of version
2958 1.1.0 means that memory for the message is allocated too early, prior to
2959 the excessive message length check. Due to way memory is allocated in
2960 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2961 to service a connection. This could lead to a Denial of Service through
2962 memory exhaustion. However, the excessive message length check still takes
2963 place, and this would cause the connection to immediately fail. Assuming
2964 that the application calls SSL_free() on the failed connection in a timely
2965 manner then the 21Mb of allocated memory will then be immediately freed
2966 again. Therefore the excessive memory allocation will be transitory in
2967 nature. This then means that there is only a security impact if:
2969 1) The application does not call SSL_free() in a timely manner in the event
2970 that the connection fails
2972 2) The application is working in a constrained environment where there is
2973 very little free memory
2975 3) The attacker initiates multiple connection attempts such that there are
2976 multiple connections in a state where memory has been allocated for the
2977 connection; SSL_free() has not yet been called; and there is insufficient
2978 memory to service the multiple requests.
2980 Except in the instance of (1) above any Denial Of Service is likely to be
2981 transitory because as soon as the connection fails the memory is
2982 subsequently freed again in the SSL_free() call. However there is an
2983 increased risk during this period of application crashes due to the lack of
2984 memory - which would then mean a more serious Denial of Service.
2986 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2987 (CVE-2016-6307 and CVE-2016-6308)
2991 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2992 had to be removed. Primary reason is that vendor assembler can't
2993 assemble our modules with -KPIC flag. As result it, assembly
2994 support, was not even available as option. But its lack means
2995 lack of side-channel resistant code, which is incompatible with
2996 security by todays standards. Fortunately gcc is readily available
2997 prepackaged option, which we firmly point at...
3001 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3003 * Windows command-line tool supports UTF-8 opt-in option for arguments
3004 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3005 (to any value) allows Windows user to access PKCS#12 file generated
3006 with Windows CryptoAPI and protected with non-ASCII password, as well
3007 as files generated under UTF-8 locale on Linux also protected with
3012 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3013 have been disabled by default and removed from DEFAULT, just like RC4.
3014 See the RC4 item below to re-enable both.
3018 * The method for finding the storage location for the Windows RAND seed file
3019 has changed. First we check %RANDFILE%. If that is not set then we check
3020 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3021 all else fails we fall back to C:\.
3025 * The EVP_EncryptUpdate() function has had its return type changed from void
3026 to int. A return of 0 indicates and error while a return of 1 indicates
3031 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3032 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3033 off the constant time implementation for RSA, DSA and DH have been made
3034 no-ops and deprecated.
3038 * Windows RAND implementation was simplified to only get entropy by
3039 calling CryptGenRandom(). Various other RAND-related tickets
3042 *Joseph Wylie Yandle, Rich Salz*
3044 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3045 and `OPENSSL_LH_`, respectively. The old names are available
3046 with API compatibility. They new names are now completely documented.
3050 * Unify TYPE_up_ref(obj) methods signature.
3051 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3052 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3053 int (instead of void) like all others TYPE_up_ref() methods.
3054 So now these methods also check the return value of CRYPTO_atomic_add(),
3055 and the validity of object reference counter.
3057 *fdasilvayy@gmail.com*
3059 * With Windows Visual Studio builds, the .pdb files are installed
3060 alongside the installed libraries and executables. For a static
3061 library installation, ossl_static.pdb is the associate compiler
3062 generated .pdb file to be used when linking programs.
3066 * Remove openssl.spec. Packaging files belong with the packagers.
3070 * Automatic Darwin/OSX configuration has had a refresh, it will now
3071 recognise x86_64 architectures automatically. You can still decide
3072 to build for a different bitness with the environment variable
3073 KERNEL_BITS (can be 32 or 64), for example:
3075 KERNEL_BITS=32 ./config
3079 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3080 256 bit AES and HMAC with SHA256.
3084 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3088 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3092 * To enable users to have their own config files and build file templates,
3093 Configure looks in the directory indicated by the environment variable
3094 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3095 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3096 name and is used as is.
3100 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3101 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3102 X509_CERT_FILE_CTX was removed.
3106 * "shared" builds are now the default. To create only static libraries use
3107 the "no-shared" Configure option.
3111 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3112 All of these option have not worked for some while and are fundamental
3117 * Make various cleanup routines no-ops and mark them as deprecated. Most
3118 global cleanup functions are no longer required because they are handled
3119 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3120 Explicitly de-initing can cause problems (e.g. where a library that uses
3121 OpenSSL de-inits, but an application is still using it). The affected
3122 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3123 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3124 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3125 COMP_zlib_cleanup().
3129 * --strict-warnings no longer enables runtime debugging options
3130 such as REF_DEBUG. Instead, debug options are automatically
3131 enabled with '--debug' builds.
3133 *Andy Polyakov, Emilia Käsper*
3135 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3136 have been moved out of the public header files. New functions for managing
3137 these have been added.
3141 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3142 objects have been moved out of the public header files. New
3143 functions for managing these have been added.
3147 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3148 have been moved out of the public header files. New functions for managing
3149 these have been added.
3153 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3154 moved out of the public header files. New functions for managing these
3159 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3163 * Removed the mk1mf build scripts.
3167 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3168 it is always safe to #include a header now.
3172 * Removed the aged BC-32 config and all its supporting scripts
3176 * Removed support for Ultrix, Netware, and OS/2.
3180 * Add support for HKDF.
3182 *Alessandro Ghedini*
3184 * Add support for blake2b and blake2s
3188 * Added support for "pipelining". Ciphers that have the
3189 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3190 encryptions/decryptions simultaneously. There are currently no built-in
3191 ciphers with this property but the expectation is that engines will be able
3192 to offer it to significantly improve throughput. Support has been extended
3193 into libssl so that multiple records for a single connection can be
3194 processed in one go (for >=TLS 1.1).
3198 * Added the AFALG engine. This is an async capable engine which is able to
3199 offload work to the Linux kernel. In this initial version it only supports
3200 AES128-CBC. The kernel must be version 4.1.0 or greater.
3204 * OpenSSL now uses a new threading API. It is no longer necessary to
3205 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3206 are two supported threading models: pthreads and windows threads. It is
3207 also possible to configure OpenSSL at compile time for "no-threads". The
3208 old threading API should no longer be used. The functions have been
3209 replaced with "no-op" compatibility macros.
3211 *Alessandro Ghedini, Matt Caswell*
3213 * Modify behavior of ALPN to invoke callback after SNI/servername
3214 callback, such that updates to the SSL_CTX affect ALPN.
3218 * Add SSL_CIPHER queries for authentication and key-exchange.
3222 * Changes to the DEFAULT cipherlist:
3223 - Prefer (EC)DHE handshakes over plain RSA.
3224 - Prefer AEAD ciphers over legacy ciphers.
3225 - Prefer ECDSA over RSA when both certificates are available.
3226 - Prefer TLSv1.2 ciphers/PRF.
3227 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3232 * Change the ECC default curve list to be this, in order: x25519,
3233 secp256r1, secp521r1, secp384r1.
3237 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3238 disabled by default. They can be re-enabled using the
3239 enable-weak-ssl-ciphers option to Configure.
3243 * If the server has ALPN configured, but supports no protocols that the
3244 client advertises, send a fatal "no_application_protocol" alert.
3245 This behaviour is SHALL in RFC 7301, though it isn't universally
3246 implemented by other servers.
3250 * Add X25519 support.
3251 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3252 for public and private key encoding using the format documented in
3253 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3254 key generation and key derivation.
3256 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3261 * Deprecate SRP_VBASE_get_by_user.
3262 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3263 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3264 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3265 seed, even if the seed is configured.
3267 Users should use SRP_VBASE_get1_by_user instead. Note that in
3268 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3269 also that even though configuring the SRP seed attempts to hide
3270 invalid usernames by continuing the handshake with fake
3271 credentials, this behaviour is not constant time and no strong
3272 guarantees are made that the handshake is indistinguishable from
3273 that of a valid user.
3277 * Configuration change; it's now possible to build dynamic engines
3278 without having to build shared libraries and vice versa. This
3279 only applies to the engines in `engines/`, those in `crypto/engine/`
3280 will always be built into libcrypto (i.e. "static").
3282 Building dynamic engines is enabled by default; to disable, use
3283 the configuration option "disable-dynamic-engine".
3285 The only requirements for building dynamic engines are the
3286 presence of the DSO module and building with position independent
3287 code, so they will also automatically be disabled if configuring
3288 with "disable-dso" or "disable-pic".
3290 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3291 are also taken away from openssl/opensslconf.h, as they are
3296 * Configuration change; if there is a known flag to compile
3297 position independent code, it will always be applied on the
3298 libcrypto and libssl object files, and never on the application
3299 object files. This means other libraries that use routines from
3300 libcrypto / libssl can be made into shared libraries regardless
3301 of how OpenSSL was configured.
3303 If this isn't desirable, the configuration options "disable-pic"
3304 or "no-pic" can be used to disable the use of PIC. This will
3305 also disable building shared libraries and dynamic engines.
3309 * Removed JPAKE code. It was experimental and has no wide use.
3313 * The INSTALL_PREFIX Makefile variable has been renamed to
3314 DESTDIR. That makes for less confusion on what this variable
3315 is for. Also, the configuration option --install_prefix is
3320 * Heartbeat for TLS has been removed and is disabled by default
3321 for DTLS; configure with enable-heartbeats. Code that uses the
3322 old #define's might need to be updated.
3324 *Emilia Käsper, Rich Salz*
3326 * Rename REF_CHECK to REF_DEBUG.
3330 * New "unified" build system
3332 The "unified" build system is aimed to be a common system for all
3333 platforms we support. With it comes new support for VMS.
3335 This system builds supports building in a different directory tree
3336 than the source tree. It produces one Makefile (for unix family
3337 or lookalikes), or one descrip.mms (for VMS).
3339 The source of information to make the Makefile / descrip.mms is
3340 small files called 'build.info', holding the necessary
3341 information for each directory with source to compile, and a
3342 template in Configurations, like unix-Makefile.tmpl or
3345 With this change, the library names were also renamed on Windows
3346 and on VMS. They now have names that are closer to the standard
3347 on Unix, and include the major version number, and in certain
3348 cases, the architecture they are built for. See "Notes on shared
3349 libraries" in INSTALL.
3351 We rely heavily on the perl module Text::Template.
3355 * Added support for auto-initialisation and de-initialisation of the library.
3356 OpenSSL no longer requires explicit init or deinit routines to be called,
3357 except in certain circumstances. See the OPENSSL_init_crypto() and
3358 OPENSSL_init_ssl() man pages for further information.
3362 * The arguments to the DTLSv1_listen function have changed. Specifically the
3363 "peer" argument is now expected to be a BIO_ADDR object.
3365 * Rewrite of BIO networking library. The BIO library lacked consistent
3366 support of IPv6, and adding it required some more extensive
3367 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3368 which hold all types of addresses and chains of address information.
3369 It also introduces a new API, with functions like BIO_socket,
3370 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3371 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3372 have been adapted accordingly.
3376 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3381 * CRIME protection: disable compression by default, even if OpenSSL is
3382 compiled with zlib enabled. Applications can still enable compression
3383 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3384 using the SSL_CONF library to configure compression.
3388 * The signature of the session callback configured with
3389 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3390 was explicitly marked as `const unsigned char*` instead of
3395 * Always DPURIFY. Remove the use of uninitialized memory in the
3396 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3400 * Removed many obsolete configuration items, including
3401 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3402 MD2_CHAR, MD2_INT, MD2_LONG
3404 IDEA_SHORT, IDEA_LONG
3405 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3407 *Rich Salz, with advice from Andy Polyakov*
3409 * Many BN internals have been moved to an internal header file.
3411 *Rich Salz with help from Andy Polyakov*
3413 * Configuration and writing out the results from it has changed.
3414 Files such as Makefile include/openssl/opensslconf.h and are now
3415 produced through general templates, such as Makefile.in and
3416 crypto/opensslconf.h.in and some help from the perl module
3419 Also, the center of configuration information is no longer
3420 Makefile. Instead, Configure produces a perl module in
3421 configdata.pm which holds most of the config data (in the hash
3422 table %config), the target data that comes from the target
3423 configuration in one of the `Configurations/*.conf` files (in
3428 * To clarify their intended purposes, the Configure options
3429 --prefix and --openssldir change their semantics, and become more
3430 straightforward and less interdependent.
3432 --prefix shall be used exclusively to give the location INSTALLTOP
3433 where programs, scripts, libraries, include files and manuals are
3434 going to be installed. The default is now /usr/local.
3436 --openssldir shall be used exclusively to give the default
3437 location OPENSSLDIR where certificates, private keys, CRLs are
3438 managed. This is also where the default openssl.cnf gets
3440 If the directory given with this option is a relative path, the
3441 values of both the --prefix value and the --openssldir value will
3442 be combined to become OPENSSLDIR.
3443 The default for --openssldir is INSTALLTOP/ssl.
3445 Anyone who uses --openssldir to specify where OpenSSL is to be
3446 installed MUST change to use --prefix instead.
3450 * The GOST engine was out of date and therefore it has been removed. An up
3451 to date GOST engine is now being maintained in an external repository.
3452 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3453 support for GOST ciphersuites (these are only activated if a GOST engine
3458 * EGD is no longer supported by default; use enable-egd when
3461 *Ben Kaduk and Rich Salz*
3463 * The distribution now has Makefile.in files, which are used to
3464 create Makefile's when Configure is run. *Configure must be run
3465 before trying to build now.*
3469 * The return value for SSL_CIPHER_description() for error conditions
3474 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3476 Obtaining and performing DNSSEC validation of TLSA records is
3477 the application's responsibility. The application provides
3478 the TLSA records of its choice to OpenSSL, and these are then
3479 used to authenticate the peer.
3481 The TLSA records need not even come from DNS. They can, for
3482 example, be used to implement local end-entity certificate or
3483 trust-anchor "pinning", where the "pin" data takes the form
3484 of TLSA records, which can augment or replace verification
3485 based on the usual WebPKI public certification authorities.
3489 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3490 continues to support deprecated interfaces in default builds.
3491 However, applications are strongly advised to compile their
3492 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3493 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3494 or the 1.1.0 releases.
3496 In environments in which all applications have been ported to
3497 not use any deprecated interfaces OpenSSL's Configure script
3498 should be used with the --api=1.1.0 option to entirely remove
3499 support for the deprecated features from the library and
3500 unconditionally disable them in the installed headers.
3501 Essentially the same effect can be achieved with the "no-deprecated"
3502 argument to Configure, except that this will always restrict
3503 the build to just the latest API, rather than a fixed API
3506 As applications are ported to future revisions of the API,
3507 they should update their compile-time OPENSSL_API_COMPAT define
3508 accordingly, but in most cases should be able to continue to
3509 compile with later releases.
3511 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3512 0x10000000L and 0x00908000L, respectively. However those
3513 versions did not support the OPENSSL_API_COMPAT feature, and
3514 so applications are not typically tested for explicit support
3515 of just the undeprecated features of either release.
3519 * Add support for setting the minimum and maximum supported protocol.
3520 It can bet set via the SSL_set_min_proto_version() and
3521 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3522 MaxProtocol. It's recommended to use the new APIs to disable
3523 protocols instead of disabling individual protocols using
3524 SSL_set_options() or SSL_CONF's Protocol. This change also
3525 removes support for disabling TLS 1.2 in the OpenSSL TLS
3526 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3530 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3534 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3535 and integrates ECDSA and ECDH functionality into EC. Implementations can
3536 now redirect key generation and no longer need to convert to or from
3539 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3540 include the ec.h header file instead.
3544 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3545 ciphers who are no longer supported and drops support the ephemeral RSA key
3546 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3550 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3551 opaque. For HMAC_CTX, the following constructors and destructors
3554 HMAC_CTX *HMAC_CTX_new(void);
3555 void HMAC_CTX_free(HMAC_CTX *ctx);
3557 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3558 destroy such methods has been added. See EVP_MD_meth_new(3) and
3559 EVP_CIPHER_meth_new(3) for documentation.
3562 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3563 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3564 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3565 an already created structure.
3566 2) For consistency with the majority of our object creators and
3567 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3568 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3569 for deprecated builds.
3573 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3574 cryptographic operations to be performed asynchronously as long as an
3575 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3576 further details. Libssl has also had this capability integrated with the
3577 introduction of the new mode SSL_MODE_ASYNC and associated error
3578 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3579 pages. This work was developed in partnership with Intel Corp.
3583 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3584 always enabled now. If you want to disable the support you should
3585 exclude it using the list of supported ciphers. This also means that the
3586 "-no_ecdhe" option has been removed from s_server.
3590 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3591 SSL_{CTX_}set1_curves() which can set a list.
3595 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3596 curve you want to support using SSL_{CTX_}set1_curves().
3600 * State machine rewrite. The state machine code has been significantly
3601 refactored in order to remove much duplication of code and solve issues
3602 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3603 further details). This change does have some associated API changes.
3604 Notably the SSL_state() function has been removed and replaced by
3605 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3606 SSL_set_state() has been removed altogether. The previous handshake states
3607 defined in ssl.h and ssl3.h have also been removed.
3611 * All instances of the string "ssleay" in the public API were replaced
3612 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3613 Some error codes related to internal RSA_eay API's were renamed.
3617 * The demo files in crypto/threads were moved to demo/threads.
3621 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3624 *Matt Caswell, Rich Salz*
3626 * New ASN.1 embed macro.
3628 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3629 structure is not allocated: it is part of the parent. That is instead of
3637 This reduces memory fragmentation and make it impossible to accidentally
3638 set a mandatory field to NULL.
3640 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3641 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3642 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3647 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3651 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3652 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3653 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3654 DES and RC4 ciphersuites.
3658 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3659 This changes the decoding behaviour for some invalid messages,
3660 though the change is mostly in the more lenient direction, and
3661 legacy behaviour is preserved as much as possible.
3665 * Fix no-stdio build.
3666 *David Woodhouse <David.Woodhouse@intel.com> and also*
3667 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3669 * New testing framework
3670 The testing framework has been largely rewritten and is now using
3671 perl and the perl modules Test::Harness and an extended variant of
3672 Test::More called OpenSSL::Test to do its work. All test scripts in
3673 test/ have been rewritten into test recipes, and all direct calls to
3674 executables in test/Makefile have become individual recipes using the
3675 simplified testing OpenSSL::Test::Simple.
3677 For documentation on our testing modules, do:
3679 perldoc test/testlib/OpenSSL/Test/Simple.pm
3680 perldoc test/testlib/OpenSSL/Test.pm
3684 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3685 are used; the latter aborts on memory leaks (usually checked on exit).
3686 Some undocumented "set malloc, etc., hooks" functions were removed
3687 and others were changed. All are now documented.
3691 * In DSA_generate_parameters_ex, if the provided seed is too short,
3694 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3696 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3697 from RFC4279, RFC4785, RFC5487, RFC5489.
3699 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3700 original RSA_PSK patch.
3704 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3705 era flag was never set throughout the codebase (only read). Also removed
3706 SSL3_FLAGS_POP_BUFFER which was only used if
3707 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3711 * Changed the default name options in the "ca", "crl", "req" and "x509"
3712 to be "oneline" instead of "compat".
3716 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3717 not aware of clients that still exhibit this bug, and the workaround
3718 hasn't been working properly for a while.
3722 * The return type of BIO_number_read() and BIO_number_written() as well as
3723 the corresponding num_read and num_write members in the BIO structure has
3724 changed from unsigned long to uint64_t. On platforms where an unsigned
3725 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3730 * Given the pervasive nature of TLS extensions it is inadvisable to run
3731 OpenSSL without support for them. It also means that maintaining
3732 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3733 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3737 * Removed support for the two export grade static DH ciphersuites
3738 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3739 were newly added (along with a number of other static DH ciphersuites) to
3740 1.0.2. However the two export ones have *never* worked since they were
3741 introduced. It seems strange in any case to be adding new export
3742 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3746 * Version negotiation has been rewritten. In particular SSLv23_method(),
3747 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3748 and turned into macros which simply call the new preferred function names
3749 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3750 should use the new names instead. Also as part of this change the ssl23.h
3751 header file has been removed.
3755 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3756 code and the associated standard is no longer considered fit-for-purpose.
3760 * RT2547 was closed. When generating a private key, try to make the
3761 output file readable only by the owner. This behavior change might
3762 be noticeable when interacting with other software.
3764 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3769 * Added HTTP GET support to the ocsp command.
3773 * Changed default digest for the dgst and enc commands from MD5 to
3778 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3782 * Added support for TLS extended master secret from
3783 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3784 initial patch which was a great help during development.
3788 * All libssl internal structures have been removed from the public header
3789 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3790 now redundant). Users should not attempt to access internal structures
3791 directly. Instead they should use the provided API functions.
3795 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3796 Access to deprecated functions can be re-enabled by running config with
3797 "enable-deprecated". In addition applications wishing to use deprecated
3798 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3799 will, by default, disable some transitive includes that previously existed
3800 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3804 * Added support for OCB mode. OpenSSL has been granted a patent license
3805 compatible with the OpenSSL license for use of OCB. Details are available
3806 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3807 for OCB can be removed by calling config with no-ocb.
3811 * SSLv2 support has been removed. It still supports receiving a SSLv2
3812 compatible client hello.
3816 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3817 done while fixing the error code for the key-too-small case.
3819 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3821 * CA.sh has been removed; use CA.pl instead.
3825 * Removed old DES API.
3829 * Remove various unsupported platforms:
3835 Sinix/ReliantUNIX RM400
3840 16-bit platforms such as WIN16
3844 * Clean up OPENSSL_NO_xxx #define's
3845 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3846 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3847 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3848 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3849 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3850 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3851 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3852 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3853 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3854 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3858 * Cleaned up dead code
3859 Remove all but one '#ifdef undef' which is to be looked at.
3863 * Clean up calling of xxx_free routines.
3864 Just like free(), fix most of the xxx_free routines to accept
3865 NULL. Remove the non-null checks from callers. Save much code.
3869 * Add secure heap for storage of private keys (when possible).
3870 Add BIO_s_secmem(), CBIGNUM, etc.
3871 Contributed by Akamai Technologies under our Corporate CLA.
3875 * Experimental support for a new, fast, unbiased prime candidate generator,
3876 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3878 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3880 * New output format NSS in the sess_id command line tool. This allows
3881 exporting the session id and the master key in NSS keylog format.
3883 *Martin Kaiser <martin@kaiser.cx>*
3885 * Harmonize version and its documentation. -f flag is used to display
3888 *mancha <mancha1@zoho.com>*
3890 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3891 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3893 *mancha <mancha1@zoho.com>*
3895 * Fix some double frees. These are not thought to be exploitable.
3897 *mancha <mancha1@zoho.com>*
3899 * A missing bounds check in the handling of the TLS heartbeat extension
3900 can be used to reveal up to 64k of memory to a connected client or
3903 Thanks for Neel Mehta of Google Security for discovering this bug and to
3904 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3905 preparing the fix ([CVE-2014-0160])
3907 *Adam Langley, Bodo Moeller*
3909 * Fix for the attack described in the paper "Recovering OpenSSL
3910 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3911 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3912 <http://eprint.iacr.org/2014/140>
3914 Thanks to Yuval Yarom and Naomi Benger for discovering this
3915 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3917 *Yuval Yarom and Naomi Benger*
3919 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3920 this fixes a limitation in previous versions of OpenSSL.
3924 * Experimental encrypt-then-mac support.
3926 Experimental support for encrypt then mac from
3927 draft-gutmann-tls-encrypt-then-mac-02.txt
3929 To enable it set the appropriate extension number (0x42 for the test
3930 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3932 For non-compliant peers (i.e. just about everything) this should have no
3935 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3939 * Add EVP support for key wrapping algorithms, to avoid problems with
3940 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3941 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3942 algorithms and include tests cases.
3946 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3951 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3952 MGF1 digest and OAEP label.
3956 * Make openssl verify return errors.
3958 *Chris Palmer <palmer@google.com> and Ben Laurie*
3960 * New function ASN1_TIME_diff to calculate the difference between two
3961 ASN1_TIME structures or one structure and the current time.
3965 * Update fips_test_suite to support multiple command line options. New
3966 test to induce all self test errors in sequence and check expected
3971 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3972 sign or verify all in one operation.
3976 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3977 test programs and fips_test_suite. Includes functionality to parse
3978 the minimal script output of fipsalgest.pl directly.
3982 * Add authorisation parameter to FIPS_module_mode_set().
3986 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3990 * Use separate DRBG fields for internal and external flags. New function
3991 FIPS_drbg_health_check() to perform on demand health checking. Add
3992 generation tests to fips_test_suite with reduced health check interval to
3993 demonstrate periodic health checking. Add "nodh" option to
3994 fips_test_suite to skip very slow DH test.
3998 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4003 * More extensive health check for DRBG checking many more failure modes.
4004 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4005 combination: call this in fips_test_suite.
4009 * Add support for canonical generation of DSA parameter 'g'. See
4012 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4013 POST to handle HMAC cases.
4017 * Add functions FIPS_module_version() and FIPS_module_version_text()
4018 to return numerical and string versions of the FIPS module number.
4022 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4023 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4024 outside the validated module in the FIPS capable OpenSSL.
4028 * Minor change to DRBG entropy callback semantics. In some cases
4029 there is no multiple of the block length between min_len and
4030 max_len. Allow the callback to return more than max_len bytes
4031 of entropy but discard any extra: it is the callback's responsibility
4032 to ensure that the extra data discarded does not impact the
4033 requested amount of entropy.
4037 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4038 information in FIPS186-3, SP800-57 and SP800-131A.
4042 * CCM support via EVP. Interface is very similar to GCM case except we
4043 must supply all data in one chunk (i.e. no update, final) and the
4044 message length must be supplied if AAD is used. Add algorithm test
4049 * Initial version of POST overhaul. Add POST callback to allow the status
4050 of POST to be monitored and/or failures induced. Modify fips_test_suite
4051 to use callback. Always run all selftests even if one fails.
4055 * XTS support including algorithm test driver in the fips_gcmtest program.
4056 Note: this does increase the maximum key length from 32 to 64 bytes but
4057 there should be no binary compatibility issues as existing applications
4058 will never use XTS mode.
4062 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4063 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4064 performs algorithm blocking for unapproved PRNG types. Also do not
4065 set PRNG type in FIPS_mode_set(): leave this to the application.
4066 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4067 the standard OpenSSL PRNG: set additional data to a date time vector.
4071 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4072 This shouldn't present any incompatibility problems because applications
4073 shouldn't be using these directly and any that are will need to rethink
4074 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4078 * Extensive self tests and health checking required by SP800-90 DRBG.
4079 Remove strength parameter from FIPS_drbg_instantiate and always
4080 instantiate at maximum supported strength.
4084 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4088 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4092 * New function DH_compute_key_padded() to compute a DH key and pad with
4093 leading zeroes if needed: this complies with SP800-56A et al.
4097 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4098 anything, incomplete, subject to change and largely untested at present.
4102 * Modify fipscanisteronly build option to only build the necessary object
4103 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4107 * Add experimental option FIPSSYMS to give all symbols in
4108 fipscanister.o and FIPS or fips prefix. This will avoid
4109 conflicts with future versions of OpenSSL. Add perl script
4110 util/fipsas.pl to preprocess assembly language source files
4111 and rename any affected symbols.
4115 * Add selftest checks and algorithm block of non-fips algorithms in
4116 FIPS mode. Remove DES2 from selftests.
4120 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4121 return internal method without any ENGINE dependencies. Add new
4122 tiny fips sign and verify functions.
4126 * New build option no-ec2m to disable characteristic 2 code.
4130 * New build option "fipscanisteronly". This only builds fipscanister.o
4131 and (currently) associated fips utilities. Uses the file Makefile.fips
4132 instead of Makefile.org as the prototype.
4136 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4137 Update fips_gcmtest to use IV generator.
4141 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4142 setting output buffer to NULL. The `*Final` function must be
4143 called although it will not retrieve any additional data. The tag
4144 can be set or retrieved with a ctrl. The IV length is by default 12
4145 bytes (96 bits) but can be set to an alternative value. If the IV
4146 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4151 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4152 underlying do_cipher function handles all cipher semantics itself
4153 including padding and finalisation. This is useful if (for example)
4154 an ENGINE cipher handles block padding itself. The behaviour of
4155 do_cipher is subtly changed if this flag is set: the return value
4156 is the number of characters written to the output buffer (zero is
4157 no longer an error code) or a negative error code. Also if the
4158 input buffer is NULL and length 0 finalisation should be performed.
4162 * If a candidate issuer certificate is already part of the constructed
4163 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4167 * Improve forward-security support: add functions
4169 void SSL_CTX_set_not_resumable_session_callback(
4170 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4171 void SSL_set_not_resumable_session_callback(
4172 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4174 for use by SSL/TLS servers; the callback function will be called whenever a
4175 new session is created, and gets to decide whether the session may be
4176 cached to make it resumable (return 0) or not (return 1). (As by the
4177 SSL/TLS protocol specifications, the session_id sent by the server will be
4178 empty to indicate that the session is not resumable; also, the server will
4179 not generate RFC 4507 (RFC 5077) session tickets.)
4181 A simple reasonable callback implementation is to return is_forward_secure.
4182 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4183 by the SSL/TLS server library, indicating whether it can provide forward
4186 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4188 * New -verify_name option in command line utilities to set verification
4193 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4194 Add CMAC pkey methods.
4198 * Experimental renegotiation in s_server -www mode. If the client
4199 browses /reneg connection is renegotiated. If /renegcert it is
4200 renegotiated requesting a certificate.
4204 * Add an "external" session cache for debugging purposes to s_server. This
4205 should help trace issues which normally are only apparent in deployed
4206 multi-process servers.
4210 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4211 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4212 BIO_set_cipher() and some obscure PEM functions were changed so they
4213 can now return an error. The RAND changes required a change to the
4214 RAND_METHOD structure.
4218 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4219 a gcc attribute to warn if the result of a function is ignored. This
4220 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4221 whose return value is often ignored.
4225 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4226 These allow SCTs (signed certificate timestamps) to be requested and
4227 validated when establishing a connection.
4229 *Rob Percival <robpercival@google.com>*
4234 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4236 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4237 used even when parsing explicit parameters, when loading a encoded key
4238 or calling `EC_GROUP_new_from_ecpkparameters()`/
4239 `EC_GROUP_new_from_ecparameters()`.
4240 This prevents bypass of security hardening and performance gains,
4241 especially for curves with specialized EC_METHODs.
4242 By default, if a key encoded with explicit parameters is loaded and later
4243 encoded, the output is still encoded with explicit parameters, even if
4244 internally a "named" EC_GROUP is used for computation.
4248 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4249 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4250 NULL. After this change, only the cofactor parameter can be NULL. It also
4251 does some minimal sanity checks on the passed order.
4256 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4257 An attack is simple, if the first CMS_recipientInfo is valid but the
4258 second CMS_recipientInfo is chosen ciphertext. If the second
4259 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4260 encryption key will be replaced by garbage, and the message cannot be
4261 decoded, but if the RSA decryption fails, the correct encryption key is
4262 used and the recipient will not notice the attack.
4263 As a work around for this potential attack the length of the decrypted
4264 key must be equal to the cipher default key length, in case the
4265 certifiate is not given and all recipientInfo are tried out.
4266 The old behaviour can be re-enabled in the CMS code by setting the
4267 CMS_DEBUG_DECRYPT flag.
4272 * Document issue with installation paths in diverse Windows builds
4274 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4275 binaries and run-time config file.
4280 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4282 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4283 This changes the size when using the `genpkey` command when no size is given.
4284 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4285 generation commands to use 2048 bits by default.
4289 * Add FIPS support for Android Arm 64-bit
4291 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4292 Module in Version 2.0.10. For some reason, the corresponding target
4293 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4294 built with FIPS support on Android Arm 64-bit. This omission has been
4297 *Matthias St. Pierre*
4299 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4301 * 0-byte record padding oracle
4303 If an application encounters a fatal protocol error and then calls
4304 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4305 then OpenSSL can respond differently to the calling application if a 0 byte
4306 record is received with invalid padding compared to if a 0 byte record is
4307 received with an invalid MAC. If the application then behaves differently
4308 based on that in a way that is detectable to the remote peer, then this
4309 amounts to a padding oracle that could be used to decrypt data.
4311 In order for this to be exploitable "non-stitched" ciphersuites must be in
4312 use. Stitched ciphersuites are optimised implementations of certain
4313 commonly used ciphersuites. Also the application must call SSL_shutdown()
4314 twice even if a protocol error has occurred (applications should not do
4315 this but some do anyway).
4317 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4318 Aviram, with additional investigation by Steven Collison and Andrew
4319 Hourselt. It was reported to OpenSSL on 10th December 2018.
4324 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4328 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4330 * Microarchitecture timing vulnerability in ECC scalar multiplication
4332 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4333 shown to be vulnerable to a microarchitecture timing side channel attack.
4334 An attacker with sufficient access to mount local timing attacks during
4335 ECDSA signature generation could recover the private key.
4337 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4338 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4344 * Timing vulnerability in DSA signature generation
4346 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4347 timing side channel attack. An attacker could use variations in the signing
4348 algorithm to recover the private key.
4350 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4355 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4356 Module, accidentally introduced while backporting security fixes from the
4357 development branch and hindering the use of ECC in FIPS mode.
4361 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4363 * Client DoS due to large DH parameter
4365 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4366 malicious server can send a very large prime value to the client. This will
4367 cause the client to spend an unreasonably long period of time generating a
4368 key for this prime resulting in a hang until the client has finished. This
4369 could be exploited in a Denial Of Service attack.
4371 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4376 * Cache timing vulnerability in RSA Key Generation
4378 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4379 a cache timing side channel attack. An attacker with sufficient access to
4380 mount cache timing attacks during the RSA key generation process could
4381 recover the private key.
4383 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4384 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4389 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4390 parameter is no longer accepted, as it leads to a corrupt table. NULL
4391 pem_str is reserved for alias entries only.
4395 * Revert blinding in ECDSA sign and instead make problematic addition
4396 length-invariant. Switch even to fixed-length Montgomery multiplication.
4400 * Change generating and checking of primes so that the error rate of not
4401 being prime depends on the intended use based on the size of the input.
4402 For larger primes this will result in more rounds of Miller-Rabin.
4403 The maximal error rate for primes with more than 1080 bits is lowered
4406 *Kurt Roeckx, Annie Yousar*
4408 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4412 * Add blinding to ECDSA and DSA signatures to protect against side channel
4413 attacks discovered by Keegan Ryan (NCC Group).
4417 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4418 now allow empty (zero character) pass phrases.
4422 * Certificate time validation (X509_cmp_time) enforces stricter
4423 compliance with RFC 5280. Fractional seconds and timezone offsets
4424 are no longer allowed.
4428 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4430 * Constructed ASN.1 types with a recursive definition could exceed the stack
4432 Constructed ASN.1 types with a recursive definition (such as can be found
4433 in PKCS7) could eventually exceed the stack given malicious input with
4434 excessive recursion. This could result in a Denial Of Service attack. There
4435 are no such structures used within SSL/TLS that come from untrusted sources
4436 so this is considered safe.
4438 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4444 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4446 * Read/write after SSL object in error state
4448 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4449 mechanism. The intent was that if a fatal error occurred during a handshake
4450 then OpenSSL would move into the error state and would immediately fail if
4451 you attempted to continue the handshake. This works as designed for the
4452 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4453 SSL_connect()), however due to a bug it does not work correctly if
4454 SSL_read() or SSL_write() is called directly. In that scenario, if the
4455 handshake fails then a fatal error will be returned in the initial function
4456 call. If SSL_read()/SSL_write() is subsequently called by the application
4457 for the same SSL object then it will succeed and the data is passed without
4458 being decrypted/encrypted directly from the SSL/TLS record layer.
4460 In order to exploit this issue an application bug would have to be present
4461 that resulted in a call to SSL_read()/SSL_write() being issued after having
4462 already received a fatal error.
4464 This issue was reported to OpenSSL by David Benjamin (Google).
4469 * rsaz_1024_mul_avx2 overflow bug on x86_64
4471 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4472 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4473 Analysis suggests that attacks against RSA and DSA as a result of this
4474 defect would be very difficult to perform and are not believed likely.
4475 Attacks against DH1024 are considered just feasible, because most of the
4476 work necessary to deduce information about a private key may be performed
4477 offline. The amount of resources required for such an attack would be
4478 significant. However, for an attack on TLS to be meaningful, the server
4479 would have to share the DH1024 private key among multiple clients, which is
4480 no longer an option since CVE-2016-0701.
4482 This only affects processors that support the AVX2 but not ADX extensions
4483 like Intel Haswell (4th generation).
4485 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4486 was originally found via the OSS-Fuzz project.
4491 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4493 * bn_sqrx8x_internal carry bug on x86_64
4495 There is a carry propagating bug in the x86_64 Montgomery squaring
4496 procedure. No EC algorithms are affected. Analysis suggests that attacks
4497 against RSA and DSA as a result of this defect would be very difficult to
4498 perform and are not believed likely. Attacks against DH are considered just
4499 feasible (although very difficult) because most of the work necessary to
4500 deduce information about a private key may be performed offline. The amount
4501 of resources required for such an attack would be very significant and
4502 likely only accessible to a limited number of attackers. An attacker would
4503 additionally need online access to an unpatched system using the target
4504 private key in a scenario with persistent DH parameters and a private
4505 key that is shared between multiple clients.
4507 This only affects processors that support the BMI1, BMI2 and ADX extensions
4508 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4510 This issue was reported to OpenSSL by the OSS-Fuzz project.
4515 * Malformed X.509 IPAddressFamily could cause OOB read
4517 If an X.509 certificate has a malformed IPAddressFamily extension,
4518 OpenSSL could do a one-byte buffer overread. The most likely result
4519 would be an erroneous display of the certificate in text format.
4521 This issue was reported to OpenSSL by the OSS-Fuzz project.
4525 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4527 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4528 platform rather than 'mingw'.
4532 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4534 * Truncated packet could crash via OOB read
4536 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4537 cipher is being used, then a truncated packet can cause that host to
4538 perform an out-of-bounds read, usually resulting in a crash.
4540 This issue was reported to OpenSSL by Robert Święcki of Google.
4545 * BN_mod_exp may produce incorrect results on x86_64
4547 There is a carry propagating bug in the x86_64 Montgomery squaring
4548 procedure. No EC algorithms are affected. Analysis suggests that attacks
4549 against RSA and DSA as a result of this defect would be very difficult to
4550 perform and are not believed likely. Attacks against DH are considered just
4551 feasible (although very difficult) because most of the work necessary to
4552 deduce information about a private key may be performed offline. The amount
4553 of resources required for such an attack would be very significant and
4554 likely only accessible to a limited number of attackers. An attacker would
4555 additionally need online access to an unpatched system using the target
4556 private key in a scenario with persistent DH parameters and a private
4557 key that is shared between multiple clients. For example this can occur by
4558 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4559 similar to CVE-2015-3193 but must be treated as a separate problem.
4561 This issue was reported to OpenSSL by the OSS-Fuzz project.
4566 * Montgomery multiplication may produce incorrect results
4568 There is a carry propagating bug in the Broadwell-specific Montgomery
4569 multiplication procedure that handles input lengths divisible by, but
4570 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4571 and DH private keys are impossible. This is because the subroutine in
4572 question is not used in operations with the private key itself and an input
4573 of the attacker's direct choice. Otherwise the bug can manifest itself as
4574 transient authentication and key negotiation failures or reproducible
4575 erroneous outcome of public-key operations with specially crafted input.
4576 Among EC algorithms only Brainpool P-512 curves are affected and one
4577 presumably can attack ECDH key negotiation. Impact was not analyzed in
4578 detail, because pre-requisites for attack are considered unlikely. Namely
4579 multiple clients have to choose the curve in question and the server has to
4580 share the private key among them, neither of which is default behaviour.
4581 Even then only clients that chose the curve will be affected.
4583 This issue was publicly reported as transient failures and was not
4584 initially recognized as a security issue. Thanks to Richard Morgan for
4585 providing reproducible case.
4590 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4591 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4592 prevent issues where no progress is being made and the peer continually
4593 sends unrecognised record types, using up resources processing them.
4597 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4599 * Missing CRL sanity check
4601 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4602 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4603 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4605 This issue only affects the OpenSSL 1.0.2i
4610 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4612 * OCSP Status Request extension unbounded memory growth
4614 A malicious client can send an excessively large OCSP Status Request
4615 extension. If that client continually requests renegotiation, sending a
4616 large OCSP Status Request extension each time, then there will be unbounded
4617 memory growth on the server. This will eventually lead to a Denial Of
4618 Service attack through memory exhaustion. Servers with a default
4619 configuration are vulnerable even if they do not support OCSP. Builds using
4620 the "no-ocsp" build time option are not affected.
4622 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4627 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4630 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4636 * OOB write in MDC2_Update()
4638 An overflow can occur in MDC2_Update() either if called directly or
4639 through the EVP_DigestUpdate() function using MDC2. If an attacker
4640 is able to supply very large amounts of input data after a previous
4641 call to EVP_EncryptUpdate() with a partial block then a length check
4642 can overflow resulting in a heap corruption.
4644 The amount of data needed is comparable to SIZE_MAX which is impractical
4647 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4652 * Malformed SHA512 ticket DoS
4654 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4655 DoS attack where a malformed ticket will result in an OOB read which will
4658 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4659 a custom server callback and ticket lookup mechanism.
4661 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4666 * OOB write in BN_bn2dec()
4668 The function BN_bn2dec() does not check the return value of BN_div_word().
4669 This can cause an OOB write if an application uses this function with an
4670 overly large BIGNUM. This could be a problem if an overly large certificate
4671 or CRL is printed out from an untrusted source. TLS is not affected because
4672 record limits will reject an oversized certificate before it is parsed.
4674 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4679 * OOB read in TS_OBJ_print_bio()
4681 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4682 the total length the OID text representation would use and not the amount
4683 of data written. This will result in OOB reads when large OIDs are
4686 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4691 * Pointer arithmetic undefined behaviour
4693 Avoid some undefined pointer arithmetic
4695 A common idiom in the codebase is to check limits in the following manner:
4698 Where "p" points to some malloc'd data of SIZE bytes and
4701 "len" here could be from some externally supplied data (e.g. from a TLS
4704 The rules of C pointer arithmetic are such that "p + len" is only well
4705 defined where len <= SIZE. Therefore the above idiom is actually
4706 undefined behaviour.
4708 For example this could cause problems if some malloc implementation
4709 provides an address for "p" such that "p + len" actually overflows for
4710 values of len that are too big and therefore p + len < limit.
4712 This issue was reported to OpenSSL by Guido Vranken
4717 * Constant time flag not preserved in DSA signing
4719 Operations in the DSA signing algorithm should run in constant time in
4720 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4721 implementation means that a non-constant time codepath is followed for
4722 certain operations. This has been demonstrated through a cache-timing
4723 attack to be sufficient for an attacker to recover the private DSA key.
4725 This issue was reported by César Pereida (Aalto University), Billy Brumley
4726 (Tampere University of Technology), and Yuval Yarom (The University of
4727 Adelaide and NICTA).
4732 * DTLS buffered message DoS
4734 In a DTLS connection where handshake messages are delivered out-of-order
4735 those messages that OpenSSL is not yet ready to process will be buffered
4736 for later use. Under certain circumstances, a flaw in the logic means that
4737 those messages do not get removed from the buffer even though the handshake
4738 has been completed. An attacker could force up to approx. 15 messages to
4739 remain in the buffer when they are no longer required. These messages will
4740 be cleared when the DTLS connection is closed. The default maximum size for
4741 a message is 100k. Therefore the attacker could force an additional 1500k
4742 to be consumed per connection. By opening many simulataneous connections an
4743 attacker could cause a DoS attack through memory exhaustion.
4745 This issue was reported to OpenSSL by Quan Luo.
4750 * DTLS replay protection DoS
4752 A flaw in the DTLS replay attack protection mechanism means that records
4753 that arrive for future epochs update the replay protection "window" before
4754 the MAC for the record has been validated. This could be exploited by an
4755 attacker by sending a record for the next epoch (which does not have to
4756 decrypt or have a valid MAC), with a very large sequence number. This means
4757 that all subsequent legitimate packets are dropped causing a denial of
4758 service for a specific DTLS connection.
4760 This issue was reported to OpenSSL by the OCAP audit team.
4765 * Certificate message OOB reads
4767 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4768 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4769 theoretical DoS risk but this has not been observed in practice on common
4772 The messages affected are client certificate, client certificate request
4773 and server certificate. As a result the attack can only be performed
4774 against a client or a server which enables client authentication.
4776 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4781 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4783 * Prevent padding oracle in AES-NI CBC MAC check
4785 A MITM attacker can use a padding oracle attack to decrypt traffic
4786 when the connection uses an AES CBC cipher and the server support
4789 This issue was introduced as part of the fix for Lucky 13 padding
4790 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4791 constant time by making sure that always the same bytes are read and
4792 compared against either the MAC or padding bytes. But it no longer
4793 checked that there was enough data to have both the MAC and padding
4796 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4800 * Fix EVP_EncodeUpdate overflow
4802 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4803 Base64 encoding of binary data. If an attacker is able to supply very large
4804 amounts of input data then a length check can overflow resulting in a heap
4807 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4808 the `PEM_write_bio*` family of functions. These are mainly used within the
4809 OpenSSL command line applications, so any application which processes data
4810 from an untrusted source and outputs it as a PEM file should be considered
4811 vulnerable to this issue. User applications that call these APIs directly
4812 with large amounts of untrusted data may also be vulnerable.
4814 This issue was reported by Guido Vranken.
4819 * Fix EVP_EncryptUpdate overflow
4821 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4822 is able to supply very large amounts of input data after a previous call to
4823 EVP_EncryptUpdate() with a partial block then a length check can overflow
4824 resulting in a heap corruption. Following an analysis of all OpenSSL
4825 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4826 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4827 the first called function after an EVP_EncryptInit(), and therefore that
4828 specific call must be safe. The second form is where the length passed to
4829 EVP_EncryptUpdate() can be seen from the code to be some small value and
4830 therefore there is no possibility of an overflow. Since all instances are
4831 one of these two forms, it is believed that there can be no overflows in
4832 internal code due to this problem. It should be noted that
4833 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4834 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4835 of these calls have also been analysed too and it is believed there are no
4836 instances in internal usage where an overflow could occur.
4838 This issue was reported by Guido Vranken.
4843 * Prevent ASN.1 BIO excessive memory allocation
4845 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4846 a short invalid encoding can cause allocation of large amounts of memory
4847 potentially consuming excessive resources or exhausting memory.
4849 Any application parsing untrusted data through d2i BIO functions is
4850 affected. The memory based functions such as d2i_X509() are *not* affected.
4851 Since the memory based functions are used by the TLS library, TLS
4852 applications are not affected.
4854 This issue was reported by Brian Carpenter.
4861 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4862 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4863 in arbitrary stack data being returned in the buffer.
4865 This issue was reported by Guido Vranken.
4870 * Modify behavior of ALPN to invoke callback after SNI/servername
4871 callback, such that updates to the SSL_CTX affect ALPN.
4875 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4880 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4881 methods are enabled and ssl2 is disabled the methods return NULL.
4885 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4887 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4888 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4889 provide any "EXPORT" or "LOW" strength ciphers.
4893 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4894 is by default disabled at build-time. Builds that are not configured with
4895 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4896 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4897 will need to explicitly call either of:
4899 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4901 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4903 as appropriate. Even if either of those is used, or the application
4904 explicitly uses the version-specific SSLv2_method() or its client and
4905 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4906 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4907 ciphers, and SSLv2 56-bit DES are no longer available.
4912 * Fix a double-free in DSA code
4914 A double free bug was discovered when OpenSSL parses malformed DSA private
4915 keys and could lead to a DoS attack or memory corruption for applications
4916 that receive DSA private keys from untrusted sources. This scenario is
4919 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4925 * Disable SRP fake user seed to address a server memory leak.
4927 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4929 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4930 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4931 was changed to ignore the "fake user" SRP seed, even if the seed
4934 Users should use SRP_VBASE_get1_by_user instead. Note that in
4935 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4936 also that even though configuring the SRP seed attempts to hide
4937 invalid usernames by continuing the handshake with fake
4938 credentials, this behaviour is not constant time and no strong
4939 guarantees are made that the handshake is indistinguishable from
4940 that of a valid user.
4945 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4947 In the BN_hex2bn function the number of hex digits is calculated using an
4948 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4949 large values of `i` this can result in `bn_expand` not allocating any
4950 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4951 field as NULL leading to a subsequent NULL ptr deref. For very large values
4952 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4953 In this case memory is allocated to the internal BIGNUM data field, but it
4954 is insufficiently sized leading to heap corruption. A similar issue exists
4955 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4956 is ever called by user applications with very large untrusted hex/dec data.
4957 This is anticipated to be a rare occurrence.
4959 All OpenSSL internal usage of these functions use data that is not expected
4960 to be untrusted, e.g. config file data or application command line
4961 arguments. If user developed applications generate config file data based
4962 on untrusted data then it is possible that this could also lead to security
4963 consequences. This is also anticipated to be rare.
4965 This issue was reported to OpenSSL by Guido Vranken.
4970 * Fix memory issues in `BIO_*printf` functions
4972 The internal `fmtstr` function used in processing a "%s" format string in
4973 the `BIO_*printf` functions could overflow while calculating the length of a
4974 string and cause an OOB read when printing very long strings.
4976 Additionally the internal `doapr_outch` function can attempt to write to an
4977 OOB memory location (at an offset from the NULL pointer) in the event of a
4978 memory allocation failure. In 1.0.2 and below this could be caused where
4979 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4980 could be in processing a very long "%s" format string. Memory leaks can
4983 The first issue may mask the second issue dependent on compiler behaviour.
4984 These problems could enable attacks where large amounts of untrusted data
4985 is passed to the `BIO_*printf` functions. If applications use these functions
4986 in this way then they could be vulnerable. OpenSSL itself uses these
4987 functions when printing out human-readable dumps of ASN.1 data. Therefore
4988 applications that print this data could be vulnerable if the data is from
4989 untrusted sources. OpenSSL command line applications could also be
4990 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4991 as command line arguments.
4993 Libssl is not considered directly vulnerable. Additionally certificates etc
4994 received via remote connections via libssl are also unlikely to be able to
4995 trigger these issues because of message size limits enforced within libssl.
4997 This issue was reported to OpenSSL Guido Vranken.
5002 * Side channel attack on modular exponentiation
5004 A side-channel attack was found which makes use of cache-bank conflicts on
5005 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5006 of RSA keys. The ability to exploit this issue is limited as it relies on
5007 an attacker who has control of code in a thread running on the same
5008 hyper-threaded core as the victim thread which is performing decryptions.
5010 This issue was reported to OpenSSL by Yuval Yarom, The University of
5011 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5012 Nadia Heninger, University of Pennsylvania with more information at
5013 <http://cachebleed.info>.
5018 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5019 if no keysize is specified with default_bits. This fixes an
5020 omission in an earlier change that changed all RSA/DSA key generation
5021 commands to use 2048 bits by default.
5025 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5027 * DH small subgroups
5029 Historically OpenSSL only ever generated DH parameters based on "safe"
5030 primes. More recently (in version 1.0.2) support was provided for
5031 generating X9.42 style parameter files such as those required for RFC 5114
5032 support. The primes used in such files may not be "safe". Where an
5033 application is using DH configured with parameters based on primes that are
5034 not "safe" then an attacker could use this fact to find a peer's private
5035 DH exponent. This attack requires that the attacker complete multiple
5036 handshakes in which the peer uses the same private DH exponent. For example
5037 this could be used to discover a TLS server's private DH exponent if it's
5038 reusing the private DH exponent or it's using a static DH ciphersuite.
5040 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5041 TLS. It is not on by default. If the option is not set then the server
5042 reuses the same private DH exponent for the life of the server process and
5043 would be vulnerable to this attack. It is believed that many popular
5044 applications do set this option and would therefore not be at risk.
5046 The fix for this issue adds an additional check where a "q" parameter is
5047 available (as is the case in X9.42 based parameters). This detects the
5048 only known attack, and is the only possible defense for static DH
5049 ciphersuites. This could have some performance impact.
5051 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5052 default and cannot be disabled. This could have some performance impact.
5054 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5059 * SSLv2 doesn't block disabled ciphers
5061 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5062 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5063 been disabled, provided that the SSLv2 protocol was not also disabled via
5066 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5067 and Sebastian Schinzel.
5072 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5074 * BN_mod_exp may produce incorrect results on x86_64
5076 There is a carry propagating bug in the x86_64 Montgomery squaring
5077 procedure. No EC algorithms are affected. Analysis suggests that attacks
5078 against RSA and DSA as a result of this defect would be very difficult to
5079 perform and are not believed likely. Attacks against DH are considered just
5080 feasible (although very difficult) because most of the work necessary to
5081 deduce information about a private key may be performed offline. The amount
5082 of resources required for such an attack would be very significant and
5083 likely only accessible to a limited number of attackers. An attacker would
5084 additionally need online access to an unpatched system using the target
5085 private key in a scenario with persistent DH parameters and a private
5086 key that is shared between multiple clients. For example this can occur by
5087 default in OpenSSL DHE based SSL/TLS ciphersuites.
5089 This issue was reported to OpenSSL by Hanno Böck.
5094 * Certificate verify crash with missing PSS parameter
5096 The signature verification routines will crash with a NULL pointer
5097 dereference if presented with an ASN.1 signature using the RSA PSS
5098 algorithm and absent mask generation function parameter. Since these
5099 routines are used to verify certificate signature algorithms this can be
5100 used to crash any certificate verification operation and exploited in a
5101 DoS attack. Any application which performs certificate verification is
5102 vulnerable including OpenSSL clients and servers which enable client
5105 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5110 * X509_ATTRIBUTE memory leak
5112 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5113 memory. This structure is used by the PKCS#7 and CMS routines so any
5114 application which reads PKCS#7 or CMS data from untrusted sources is
5115 affected. SSL/TLS is not affected.
5117 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5123 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5124 This changes the decoding behaviour for some invalid messages,
5125 though the change is mostly in the more lenient direction, and
5126 legacy behaviour is preserved as much as possible.
5130 * In DSA_generate_parameters_ex, if the provided seed is too short,
5133 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5135 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5137 * Alternate chains certificate forgery
5139 During certificate verification, OpenSSL will attempt to find an
5140 alternative certificate chain if the first attempt to build such a chain
5141 fails. An error in the implementation of this logic can mean that an
5142 attacker could cause certain checks on untrusted certificates to be
5143 bypassed, such as the CA flag, enabling them to use a valid leaf
5144 certificate to act as a CA and "issue" an invalid certificate.
5146 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5151 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5153 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5154 incompatibility in the handling of HMAC. The previous ABI has now been
5159 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5161 * Malformed ECParameters causes infinite loop
5163 When processing an ECParameters structure OpenSSL enters an infinite loop
5164 if the curve specified is over a specially malformed binary polynomial
5167 This can be used to perform denial of service against any
5168 system which processes public keys, certificate requests or
5169 certificates. This includes TLS clients and TLS servers with
5170 client authentication enabled.
5172 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5177 * Exploitable out-of-bounds read in X509_cmp_time
5179 X509_cmp_time does not properly check the length of the ASN1_TIME
5180 string and can read a few bytes out of bounds. In addition,
5181 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5184 An attacker can use this to craft malformed certificates and CRLs of
5185 various sizes and potentially cause a segmentation fault, resulting in
5186 a DoS on applications that verify certificates or CRLs. TLS clients
5187 that verify CRLs are affected. TLS clients and servers with client
5188 authentication enabled may be affected if they use custom verification
5191 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5192 independently by Hanno Böck.
5197 * PKCS7 crash with missing EnvelopedContent
5199 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5200 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5201 with missing content and trigger a NULL pointer dereference on parsing.
5203 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5204 structures from untrusted sources are affected. OpenSSL clients and
5205 servers are not affected.
5207 This issue was reported to OpenSSL by Michal Zalewski (Google).
5212 * CMS verify infinite loop with unknown hash function
5214 When verifying a signedData message the CMS code can enter an infinite loop
5215 if presented with an unknown hash function OID. This can be used to perform
5216 denial of service against any system which verifies signedData messages using
5218 This issue was reported to OpenSSL by Johannes Bauer.
5223 * Race condition handling NewSessionTicket
5225 If a NewSessionTicket is received by a multi-threaded client when attempting to
5226 reuse a previous ticket then a race condition can occur potentially leading to
5227 a double free of the ticket data.
5232 * Only support 256-bit or stronger elliptic curves with the
5233 'ecdh_auto' setting (server) or by default (client). Of supported
5234 curves, prefer P-256 (both).
5238 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5240 * ClientHello sigalgs DoS fix
5242 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5243 invalid signature algorithms extension a NULL pointer dereference will
5244 occur. This can be exploited in a DoS attack against the server.
5246 This issue was was reported to OpenSSL by David Ramos of Stanford
5250 *Stephen Henson and Matt Caswell*
5252 * Multiblock corrupted pointer fix
5254 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5255 feature only applies on 64 bit x86 architecture platforms that support AES
5256 NI instructions. A defect in the implementation of "multiblock" can cause
5257 OpenSSL's internal write buffer to become incorrectly set to NULL when
5258 using non-blocking IO. Typically, when the user application is using a
5259 socket BIO for writing, this will only result in a failed connection.
5260 However if some other BIO is used then it is likely that a segmentation
5261 fault will be triggered, thus enabling a potential DoS attack.
5263 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5268 * Segmentation fault in DTLSv1_listen fix
5270 The DTLSv1_listen function is intended to be stateless and processes the
5271 initial ClientHello from many peers. It is common for user code to loop
5272 over the call to DTLSv1_listen until a valid ClientHello is received with
5273 an associated cookie. A defect in the implementation of DTLSv1_listen means
5274 that state is preserved in the SSL object from one invocation to the next
5275 that can lead to a segmentation fault. Errors processing the initial
5276 ClientHello can trigger this scenario. An example of such an error could be
5277 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5280 This issue was reported to OpenSSL by Per Allansson.
5285 * Segmentation fault in ASN1_TYPE_cmp fix
5287 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5288 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5289 certificate signature algorithm consistency this can be used to crash any
5290 certificate verification operation and exploited in a DoS attack. Any
5291 application which performs certificate verification is vulnerable including
5292 OpenSSL clients and servers which enable client authentication.
5297 * Segmentation fault for invalid PSS parameters fix
5299 The signature verification routines will crash with a NULL pointer
5300 dereference if presented with an ASN.1 signature using the RSA PSS
5301 algorithm and invalid parameters. Since these routines are used to verify
5302 certificate signature algorithms this can be used to crash any
5303 certificate verification operation and exploited in a DoS attack. Any
5304 application which performs certificate verification is vulnerable including
5305 OpenSSL clients and servers which enable client authentication.
5307 This issue was was reported to OpenSSL by Brian Carpenter.
5312 * ASN.1 structure reuse memory corruption fix
5314 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5315 memory corruption via an invalid write. Such reuse is and has been
5316 strongly discouraged and is believed to be rare.
5318 Applications that parse structures containing CHOICE or ANY DEFINED BY
5319 components may be affected. Certificate parsing (d2i_X509 and related
5320 functions) are however not affected. OpenSSL clients and servers are
5326 * PKCS7 NULL pointer dereferences fix
5328 The PKCS#7 parsing code does not handle missing outer ContentInfo
5329 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5330 missing content and trigger a NULL pointer dereference on parsing.
5332 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5333 otherwise parse PKCS#7 structures from untrusted sources are
5334 affected. OpenSSL clients and servers are not affected.
5336 This issue was reported to OpenSSL by Michal Zalewski (Google).
5341 * DoS via reachable assert in SSLv2 servers fix
5343 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5344 servers that both support SSLv2 and enable export cipher suites by sending
5345 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5347 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5348 (OpenSSL development team).
5353 * Empty CKE with client auth and DHE fix
5355 If client auth is used then a server can seg fault in the event of a DHE
5356 ciphersuite being selected and a zero length ClientKeyExchange message
5357 being sent by the client. This could be exploited in a DoS attack.
5362 * Handshake with unseeded PRNG fix
5364 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5365 with an unseeded PRNG. The conditions are:
5366 - The client is on a platform where the PRNG has not been seeded
5367 automatically, and the user has not seeded manually
5368 - A protocol specific client method version has been used (i.e. not
5369 SSL_client_methodv23)
5370 - A ciphersuite is used that does not require additional random data from
5371 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5373 If the handshake succeeds then the client random that has been used will
5374 have been generated from a PRNG with insufficient entropy and therefore the
5375 output may be predictable.
5377 For example using the following command with an unseeded openssl will
5378 succeed on an unpatched platform:
5380 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5385 * Use After Free following d2i_ECPrivatekey error fix
5387 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5388 could cause a use after free condition. This, in turn, could cause a double
5389 free in several private key parsing functions (such as d2i_PrivateKey
5390 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5391 for applications that receive EC private keys from untrusted
5392 sources. This scenario is considered rare.
5394 This issue was discovered by the BoringSSL project and fixed in their
5400 * X509_to_X509_REQ NULL pointer deref fix
5402 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5403 the certificate key is invalid. This function is rarely used in practice.
5405 This issue was discovered by Brian Carpenter.
5410 * Removed the export ciphers from the DEFAULT ciphers
5414 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5416 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5417 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5418 So far those who have to target multiple platforms would compromise
5419 and argue that binary targeting say ARMv5 would still execute on
5420 ARMv8. "Universal" build resolves this compromise by providing
5421 near-optimal performance even on newer platforms.
5425 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5426 (other platforms pending).
5428 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5430 * Add support for the SignedCertificateTimestampList certificate and
5431 OCSP response extensions from RFC6962.
5435 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5436 for corner cases. (Certain input points at infinity could lead to
5437 bogus results, with non-infinity inputs mapped to infinity too.)
5441 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5442 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5443 common cases are optimized and there still is room for further
5444 improvements. Vector Permutation AES for Altivec is also added.
5448 * Add support for little-endian ppc64 Linux target.
5450 *Marcelo Cerri (IBM)*
5452 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5453 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5454 are optimized and there still is room for further improvements.
5455 Both 32- and 64-bit modes are supported.
5457 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5459 * Improved ARMv7 NEON support.
5463 * Support for SPARC Architecture 2011 crypto extensions, first
5464 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5465 SHA256/512, MD5, GHASH and modular exponentiation.
5467 *Andy Polyakov, David Miller*
5469 * Accelerated modular exponentiation for Intel processors, a.k.a.
5472 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5474 * Support for new and upcoming Intel processors, including AVX2,
5475 BMI and SHA ISA extensions. This includes additional "stitched"
5476 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5479 This work was sponsored by Intel Corp.
5483 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5484 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5485 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5489 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5490 this fixes a limitation in previous versions of OpenSSL.
5494 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5495 MGF1 digest and OAEP label.
5499 * Add EVP support for key wrapping algorithms, to avoid problems with
5500 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5501 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5502 algorithms and include tests cases.
5506 * Add functions to allocate and set the fields of an ECDSA_METHOD
5509 *Douglas E. Engert, Steve Henson*
5511 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5512 difference in days and seconds between two tm or ASN1_TIME structures.
5516 * Add -rev test option to s_server to just reverse order of characters
5517 received by client and send back to server. Also prints an abbreviated
5518 summary of the connection parameters.
5522 * New option -brief for s_client and s_server to print out a brief summary
5523 of connection parameters.
5527 * Add callbacks for arbitrary TLS extensions.
5529 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5531 * New option -crl_download in several openssl utilities to download CRLs
5532 from CRLDP extension in certificates.
5536 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5540 * New function X509_CRL_diff to generate a delta CRL from the difference
5541 of two full CRLs. Add support to "crl" utility.
5545 * New functions to set lookup_crls function and to retrieve
5546 X509_STORE from X509_STORE_CTX.
5550 * Print out deprecated issuer and subject unique ID fields in
5555 * Extend OCSP I/O functions so they can be used for simple general purpose
5556 HTTP as well as OCSP. New wrapper function which can be used to download
5557 CRLs using the OCSP API.
5561 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5565 * `SSL_CONF*` functions. These provide a common framework for application
5566 configuration using configuration files or command lines.
5570 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5571 message callback and prints the results. Needs compile time option
5572 "enable-ssl-trace". New options to s_client and s_server to enable
5577 * New ctrl and macro to retrieve supported points extensions.
5578 Print out extension in s_server and s_client.
5582 * New functions to retrieve certificate signature and signature
5587 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5592 * New Suite B modes for TLS code. These use and enforce the requirements
5593 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5594 only use Suite B curves. The Suite B modes can be set by using the
5595 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5599 * New chain verification flags for Suite B levels of security. Check
5600 algorithms are acceptable when flags are set in X509_verify_cert.
5604 * Make tls1_check_chain return a set of flags indicating checks passed
5605 by a certificate chain. Add additional tests to handle client
5606 certificates: checks for matching certificate type and issuer name
5611 * If an attempt is made to use a signature algorithm not in the peer
5612 preference list abort the handshake. If client has no suitable
5613 signature algorithms in response to a certificate request do not
5614 use the certificate.
5618 * If server EC tmp key is not in client preference list abort handshake.
5622 * Add support for certificate stores in CERT structure. This makes it
5623 possible to have different stores per SSL structure or one store in
5624 the parent SSL_CTX. Include distinct stores for certificate chain
5625 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5626 to build and store a certificate chain in CERT structure: returning
5627 an error if the chain cannot be built: this will allow applications
5628 to test if a chain is correctly configured.
5630 Note: if the CERT based stores are not set then the parent SSL_CTX
5631 store is used to retain compatibility with existing behaviour.
5635 * New function ssl_set_client_disabled to set a ciphersuite disabled
5636 mask based on the current session, check mask when sending client
5637 hello and checking the requested ciphersuite.
5641 * New ctrls to retrieve and set certificate types in a certificate
5642 request message. Print out received values in s_client. If certificate
5643 types is not set with custom values set sensible values based on
5644 supported signature algorithms.
5648 * Support for distinct client and server supported signature algorithms.
5652 * Add certificate callback. If set this is called whenever a certificate
5653 is required by client or server. An application can decide which
5654 certificate chain to present based on arbitrary criteria: for example
5655 supported signature algorithms. Add very simple example to s_server.
5656 This fixes many of the problems and restrictions of the existing client
5657 certificate callback: for example you can now clear an existing
5658 certificate and specify the whole chain.
5662 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5663 the certificate can be used for (if anything). Set valid_flags field
5664 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5665 to have similar checks in it.
5667 Add new "cert_flags" field to CERT structure and include a "strict mode".
5668 This enforces some TLS certificate requirements (such as only permitting
5669 certificate signature algorithms contained in the supported algorithms
5670 extension) which some implementations ignore: this option should be used
5671 with caution as it could cause interoperability issues.
5675 * Update and tidy signature algorithm extension processing. Work out
5676 shared signature algorithms based on preferences and peer algorithms
5677 and print them out in s_client and s_server. Abort handshake if no
5678 shared signature algorithms.
5682 * Add new functions to allow customised supported signature algorithms
5683 for SSL and SSL_CTX structures. Add options to s_client and s_server
5688 * New function SSL_certs_clear() to delete all references to certificates
5689 from an SSL structure. Before this once a certificate had been added
5690 it couldn't be removed.
5694 * Integrate hostname, email address and IP address checking with certificate
5695 verification. New verify options supporting checking in openssl utility.
5699 * Fixes and wildcard matching support to hostname and email checking
5700 functions. Add manual page.
5702 *Florian Weimer (Red Hat Product Security Team)*
5704 * New functions to check a hostname email or IP address against a
5705 certificate. Add options x509 utility to print results of checks against
5710 * Fix OCSP checking.
5712 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5714 * Initial experimental support for explicitly trusted non-root CAs.
5715 OpenSSL still tries to build a complete chain to a root but if an
5716 intermediate CA has a trust setting included that is used. The first
5717 setting is used: whether to trust (e.g., -addtrust option to the x509
5722 * Add -trusted_first option which attempts to find certificates in the
5723 trusted store even if an untrusted chain is also supplied.
5727 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5728 platform support for Linux and Android.
5732 * Support for linux-x32, ILP32 environment in x86_64 framework.
5736 * Experimental multi-implementation support for FIPS capable OpenSSL.
5737 When in FIPS mode the approved implementations are used as normal,
5738 when not in FIPS mode the internal unapproved versions are used instead.
5739 This means that the FIPS capable OpenSSL isn't forced to use the
5740 (often lower performance) FIPS implementations outside FIPS mode.
5744 * Transparently support X9.42 DH parameters when calling
5745 PEM_read_bio_DHparameters. This means existing applications can handle
5746 the new parameter format automatically.
5750 * Initial experimental support for X9.42 DH parameter format: mainly
5751 to support use of 'q' parameter for RFC5114 parameters.
5755 * Add DH parameters from RFC5114 including test data to dhtest.
5759 * Support for automatic EC temporary key parameter selection. If enabled
5760 the most preferred EC parameters are automatically used instead of
5761 hardcoded fixed parameters. Now a server just has to call:
5762 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5763 support ECDH and use the most appropriate parameters.
5767 * Enhance and tidy EC curve and point format TLS extension code. Use
5768 static structures instead of allocation if default values are used.
5769 New ctrls to set curves we wish to support and to retrieve shared curves.
5770 Print out shared curves in s_server. New options to s_server and s_client
5771 to set list of supported curves.
5775 * New ctrls to retrieve supported signature algorithms and
5776 supported curve values as an array of NIDs. Extend openssl utility
5777 to print out received values.
5781 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5782 between NIDs and the more common NIST names such as "P-256". Enhance
5783 ecparam utility and ECC method to recognise the NIST names for curves.
5787 * Enhance SSL/TLS certificate chain handling to support different
5788 chains for each certificate instead of one chain in the parent SSL_CTX.
5792 * Support for fixed DH ciphersuite client authentication: where both
5793 server and client use DH certificates with common parameters.
5797 * Support for fixed DH ciphersuites: those requiring DH server
5802 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5804 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5805 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5806 X509_CINF_get_signature were reverted post internal team review.
5811 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5813 * OCSP Status Request extension unbounded memory growth
5815 A malicious client can send an excessively large OCSP Status Request
5816 extension. If that client continually requests renegotiation, sending a
5817 large OCSP Status Request extension each time, then there will be unbounded
5818 memory growth on the server. This will eventually lead to a Denial Of
5819 Service attack through memory exhaustion. Servers with a default
5820 configuration are vulnerable even if they do not support OCSP. Builds using
5821 the "no-ocsp" build time option are not affected.
5823 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5828 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5831 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5837 * OOB write in MDC2_Update()
5839 An overflow can occur in MDC2_Update() either if called directly or
5840 through the EVP_DigestUpdate() function using MDC2. If an attacker
5841 is able to supply very large amounts of input data after a previous
5842 call to EVP_EncryptUpdate() with a partial block then a length check
5843 can overflow resulting in a heap corruption.
5845 The amount of data needed is comparable to SIZE_MAX which is impractical
5848 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5853 * Malformed SHA512 ticket DoS
5855 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5856 DoS attack where a malformed ticket will result in an OOB read which will
5859 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5860 a custom server callback and ticket lookup mechanism.
5862 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5867 * OOB write in BN_bn2dec()
5869 The function BN_bn2dec() does not check the return value of BN_div_word().
5870 This can cause an OOB write if an application uses this function with an
5871 overly large BIGNUM. This could be a problem if an overly large certificate
5872 or CRL is printed out from an untrusted source. TLS is not affected because
5873 record limits will reject an oversized certificate before it is parsed.
5875 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5880 * OOB read in TS_OBJ_print_bio()
5882 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5883 the total length the OID text representation would use and not the amount
5884 of data written. This will result in OOB reads when large OIDs are
5887 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5892 * Pointer arithmetic undefined behaviour
5894 Avoid some undefined pointer arithmetic
5896 A common idiom in the codebase is to check limits in the following manner:
5899 Where "p" points to some malloc'd data of SIZE bytes and
5902 "len" here could be from some externally supplied data (e.g. from a TLS
5905 The rules of C pointer arithmetic are such that "p + len" is only well
5906 defined where len <= SIZE. Therefore the above idiom is actually
5907 undefined behaviour.
5909 For example this could cause problems if some malloc implementation
5910 provides an address for "p" such that "p + len" actually overflows for
5911 values of len that are too big and therefore p + len < limit.
5913 This issue was reported to OpenSSL by Guido Vranken
5918 * Constant time flag not preserved in DSA signing
5920 Operations in the DSA signing algorithm should run in constant time in
5921 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5922 implementation means that a non-constant time codepath is followed for
5923 certain operations. This has been demonstrated through a cache-timing
5924 attack to be sufficient for an attacker to recover the private DSA key.
5926 This issue was reported by César Pereida (Aalto University), Billy Brumley
5927 (Tampere University of Technology), and Yuval Yarom (The University of
5928 Adelaide and NICTA).
5933 * DTLS buffered message DoS
5935 In a DTLS connection where handshake messages are delivered out-of-order
5936 those messages that OpenSSL is not yet ready to process will be buffered
5937 for later use. Under certain circumstances, a flaw in the logic means that
5938 those messages do not get removed from the buffer even though the handshake
5939 has been completed. An attacker could force up to approx. 15 messages to
5940 remain in the buffer when they are no longer required. These messages will
5941 be cleared when the DTLS connection is closed. The default maximum size for
5942 a message is 100k. Therefore the attacker could force an additional 1500k
5943 to be consumed per connection. By opening many simulataneous connections an
5944 attacker could cause a DoS attack through memory exhaustion.
5946 This issue was reported to OpenSSL by Quan Luo.
5951 * DTLS replay protection DoS
5953 A flaw in the DTLS replay attack protection mechanism means that records
5954 that arrive for future epochs update the replay protection "window" before
5955 the MAC for the record has been validated. This could be exploited by an
5956 attacker by sending a record for the next epoch (which does not have to
5957 decrypt or have a valid MAC), with a very large sequence number. This means
5958 that all subsequent legitimate packets are dropped causing a denial of
5959 service for a specific DTLS connection.
5961 This issue was reported to OpenSSL by the OCAP audit team.
5966 * Certificate message OOB reads
5968 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5969 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5970 theoretical DoS risk but this has not been observed in practice on common
5973 The messages affected are client certificate, client certificate request
5974 and server certificate. As a result the attack can only be performed
5975 against a client or a server which enables client authentication.
5977 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5982 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5984 * Prevent padding oracle in AES-NI CBC MAC check
5986 A MITM attacker can use a padding oracle attack to decrypt traffic
5987 when the connection uses an AES CBC cipher and the server support
5990 This issue was introduced as part of the fix for Lucky 13 padding
5991 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5992 constant time by making sure that always the same bytes are read and
5993 compared against either the MAC or padding bytes. But it no longer
5994 checked that there was enough data to have both the MAC and padding
5997 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6002 * Fix EVP_EncodeUpdate overflow
6004 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6005 Base64 encoding of binary data. If an attacker is able to supply very large
6006 amounts of input data then a length check can overflow resulting in a heap
6009 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6010 the `PEM_write_bio*` family of functions. These are mainly used within the
6011 OpenSSL command line applications, so any application which processes data
6012 from an untrusted source and outputs it as a PEM file should be considered
6013 vulnerable to this issue. User applications that call these APIs directly
6014 with large amounts of untrusted data may also be vulnerable.
6016 This issue was reported by Guido Vranken.
6021 * Fix EVP_EncryptUpdate overflow
6023 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6024 is able to supply very large amounts of input data after a previous call to
6025 EVP_EncryptUpdate() with a partial block then a length check can overflow
6026 resulting in a heap corruption. Following an analysis of all OpenSSL
6027 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6028 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6029 the first called function after an EVP_EncryptInit(), and therefore that
6030 specific call must be safe. The second form is where the length passed to
6031 EVP_EncryptUpdate() can be seen from the code to be some small value and
6032 therefore there is no possibility of an overflow. Since all instances are
6033 one of these two forms, it is believed that there can be no overflows in
6034 internal code due to this problem. It should be noted that
6035 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6036 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6037 of these calls have also been analysed too and it is believed there are no
6038 instances in internal usage where an overflow could occur.
6040 This issue was reported by Guido Vranken.
6045 * Prevent ASN.1 BIO excessive memory allocation
6047 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6048 a short invalid encoding can casuse allocation of large amounts of memory
6049 potentially consuming excessive resources or exhausting memory.
6051 Any application parsing untrusted data through d2i BIO functions is
6052 affected. The memory based functions such as d2i_X509() are *not* affected.
6053 Since the memory based functions are used by the TLS library, TLS
6054 applications are not affected.
6056 This issue was reported by Brian Carpenter.
6063 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6064 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6065 in arbitrary stack data being returned in the buffer.
6067 This issue was reported by Guido Vranken.
6072 * Modify behavior of ALPN to invoke callback after SNI/servername
6073 callback, such that updates to the SSL_CTX affect ALPN.
6077 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6082 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6083 methods are enabled and ssl2 is disabled the methods return NULL.
6087 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6089 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6090 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6091 provide any "EXPORT" or "LOW" strength ciphers.
6095 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6096 is by default disabled at build-time. Builds that are not configured with
6097 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6098 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6099 will need to explicitly call either of:
6101 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6103 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6105 as appropriate. Even if either of those is used, or the application
6106 explicitly uses the version-specific SSLv2_method() or its client and
6107 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6108 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6109 ciphers, and SSLv2 56-bit DES are no longer available.
6114 * Fix a double-free in DSA code
6116 A double free bug was discovered when OpenSSL parses malformed DSA private
6117 keys and could lead to a DoS attack or memory corruption for applications
6118 that receive DSA private keys from untrusted sources. This scenario is
6121 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6127 * Disable SRP fake user seed to address a server memory leak.
6129 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6131 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6132 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6133 was changed to ignore the "fake user" SRP seed, even if the seed
6136 Users should use SRP_VBASE_get1_by_user instead. Note that in
6137 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6138 also that even though configuring the SRP seed attempts to hide
6139 invalid usernames by continuing the handshake with fake
6140 credentials, this behaviour is not constant time and no strong
6141 guarantees are made that the handshake is indistinguishable from
6142 that of a valid user.
6147 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6149 In the BN_hex2bn function the number of hex digits is calculated using an
6150 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6151 large values of `i` this can result in `bn_expand` not allocating any
6152 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6153 field as NULL leading to a subsequent NULL ptr deref. For very large values
6154 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6155 In this case memory is allocated to the internal BIGNUM data field, but it
6156 is insufficiently sized leading to heap corruption. A similar issue exists
6157 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6158 is ever called by user applications with very large untrusted hex/dec data.
6159 This is anticipated to be a rare occurrence.
6161 All OpenSSL internal usage of these functions use data that is not expected
6162 to be untrusted, e.g. config file data or application command line
6163 arguments. If user developed applications generate config file data based
6164 on untrusted data then it is possible that this could also lead to security
6165 consequences. This is also anticipated to be rare.
6167 This issue was reported to OpenSSL by Guido Vranken.
6172 * Fix memory issues in `BIO_*printf` functions
6174 The internal `fmtstr` function used in processing a "%s" format string in
6175 the `BIO_*printf` functions could overflow while calculating the length of a
6176 string and cause an OOB read when printing very long strings.
6178 Additionally the internal `doapr_outch` function can attempt to write to an
6179 OOB memory location (at an offset from the NULL pointer) in the event of a
6180 memory allocation failure. In 1.0.2 and below this could be caused where
6181 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6182 could be in processing a very long "%s" format string. Memory leaks can
6185 The first issue may mask the second issue dependent on compiler behaviour.
6186 These problems could enable attacks where large amounts of untrusted data
6187 is passed to the `BIO_*printf` functions. If applications use these functions
6188 in this way then they could be vulnerable. OpenSSL itself uses these
6189 functions when printing out human-readable dumps of ASN.1 data. Therefore
6190 applications that print this data could be vulnerable if the data is from
6191 untrusted sources. OpenSSL command line applications could also be
6192 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6193 as command line arguments.
6195 Libssl is not considered directly vulnerable. Additionally certificates etc
6196 received via remote connections via libssl are also unlikely to be able to
6197 trigger these issues because of message size limits enforced within libssl.
6199 This issue was reported to OpenSSL Guido Vranken.
6204 * Side channel attack on modular exponentiation
6206 A side-channel attack was found which makes use of cache-bank conflicts on
6207 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6208 of RSA keys. The ability to exploit this issue is limited as it relies on
6209 an attacker who has control of code in a thread running on the same
6210 hyper-threaded core as the victim thread which is performing decryptions.
6212 This issue was reported to OpenSSL by Yuval Yarom, The University of
6213 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6214 Nadia Heninger, University of Pennsylvania with more information at
6215 <http://cachebleed.info>.
6220 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6221 if no keysize is specified with default_bits. This fixes an
6222 omission in an earlier change that changed all RSA/DSA key generation
6223 commands to use 2048 bits by default.
6227 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6229 * Protection for DH small subgroup attacks
6231 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6232 switched on by default and cannot be disabled. This could have some
6237 * SSLv2 doesn't block disabled ciphers
6239 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6240 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6241 been disabled, provided that the SSLv2 protocol was not also disabled via
6244 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6245 and Sebastian Schinzel.
6250 * Reject DH handshakes with parameters shorter than 1024 bits.
6254 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6256 * Certificate verify crash with missing PSS parameter
6258 The signature verification routines will crash with a NULL pointer
6259 dereference if presented with an ASN.1 signature using the RSA PSS
6260 algorithm and absent mask generation function parameter. Since these
6261 routines are used to verify certificate signature algorithms this can be
6262 used to crash any certificate verification operation and exploited in a
6263 DoS attack. Any application which performs certificate verification is
6264 vulnerable including OpenSSL clients and servers which enable client
6267 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6272 * X509_ATTRIBUTE memory leak
6274 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6275 memory. This structure is used by the PKCS#7 and CMS routines so any
6276 application which reads PKCS#7 or CMS data from untrusted sources is
6277 affected. SSL/TLS is not affected.
6279 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6285 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6286 This changes the decoding behaviour for some invalid messages,
6287 though the change is mostly in the more lenient direction, and
6288 legacy behaviour is preserved as much as possible.
6292 * In DSA_generate_parameters_ex, if the provided seed is too short,
6293 use a random seed, as already documented.
6295 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6297 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6299 * Alternate chains certificate forgery
6301 During certificate verfification, OpenSSL will attempt to find an
6302 alternative certificate chain if the first attempt to build such a chain
6303 fails. An error in the implementation of this logic can mean that an
6304 attacker could cause certain checks on untrusted certificates to be
6305 bypassed, such as the CA flag, enabling them to use a valid leaf
6306 certificate to act as a CA and "issue" an invalid certificate.
6308 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6314 * Race condition handling PSK identify hint
6316 If PSK identity hints are received by a multi-threaded client then
6317 the values are wrongly updated in the parent SSL_CTX structure. This can
6318 result in a race condition potentially leading to a double free of the
6324 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6326 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6327 incompatibility in the handling of HMAC. The previous ABI has now been
6330 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6332 * Malformed ECParameters causes infinite loop
6334 When processing an ECParameters structure OpenSSL enters an infinite loop
6335 if the curve specified is over a specially malformed binary polynomial
6338 This can be used to perform denial of service against any
6339 system which processes public keys, certificate requests or
6340 certificates. This includes TLS clients and TLS servers with
6341 client authentication enabled.
6343 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6348 * Exploitable out-of-bounds read in X509_cmp_time
6350 X509_cmp_time does not properly check the length of the ASN1_TIME
6351 string and can read a few bytes out of bounds. In addition,
6352 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6355 An attacker can use this to craft malformed certificates and CRLs of
6356 various sizes and potentially cause a segmentation fault, resulting in
6357 a DoS on applications that verify certificates or CRLs. TLS clients
6358 that verify CRLs are affected. TLS clients and servers with client
6359 authentication enabled may be affected if they use custom verification
6362 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6363 independently by Hanno Böck.
6368 * PKCS7 crash with missing EnvelopedContent
6370 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6371 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6372 with missing content and trigger a NULL pointer dereference on parsing.
6374 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6375 structures from untrusted sources are affected. OpenSSL clients and
6376 servers are not affected.
6378 This issue was reported to OpenSSL by Michal Zalewski (Google).
6383 * CMS verify infinite loop with unknown hash function
6385 When verifying a signedData message the CMS code can enter an infinite loop
6386 if presented with an unknown hash function OID. This can be used to perform
6387 denial of service against any system which verifies signedData messages using
6389 This issue was reported to OpenSSL by Johannes Bauer.
6394 * Race condition handling NewSessionTicket
6396 If a NewSessionTicket is received by a multi-threaded client when attempting to
6397 reuse a previous ticket then a race condition can occur potentially leading to
6398 a double free of the ticket data.
6403 * Reject DH handshakes with parameters shorter than 768 bits.
6405 *Kurt Roeckx and Emilia Kasper*
6407 * dhparam: generate 2048-bit parameters by default.
6409 *Kurt Roeckx and Emilia Kasper*
6411 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6413 * Segmentation fault in ASN1_TYPE_cmp fix
6415 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6416 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6417 certificate signature algorithm consistency this can be used to crash any
6418 certificate verification operation and exploited in a DoS attack. Any
6419 application which performs certificate verification is vulnerable including
6420 OpenSSL clients and servers which enable client authentication.
6425 * ASN.1 structure reuse memory corruption fix
6427 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6428 memory corruption via an invalid write. Such reuse is and has been
6429 strongly discouraged and is believed to be rare.
6431 Applications that parse structures containing CHOICE or ANY DEFINED BY
6432 components may be affected. Certificate parsing (d2i_X509 and related
6433 functions) are however not affected. OpenSSL clients and servers are
6439 * PKCS7 NULL pointer dereferences fix
6441 The PKCS#7 parsing code does not handle missing outer ContentInfo
6442 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6443 missing content and trigger a NULL pointer dereference on parsing.
6445 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6446 otherwise parse PKCS#7 structures from untrusted sources are
6447 affected. OpenSSL clients and servers are not affected.
6449 This issue was reported to OpenSSL by Michal Zalewski (Google).
6454 * DoS via reachable assert in SSLv2 servers fix
6456 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6457 servers that both support SSLv2 and enable export cipher suites by sending
6458 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6460 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6461 (OpenSSL development team).
6466 * Use After Free following d2i_ECPrivatekey error fix
6468 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6469 could cause a use after free condition. This, in turn, could cause a double
6470 free in several private key parsing functions (such as d2i_PrivateKey
6471 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6472 for applications that receive EC private keys from untrusted
6473 sources. This scenario is considered rare.
6475 This issue was discovered by the BoringSSL project and fixed in their
6481 * X509_to_X509_REQ NULL pointer deref fix
6483 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6484 the certificate key is invalid. This function is rarely used in practice.
6486 This issue was discovered by Brian Carpenter.
6491 * Removed the export ciphers from the DEFAULT ciphers
6495 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6497 * Build fixes for the Windows and OpenVMS platforms
6499 *Matt Caswell and Richard Levitte*
6501 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6503 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6504 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6505 dereference. This could lead to a Denial Of Service attack. Thanks to
6506 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6511 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6512 dtls1_buffer_record function under certain conditions. In particular this
6513 could occur if an attacker sent repeated DTLS records with the same
6514 sequence number but for the next epoch. The memory leak could be exploited
6515 by an attacker in a Denial of Service attack through memory exhaustion.
6516 Thanks to Chris Mueller for reporting this issue.
6521 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6522 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6523 method would be set to NULL which could later result in a NULL pointer
6524 dereference. Thanks to Frank Schmirler for reporting this issue.
6529 * Abort handshake if server key exchange message is omitted for ephemeral
6532 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6533 reporting this issue.
6538 * Remove non-export ephemeral RSA code on client and server. This code
6539 violated the TLS standard by allowing the use of temporary RSA keys in
6540 non-export ciphersuites and could be used by a server to effectively
6541 downgrade the RSA key length used to a value smaller than the server
6542 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6543 INRIA or reporting this issue.
6548 * Fixed issue where DH client certificates are accepted without verification.
6549 An OpenSSL server will accept a DH certificate for client authentication
6550 without the certificate verify message. This effectively allows a client to
6551 authenticate without the use of a private key. This only affects servers
6552 which trust a client certificate authority which issues certificates
6553 containing DH keys: these are extremely rare and hardly ever encountered.
6554 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6560 * Ensure that the session ID context of an SSL is updated when its
6561 SSL_CTX is updated via SSL_set_SSL_CTX.
6563 The session ID context is typically set from the parent SSL_CTX,
6564 and can vary with the CTX.
6568 * Fix various certificate fingerprint issues.
6570 By using non-DER or invalid encodings outside the signed portion of a
6571 certificate the fingerprint can be changed without breaking the signature.
6572 Although no details of the signed portion of the certificate can be changed
6573 this can cause problems with some applications: e.g. those using the
6574 certificate fingerprint for blacklists.
6576 1. Reject signatures with non zero unused bits.
6578 If the BIT STRING containing the signature has non zero unused bits reject
6579 the signature. All current signature algorithms require zero unused bits.
6581 2. Check certificate algorithm consistency.
6583 Check the AlgorithmIdentifier inside TBS matches the one in the
6584 certificate signature. NB: this will result in signature failure
6585 errors for some broken certificates.
6587 Thanks to Konrad Kraszewski from Google for reporting this issue.
6589 3. Check DSA/ECDSA signatures use DER.
6591 Re-encode DSA/ECDSA signatures and compare with the original received
6592 signature. Return an error if there is a mismatch.
6594 This will reject various cases including garbage after signature
6595 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6596 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6597 (negative or with leading zeroes).
6599 Further analysis was conducted and fixes were developed by Stephen Henson
6600 of the OpenSSL core team.
6606 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6607 results on some platforms, including x86_64. This bug occurs at random
6608 with a very low probability, and is not known to be exploitable in any
6609 way, though its exact impact is difficult to determine. Thanks to Pieter
6610 Wuille (Blockstream) who reported this issue and also suggested an initial
6611 fix. Further analysis was conducted by the OpenSSL development team and
6612 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6613 the OpenSSL core team.
6618 * Do not resume sessions on the server if the negotiated protocol
6619 version does not match the session's version. Resuming with a different
6620 version, while not strictly forbidden by the RFC, is of questionable
6621 sanity and breaks all known clients.
6623 *David Benjamin, Emilia Käsper*
6625 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6626 early CCS messages during renegotiation. (Note that because
6627 renegotiation is encrypted, this early CCS was not exploitable.)
6631 * Tighten client-side session ticket handling during renegotiation:
6632 ensure that the client only accepts a session ticket if the server sends
6633 the extension anew in the ServerHello. Previously, a TLS client would
6634 reuse the old extension state and thus accept a session ticket if one was
6635 announced in the initial ServerHello.
6637 Similarly, ensure that the client requires a session ticket if one
6638 was advertised in the ServerHello. Previously, a TLS client would
6639 ignore a missing NewSessionTicket message.
6643 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6647 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6648 sends a carefully crafted handshake message, to cause OpenSSL to fail
6649 to free up to 64k of memory causing a memory leak. This could be
6650 exploited in a Denial Of Service attack. This issue affects OpenSSL
6651 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6652 whether SRTP is used or configured. Implementations of OpenSSL that
6653 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6655 The fix was developed by the OpenSSL team.
6660 * Session Ticket Memory Leak.
6662 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6663 integrity of that ticket is first verified. In the event of a session
6664 ticket integrity check failing, OpenSSL will fail to free memory
6665 causing a memory leak. By sending a large number of invalid session
6666 tickets an attacker could exploit this issue in a Denial Of Service
6672 * Build option no-ssl3 is incomplete.
6674 When OpenSSL is configured with "no-ssl3" as a build option, servers
6675 could accept and complete a SSL 3.0 handshake, and clients could be
6676 configured to send them.
6679 *Akamai and the OpenSSL team*
6681 * Add support for TLS_FALLBACK_SCSV.
6682 Client applications doing fallback retries should call
6683 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6686 *Adam Langley, Bodo Moeller*
6688 * Add additional DigestInfo checks.
6690 Re-encode DigestInto in DER and check against the original when
6691 verifying RSA signature: this will reject any improperly encoded
6692 DigestInfo structures.
6694 Note: this is a precautionary measure and no attacks are currently known.
6698 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6700 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6701 SRP code can be overrun an internal buffer. Add sanity check that
6702 g, A, B < N to SRP code.
6704 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6705 Group for discovering this issue.
6710 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6711 TLS 1.0 instead of higher protocol versions when the ClientHello message
6712 is badly fragmented. This allows a man-in-the-middle attacker to force a
6713 downgrade to TLS 1.0 even if both the server and the client support a
6714 higher protocol version, by modifying the client's TLS records.
6716 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6717 researching this issue.
6722 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6723 to a denial of service attack. A malicious server can crash the client
6724 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6725 ciphersuite and sending carefully crafted handshake messages.
6727 Thanks to Felix Gröbert (Google) for discovering and researching this
6733 * By sending carefully crafted DTLS packets an attacker could cause openssl
6734 to leak memory. This can be exploited through a Denial of Service attack.
6735 Thanks to Adam Langley for discovering and researching this issue.
6740 * An attacker can force openssl to consume large amounts of memory whilst
6741 processing DTLS handshake messages. This can be exploited through a
6742 Denial of Service attack.
6743 Thanks to Adam Langley for discovering and researching this issue.
6748 * An attacker can force an error condition which causes openssl to crash
6749 whilst processing DTLS packets due to memory being freed twice. This
6750 can be exploited through a Denial of Service attack.
6751 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6757 * If a multithreaded client connects to a malicious server using a resumed
6758 session and the server sends an ec point format extension it could write
6759 up to 255 bytes to freed memory.
6761 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6767 * A malicious server can crash an OpenSSL client with a null pointer
6768 dereference (read) by specifying an SRP ciphersuite even though it was not
6769 properly negotiated with the client. This can be exploited through a
6770 Denial of Service attack.
6772 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6773 discovering and researching this issue.
6778 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6779 X509_name_oneline, X509_name_print_ex et al. to leak some information
6780 from the stack. Applications may be affected if they echo pretty printing
6781 output to the attacker.
6783 Thanks to Ivan Fratric (Google) for discovering this issue.
6786 *Emilia Käsper, and Steve Henson*
6788 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6789 for corner cases. (Certain input points at infinity could lead to
6790 bogus results, with non-infinity inputs mapped to infinity too.)
6794 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6796 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6797 handshake can force the use of weak keying material in OpenSSL
6798 SSL/TLS clients and servers.
6800 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6801 researching this issue. ([CVE-2014-0224])
6803 *KIKUCHI Masashi, Steve Henson*
6805 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6806 OpenSSL DTLS client the code can be made to recurse eventually crashing
6809 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6812 *Imre Rad, Steve Henson*
6814 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6815 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6816 client or server. This is potentially exploitable to run arbitrary
6817 code on a vulnerable client or server.
6819 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6821 *Jüri Aedla, Steve Henson*
6823 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6824 are subject to a denial of service attack.
6826 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6827 this issue. ([CVE-2014-3470])
6829 *Felix Gröbert, Ivan Fratric, Steve Henson*
6831 * Harmonize version and its documentation. -f flag is used to display
6834 *mancha <mancha1@zoho.com>*
6836 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6837 in i2d_ECPrivateKey.
6839 *mancha <mancha1@zoho.com>*
6841 * Fix some double frees. These are not thought to be exploitable.
6843 *mancha <mancha1@zoho.com>*
6845 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6847 * A missing bounds check in the handling of the TLS heartbeat extension
6848 can be used to reveal up to 64k of memory to a connected client or
6851 Thanks for Neel Mehta of Google Security for discovering this bug and to
6852 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6853 preparing the fix ([CVE-2014-0160])
6855 *Adam Langley, Bodo Moeller*
6857 * Fix for the attack described in the paper "Recovering OpenSSL
6858 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6859 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6860 <http://eprint.iacr.org/2014/140>
6862 Thanks to Yuval Yarom and Naomi Benger for discovering this
6863 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6865 *Yuval Yarom and Naomi Benger*
6867 * TLS pad extension: draft-agl-tls-padding-03
6869 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6870 TLS client Hello record length value would otherwise be > 255 and
6871 less that 512 pad with a dummy extension containing zeroes so it
6872 is at least 512 bytes long.
6874 *Adam Langley, Steve Henson*
6876 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6878 * Fix for TLS record tampering bug. A carefully crafted invalid
6879 handshake could crash OpenSSL with a NULL pointer exception.
6880 Thanks to Anton Johansson for reporting this issues.
6883 * Keep original DTLS digest and encryption contexts in retransmission
6884 structures so we can use the previous session parameters if they need
6885 to be resent. ([CVE-2013-6450])
6889 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6890 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6891 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6892 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6893 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6894 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6896 *Rob Stradling, Adam Langley*
6898 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6900 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6901 supporting platforms or when small records were transferred.
6903 *Andy Polyakov, Steve Henson*
6905 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6907 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6909 This addresses the flaw in CBC record processing discovered by
6910 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6911 at: <http://www.isg.rhul.ac.uk/tls/>
6913 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6914 Security Group at Royal Holloway, University of London
6915 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6916 Emilia Käsper for the initial patch.
6919 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6921 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6922 ciphersuites which can be exploited in a denial of service attack.
6923 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6924 and detecting this bug and to Wolfgang Ettlinger
6925 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6930 * Return an error when checking OCSP signatures when key is NULL.
6931 This fixes a DoS attack. ([CVE-2013-0166])
6935 * Make openssl verify return errors.
6937 *Chris Palmer <palmer@google.com> and Ben Laurie*
6939 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6940 the right response is stapled. Also change SSL_get_certificate()
6941 so it returns the certificate actually sent.
6942 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6944 *Rob Stradling <rob.stradling@comodo.com>*
6946 * Fix possible deadlock when decoding public keys.
6950 * Don't use TLS 1.0 record version number in initial client hello
6955 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6957 * Sanity check record length before skipping explicit IV in TLS
6958 1.2, 1.1 and DTLS to fix DoS attack.
6960 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6961 fuzzing as a service testing platform.
6966 * Initialise tkeylen properly when encrypting CMS messages.
6967 Thanks to Solar Designer of Openwall for reporting this issue.
6971 * In FIPS mode don't try to use composite ciphers as they are not
6976 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6978 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6979 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6980 mean any application compiled against OpenSSL 1.0.0 headers setting
6981 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6982 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6983 0x10000000L Any application which was previously compiled against
6984 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6985 will need to be recompiled as a result. Letting be results in
6986 inability to disable specifically TLS 1.1 and in client context,
6987 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6991 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6992 disable just protocol X, but all protocols above X *if* there are
6993 protocols *below* X still enabled. In more practical terms it means
6994 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6995 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6996 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7001 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7003 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7004 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7005 in CRYPTO_realloc_clean.
7007 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7008 issue and to Adam Langley <agl@chromium.org> for fixing it.
7011 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7013 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7017 * Workarounds for some broken servers that "hang" if a client hello
7018 record length exceeds 255 bytes.
7020 1. Do not use record version number > TLS 1.0 in initial client
7021 hello: some (but not all) hanging servers will now work.
7022 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7023 the number of ciphers sent in the client hello. This should be
7024 set to an even number, such as 50, for example by passing:
7025 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7026 Most broken servers should now work.
7027 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7028 TLS 1.2 client support entirely.
7032 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7036 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7038 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7039 STRING form instead of a DigestInfo.
7043 * The format used for MDC2 RSA signatures is inconsistent between EVP
7044 and the RSA_sign/RSA_verify functions. This was made more apparent when
7045 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7046 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7047 the correct format in RSA_verify so both forms transparently work.
7051 * Some servers which support TLS 1.0 can choke if we initially indicate
7052 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7053 encrypted premaster secret. As a workaround use the maximum permitted
7054 client version in client hello, this should keep such servers happy
7055 and still work with previous versions of OpenSSL.
7059 * Add support for TLS/DTLS heartbeats.
7061 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7063 * Add support for SCTP.
7065 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7067 * Improved PRNG seeding for VOS.
7069 *Paul Green <Paul.Green@stratus.com>*
7071 * Extensive assembler packs updates, most notably:
7073 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7074 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7075 - x86_64: bit-sliced AES implementation;
7076 - ARM: NEON support, contemporary platforms optimizations;
7077 - s390x: z196 support;
7078 - `*`: GHASH and GF(2^m) multiplication implementations;
7082 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7083 (removal of unnecessary code)
7085 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7087 * Add TLS key material exporter from RFC 5705.
7091 * Add DTLS-SRTP negotiation from RFC 5764.
7095 * Add Next Protocol Negotiation,
7096 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7097 disabled with a no-npn flag to config or Configure. Code donated
7100 *Adam Langley <agl@google.com> and Ben Laurie*
7102 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7103 NIST-P256, NIST-P521, with constant-time single point multiplication on
7104 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7105 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7106 Code made available under Apache License version 2.0.
7108 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7109 line to include this in your build of OpenSSL, and run "make depend" (or
7110 "make update"). This enables the following EC_METHODs:
7112 EC_GFp_nistp224_method()
7113 EC_GFp_nistp256_method()
7114 EC_GFp_nistp521_method()
7116 EC_GROUP_new_by_curve_name() will automatically use these (while
7117 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7120 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7122 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7123 all platforms. Move ssize_t definition from e_os.h to the public
7124 header file e_os2.h as it now appears in public header file cms.h
7128 * New -sigopt option to the ca, req and x509 utilities. Additional
7129 signature parameters can be passed using this option and in
7134 * Add RSA PSS signing function. This will generate and set the
7135 appropriate AlgorithmIdentifiers for PSS based on those in the
7136 corresponding EVP_MD_CTX structure. No application support yet.
7140 * Support for companion algorithm specific ASN1 signing routines.
7141 New function ASN1_item_sign_ctx() signs a pre-initialised
7142 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7143 the appropriate parameters.
7147 * Add new algorithm specific ASN1 verification initialisation function
7148 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7149 handling will be the same no matter what EVP_PKEY_METHOD is used.
7150 Add a PSS handler to support verification of PSS signatures: checked
7151 against a number of sample certificates.
7155 * Add signature printing for PSS. Add PSS OIDs.
7157 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7159 * Add algorithm specific signature printing. An individual ASN1 method
7160 can now print out signatures instead of the standard hex dump.
7162 More complex signatures (e.g. PSS) can print out more meaningful
7163 information. Include DSA version that prints out the signature
7168 * Password based recipient info support for CMS library: implementing
7173 * Split password based encryption into PBES2 and PBKDF2 functions. This
7174 neatly separates the code into cipher and PBE sections and is required
7175 for some algorithms that split PBES2 into separate pieces (such as
7176 password based CMS).
7180 * Session-handling fixes:
7181 - Fix handling of connections that are resuming with a session ID,
7182 but also support Session Tickets.
7183 - Fix a bug that suppressed issuing of a new ticket if the client
7184 presented a ticket with an expired session.
7185 - Try to set the ticket lifetime hint to something reasonable.
7186 - Make tickets shorter by excluding irrelevant information.
7187 - On the client side, don't ignore renewed tickets.
7189 *Adam Langley, Bodo Moeller (Google)*
7191 * Fix PSK session representation.
7195 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7197 This work was sponsored by Intel.
7201 * Add GCM support to TLS library. Some custom code is needed to split
7202 the IV between the fixed (from PRF) and explicit (from TLS record)
7203 portions. This adds all GCM ciphersuites supported by RFC5288 and
7204 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7205 add a special AESGCM string for GCM only.
7209 * Expand range of ctrls for AES GCM. Permit setting invocation
7210 field on decrypt and retrieval of invocation field only on encrypt.
7214 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7215 As required by RFC5289 these ciphersuites cannot be used if for
7216 versions of TLS earlier than 1.2.
7220 * For FIPS capable OpenSSL interpret a NULL default public key method
7221 as unset and return the appropriate default but do *not* set the default.
7222 This means we can return the appropriate method in applications that
7223 switch between FIPS and non-FIPS modes.
7227 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7228 ENGINE is used then we cannot handle that in the FIPS module so we
7229 keep original code iff non-FIPS operations are allowed.
7233 * Add -attime option to openssl utilities.
7235 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7237 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7241 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7242 FIPS EC methods unconditionally for now.
7246 * New build option no-ec2m to disable characteristic 2 code.
7250 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7251 all cases can be covered as some introduce binary incompatibilities.
7255 * Redirect RSA operations to FIPS module including keygen,
7256 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7260 * Add similar low-level API blocking to ciphers.
7264 * low-level digest APIs are not approved in FIPS mode: any attempt
7265 to use these will cause a fatal error. Applications that *really* want
7266 to use them can use the `private_*` version instead.
7270 * Redirect cipher operations to FIPS module for FIPS builds.
7274 * Redirect digest operations to FIPS module for FIPS builds.
7278 * Update build system to add "fips" flag which will link in fipscanister.o
7279 for static and shared library builds embedding a signature if needed.
7283 * Output TLS supported curves in preference order instead of numerical
7284 order. This is currently hardcoded for the highest order curves first.
7285 This should be configurable so applications can judge speed vs strength.
7289 * Add TLS v1.2 server support for client authentication.
7293 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7298 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7299 FIPS modules versions.
7303 * Add TLS v1.2 client side support for client authentication. Keep cache
7304 of handshake records longer as we don't know the hash algorithm to use
7305 until after the certificate request message is received.
7309 * Initial TLS v1.2 client support. Add a default signature algorithms
7310 extension including all the algorithms we support. Parse new signature
7311 format in client key exchange. Relax some ECC signing restrictions for
7312 TLS v1.2 as indicated in RFC5246.
7316 * Add server support for TLS v1.2 signature algorithms extension. Switch
7317 to new signature format when needed using client digest preference.
7318 All server ciphersuites should now work correctly in TLS v1.2. No client
7319 support yet and no support for client certificates.
7323 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7324 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7325 ciphersuites. At present only RSA key exchange ciphersuites work with
7326 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7327 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7328 and version checking.
7332 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7333 with this defined it will not be affected by any changes to ssl internal
7334 structures. Add several utility functions to allow openssl application
7335 to work with OPENSSL_NO_SSL_INTERN defined.
7339 * A long standing patch to add support for SRP from EdelWeb (Peter
7340 Sylvester and Christophe Renou) was integrated.
7341 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7342 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7345 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7349 * Permit abbreviated handshakes when renegotiating using the function
7350 SSL_renegotiate_abbreviated().
7352 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7354 * Add call to ENGINE_register_all_complete() to
7355 ENGINE_load_builtin_engines(), so some implementations get used
7356 automatically instead of needing explicit application support.
7360 * Add support for TLS key exporter as described in RFC5705.
7362 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7364 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7365 a few changes are required:
7367 Add SSL_OP_NO_TLSv1_1 flag.
7368 Add TLSv1_1 methods.
7369 Update version checking logic to handle version 1.1.
7370 Add explicit IV handling (ported from DTLS code).
7371 Add command line options to s_client/s_server.
7378 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7380 * X509_ATTRIBUTE memory leak
7382 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7383 memory. This structure is used by the PKCS#7 and CMS routines so any
7384 application which reads PKCS#7 or CMS data from untrusted sources is
7385 affected. SSL/TLS is not affected.
7387 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7393 * Race condition handling PSK identify hint
7395 If PSK identity hints are received by a multi-threaded client then
7396 the values are wrongly updated in the parent SSL_CTX structure. This can
7397 result in a race condition potentially leading to a double free of the
7403 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7405 * Malformed ECParameters causes infinite loop
7407 When processing an ECParameters structure OpenSSL enters an infinite loop
7408 if the curve specified is over a specially malformed binary polynomial
7411 This can be used to perform denial of service against any
7412 system which processes public keys, certificate requests or
7413 certificates. This includes TLS clients and TLS servers with
7414 client authentication enabled.
7416 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7421 * Exploitable out-of-bounds read in X509_cmp_time
7423 X509_cmp_time does not properly check the length of the ASN1_TIME
7424 string and can read a few bytes out of bounds. In addition,
7425 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7428 An attacker can use this to craft malformed certificates and CRLs of
7429 various sizes and potentially cause a segmentation fault, resulting in
7430 a DoS on applications that verify certificates or CRLs. TLS clients
7431 that verify CRLs are affected. TLS clients and servers with client
7432 authentication enabled may be affected if they use custom verification
7435 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7436 independently by Hanno Böck.
7441 * PKCS7 crash with missing EnvelopedContent
7443 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7444 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7445 with missing content and trigger a NULL pointer dereference on parsing.
7447 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7448 structures from untrusted sources are affected. OpenSSL clients and
7449 servers are not affected.
7451 This issue was reported to OpenSSL by Michal Zalewski (Google).
7456 * CMS verify infinite loop with unknown hash function
7458 When verifying a signedData message the CMS code can enter an infinite loop
7459 if presented with an unknown hash function OID. This can be used to perform
7460 denial of service against any system which verifies signedData messages using
7462 This issue was reported to OpenSSL by Johannes Bauer.
7467 * Race condition handling NewSessionTicket
7469 If a NewSessionTicket is received by a multi-threaded client when attempting to
7470 reuse a previous ticket then a race condition can occur potentially leading to
7471 a double free of the ticket data.
7476 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7478 * Segmentation fault in ASN1_TYPE_cmp fix
7480 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7481 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7482 certificate signature algorithm consistency this can be used to crash any
7483 certificate verification operation and exploited in a DoS attack. Any
7484 application which performs certificate verification is vulnerable including
7485 OpenSSL clients and servers which enable client authentication.
7490 * ASN.1 structure reuse memory corruption fix
7492 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7493 memory corruption via an invalid write. Such reuse is and has been
7494 strongly discouraged and is believed to be rare.
7496 Applications that parse structures containing CHOICE or ANY DEFINED BY
7497 components may be affected. Certificate parsing (d2i_X509 and related
7498 functions) are however not affected. OpenSSL clients and servers are
7504 * PKCS7 NULL pointer dereferences fix
7506 The PKCS#7 parsing code does not handle missing outer ContentInfo
7507 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7508 missing content and trigger a NULL pointer dereference on parsing.
7510 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7511 otherwise parse PKCS#7 structures from untrusted sources are
7512 affected. OpenSSL clients and servers are not affected.
7514 This issue was reported to OpenSSL by Michal Zalewski (Google).
7519 * DoS via reachable assert in SSLv2 servers fix
7521 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7522 servers that both support SSLv2 and enable export cipher suites by sending
7523 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7525 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7526 (OpenSSL development team).
7531 * Use After Free following d2i_ECPrivatekey error fix
7533 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7534 could cause a use after free condition. This, in turn, could cause a double
7535 free in several private key parsing functions (such as d2i_PrivateKey
7536 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7537 for applications that receive EC private keys from untrusted
7538 sources. This scenario is considered rare.
7540 This issue was discovered by the BoringSSL project and fixed in their
7546 * X509_to_X509_REQ NULL pointer deref fix
7548 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7549 the certificate key is invalid. This function is rarely used in practice.
7551 This issue was discovered by Brian Carpenter.
7556 * Removed the export ciphers from the DEFAULT ciphers
7560 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7562 * Build fixes for the Windows and OpenVMS platforms
7564 *Matt Caswell and Richard Levitte*
7566 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7568 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7569 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7570 dereference. This could lead to a Denial Of Service attack. Thanks to
7571 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7576 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7577 dtls1_buffer_record function under certain conditions. In particular this
7578 could occur if an attacker sent repeated DTLS records with the same
7579 sequence number but for the next epoch. The memory leak could be exploited
7580 by an attacker in a Denial of Service attack through memory exhaustion.
7581 Thanks to Chris Mueller for reporting this issue.
7586 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7587 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7588 method would be set to NULL which could later result in a NULL pointer
7589 dereference. Thanks to Frank Schmirler for reporting this issue.
7594 * Abort handshake if server key exchange message is omitted for ephemeral
7597 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7598 reporting this issue.
7603 * Remove non-export ephemeral RSA code on client and server. This code
7604 violated the TLS standard by allowing the use of temporary RSA keys in
7605 non-export ciphersuites and could be used by a server to effectively
7606 downgrade the RSA key length used to a value smaller than the server
7607 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7608 INRIA or reporting this issue.
7613 * Fixed issue where DH client certificates are accepted without verification.
7614 An OpenSSL server will accept a DH certificate for client authentication
7615 without the certificate verify message. This effectively allows a client to
7616 authenticate without the use of a private key. This only affects servers
7617 which trust a client certificate authority which issues certificates
7618 containing DH keys: these are extremely rare and hardly ever encountered.
7619 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7625 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7626 results on some platforms, including x86_64. This bug occurs at random
7627 with a very low probability, and is not known to be exploitable in any
7628 way, though its exact impact is difficult to determine. Thanks to Pieter
7629 Wuille (Blockstream) who reported this issue and also suggested an initial
7630 fix. Further analysis was conducted by the OpenSSL development team and
7631 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7632 the OpenSSL core team.
7637 * Fix various certificate fingerprint issues.
7639 By using non-DER or invalid encodings outside the signed portion of a
7640 certificate the fingerprint can be changed without breaking the signature.
7641 Although no details of the signed portion of the certificate can be changed
7642 this can cause problems with some applications: e.g. those using the
7643 certificate fingerprint for blacklists.
7645 1. Reject signatures with non zero unused bits.
7647 If the BIT STRING containing the signature has non zero unused bits reject
7648 the signature. All current signature algorithms require zero unused bits.
7650 2. Check certificate algorithm consistency.
7652 Check the AlgorithmIdentifier inside TBS matches the one in the
7653 certificate signature. NB: this will result in signature failure
7654 errors for some broken certificates.
7656 Thanks to Konrad Kraszewski from Google for reporting this issue.
7658 3. Check DSA/ECDSA signatures use DER.
7660 Reencode DSA/ECDSA signatures and compare with the original received
7661 signature. Return an error if there is a mismatch.
7663 This will reject various cases including garbage after signature
7664 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7665 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7666 (negative or with leading zeroes).
7668 Further analysis was conducted and fixes were developed by Stephen Henson
7669 of the OpenSSL core team.
7675 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7677 * Session Ticket Memory Leak.
7679 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7680 integrity of that ticket is first verified. In the event of a session
7681 ticket integrity check failing, OpenSSL will fail to free memory
7682 causing a memory leak. By sending a large number of invalid session
7683 tickets an attacker could exploit this issue in a Denial Of Service
7689 * Build option no-ssl3 is incomplete.
7691 When OpenSSL is configured with "no-ssl3" as a build option, servers
7692 could accept and complete a SSL 3.0 handshake, and clients could be
7693 configured to send them.
7696 *Akamai and the OpenSSL team*
7698 * Add support for TLS_FALLBACK_SCSV.
7699 Client applications doing fallback retries should call
7700 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7703 *Adam Langley, Bodo Moeller*
7705 * Add additional DigestInfo checks.
7707 Reencode DigestInto in DER and check against the original when
7708 verifying RSA signature: this will reject any improperly encoded
7709 DigestInfo structures.
7711 Note: this is a precautionary measure and no attacks are currently known.
7715 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7717 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7718 to a denial of service attack. A malicious server can crash the client
7719 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7720 ciphersuite and sending carefully crafted handshake messages.
7722 Thanks to Felix Gröbert (Google) for discovering and researching this
7728 * By sending carefully crafted DTLS packets an attacker could cause openssl
7729 to leak memory. This can be exploited through a Denial of Service attack.
7730 Thanks to Adam Langley for discovering and researching this issue.
7735 * An attacker can force openssl to consume large amounts of memory whilst
7736 processing DTLS handshake messages. This can be exploited through a
7737 Denial of Service attack.
7738 Thanks to Adam Langley for discovering and researching this issue.
7743 * An attacker can force an error condition which causes openssl to crash
7744 whilst processing DTLS packets due to memory being freed twice. This
7745 can be exploited through a Denial of Service attack.
7746 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7752 * If a multithreaded client connects to a malicious server using a resumed
7753 session and the server sends an ec point format extension it could write
7754 up to 255 bytes to freed memory.
7756 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7762 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7763 X509_name_oneline, X509_name_print_ex et al. to leak some information
7764 from the stack. Applications may be affected if they echo pretty printing
7765 output to the attacker.
7767 Thanks to Ivan Fratric (Google) for discovering this issue.
7770 *Emilia Käsper, and Steve Henson*
7772 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7773 for corner cases. (Certain input points at infinity could lead to
7774 bogus results, with non-infinity inputs mapped to infinity too.)
7778 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7780 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7781 handshake can force the use of weak keying material in OpenSSL
7782 SSL/TLS clients and servers.
7784 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7785 researching this issue. ([CVE-2014-0224])
7787 *KIKUCHI Masashi, Steve Henson*
7789 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7790 OpenSSL DTLS client the code can be made to recurse eventually crashing
7793 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7796 *Imre Rad, Steve Henson*
7798 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7799 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7800 client or server. This is potentially exploitable to run arbitrary
7801 code on a vulnerable client or server.
7803 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7805 *Jüri Aedla, Steve Henson*
7807 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7808 are subject to a denial of service attack.
7810 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7811 this issue. ([CVE-2014-3470])
7813 *Felix Gröbert, Ivan Fratric, Steve Henson*
7815 * Harmonize version and its documentation. -f flag is used to display
7818 *mancha <mancha1@zoho.com>*
7820 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7821 in i2d_ECPrivateKey.
7823 *mancha <mancha1@zoho.com>*
7825 * Fix some double frees. These are not thought to be exploitable.
7827 *mancha <mancha1@zoho.com>*
7829 * Fix for the attack described in the paper "Recovering OpenSSL
7830 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7831 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7832 <http://eprint.iacr.org/2014/140>
7834 Thanks to Yuval Yarom and Naomi Benger for discovering this
7835 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7837 *Yuval Yarom and Naomi Benger*
7839 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7841 * Keep original DTLS digest and encryption contexts in retransmission
7842 structures so we can use the previous session parameters if they need
7843 to be resent. ([CVE-2013-6450])
7847 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7848 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7849 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7850 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7851 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7852 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7854 *Rob Stradling, Adam Langley*
7856 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7858 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7860 This addresses the flaw in CBC record processing discovered by
7861 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7862 at: <http://www.isg.rhul.ac.uk/tls/>
7864 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7865 Security Group at Royal Holloway, University of London
7866 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7867 Emilia Käsper for the initial patch.
7870 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7872 * Return an error when checking OCSP signatures when key is NULL.
7873 This fixes a DoS attack. ([CVE-2013-0166])
7877 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7878 the right response is stapled. Also change SSL_get_certificate()
7879 so it returns the certificate actually sent.
7880 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7881 (This is a backport)
7883 *Rob Stradling <rob.stradling@comodo.com>*
7885 * Fix possible deadlock when decoding public keys.
7889 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7891 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7894 * Sanity check record length before skipping explicit IV in DTLS
7897 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7898 fuzzing as a service testing platform.
7903 * Initialise tkeylen properly when encrypting CMS messages.
7904 Thanks to Solar Designer of Openwall for reporting this issue.
7908 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7910 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7911 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7912 in CRYPTO_realloc_clean.
7914 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7915 issue and to Adam Langley <agl@chromium.org> for fixing it.
7918 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7920 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7922 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7923 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7924 content decryption and always return the same error. Note: this attack
7925 needs on average 2^20 messages so it only affects automated senders. The
7926 old behaviour can be re-enabled in the CMS code by setting the
7927 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7928 an MMA defence is not necessary.
7929 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7930 this issue. ([CVE-2012-0884])
7934 * Fix CVE-2011-4619: make sure we really are receiving a
7935 client hello before rejecting multiple SGC restarts. Thanks to
7936 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7940 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7942 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7943 Thanks to Antonio Martin, Enterprise Secure Access Research and
7944 Development, Cisco Systems, Inc. for discovering this bug and
7945 preparing a fix. ([CVE-2012-0050])
7949 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7951 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7952 of the Vaudenay padding oracle attack on CBC mode encryption
7953 which enables an efficient plaintext recovery attack against
7954 the OpenSSL implementation of DTLS. Their attack exploits timing
7955 differences arising during decryption processing. A research
7956 paper describing this attack can be found at:
7957 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7958 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7959 Security Group at Royal Holloway, University of London
7960 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7961 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7962 for preparing the fix. ([CVE-2011-4108])
7964 *Robin Seggelmann, Michael Tuexen*
7966 * Clear bytes used for block padding of SSL 3.0 records.
7969 *Adam Langley (Google)*
7971 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7972 Kadianakis <desnacked@gmail.com> for discovering this issue and
7973 Adam Langley for preparing the fix. ([CVE-2011-4619])
7975 *Adam Langley (Google)*
7977 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7979 *Andrey Kulikov <amdeich@gmail.com>*
7981 * Prevent malformed RFC3779 data triggering an assertion failure.
7982 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7983 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7985 *Rob Austein <sra@hactrn.net>*
7987 * Improved PRNG seeding for VOS.
7989 *Paul Green <Paul.Green@stratus.com>*
7991 * Fix ssl_ciph.c set-up race.
7993 *Adam Langley (Google)*
7995 * Fix spurious failures in ecdsatest.c.
7997 *Emilia Käsper (Google)*
7999 * Fix the BIO_f_buffer() implementation (which was mixing different
8000 interpretations of the `..._len` fields).
8002 *Adam Langley (Google)*
8004 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8005 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8006 threads won't reuse the same blinding coefficients.
8008 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8009 lock to call BN_BLINDING_invert_ex, and avoids one use of
8010 BN_BLINDING_update for each BN_BLINDING structure (previously,
8011 the last update always remained unused).
8013 *Emilia Käsper (Google)*
8015 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8017 *Bob Buckholz (Google)*
8019 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8021 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8022 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8024 *Kaspar Brand <ossl@velox.ch>*
8026 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8027 for multi-threaded use of ECDH. ([CVE-2011-3210])
8029 *Adam Langley (Google)*
8031 * Fix x509_name_ex_d2i memory leak on bad inputs.
8035 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8036 signature public key algorithm by using OID xref utilities instead.
8037 Before this you could only use some ECC ciphersuites with SHA1 only.
8041 * Add protection against ECDSA timing attacks as mentioned in the paper
8042 by Billy Bob Brumley and Nicola Tuveri, see:
8043 <http://eprint.iacr.org/2011/232.pdf>
8045 *Billy Bob Brumley and Nicola Tuveri*
8047 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8049 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8051 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8053 * Fix bug in string printing code: if *any* escaping is enabled we must
8054 escape the escape character (backslash) or the resulting string is
8059 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8061 * Disable code workaround for ancient and obsolete Netscape browsers
8062 and servers: an attacker can use it in a ciphersuite downgrade attack.
8063 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8067 * Fixed J-PAKE implementation error, originally discovered by
8068 Sebastien Martini, further info and confirmation from Stefan
8069 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8073 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8075 * Fix extension code to avoid race conditions which can result in a buffer
8076 overrun vulnerability: resumed sessions must not be modified as they can
8077 be shared by multiple threads. CVE-2010-3864
8081 * Fix WIN32 build system to correctly link an ENGINE directory into
8086 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8088 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8091 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8093 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8095 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8096 context. The operation can be customised via the ctrl mechanism in
8097 case ENGINEs want to include additional functionality.
8101 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8105 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8106 output hashes compatible with older versions of OpenSSL.
8108 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8110 * Fix compression algorithm handling: if resuming a session use the
8111 compression algorithm of the resumed session instead of determining
8112 it from client hello again. Don't allow server to change algorithm.
8116 * Add load_crls() function to commands tidying load_certs() too. Add option
8117 to verify utility to allow additional CRLs to be included.
8121 * Update OCSP request code to permit adding custom headers to the request:
8122 some responders need this.
8126 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8129 *Julia Lawall <julia@diku.dk>*
8131 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8132 needlessly dereferenced structures, used obsolete functions and
8133 didn't handle all updated verify codes correctly.
8137 * Disable MD2 in the default configuration.
8141 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8142 indicate the initial BIO being pushed or popped. This makes it possible
8143 to determine whether the BIO is the one explicitly called or as a result
8144 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8145 it handles reference counts correctly and doesn't zero out the I/O bio
8146 when it is not being explicitly popped. WARNING: applications which
8147 included workarounds for the old buggy behaviour will need to be modified
8148 or they could free up already freed BIOs.
8152 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8153 renaming to all platforms (within the 0.9.8 branch, this was
8154 done conditionally on Netware platforms to avoid a name clash).
8156 *Guenter <lists@gknw.net>*
8158 * Add ECDHE and PSK support to DTLS.
8160 *Michael Tuexen <tuexen@fh-muenster.de>*
8162 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8167 * Add "missing" function EVP_MD_flags() (without this the only way to
8168 retrieve a digest flags is by accessing the structure directly. Update
8169 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8170 or cipher is registered as in the "from" argument. Print out all
8171 registered digests in the dgst usage message instead of manually
8172 attempting to work them out.
8176 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8177 this allows the use of compression and extensions. Change default cipher
8178 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8179 by default unless an application cipher string requests it.
8183 * Alter match criteria in PKCS12_parse(). It used to try to use local
8184 key ids to find matching certificates and keys but some PKCS#12 files
8185 don't follow the (somewhat unwritten) rules and this strategy fails.
8186 Now just gather all certificates together and the first private key
8187 then look for the first certificate that matches the key.
8191 * Support use of registered digest and cipher names for dgst and cipher
8192 commands instead of having to add each one as a special case. So now
8199 openssl dgst -sha256 foo
8201 and this works for ENGINE based algorithms too.
8205 * Update Gost ENGINE to support parameter files.
8207 *Victor B. Wagner <vitus@cryptocom.ru>*
8209 * Support GeneralizedTime in ca utility.
8211 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8213 * Enhance the hash format used for certificate directory links. The new
8214 form uses the canonical encoding (meaning equivalent names will work
8215 even if they aren't identical) and uses SHA1 instead of MD5. This form
8216 is incompatible with the older format and as a result c_rehash should
8217 be used to rebuild symbolic links.
8221 * Make PKCS#8 the default write format for private keys, replacing the
8222 traditional format. This form is standardised, more secure and doesn't
8223 include an implicit MD5 dependency.
8227 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8228 committed to OpenSSL should pass this lot as a minimum.
8232 * Add session ticket override functionality for use by EAP-FAST.
8234 *Jouni Malinen <j@w1.fi>*
8236 * Modify HMAC functions to return a value. Since these can be implemented
8237 in an ENGINE errors can occur.
8241 * Type-checked OBJ_bsearch_ex.
8245 * Type-checked OBJ_bsearch. Also some constification necessitated
8246 by type-checking. Still to come: TXT_DB, bsearch(?),
8247 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8252 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8253 seconds to a tm structure directly, instead of going through OS
8254 specific date routines. This avoids any issues with OS routines such
8255 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8256 and X509_time_adj_ex() to cover the extended range. The existing
8257 X509_time_adj() is still usable and will no longer have any date issues.
8261 * Delta CRL support. New use deltas option which will attempt to locate
8262 and search any appropriate delta CRLs available.
8264 This work was sponsored by Google.
8268 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8269 code and add additional score elements. Validate alternate CRL paths
8270 as part of the CRL checking and indicate a new error "CRL path validation
8271 error" in this case. Applications wanting additional details can use
8272 the verify callback and check the new "parent" field. If this is not
8273 NULL CRL path validation is taking place. Existing applications won't
8274 see this because it requires extended CRL support which is off by
8277 This work was sponsored by Google.
8281 * Support for freshest CRL extension.
8283 This work was sponsored by Google.
8287 * Initial indirect CRL support. Currently only supported in the CRLs
8288 passed directly and not via lookup. Process certificate issuer
8289 CRL entry extension and lookup CRL entries by bother issuer name
8290 and serial number. Check and process CRL issuer entry in IDP extension.
8292 This work was sponsored by Google.
8296 * Add support for distinct certificate and CRL paths. The CRL issuer
8297 certificate is validated separately in this case. Only enabled if
8298 an extended CRL support flag is set: this flag will enable additional
8299 CRL functionality in future.
8301 This work was sponsored by Google.
8305 * Add support for policy mappings extension.
8307 This work was sponsored by Google.
8311 * Fixes to pathlength constraint, self issued certificate handling,
8312 policy processing to align with RFC3280 and PKITS tests.
8314 This work was sponsored by Google.
8318 * Support for name constraints certificate extension. DN, email, DNS
8319 and URI types are currently supported.
8321 This work was sponsored by Google.
8325 * To cater for systems that provide a pointer-based thread ID rather
8326 than numeric, deprecate the current numeric thread ID mechanism and
8327 replace it with a structure and associated callback type. This
8328 mechanism allows a numeric "hash" to be extracted from a thread ID in
8329 either case, and on platforms where pointers are larger than 'long',
8330 mixing is done to help ensure the numeric 'hash' is usable even if it
8331 can't be guaranteed unique. The default mechanism is to use "&errno"
8332 as a pointer-based thread ID to distinguish between threads.
8334 Applications that want to provide their own thread IDs should now use
8335 CRYPTO_THREADID_set_callback() to register a callback that will call
8336 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8338 Note that ERR_remove_state() is now deprecated, because it is tied
8339 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8340 to free the current thread's error state should be replaced by
8341 ERR_remove_thread_state(NULL).
8343 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8344 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8345 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8346 application was previously providing a numeric thread callback that
8347 was inappropriate for distinguishing threads, then uniqueness might
8348 have been obtained with &errno that happened immediately in the
8349 intermediate development versions of OpenSSL; this is no longer the
8350 case, the numeric thread callback will now override the automatic use
8353 *Geoff Thorpe, with help from Bodo Moeller*
8355 * Initial support for different CRL issuing certificates. This covers a
8356 simple case where the self issued certificates in the chain exist and
8357 the real CRL issuer is higher in the existing chain.
8359 This work was sponsored by Google.
8363 * Removed effectively defunct crypto/store from the build.
8367 * Revamp of STACK to provide stronger type-checking. Still to come:
8368 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8369 ASN1_STRING, CONF_VALUE.
8373 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8374 RAM on SSL connections. This option can save about 34k per idle SSL.
8378 * Revamp of LHASH to provide stronger type-checking. Still to come:
8379 STACK, TXT_DB, bsearch, qsort.
8383 * Initial support for Cryptographic Message Syntax (aka CMS) based
8384 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8385 support for data, signedData, compressedData, digestedData and
8386 encryptedData, envelopedData types included. Scripts to check against
8387 RFC4134 examples draft and interop and consistency checks of many
8388 content types and variants.
8392 * Add options to enc utility to support use of zlib compression BIO.
8396 * Extend mk1mf to support importing of options and assembly language
8397 files from Configure script, currently only included in VC-WIN32.
8398 The assembly language rules can now optionally generate the source
8399 files from the associated perl scripts.
8403 * Implement remaining functionality needed to support GOST ciphersuites.
8404 Interop testing has been performed using CryptoPro implementations.
8406 *Victor B. Wagner <vitus@cryptocom.ru>*
8408 * s390x assembler pack.
8412 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8417 * Implement Opaque PRF Input TLS extension as specified in
8418 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8419 official specification yet and no extension type assignment by
8420 IANA exists, this extension (for now) will have to be explicitly
8421 enabled when building OpenSSL by providing the extension number
8422 to use. For example, specify an option
8424 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8426 to the "config" or "Configure" script to enable the extension,
8427 assuming extension number 0x9527 (which is a completely arbitrary
8428 and unofficial assignment based on the MD5 hash of the Internet
8429 Draft). Note that by doing so, you potentially lose
8430 interoperability with other TLS implementations since these might
8431 be using the same extension number for other purposes.
8433 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8434 opaque PRF input value to use in the handshake. This will create
8435 an internal copy of the length-'len' string at 'src', and will
8436 return non-zero for success.
8438 To get more control and flexibility, provide a callback function
8441 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8442 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8446 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8449 Callback function 'cb' will be called in handshakes, and is
8450 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8451 Argument 'arg' is for application purposes (the value as given to
8452 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8453 be provided to the callback function). The callback function
8454 has to return non-zero to report success: usually 1 to use opaque
8455 PRF input just if possible, or 2 to enforce use of the opaque PRF
8456 input. In the latter case, the library will abort the handshake
8457 if opaque PRF input is not successfully negotiated.
8459 Arguments 'peerinput' and 'len' given to the callback function
8460 will always be NULL and 0 in the case of a client. A server will
8461 see the client's opaque PRF input through these variables if
8462 available (NULL and 0 otherwise). Note that if the server
8463 provides an opaque PRF input, the length must be the same as the
8464 length of the client's opaque PRF input.
8466 Note that the callback function will only be called when creating
8467 a new session (session resumption can resume whatever was
8468 previously negotiated), and will not be called in SSL 2.0
8469 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8470 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8471 for applications that need to enforce opaque PRF input.
8475 * Update ssl code to support digests other than SHA1+MD5 for handshake
8478 *Victor B. Wagner <vitus@cryptocom.ru>*
8480 * Add RFC4507 support to OpenSSL. This includes the corrections in
8481 RFC4507bis. The encrypted ticket format is an encrypted encoded
8482 SSL_SESSION structure, that way new session features are automatically
8485 If a client application caches session in an SSL_SESSION structure
8486 support is transparent because tickets are now stored in the encoded
8489 The SSL_CTX structure automatically generates keys for ticket
8490 protection in servers so again support should be possible
8491 with no application modification.
8493 If a client or server wishes to disable RFC4507 support then the option
8494 SSL_OP_NO_TICKET can be set.
8496 Add a TLS extension debugging callback to allow the contents of any client
8497 or server extensions to be examined.
8499 This work was sponsored by Google.
8503 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8504 OpenSSL should now compile cleanly on gcc 4.2
8506 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8508 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8509 support including streaming MAC support: this is required for GOST
8510 ciphersuite support.
8512 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8514 * Add option -stream to use PKCS#7 streaming in smime utility. New
8515 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8516 to output in BER and PEM format.
8520 * Experimental support for use of HMAC via EVP_PKEY interface. This
8521 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8522 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8523 ENGINE support for HMAC keys which are unextractable. New -mac and
8524 -macopt options to dgst utility.
8528 * New option -sigopt to dgst utility. Update dgst to use
8529 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8530 alternative signing parameters such as X9.31 or PSS in the dgst
8535 * Change ssl_cipher_apply_rule(), the internal function that does
8536 the work each time a ciphersuite string requests enabling
8537 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8538 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8539 the order of disabled ciphersuites such that those ciphersuites
8540 that most recently went from enabled to disabled not only stay
8541 in order with respect to each other, but also have higher priority
8542 than other disabled ciphersuites the next time ciphersuites are
8545 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8546 the same ciphersuites as with "HIGH" alone, but in a specific
8547 order where the PSK ciphersuites come first (since they are the
8548 most recently disabled ciphersuites when "HIGH" is parsed).
8550 Also, change ssl_create_cipher_list() (using this new
8551 functionality) such that between otherwise identical
8552 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8557 * Change ssl_create_cipher_list() so that it automatically
8558 arranges the ciphersuites in reasonable order before starting
8559 to process the rule string. Thus, the definition for "DEFAULT"
8560 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8561 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8562 This makes it much easier to arrive at a reasonable default order
8563 in applications for which anonymous ciphers are OK (meaning
8564 that you can't actually use DEFAULT).
8566 *Bodo Moeller; suggested by Victor Duchovni*
8568 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8569 processing) into multiple integers instead of setting
8570 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8571 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8572 (These masks as well as the individual bit definitions are hidden
8573 away into the non-exported interface ssl/ssl_locl.h, so this
8574 change to the definition of the SSL_CIPHER structure shouldn't
8575 affect applications.) This give us more bits for each of these
8576 categories, so there is no longer a need to coagulate AES128 and
8577 AES256 into a single algorithm bit, and to coagulate Camellia128
8578 and Camellia256 into a single algorithm bit, which has led to all
8581 Thus, among other things, the kludge introduced in 0.9.7m and
8582 0.9.8e for masking out AES256 independently of AES128 or masking
8583 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8585 With the change, we also introduce new ciphersuite aliases that
8586 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8591 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8592 Use the leftmost N bytes of the signature input if the input is
8593 larger than the prime q (with N being the size in bytes of q).
8597 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8598 it yet and it is largely untested.
8602 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8606 * Initial incomplete changes to avoid need for function casts in OpenSSL
8607 some compilers (gcc 4.2 and later) reject their use. Safestack is
8608 reimplemented. Update ASN1 to avoid use of legacy functions.
8612 * Win32/64 targets are linked with Winsock2.
8616 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8617 to external functions. This can be used to increase CRL handling
8618 efficiency especially when CRLs are very large by (for example) storing
8619 the CRL revoked certificates in a database.
8623 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8624 new CRLs added to a directory can be used. New command line option
8625 -verify_return_error to s_client and s_server. This causes real errors
8626 to be returned by the verify callback instead of carrying on no matter
8627 what. This reflects the way a "real world" verify callback would behave.
8631 * GOST engine, supporting several GOST algorithms and public key formats.
8632 Kindly donated by Cryptocom.
8636 * Partial support for Issuing Distribution Point CRL extension. CRLs
8637 partitioned by DP are handled but no indirect CRL or reason partitioning
8638 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8639 selected via a scoring technique which handles IDP and AKID in CRLs.
8643 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8644 will ultimately be used for all verify operations: this will remove the
8645 X509_STORE dependency on certificate verification and allow alternative
8646 lookup methods. X509_STORE based implementations of these two callbacks.
8650 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8651 Modify get_crl() to find a valid (unexpired) CRL if possible.
8655 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8656 this would be called X509_CRL_cmp() but that name is already used by
8657 a function that just compares CRL issuer names. Cache several CRL
8658 extensions in X509_CRL structure and cache CRLDP in X509.
8662 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8663 this maps equivalent X509_NAME structures into a consistent structure.
8664 Name comparison can then be performed rapidly using memcmp().
8668 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8673 * Allow digests to supply their own micalg string for S/MIME type using
8674 the ctrl EVP_MD_CTRL_MICALG.
8678 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8679 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8680 ctrl. It can then customise the structure before and/or after signing
8685 * New function OBJ_add_sigid() to allow application defined signature OIDs
8686 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8687 to free up any added signature OIDs.
8691 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8692 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8693 digest and cipher tables. New options added to openssl utility:
8694 list-message-digest-algorithms and list-cipher-algorithms.
8698 * Change the array representation of binary polynomials: the list
8699 of degrees of non-zero coefficients is now terminated with -1.
8700 Previously it was terminated with 0, which was also part of the
8701 value; thus, the array representation was not applicable to
8702 polynomials where t^0 has coefficient zero. This change makes
8703 the array representation useful in a more general context.
8707 * Various modifications and fixes to SSL/TLS cipher string
8708 handling. For ECC, the code now distinguishes between fixed ECDH
8709 with RSA certificates on the one hand and with ECDSA certificates
8710 on the other hand, since these are separate ciphersuites. The
8711 unused code for Fortezza ciphersuites has been removed.
8713 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8714 (not "ECDHE"). For consistency with the code for DH
8715 certificates, use of ECDH certificates is now considered ECDH
8716 authentication, not RSA or ECDSA authentication (the latter is
8717 merely the CA's signing algorithm and not actively used in the
8720 The temporary ciphersuite alias "ECCdraft" is no longer
8721 available, and ECC ciphersuites are no longer excluded from "ALL"
8722 and "DEFAULT". The following aliases now exist for RFC 4492
8723 ciphersuites, most of these by analogy with the DH case:
8725 kECDHr - ECDH cert, signed with RSA
8726 kECDHe - ECDH cert, signed with ECDSA
8727 kECDH - ECDH cert (signed with either RSA or ECDSA)
8728 kEECDH - ephemeral ECDH
8729 ECDH - ECDH cert or ephemeral ECDH
8735 AECDH - anonymous ECDH
8736 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8740 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8741 Use correct micalg parameters depending on digest(s) in signed message.
8745 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8746 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8750 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8751 an engine to register a method. Add ENGINE lookups for methods and
8752 functional reference processing.
8756 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8757 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8762 * New -resign option to smime utility. This adds one or more signers
8763 to an existing PKCS#7 signedData structure. Also -md option to use an
8764 alternative message digest algorithm for signing.
8768 * Tidy up PKCS#7 routines and add new functions to make it easier to
8769 create PKCS7 structures containing multiple signers. Update smime
8770 application to support multiple signers.
8774 * New -macalg option to pkcs12 utility to allow setting of an alternative
8779 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8780 Reorganize PBE internals to lookup from a static table using NIDs,
8781 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8782 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8783 PRF which will be automatically used with PBES2.
8787 * Replace the algorithm specific calls to generate keys in "req" with the
8792 * Update PKCS#7 enveloped data routines to use new API. This is now
8793 supported by any public key method supporting the encrypt operation. A
8794 ctrl is added to allow the public key algorithm to examine or modify
8795 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8800 * Add a ctrl to asn1 method to allow a public key algorithm to express
8801 a default digest type to use. In most cases this will be SHA1 but some
8802 algorithms (such as GOST) need to specify an alternative digest. The
8803 return value indicates how strong the preference is 1 means optional and
8804 2 is mandatory (that is it is the only supported type). Modify
8805 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8806 use the default md. Update openssl utilities to use the default digest
8807 type for signing if it is not explicitly indicated.
8811 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8812 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8813 signing method from the key type. This effectively removes the link
8814 between digests and public key types.
8818 * Add an OID cross reference table and utility functions. Its purpose is to
8819 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8820 rsaEncryption. This will allow some of the algorithm specific hackery
8821 needed to use the correct OID to be removed.
8825 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8826 structures for PKCS7_sign(). They are now set up by the relevant public
8831 * Add provisional EC pkey method with support for ECDSA and ECDH.
8835 * Add support for key derivation (agreement) in the API, DH method and
8840 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8841 public and private key formats. As a side effect these add additional
8842 command line functionality not previously available: DSA signatures can be
8843 generated and verified using pkeyutl and DH key support and generation in
8850 *Oliver Tappe <zooey@hirschkaefer.de>*
8852 * New make target "install_html_docs" installs HTML renditions of the
8855 *Oliver Tappe <zooey@hirschkaefer.de>*
8857 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8858 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8859 support key and parameter generation and add initial key generation
8860 functionality for RSA.
8864 * Add functions for main EVP_PKEY_method operations. The undocumented
8865 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8866 `EVP_PKEY_{encrypt,decrypt}_old`.
8870 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8871 key API, doesn't do much yet.
8875 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8876 public key algorithms. New option to openssl utility:
8877 "list-public-key-algorithms" to print out info.
8881 * Implement the Supported Elliptic Curves Extension for
8882 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8886 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8887 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8891 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8892 utilities such as rsa, dsa, dsaparam etc except they process any key
8897 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8898 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8899 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8904 * Initial support for pluggable public key ASN1.
8905 De-spaghettify the public key ASN1 handling. Move public and private
8906 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8907 algorithm specific handling to a single module within the relevant
8908 algorithm directory. Add functions to allow (near) opaque processing
8909 of public and private key structures.
8913 * Implement the Supported Point Formats Extension for
8914 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8918 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8919 for the psk identity [hint] and the psk callback functions to the
8920 SSL_SESSION, SSL and SSL_CTX structure.
8923 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8927 SSL_CTX_use_psk_identity_hint
8928 SSL_get_psk_identity_hint
8929 SSL_get_psk_identity
8930 SSL_use_psk_identity_hint
8932 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8934 * Add RFC 3161 compliant time stamp request creation, response generation
8935 and response verification functionality.
8937 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8939 * Add initial support for TLS extensions, specifically for the server_name
8940 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8941 have new members for a host name. The SSL data structure has an
8942 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8943 stored in that context to allow for session resumption, even after the
8944 SSL has been switched to a new SSL_CTX in reaction to a client's
8945 server_name extension.
8947 New functions (subject to change):
8949 SSL_get_servername()
8950 SSL_get_servername_type()
8953 New CTRL codes and macros (subject to change):
8955 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8956 - SSL_CTX_set_tlsext_servername_callback()
8957 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8958 - SSL_CTX_set_tlsext_servername_arg()
8959 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8961 openssl s_client has a new '-servername ...' option.
8963 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8964 '-key2 ...', '-servername_fatal' (subject to change). This allows
8965 testing the HostName extension for a specific single host name ('-cert'
8966 and '-key' remain fallbacks for handshakes without HostName
8967 negotiation). If the unrecognized_name alert has to be sent, this by
8968 default is a warning; it becomes fatal with the '-servername_fatal'
8971 *Peter Sylvester, Remy Allais, Christophe Renou*
8973 * Whirlpool hash implementation is added.
8977 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8978 bn(64,32). Because of instruction set limitations it doesn't have
8979 any negative impact on performance. This was done mostly in order
8980 to make it possible to share assembler modules, such as bn_mul_mont
8981 implementations, between 32- and 64-bit builds without hassle.
8985 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8986 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8991 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8992 dedicated Montgomery multiplication procedure, is introduced.
8993 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8994 "64-bit" performance on certain 32-bit targets.
8998 * New option SSL_OP_NO_COMP to disable use of compression selectively
8999 in SSL structures. New SSL ctrl to set maximum send fragment size.
9000 Save memory by setting the I/O buffer sizes dynamically instead of
9001 using the maximum available value.
9005 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9006 in addition to the text details.
9010 * Very, very preliminary EXPERIMENTAL support for printing of general
9011 ASN1 structures. This currently produces rather ugly output and doesn't
9012 handle several customised structures at all.
9016 * Integrated support for PVK file format and some related formats such
9017 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9018 these in the 'rsa' and 'dsa' utilities.
9022 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9026 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9027 place for the (very old) "NETSCAPE" format certificates which are now
9028 handled using new ASN1 code equivalents.
9032 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9033 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9034 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9038 * Modify CRL distribution points extension code to print out previously
9039 unsupported fields. Enhance extension setting code to allow setting of
9044 * Add print and set support for Issuing Distribution Point CRL extension.
9048 * Change 'Configure' script to enable Camellia by default.
9055 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9057 * When rejecting SSL/TLS records due to an incorrect version number, never
9058 update s->server with a new major version number. As of
9059 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9060 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9061 the previous behavior could result in a read attempt at NULL when
9062 receiving specific incorrect SSL/TLS records once record payload
9063 protection is active. ([CVE-2010-0740])
9065 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9067 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9068 could be crashed if the relevant tables were not present (e.g. chrooted).
9070 *Tomas Hoger <thoger@redhat.com>*
9072 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9074 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9076 *Martin Olsson, Neel Mehta*
9078 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9079 accommodate for stack sorting, always a write lock!).
9083 * On some versions of WIN32 Heap32Next is very slow. This can cause
9084 excessive delays in the RAND_poll(): over a minute. As a workaround
9085 include a time check in the inner Heap32Next loop too.
9089 * The code that handled flushing of data in SSL/TLS originally used the
9090 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9091 the problem outlined in PR#1949. The fix suggested there however can
9092 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9093 of Apache). So instead simplify the code to flush unconditionally.
9094 This should be fine since flushing with no data to flush is a no op.
9098 * Handle TLS versions 2.0 and later properly and correctly use the
9099 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9100 off ancient servers have a habit of sticking around for a while...
9104 * Modify compression code so it frees up structures without using the
9105 ex_data callbacks. This works around a problem where some applications
9106 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9107 restarting) then use compression (e.g. SSL with compression) later.
9108 This results in significant per-connection memory leaks and
9109 has caused some security issues including CVE-2008-1678 and
9114 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9115 change when encrypting or decrypting.
9119 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9120 connect and renegotiate with servers which do not support RI.
9121 Until RI is more widely deployed this option is enabled by default.
9125 * Add "missing" ssl ctrls to clear options and mode.
9129 * If client attempts to renegotiate and doesn't support RI respond with
9130 a no_renegotiation alert as required by RFC5746. Some renegotiating
9131 TLS clients will continue a connection gracefully when they receive
9132 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9133 waiting for a server hello which it will never receive. Now we treat a
9134 received no_renegotiation alert as a fatal error. This is because
9135 applications requesting a renegotiation might well expect it to succeed
9136 and would have no code in place to handle the server denying it so the
9137 only safe thing to do is to terminate the connection.
9141 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9142 peer supports secure renegotiation and 0 otherwise. Print out peer
9143 renegotiation support in s_client/s_server.
9147 * Replace the highly broken and deprecated SPKAC certification method with
9148 the updated NID creation version. This should correctly handle UTF8.
9152 * Implement RFC5746. Re-enable renegotiation but require the extension
9153 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9154 turns out to be a bad idea. It has been replaced by
9155 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9156 SSL_CTX_set_options(). This is really not recommended unless you
9157 know what you are doing.
9159 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9161 * Fixes to stateless session resumption handling. Use initial_ctx when
9162 issuing and attempting to decrypt tickets in case it has changed during
9163 servername handling. Use a non-zero length session ID when attempting
9164 stateless session resumption: this makes it possible to determine if
9165 a resumption has occurred immediately after receiving server hello
9166 (several places in OpenSSL subtly assume this) instead of later in
9171 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9172 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9173 fixes for a few places where the return code is not checked
9176 *Julia Lawall <julia@diku.dk>*
9178 * Add --strict-warnings option to Configure script to include devteam
9179 warnings in other configurations.
9183 * Add support for --libdir option and LIBDIR variable in makefiles. This
9184 makes it possible to install openssl libraries in locations which
9185 have names other than "lib", for example "/usr/lib64" which some
9188 *Steve Henson, based on patch from Jeremy Utley*
9190 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9191 X690 8.9.12 and can produce some misleading textual output of OIDs.
9193 *Steve Henson, reported by Dan Kaminsky*
9195 * Delete MD2 from algorithm tables. This follows the recommendation in
9196 several standards that it is not used in new applications due to
9197 several cryptographic weaknesses. For binary compatibility reasons
9198 the MD2 API is still compiled in by default.
9202 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9207 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9208 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9211 *Guenter <lists@gknw.net>*
9213 * Fix the server certificate chain building code to use X509_verify_cert(),
9214 it used to have an ad-hoc builder which was unable to cope with anything
9215 other than a simple chain.
9217 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9219 * Don't check self signed certificate signatures in X509_verify_cert()
9220 by default (a flag can override this): it just wastes time without
9221 adding any security. As a useful side effect self signed root CAs
9222 with non-FIPS digests are now usable in FIPS mode.
9226 * In dtls1_process_out_of_seq_message() the check if the current message
9227 is already buffered was missing. For every new message was memory
9228 allocated, allowing an attacker to perform an denial of service attack
9229 with sending out of seq handshake messages until there is no memory
9230 left. Additionally every future message was buffered, even if the
9231 sequence number made no sense and would be part of another handshake.
9232 So only messages with sequence numbers less than 10 in advance will be
9233 buffered. ([CVE-2009-1378])
9235 *Robin Seggelmann, discovered by Daniel Mentz*
9237 * Records are buffered if they arrive with a future epoch to be
9238 processed after finishing the corresponding handshake. There is
9239 currently no limitation to this buffer allowing an attacker to perform
9240 a DOS attack with sending records with future epochs until there is no
9241 memory left. This patch adds the pqueue_size() function to determine
9242 the size of a buffer and limits the record buffer to 100 entries.
9245 *Robin Seggelmann, discovered by Daniel Mentz*
9247 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9248 parent structure is freed. ([CVE-2009-1379])
9252 * Handle non-blocking I/O properly in SSL_shutdown() call.
9254 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9256 * Add `2.5.4.*` OIDs
9258 *Ilya O. <vrghost@gmail.com>*
9260 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9262 * Disable renegotiation completely - this fixes a severe security
9263 problem ([CVE-2009-3555]) at the cost of breaking all
9264 renegotiation. Renegotiation can be re-enabled by setting
9265 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9266 run-time. This is really not recommended unless you know what
9271 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9273 * Don't set val to NULL when freeing up structures, it is freed up by
9274 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9275 zeroing past the valid field. ([CVE-2009-0789])
9277 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9279 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9280 checked correctly. This would allow some invalid signed attributes to
9281 appear to verify correctly. ([CVE-2009-0591])
9283 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9285 * Reject UniversalString and BMPString types with invalid lengths. This
9286 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9287 a legal length. ([CVE-2009-0590])
9291 * Set S/MIME signing as the default purpose rather than setting it
9292 unconditionally. This allows applications to override it at the store
9297 * Permit restricted recursion of ASN1 strings. This is needed in practice
9298 to handle some structures.
9302 * Improve efficiency of mem_gets: don't search whole buffer each time
9305 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9307 * New -hex option for openssl rand.
9311 * Print out UTF8String and NumericString when parsing ASN1.
9315 * Support NumericString type for name components.
9319 * Allow CC in the environment to override the automatically chosen
9320 compiler. Note that nothing is done to ensure flags work with the
9325 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9327 * Properly check EVP_VerifyFinal() and similar return values
9330 *Ben Laurie, Bodo Moeller, Google Security Team*
9332 * Enable TLS extensions by default.
9336 * Allow the CHIL engine to be loaded, whether the application is
9337 multithreaded or not. (This does not release the developer from the
9338 obligation to set up the dynamic locking callbacks.)
9340 *Sander Temme <sander@temme.net>*
9342 * Use correct exit code if there is an error in dgst command.
9344 *Steve Henson; problem pointed out by Roland Dirlewanger*
9346 * Tweak Configure so that you need to say "experimental-jpake" to enable
9347 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9351 * Add experimental JPAKE support, including demo authentication in
9352 s_client and s_server.
9356 * Set the comparison function in v3_addr_canonize().
9358 *Rob Austein <sra@hactrn.net>*
9360 * Add support for XMPP STARTTLS in s_client.
9362 *Philip Paeps <philip@freebsd.org>*
9364 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9365 to ensure that even with this option, only ciphersuites in the
9366 server's preference list will be accepted. (Note that the option
9367 applies only when resuming a session, so the earlier behavior was
9368 just about the algorithm choice for symmetric cryptography.)
9372 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9374 * Fix NULL pointer dereference if a DTLS server received
9375 ChangeCipherSpec as first record ([CVE-2009-1386]).
9379 * Fix a state transition in s3_srvr.c and d1_srvr.c
9380 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9384 * The fix in 0.9.8c that supposedly got rid of unsafe
9385 double-checked locking was incomplete for RSA blinding,
9386 addressing just one layer of what turns out to have been
9387 doubly unsafe triple-checked locking.
9389 So now fix this for real by retiring the MONT_HELPER macro
9390 in crypto/rsa/rsa_eay.c.
9392 *Bodo Moeller; problem pointed out by Marius Schilder*
9394 * Various precautionary measures:
9396 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9398 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9399 (NB: This would require knowledge of the secret session ticket key
9400 to exploit, in which case you'd be SOL either way.)
9402 - Change bn_nist.c so that it will properly handle input BIGNUMs
9403 outside the expected range.
9405 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9408 *Neel Mehta, Bodo Moeller*
9410 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9411 the load fails. Useful for distros.
9413 *Ben Laurie and the FreeBSD team*
9415 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9419 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9423 * Expand ENGINE to support engine supplied SSL client certificate functions.
9425 This work was sponsored by Logica.
9429 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9430 keystores. Support for SSL/TLS client authentication too.
9431 Not compiled unless enable-capieng specified to Configure.
9433 This work was sponsored by Logica.
9437 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9438 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9439 attribute creation routines such as certificate requests and PKCS#12
9444 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9446 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9447 handshake which could lead to a client crash as found using the
9448 Codenomicon TLS test suite ([CVE-2008-1672])
9450 *Steve Henson, Mark Cox*
9452 * Fix double free in TLS server name extensions which could lead to
9453 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9457 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9459 Clear the error queue to ensure that error entries left from
9460 older function calls do not interfere with the correct operation.
9462 *Lutz Jaenicke, Erik de Castro Lopo*
9464 * Remove root CA certificates of commercial CAs:
9466 The OpenSSL project does not recommend any specific CA and does not
9467 have any policy with respect to including or excluding any CA.
9468 Therefore it does not make any sense to ship an arbitrary selection
9469 of root CA certificates with the OpenSSL software.
9473 * RSA OAEP patches to fix two separate invalid memory reads.
9474 The first one involves inputs when 'lzero' is greater than
9475 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9476 before the beginning of from). The second one involves inputs where
9477 the 'db' section contains nothing but zeroes (there is a one-byte
9478 invalid read after the end of 'db').
9480 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9482 * Partial backport from 0.9.9-dev:
9484 Introduce bn_mul_mont (dedicated Montgomery multiplication
9485 procedure) as a candidate for BIGNUM assembler implementation.
9486 While 0.9.9-dev uses assembler for various architectures, only
9487 x86_64 is available by default here in the 0.9.8 branch, and
9488 32-bit x86 is available through a compile-time setting.
9490 To try the 32-bit x86 assembler implementation, use Configure
9491 option "enable-montasm" (which exists only for this backport).
9493 As "enable-montasm" for 32-bit x86 disclaims code stability
9494 anyway, in this constellation we activate additional code
9495 backported from 0.9.9-dev for further performance improvements,
9496 namely BN_from_montgomery_word. (To enable this otherwise,
9497 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9499 *Andy Polyakov (backport partially by Bodo Moeller)*
9501 * Add TLS session ticket callback. This allows an application to set
9502 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9503 values. This is useful for key rollover for example where several key
9504 sets may exist with different names.
9508 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9509 This was broken until now in 0.9.8 releases, such that the only way
9510 a registered ENGINE could be used (assuming it initialises
9511 successfully on the host) was to explicitly set it as the default
9512 for the relevant algorithms. This is in contradiction with 0.9.7
9513 behaviour and the documentation. With this fix, when an ENGINE is
9514 registered into a given algorithm's table of implementations, the
9515 'uptodate' flag is reset so that auto-discovery will be used next
9516 time a new context for that algorithm attempts to select an
9519 *Ian Lister (tweaked by Geoff Thorpe)*
9521 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9522 implementation in the following ways:
9524 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9527 Lack of BER streaming support means one pass streaming processing is
9528 only supported if data is detached: setting the streaming flag is
9529 ignored for embedded content.
9531 CMS support is disabled by default and must be explicitly enabled
9532 with the enable-cms configuration option.
9536 * Update the GMP engine glue to do direct copies between BIGNUM and
9537 mpz_t when openssl and GMP use the same limb size. Otherwise the
9538 existing "conversion via a text string export" trick is still used.
9540 *Paul Sheer <paulsheer@gmail.com>*
9542 * Zlib compression BIO. This is a filter BIO which compressed and
9543 uncompresses any data passed through it.
9547 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9548 RFC3394 compatible AES key wrapping.
9552 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9553 sets string data without copying. X509_ALGOR_set0() and
9554 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9555 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9556 from an X509_ATTRIBUTE structure optionally checking it occurs only
9557 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9562 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9563 to get the expected BN_FLG_CONSTTIME behavior.
9565 *Bodo Moeller (Google)*
9569 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9570 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9571 - added some more tests to do_tests.pl
9572 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9573 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9574 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9575 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9576 - various changes to netware.pl to enable gcc-cross builds on Win32
9578 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9579 - various changes to fix missing prototype warnings
9580 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9581 - added AES, WHIRLPOOL and CPUID assembler code to build files
9582 - added missing AES assembler make rules to mk1mf.pl
9583 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9585 *Guenter Knauf <eflash@gmx.net>*
9587 * Implement certificate status request TLS extension defined in RFC3546.
9588 A client can set the appropriate parameters and receive the encoded
9589 OCSP response via a callback. A server can query the supplied parameters
9590 and set the encoded OCSP response in the callback. Add simplified examples
9591 to s_client and s_server.
9595 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9598 + Binary incompatibility of ssl_ctx_st structure
9599 + DTLS interoperation with non-compliant servers
9600 + Don't call get_session_cb() without proposed session
9601 + Fix ia64 assembler code
9603 *Andy Polyakov, Steve Henson*
9605 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9607 * DTLS Handshake overhaul. There were longstanding issues with
9608 OpenSSL DTLS implementation, which were making it impossible for
9609 RFC 4347 compliant client to communicate with OpenSSL server.
9610 Unfortunately just fixing these incompatibilities would "cut off"
9611 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9612 server keeps tolerating non RFC compliant syntax. The opposite is
9613 not true, 0.9.8f client can not communicate with earlier server.
9614 This update even addresses CVE-2007-4995.
9618 * Changes to avoid need for function casts in OpenSSL: some compilers
9619 (gcc 4.2 and later) reject their use.
9620 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9623 * Add RFC4507 support to OpenSSL. This includes the corrections in
9624 RFC4507bis. The encrypted ticket format is an encrypted encoded
9625 SSL_SESSION structure, that way new session features are automatically
9628 If a client application caches session in an SSL_SESSION structure
9629 support is transparent because tickets are now stored in the encoded
9632 The SSL_CTX structure automatically generates keys for ticket
9633 protection in servers so again support should be possible
9634 with no application modification.
9636 If a client or server wishes to disable RFC4507 support then the option
9637 SSL_OP_NO_TICKET can be set.
9639 Add a TLS extension debugging callback to allow the contents of any client
9640 or server extensions to be examined.
9642 This work was sponsored by Google.
9646 * Add initial support for TLS extensions, specifically for the server_name
9647 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9648 have new members for a host name. The SSL data structure has an
9649 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9650 stored in that context to allow for session resumption, even after the
9651 SSL has been switched to a new SSL_CTX in reaction to a client's
9652 server_name extension.
9654 New functions (subject to change):
9656 SSL_get_servername()
9657 SSL_get_servername_type()
9660 New CTRL codes and macros (subject to change):
9662 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9663 - SSL_CTX_set_tlsext_servername_callback()
9664 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9665 - SSL_CTX_set_tlsext_servername_arg()
9666 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9668 openssl s_client has a new '-servername ...' option.
9670 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9671 '-key2 ...', '-servername_fatal' (subject to change). This allows
9672 testing the HostName extension for a specific single host name ('-cert'
9673 and '-key' remain fallbacks for handshakes without HostName
9674 negotiation). If the unrecognized_name alert has to be sent, this by
9675 default is a warning; it becomes fatal with the '-servername_fatal'
9678 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9680 * Add AES and SSE2 assembly language support to VC++ build.
9684 * Mitigate attack on final subtraction in Montgomery reduction.
9688 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9689 (which previously caused an internal error).
9693 * Squeeze another 10% out of IGE mode when in != out.
9697 * AES IGE mode speedup.
9699 *Dean Gaudet (Google)*
9701 * Add the Korean symmetric 128-bit cipher SEED (see
9702 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9703 add SEED ciphersuites from RFC 4162:
9705 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9706 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9707 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9708 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9710 To minimize changes between patchlevels in the OpenSSL 0.9.8
9711 series, SEED remains excluded from compilation unless OpenSSL
9712 is configured with 'enable-seed'.
9714 *KISA, Bodo Moeller*
9716 * Mitigate branch prediction attacks, which can be practical if a
9717 single processor is shared, allowing a spy process to extract
9718 information. For detailed background information, see
9719 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9720 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9721 and Necessary Software Countermeasures"). The core of the change
9722 are new versions BN_div_no_branch() and
9723 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9724 respectively, which are slower, but avoid the security-relevant
9725 conditional branches. These are automatically called by BN_div()
9726 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9727 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9728 remove a conditional branch.
9730 BN_FLG_CONSTTIME is the new name for the previous
9731 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9732 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9733 in the exponent causes BN_mod_exp_mont() to use the alternative
9734 implementation in BN_mod_exp_mont_consttime().) The old name
9735 remains as a deprecated alias.
9737 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9738 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9739 constant-time implementations for more than just exponentiation.
9740 Here too the old name is kept as a deprecated alias.
9742 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9743 the BN_BLINDING structure gets an independent copy of the
9744 modulus. This means that the previous `BIGNUM *m` argument to
9745 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9746 essentially becomes `const BIGNUM *m`, although we can't actually
9747 change this in the header file before 0.9.9. It allows
9748 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9749 enable BN_FLG_CONSTTIME.
9751 *Matthew D Wood (Intel Corp)*
9753 * In the SSL/TLS server implementation, be strict about session ID
9754 context matching (which matters if an application uses a single
9755 external cache for different purposes). Previously,
9756 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9757 set. This did ensure strict client verification, but meant that,
9758 with applications using a single external cache for quite
9759 different requirements, clients could circumvent ciphersuite
9760 restrictions for a given session ID context by starting a session
9761 in a different context.
9765 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9766 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9767 authentication-only ciphersuites.
9771 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9772 not complete and could lead to a possible single byte overflow
9773 ([CVE-2007-5135]) [Ben Laurie]
9775 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9777 * Since AES128 and AES256 (and similarly Camellia128 and
9778 Camellia256) share a single mask bit in the logic of
9779 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9780 kludge to work properly if AES128 is available and AES256 isn't
9781 (or if Camellia128 is available and Camellia256 isn't).
9785 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9786 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9787 When a point or a seed is encoded in a BIT STRING, we need to
9788 prevent the removal of trailing zero bits to get the proper DER
9789 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9790 of a NamedBitList, for which trailing 0 bits need to be removed.)
9794 * Have SSL/TLS server implementation tolerate "mismatched" record
9795 protocol version while receiving ClientHello even if the
9796 ClientHello is fragmented. (The server can't insist on the
9797 particular protocol version it has chosen before the ServerHello
9798 message has informed the client about his choice.)
9802 * Add RFC 3779 support.
9804 *Rob Austein for ARIN, Ben Laurie*
9806 * Load error codes if they are not already present instead of using a
9807 static variable. This allows them to be cleanly unloaded and reloaded.
9808 Improve header file function name parsing.
9812 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9813 or CAPABILITY handshake as required by RFCs.
9817 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9819 * Introduce limits to prevent malicious keys being able to
9820 cause a denial of service. ([CVE-2006-2940])
9822 *Steve Henson, Bodo Moeller*
9824 * Fix ASN.1 parsing of certain invalid structures that can result
9825 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9827 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9828 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9830 * Fix SSL client code which could crash if connecting to a
9831 malicious SSLv2 server. ([CVE-2006-4343])
9833 *Tavis Ormandy and Will Drewry, Google Security Team*
9835 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9836 match only those. Before that, "AES256-SHA" would be interpreted
9837 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9838 the same strength classification in 0.9.7h) as we currently only
9839 have a single AES bit in the ciphersuite description bitmap.
9840 That change, however, also applied to ciphersuite strings such as
9841 "RC4-MD5" that intentionally matched multiple ciphersuites --
9842 namely, SSL 2.0 ciphersuites in addition to the more common ones
9843 from SSL 3.0/TLS 1.0.
9845 So we change the selection algorithm again: Naming an explicit
9846 ciphersuite selects this one ciphersuite, and any other similar
9847 ciphersuite (same bitmap) from *other* protocol versions.
9848 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9849 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9851 Since SSL 2.0 does not have any ciphersuites for which the
9852 128/256 bit distinction would be relevant, this works for now.
9853 The proper fix will be to use different bits for AES128 and
9854 AES256, which would have avoided the problems from the beginning;
9855 however, bits are scarce, so we can only do this in a new release
9856 (not just a patchlevel) when we can change the SSL_CIPHER
9857 definition to split the single 'unsigned long mask' bitmap into
9858 multiple values to extend the available space.
9862 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9864 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9865 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9867 * Add AES IGE and biIGE modes.
9871 * Change the Unix randomness entropy gathering to use poll() when
9872 possible instead of select(), since the latter has some
9873 undesirable limitations.
9875 *Darryl Miles via Richard Levitte and Bodo Moeller*
9877 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9878 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9879 cannot be implicitly activated as part of, e.g., the "AES" alias.
9880 However, please upgrade to OpenSSL 0.9.9[-dev] for
9881 non-experimental use of the ECC ciphersuites to get TLS extension
9882 support, which is required for curve and point format negotiation
9883 to avoid potential handshake problems.
9887 * Disable rogue ciphersuites:
9889 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9890 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9891 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9893 The latter two were purportedly from
9894 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9897 Also deactivate the remaining ciphersuites from
9898 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9899 unofficial, and the ID has long expired.
9903 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9904 dual-core machines) and other potential thread-safety issues.
9908 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9909 versions), which is now available for royalty-free use
9910 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9911 Also, add Camellia TLS ciphersuites from RFC 4132.
9913 To minimize changes between patchlevels in the OpenSSL 0.9.8
9914 series, Camellia remains excluded from compilation unless OpenSSL
9915 is configured with 'enable-camellia'.
9919 * Disable the padding bug check when compression is in use. The padding
9920 bug check assumes the first packet is of even length, this is not
9921 necessarily true if compression is enabled and can result in false
9922 positives causing handshake failure. The actual bug test is ancient
9923 code so it is hoped that implementations will either have fixed it by
9924 now or any which still have the bug do not support compression.
9928 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9930 * When applying a cipher rule check to see if string match is an explicit
9931 cipher suite and only match that one cipher suite if it is.
9935 * Link in manifests for VC++ if needed.
9937 *Austin Ziegler <halostatue@gmail.com>*
9939 * Update support for ECC-based TLS ciphersuites according to
9940 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9941 TLS extensions, which are supported starting with the 0.9.9
9942 branch, not in the OpenSSL 0.9.8 branch).
9946 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9947 opaque EVP_CIPHER_CTX handling.
9951 * Fixes and enhancements to zlib compression code. We now only use
9952 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9953 to conform with the standards mentioned here:
9954 <http://www.zlib.net/DLL_FAQ.txt>
9955 Static zlib linking now works on Windows and the new --with-zlib-include
9956 --with-zlib-lib options to Configure can be used to supply the location
9957 of the headers and library. Gracefully handle case where zlib library
9962 * Several fixes and enhancements to the OID generation code. The old code
9963 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9964 handle numbers larger than ULONG_MAX, truncated printing and had a
9965 non standard OBJ_obj2txt() behaviour.
9969 * Add support for building of engines under engine/ as shared libraries
9970 under VC++ build system.
9974 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9975 Hopefully, we will not see any false combination of paths any more.
9979 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9981 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9982 (part of SSL_OP_ALL). This option used to disable the
9983 countermeasure against man-in-the-middle protocol-version
9984 rollback in the SSL 2.0 server implementation, which is a bad
9985 idea. ([CVE-2005-2969])
9987 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9988 for Information Security, National Institute of Advanced Industrial
9989 Science and Technology [AIST], Japan)*
9991 * Add two function to clear and return the verify parameter flags.
9995 * Keep cipherlists sorted in the source instead of sorting them at
9996 runtime, thus removing the need for a lock.
10000 * Avoid some small subgroup attacks in Diffie-Hellman.
10002 *Nick Mathewson and Ben Laurie*
10004 * Add functions for well-known primes.
10008 * Extended Windows CE support.
10010 *Satoshi Nakamura and Andy Polyakov*
10012 * Initialize SSL_METHOD structures at compile time instead of during
10013 runtime, thus removing the need for a lock.
10017 * Make PKCS7_decrypt() work even if no certificate is supplied by
10018 attempting to decrypt each encrypted key in turn. Add support to
10023 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10025 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10028 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10032 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10033 key into the same file any more.
10037 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10041 * Add -utf8 command line and config file option to 'ca'.
10043 *Stefan <stf@udoma.org*
10045 * Removed the macro des_crypt(), as it seems to conflict with some
10046 libraries. Use DES_crypt().
10050 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10051 involves renaming the source and generated shared-libs for
10052 both. The engines will accept the corrected or legacy ids
10053 ('ncipher' and '4758_cca' respectively) when binding. NB,
10054 this only applies when building 'shared'.
10056 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10058 * Add attribute functions to EVP_PKEY structure. Modify
10059 PKCS12_create() to recognize a CSP name attribute and
10060 use it. Make -CSP option work again in pkcs12 utility.
10064 * Add new functionality to the bn blinding code:
10065 - automatic re-creation of the BN_BLINDING parameters after
10066 a fixed number of uses (currently 32)
10067 - add new function for parameter creation
10068 - introduce flags to control the update behaviour of the
10069 BN_BLINDING parameters
10070 - hide BN_BLINDING structure
10071 Add a second BN_BLINDING slot to the RSA structure to improve
10072 performance when a single RSA object is shared among several
10077 * Add support for DTLS.
10079 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10081 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10082 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10086 * Remove buggy and incomplete DH cert support from
10087 ssl/ssl_rsa.c and ssl/s3_both.c
10091 * Use SHA-1 instead of MD5 as the default digest algorithm for
10092 the `apps/openssl` commands.
10096 * Compile clean with "-Wall -Wmissing-prototypes
10097 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10098 DEBUG_SAFESTACK must also be set.
10102 * Change ./Configure so that certain algorithms can be disabled by default.
10103 The new counterpiece to "no-xxx" is "enable-xxx".
10105 The patented RC5 and MDC2 algorithms will now be disabled unless
10106 "enable-rc5" and "enable-mdc2", respectively, are specified.
10108 (IDEA remains enabled despite being patented. This is because IDEA
10109 is frequently required for interoperability, and there is no license
10110 fee for non-commercial use. As before, "no-idea" can be used to
10111 avoid this algorithm.)
10115 * Add processing of proxy certificates (see RFC 3820). This work was
10116 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10117 EGEE (Enabling Grids for E-science in Europe).
10121 * RC4 performance overhaul on modern architectures/implementations, such
10122 as Intel P4, IA-64 and AMD64.
10126 * New utility extract-section.pl. This can be used specify an alternative
10127 section number in a pod file instead of having to treat each file as
10128 a separate case in Makefile. This can be done by adding two lines to the
10131 =for comment openssl_section:XXX
10133 The blank line is mandatory.
10137 * New arguments -certform, -keyform and -pass for s_client and s_server
10138 to allow alternative format key and certificate files and passphrase
10143 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10144 update associated structures and add various utility functions.
10146 Add new policy related verify parameters, include policy checking in
10147 standard verify code. Enhance 'smime' application with extra parameters
10148 to support policy checking and print out.
10152 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10153 Nehemiah processors. These extensions support AES encryption in hardware
10154 as well as RNG (though RNG support is currently disabled).
10156 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10158 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10162 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10164 *Andy Polyakov and a number of other people*
10166 * Improved PowerPC platform support. Most notably BIGNUM assembler
10167 implementation contributed by IBM.
10169 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10171 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10172 exponent rather than 'unsigned long'. There is a corresponding change to
10173 the new 'rsa_keygen' element of the RSA_METHOD structure.
10175 *Jelte Jansen, Geoff Thorpe*
10177 * Functionality for creating the initial serial number file is now
10178 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10180 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10181 number file to 1, which is bound to cause problems. To avoid
10182 the problems while respecting compatibility between different 0.9.7
10183 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10184 CA.pl for serial number initialization. With the new release 0.9.8,
10185 we can fix the problem directly in the 'ca' utility.)
10189 * Reduced header interdependencies by declaring more opaque objects in
10190 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10191 give fewer recursive includes, which could break lazy source code - so
10192 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10193 developers should define this symbol when building and using openssl to
10194 ensure they track the recommended behaviour, interfaces, [etc], but
10195 backwards-compatible behaviour prevails when this isn't defined.
10199 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10203 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10204 This will generate a random key of the appropriate length based on the
10205 cipher context. The EVP_CIPHER can provide its own random key generation
10206 routine to support keys of a specific form. This is used in the des and
10207 3des routines to generate a key of the correct parity. Update S/MIME
10208 code to use new functions and hence generate correct parity DES keys.
10209 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10210 valid (weak or incorrect parity).
10214 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10215 as looking them up. This is useful when the verified structure may contain
10216 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10217 present unless the new PKCS7_NO_CRL flag is asserted.
10221 * Extend ASN1 oid configuration module. It now additionally accepts the
10224 shortName = some long name, 1.2.3.4
10228 * Reimplemented the BN_CTX implementation. There is now no more static
10229 limitation on the number of variables it can handle nor the depth of the
10230 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10231 information can now expand as required, and rather than having a single
10232 static array of bignums, BN_CTX now uses a linked-list of such arrays
10233 allowing it to expand on demand whilst maintaining the usefulness of
10234 BN_CTX's "bundling".
10238 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10239 to allow all RSA operations to function using a single BN_CTX.
10243 * Preliminary support for certificate policy evaluation and checking. This
10244 is initially intended to pass the tests outlined in "Conformance Testing
10245 of Relying Party Client Certificate Path Processing Logic" v1.07.
10249 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10250 remained unused and not that useful. A variety of other little bignum
10251 tweaks and fixes have also been made continuing on from the audit (see
10256 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10257 associated ASN1, EVP and SSL functions and old ASN1 macros.
10261 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10262 and this should never fail. So the return value from the use of
10263 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10264 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10268 * BN_CTX_get() should return zero-valued bignums, providing the same
10269 initialised value as BN_new().
10271 *Geoff Thorpe, suggested by Ulf Möller*
10273 * Support for inhibitAnyPolicy certificate extension.
10277 * An audit of the BIGNUM code is underway, for which debugging code is
10278 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10279 is considered valid when processing BIGNUMs, and causes execution to
10280 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10281 further steps are taken to deliberately pollute unused data in BIGNUM
10282 structures to try and expose faulty code further on. For now, openssl will
10283 (in its default mode of operation) continue to tolerate the inconsistent
10284 forms that it has tolerated in the past, but authors and packagers should
10285 consider trying openssl and their own applications when compiled with
10286 these debugging symbols defined. It will help highlight potential bugs in
10287 their own code, and will improve the test coverage for OpenSSL itself. At
10288 some point, these tighter rules will become openssl's default to improve
10289 maintainability, though the assert()s and other overheads will remain only
10290 in debugging configurations. See bn.h for more details.
10292 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10294 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10295 that can only be obtained through BN_CTX_new() (which implicitly
10296 initialises it). The presence of this function only made it possible
10297 to overwrite an existing structure (and cause memory leaks).
10301 * Because of the callback-based approach for implementing LHASH as a
10302 template type, lh_insert() adds opaque objects to hash-tables and
10303 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10304 to clean up those corresponding objects before destroying the hash table
10305 (and losing the object pointers). So some over-zealous constifications in
10306 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10307 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10308 prototyped to have "const" restrictions on the object pointers they are
10309 given (and so aren't required to cast them away any more).
10313 * The tmdiff.h API was so ugly and minimal that our own timing utility
10314 (speed) prefers to use its own implementation. The two implementations
10315 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10316 its object type properly exposed (MS_TM) instead of casting to/from
10317 `char *`. This may still change yet if someone realises MS_TM and
10319 aren't necessarily the greatest nomenclatures - but this is what was used
10320 internally to the implementation so I've used that for now.
10324 * Ensure that deprecated functions do not get compiled when
10325 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10326 the self-tests were still using deprecated key-generation functions so
10327 these have been updated also.
10331 * Reorganise PKCS#7 code to separate the digest location functionality
10332 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10333 New function PKCS7_set_digest() to set the digest type for PKCS#7
10334 digestedData type. Add additional code to correctly generate the
10335 digestedData type and add support for this type in PKCS7 initialization
10340 * New function PKCS7_set0_type_other() this initializes a PKCS7
10341 structure of type "other".
10345 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10346 sure the loop does correctly stop and breaking ("division by zero")
10347 modulus operations are not performed. The (pre-generated) prime
10348 table crypto/bn/bn_prime.h was already correct, but it could not be
10349 re-generated on some platforms because of the "division by zero"
10350 situation in the script.
10352 *Ralf S. Engelschall*
10354 * Update support for ECC-based TLS ciphersuites according to
10355 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10356 SHA-1 now is only used for "small" curves (where the
10357 representation of a field element takes up to 24 bytes); for
10358 larger curves, the field element resulting from ECDH is directly
10359 used as premaster secret.
10361 *Douglas Stebila (Sun Microsystems Laboratories)*
10363 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10364 curve secp160r1 to the tests.
10366 *Douglas Stebila (Sun Microsystems Laboratories)*
10368 * Add the possibility to load symbols globally with DSO.
10370 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10372 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10373 control of the error stack.
10377 * Add support for STORE in ENGINE.
10381 * Add the STORE type. The intention is to provide a common interface
10382 to certificate and key stores, be they simple file-based stores, or
10383 HSM-type store, or LDAP stores, or...
10384 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10388 * Add a generic structure called OPENSSL_ITEM. This can be used to
10389 pass a list of arguments to any function as well as provide a way
10390 for a function to pass data back to the caller.
10394 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10395 works like BUF_strdup() but can be used to duplicate a portion of
10396 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10401 * Add the function sk_find_ex() which works like sk_find(), but will
10402 return an index to an element even if an exact match couldn't be
10403 found. The index is guaranteed to point at the element where the
10404 searched-for key would be inserted to preserve sorting order.
10408 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10409 takes an extra flags argument for optional functionality. Currently,
10410 the following flags are defined:
10412 OBJ_BSEARCH_VALUE_ON_NOMATCH
10413 This one gets OBJ_bsearch_ex() to return a pointer to the first
10414 element where the comparing function returns a negative or zero
10417 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10418 This one gets OBJ_bsearch_ex() to return a pointer to the first
10419 element where the comparing function returns zero. This is useful
10420 if there are more than one element where the comparing function
10425 * Make it possible to create self-signed certificates with 'openssl ca'
10426 in such a way that the self-signed certificate becomes part of the
10427 CA database and uses the same mechanisms for serial number generation
10428 as all other certificate signing. The new flag '-selfsign' enables
10429 this functionality. Adapt CA.sh and CA.pl.in.
10433 * Add functionality to check the public key of a certificate request
10434 against a given private. This is useful to check that a certificate
10435 request can be signed by that key (self-signing).
10439 * Make it possible to have multiple active certificates with the same
10440 subject in the CA index file. This is done only if the keyword
10441 'unique_subject' is set to 'no' in the main CA section (default
10442 if 'CA_default') of the configuration file. The value is saved
10443 with the database itself in a separate index attribute file,
10444 named like the index file with '.attr' appended to the name.
10448 * Generate multi-valued AVAs using '+' notation in config files for
10453 * Support for nameConstraints certificate extension.
10457 * Support for policyConstraints certificate extension.
10461 * Support for policyMappings certificate extension.
10465 * Make sure the default DSA_METHOD implementation only uses its
10466 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10467 and change its own handlers to be NULL so as to remove unnecessary
10468 indirection. This lets alternative implementations fallback to the
10469 default implementation more easily.
10473 * Support for directoryName in GeneralName related extensions
10478 * Make it possible to link applications using Makefile.shared.
10479 Make that possible even when linking against static libraries!
10483 * Support for single pass processing for S/MIME signing. This now
10484 means that S/MIME signing can be done from a pipe, in addition
10485 cleartext signing (multipart/signed type) is effectively streaming
10486 and the signed data does not need to be all held in memory.
10488 This is done with a new flag PKCS7_STREAM. When this flag is set
10489 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10490 is done after the data is output (and digests calculated) in
10491 SMIME_write_PKCS7().
10495 * Add full support for -rpath/-R, both in shared libraries and
10496 applications, at least on the platforms where it's known how
10501 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10502 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10503 will now compute a table of multiples of the generator that
10504 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10505 faster (notably in the case of a single point multiplication,
10506 scalar * generator).
10508 *Nils Larsch, Bodo Moeller*
10510 * IPv6 support for certificate extensions. The various extensions
10511 which use the IP:a.b.c.d can now take IPv6 addresses using the
10512 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10517 * Added an ENGINE that implements RSA by performing private key
10518 exponentiations with the GMP library. The conversions to and from
10519 GMP's mpz_t format aren't optimised nor are any montgomery forms
10520 cached, and on x86 it appears OpenSSL's own performance has caught up.
10521 However there are likely to be other architectures where GMP could
10522 provide a boost. This ENGINE is not built in by default, but it can be
10523 specified at Configure time and should be accompanied by the necessary
10524 linker additions, eg;
10525 ./config -DOPENSSL_USE_GMP -lgmp
10529 * "openssl engine" will not display ENGINE/DSO load failure errors when
10530 testing availability of engines with "-t" - the old behaviour is
10531 produced by increasing the feature's verbosity with "-tt".
10535 * ECDSA routines: under certain error conditions uninitialized BN objects
10536 could be freed. Solution: make sure initialization is performed early
10537 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10542 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10543 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10544 software implementations. For DSA and DH, parameter generation can
10545 also be overridden by providing the appropriate method callbacks.
10549 * Change the "progress" mechanism used in key-generation and
10550 primality testing to functions that take a new BN_GENCB pointer in
10551 place of callback/argument pairs. The new API functions have `_ex`
10552 postfixes and the older functions are reimplemented as wrappers for
10553 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10554 declarations of the old functions to help (graceful) attempts to
10555 migrate to the new functions. Also, the new key-generation API
10556 functions operate on a caller-supplied key-structure and return
10557 success/failure rather than returning a key or NULL - this is to
10558 help make "keygen" another member function of RSA_METHOD etc.
10560 Example for using the new callback interface:
10562 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10563 void *my_arg = ...;
10566 BN_GENCB_set(&my_cb, my_callback, my_arg);
10568 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10569 /* For the meaning of a, b in calls to my_callback(), see the
10570 * documentation of the function that calls the callback.
10571 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10572 * my_callback should return 1 if it wants BN_is_prime_ex()
10573 * to continue, or 0 to stop.
10578 * Change the ZLIB compression method to be stateful, and make it
10579 available to TLS with the number defined in
10580 draft-ietf-tls-compression-04.txt.
10584 * Add the ASN.1 structures and functions for CertificatePair, which
10585 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10587 CertificatePair ::= SEQUENCE {
10588 forward [0] Certificate OPTIONAL,
10589 reverse [1] Certificate OPTIONAL,
10590 -- at least one of the pair shall be present -- }
10592 Also implement the PEM functions to read and write certificate
10593 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10595 This needed to be defined, mostly for the sake of the LDAP
10596 attribute crossCertificatePair, but may prove useful elsewhere as
10601 * Make it possible to inhibit symlinking of shared libraries in
10602 Makefile.shared, for Cygwin's sake.
10606 * Extend the BIGNUM API by creating a function
10607 void BN_set_negative(BIGNUM *a, int neg);
10608 and a macro that behave like
10609 int BN_is_negative(const BIGNUM *a);
10611 to avoid the need to access 'a->neg' directly in applications.
10615 * Implement fast modular reduction for pseudo-Mersenne primes
10616 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10617 EC_GROUP_new_curve_GFp() will now automatically use this
10620 *Nils Larsch <nla@trustcenter.de>*
10622 * Add new lock type (CRYPTO_LOCK_BN).
10626 * Change the ENGINE framework to automatically load engines
10627 dynamically from specific directories unless they could be
10628 found to already be built in or loaded. Move all the
10629 current engines except for the cryptodev one to a new
10630 directory engines/.
10631 The engines in engines/ are built as shared libraries if
10632 the "shared" options was given to ./Configure or ./config.
10633 Otherwise, they are inserted in libcrypto.a.
10634 /usr/local/ssl/engines is the default directory for dynamic
10635 engines, but that can be overridden at configure time through
10636 the usual use of --prefix and/or --openssldir, and at run
10637 time with the environment variable OPENSSL_ENGINES.
10639 *Geoff Thorpe and Richard Levitte*
10641 * Add Makefile.shared, a helper makefile to build shared
10642 libraries. Adapt Makefile.org.
10646 * Add version info to Win32 DLLs.
10648 *Peter 'Luna' Runestig" <peter@runestig.com>*
10650 * Add new 'medium level' PKCS#12 API. Certificates and keys
10651 can be added using this API to created arbitrary PKCS#12
10652 files while avoiding the low-level API.
10654 New options to PKCS12_create(), key or cert can be NULL and
10655 will then be omitted from the output file. The encryption
10656 algorithm NIDs can be set to -1 for no encryption, the mac
10657 iteration count can be set to 0 to omit the mac.
10659 Enhance pkcs12 utility by making the -nokeys and -nocerts
10660 options work when creating a PKCS#12 file. New option -nomac
10661 to omit the mac, NONE can be set for an encryption algorithm.
10662 New code is modified to use the enhanced PKCS12_create()
10663 instead of the low-level API.
10667 * Extend ASN1 encoder to support indefinite length constructed
10668 encoding. This can output sequences tags and octet strings in
10669 this form. Modify pk7_asn1.c to support indefinite length
10670 encoding. This is experimental and needs additional code to
10671 be useful, such as an ASN1 bio and some enhanced streaming
10674 Extend template encode functionality so that tagging is passed
10675 down to the template encoder.
10679 * Let 'openssl req' fail if an argument to '-newkey' is not
10680 recognized instead of using RSA as a default.
10684 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10685 As these are not official, they are not included in "ALL";
10686 the "ECCdraft" ciphersuite group alias can be used to select them.
10688 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10690 * Add ECDH engine support.
10692 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10694 * Add ECDH in new directory crypto/ecdh/.
10696 *Douglas Stebila (Sun Microsystems Laboratories)*
10698 * Let BN_rand_range() abort with an error after 100 iterations
10699 without success (which indicates a broken PRNG).
10703 * Change BN_mod_sqrt() so that it verifies that the input value
10704 is really the square of the return value. (Previously,
10705 BN_mod_sqrt would show GIGO behaviour.)
10709 * Add named elliptic curves over binary fields from X9.62, SECG,
10710 and WAP/WTLS; add OIDs that were still missing.
10712 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10714 * Extend the EC library for elliptic curves over binary fields
10715 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10718 EC_GF2m_simple_method
10722 EC_GROUP_new_curve_GF2m
10723 EC_GROUP_set_curve_GF2m
10724 EC_GROUP_get_curve_GF2m
10725 EC_POINT_set_affine_coordinates_GF2m
10726 EC_POINT_get_affine_coordinates_GF2m
10727 EC_POINT_set_compressed_coordinates_GF2m
10729 Point compression for binary fields is disabled by default for
10730 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10733 As binary polynomials are represented as BIGNUMs, various members
10734 of the EC_GROUP and EC_POINT data structures can be shared
10735 between the implementations for prime fields and binary fields;
10736 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10737 are essentially identical to their `..._GFp` counterparts.
10738 (For simplicity, the `..._GFp` prefix has been dropped from
10739 various internal method names.)
10741 An internal 'field_div' method (similar to 'field_mul' and
10742 'field_sqr') has been added; this is used only for binary fields.
10744 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10746 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10747 through methods ('mul', 'precompute_mult').
10749 The generic implementations (now internally called 'ec_wNAF_mul'
10750 and 'ec_wNAF_precomputed_mult') remain the default if these
10751 methods are undefined.
10753 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10755 * New function EC_GROUP_get_degree, which is defined through
10756 EC_METHOD. For curves over prime fields, this returns the bit
10757 length of the modulus.
10759 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10761 * New functions EC_GROUP_dup, EC_POINT_dup.
10762 (These simply call ..._new and ..._copy).
10764 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10766 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10767 Polynomials are represented as BIGNUMs (where the sign bit is not
10768 used) in the following functions [macros]:
10771 BN_GF2m_sub [= BN_GF2m_add]
10772 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10773 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10774 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10776 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10777 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10778 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10779 BN_GF2m_cmp [= BN_ucmp]
10781 (Note that only the 'mod' functions are actually for fields GF(2^m).
10782 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10784 For some functions, an the irreducible polynomial defining a
10785 field can be given as an 'unsigned int[]' with strictly
10786 decreasing elements giving the indices of those bits that are set;
10787 i.e., p[] represents the polynomial
10788 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10790 p[0] > p[1] > ... > p[k] = 0.
10791 This applies to the following functions:
10794 BN_GF2m_mod_mul_arr
10795 BN_GF2m_mod_sqr_arr
10796 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10797 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10798 BN_GF2m_mod_exp_arr
10799 BN_GF2m_mod_sqrt_arr
10800 BN_GF2m_mod_solve_quad_arr
10804 Conversion can be performed by the following functions:
10809 bntest.c has additional tests for binary polynomial arithmetic.
10811 Two implementations for BN_GF2m_mod_div() are available.
10812 The default algorithm simply uses BN_GF2m_mod_inv() and
10813 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10814 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10815 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10817 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10819 * Add new error code 'ERR_R_DISABLED' that can be used when some
10820 functionality is disabled at compile-time.
10822 *Douglas Stebila <douglas.stebila@sun.com>*
10824 * Change default behaviour of 'openssl asn1parse' so that more
10825 information is visible when viewing, e.g., a certificate:
10827 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10828 mode the content of non-printable OCTET STRINGs is output in a
10829 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10830 avoid the appearance of a printable string.
10832 *Nils Larsch <nla@trustcenter.de>*
10834 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10836 EC_GROUP_set_asn1_flag()
10837 EC_GROUP_get_asn1_flag()
10838 EC_GROUP_set_point_conversion_form()
10839 EC_GROUP_get_point_conversion_form()
10840 These control ASN1 encoding details:
10841 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10842 has been set to OPENSSL_EC_NAMED_CURVE.
10843 - Points are encoded in uncompressed form by default; options for
10844 asn1_for are as for point2oct, namely
10845 POINT_CONVERSION_COMPRESSED
10846 POINT_CONVERSION_UNCOMPRESSED
10847 POINT_CONVERSION_HYBRID
10849 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10851 EC_GROUP_set_seed()
10852 EC_GROUP_get0_seed()
10853 EC_GROUP_get_seed_len()
10854 This is used only for ASN1 purposes (so far).
10856 *Nils Larsch <nla@trustcenter.de>*
10858 * Add 'field_type' member to EC_METHOD, which holds the NID
10859 of the appropriate field type OID. The new function
10860 EC_METHOD_get_field_type() returns this value.
10862 *Nils Larsch <nla@trustcenter.de>*
10865 EC_POINT_point2bn()
10866 EC_POINT_bn2point()
10867 EC_POINT_point2hex()
10868 EC_POINT_hex2point()
10869 providing useful interfaces to EC_POINT_point2oct() and
10870 EC_POINT_oct2point().
10872 *Nils Larsch <nla@trustcenter.de>*
10874 * Change internals of the EC library so that the functions
10875 EC_GROUP_set_generator()
10876 EC_GROUP_get_generator()
10877 EC_GROUP_get_order()
10878 EC_GROUP_get_cofactor()
10879 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10880 to methods, which would lead to unnecessary code duplication when
10881 adding different types of curves.
10883 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10885 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10886 arithmetic, and such that modified wNAFs are generated
10887 (which avoid length expansion in many cases).
10891 * Add a function EC_GROUP_check_discriminant() (defined via
10892 EC_METHOD) that verifies that the curve discriminant is non-zero.
10894 Add a function EC_GROUP_check() that makes some sanity tests
10895 on a EC_GROUP, its generator and order. This includes
10896 EC_GROUP_check_discriminant().
10898 *Nils Larsch <nla@trustcenter.de>*
10900 * Add ECDSA in new directory crypto/ecdsa/.
10902 Add applications 'openssl ecparam' and 'openssl ecdsa'
10903 (these are based on 'openssl dsaparam' and 'openssl dsa').
10905 ECDSA support is also included in various other files across the
10906 library. Most notably,
10907 - 'openssl req' now has a '-newkey ecdsa:file' option;
10908 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10909 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10910 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10911 them suitable for ECDSA where domain parameters must be
10912 extracted before the specific public key;
10913 - ECDSA engine support has been added.
10915 *Nils Larsch <nla@trustcenter.de>*
10917 * Include some named elliptic curves, and add OIDs from X9.62,
10918 SECG, and WAP/WTLS. Each curve can be obtained from the new
10920 EC_GROUP_new_by_curve_name(),
10921 and the list of available named curves can be obtained with
10922 EC_get_builtin_curves().
10923 Also add a 'curve_name' member to EC_GROUP objects, which can be
10925 EC_GROUP_set_curve_name()
10926 EC_GROUP_get_curve_name()
10928 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10930 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10931 was actually never needed) and in BN_mul(). The removal in BN_mul()
10932 required a small change in bn_mul_part_recursive() and the addition
10933 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10934 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10935 bn_sub_words() and bn_add_words() except they take arrays with
10940 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10942 * Cleanse PEM buffers before freeing them since they may contain
10945 *Benjamin Bennett <ben@psc.edu>*
10947 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10948 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10949 authentication-only ciphersuites.
10953 * Since AES128 and AES256 share a single mask bit in the logic of
10954 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10955 kludge to work properly if AES128 is available and AES256 isn't.
10959 * Expand security boundary to match 1.1.1 module.
10963 * Remove redundant features: hash file source, editing of test vectors
10964 modify fipsld to use external fips_premain.c signature.
10968 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10969 run algorithm test programs.
10973 * Make algorithm test programs more tolerant of whitespace.
10977 * Have SSL/TLS server implementation tolerate "mismatched" record
10978 protocol version while receiving ClientHello even if the
10979 ClientHello is fragmented. (The server can't insist on the
10980 particular protocol version it has chosen before the ServerHello
10981 message has informed the client about his choice.)
10985 * Load error codes if they are not already present instead of using a
10986 static variable. This allows them to be cleanly unloaded and reloaded.
10990 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10992 * Introduce limits to prevent malicious keys being able to
10993 cause a denial of service. ([CVE-2006-2940])
10995 *Steve Henson, Bodo Moeller*
10997 * Fix ASN.1 parsing of certain invalid structures that can result
10998 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11000 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11001 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11003 * Fix SSL client code which could crash if connecting to a
11004 malicious SSLv2 server. ([CVE-2006-4343])
11006 *Tavis Ormandy and Will Drewry, Google Security Team*
11008 * Change ciphersuite string processing so that an explicit
11009 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11010 will no longer include "AES128-SHA"), and any other similar
11011 ciphersuite (same bitmap) from *other* protocol versions (so that
11012 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11013 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11014 changes from 0.9.8b and 0.9.8d.
11018 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11020 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11021 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11023 * Change the Unix randomness entropy gathering to use poll() when
11024 possible instead of select(), since the latter has some
11025 undesirable limitations.
11027 *Darryl Miles via Richard Levitte and Bodo Moeller*
11029 * Disable rogue ciphersuites:
11031 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11032 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11033 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11035 The latter two were purportedly from
11036 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11039 Also deactivate the remaining ciphersuites from
11040 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11041 unofficial, and the ID has long expired.
11045 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11046 dual-core machines) and other potential thread-safety issues.
11050 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11052 * Adapt fipsld and the build system to link against the validated FIPS
11053 module in FIPS mode.
11057 * Fixes for VC++ 2005 build under Windows.
11061 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11062 from a Windows bash shell such as MSYS. It is autodetected from the
11063 "config" script when run from a VC++ environment. Modify standard VC++
11064 build to use fipscanister.o from the GNU make build.
11068 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11070 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11071 The value now differs depending on if you build for FIPS or not.
11072 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11073 safely run with a non-FIPSed libcrypto, as it may crash because of
11074 the difference induced by this change.
11078 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11080 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11081 (part of SSL_OP_ALL). This option used to disable the
11082 countermeasure against man-in-the-middle protocol-version
11083 rollback in the SSL 2.0 server implementation, which is a bad
11084 idea. ([CVE-2005-2969])
11086 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11087 for Information Security, National Institute of Advanced Industrial
11088 Science and Technology [AIST, Japan)]*
11090 * Minimal support for X9.31 signatures and PSS padding modes. This is
11091 mainly for FIPS compliance and not fully integrated at this stage.
11095 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11096 the exponentiation using a fixed-length exponent. (Otherwise,
11097 the information leaked through timing could expose the secret key
11098 after many signatures; cf. Bleichenbacher's attack on DSA with
11103 * Make a new fixed-window mod_exp implementation the default for
11104 RSA, DSA, and DH private-key operations so that the sequence of
11105 squares and multiplies and the memory access pattern are
11106 independent of the particular secret key. This will mitigate
11107 cache-timing and potential related attacks.
11109 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11110 and this is automatically used by BN_mod_exp_mont() if the new flag
11111 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11112 will use this BN flag for private exponents unless the flag
11113 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11114 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11116 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11118 * Change the client implementation for SSLv23_method() and
11119 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11120 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11121 (Previously, the SSL 2.0 backwards compatible Client Hello
11122 message format would be used even with SSL_OP_NO_SSLv2.)
11126 * Add support for smime-type MIME parameter in S/MIME messages which some
11131 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11132 a threadsafe manner. Modify rsa code to use new function and add calls
11133 to dsa and dh code (which had race conditions before).
11137 * Include the fixed error library code in the C error file definitions
11138 instead of fixing them up at runtime. This keeps the error code
11139 structures constant.
11143 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11145 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11148 * Fixes for newer kerberos headers. NB: the casts are needed because
11149 the 'length' field is signed on one version and unsigned on another
11150 with no (?) obvious way to tell the difference, without these VC++
11151 complains. Also the "definition" of FAR (blank) is no longer included
11152 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11153 some needed definitions.
11157 * Undo Cygwin change.
11161 * Added support for proxy certificates according to RFC 3820.
11162 Because they may be a security thread to unaware applications,
11163 they must be explicitly allowed in run-time. See
11164 docs/HOWTO/proxy_certificates.txt for further information.
11168 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11170 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11171 server and client random values. Previously
11172 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11173 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11175 This change has negligible security impact because:
11177 1. Server and client random values still have 24 bytes of pseudo random
11180 2. Server and client random values are sent in the clear in the initial
11183 3. The master secret is derived using the premaster secret (48 bytes in
11184 size for static RSA ciphersuites) as well as client server and random
11187 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11190 *Stephen Henson, reported by UK NISCC*
11192 * Use Windows randomness collection on Cygwin.
11196 * Fix hang in EGD/PRNGD query when communication socket is closed
11197 prematurely by EGD/PRNGD.
11199 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11201 * Prompt for pass phrases when appropriate for PKCS12 input format.
11205 * Back-port of selected performance improvements from development
11206 branch, as well as improved support for PowerPC platforms.
11210 * Add lots of checks for memory allocation failure, error codes to indicate
11211 failure and freeing up memory if a failure occurs.
11213 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11215 * Add new -passin argument to dgst.
11219 * Perform some character comparisons of different types in X509_NAME_cmp:
11220 this is needed for some certificates that re-encode DNs into UTF8Strings
11221 (in violation of RFC3280) and can't or won't issue name rollover
11226 * Make an explicit check during certificate validation to see that
11227 the CA setting in each certificate on the chain is correct. As a
11228 side effect always do the following basic checks on extensions,
11229 not just when there's an associated purpose to the check:
11231 - if there is an unhandled critical extension (unless the user
11232 has chosen to ignore this fault)
11233 - if the path length has been exceeded (if one is set at all)
11234 - that certain extensions fit the associated purpose (if one has
11239 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11241 * Avoid a race condition when CRLs are checked in a multi threaded
11242 environment. This would happen due to the reordering of the revoked
11243 entries during signature checking and serial number lookup. Now the
11244 encoding is cached and the serial number sort performed under a lock.
11245 Add new STACK function sk_is_sorted().
11249 * Add Delta CRL to the extension code.
11253 * Various fixes to s3_pkt.c so alerts are sent properly.
11255 *David Holmes <d.holmes@f5.com>*
11257 * Reduce the chances of duplicate issuer name and serial numbers (in
11258 violation of RFC3280) using the OpenSSL certificate creation utilities.
11259 This is done by creating a random 64 bit value for the initial serial
11260 number when a serial number file is created or when a self signed
11261 certificate is created using 'openssl req -x509'. The initial serial
11262 number file is created using 'openssl x509 -next_serial' in CA.pl
11263 rather than being initialized to 1.
11267 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11269 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11270 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11272 *Joe Orton, Steve Henson*
11274 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11277 *Joe Orton, Steve Henson*
11279 * Make it possible to have multiple active certificates with the same
11280 subject in the CA index file. This is done only if the keyword
11281 'unique_subject' is set to 'no' in the main CA section (default
11282 if 'CA_default') of the configuration file. The value is saved
11283 with the database itself in a separate index attribute file,
11284 named like the index file with '.attr' appended to the name.
11288 * X509 verify fixes. Disable broken certificate workarounds when
11289 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11290 keyUsage extension present. Don't accept CRLs with unhandled critical
11291 extensions: since verify currently doesn't process CRL extensions this
11292 rejects a CRL with *any* critical extensions. Add new verify error codes
11297 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11298 A clarification of RFC2560 will require the use of OCTET STRINGs and
11299 some implementations cannot handle the current raw format. Since OpenSSL
11300 copies and compares OCSP nonces as opaque blobs without any attempt at
11301 parsing them this should not create any compatibility issues.
11305 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11306 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11307 this HMAC (and other) operations are several times slower than OpenSSL
11312 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11314 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11316 * Use the correct content when signing type "other".
11320 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11322 * Fix various bugs revealed by running the NISCC test suite:
11324 Stop out of bounds reads in the ASN1 code when presented with
11325 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11327 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11329 If verify callback ignores invalid public key errors don't try to check
11330 certificate signature with the NULL public key.
11334 * New -ignore_err option in ocsp application to stop the server
11335 exiting on the first error in a request.
11339 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11340 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11345 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11346 extra data after the compression methods not only for TLS 1.0
11347 but also for SSL 3.0 (as required by the specification).
11349 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11351 * Change X509_certificate_type() to mark the key as exported/exportable
11352 when it's 512 *bits* long, not 512 bytes.
11356 * Change AES_cbc_encrypt() so it outputs exact multiple of
11357 blocks during encryption.
11361 * Various fixes to base64 BIO and non blocking I/O. On write
11362 flushes were not handled properly if the BIO retried. On read
11363 data was not being buffered properly and had various logic bugs.
11364 This also affects blocking I/O when the data being decoded is a
11369 * Various S/MIME bugfixes and compatibility changes:
11370 output correct application/pkcs7 MIME type if
11371 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11372 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11373 of files as .eml work). Correctly handle very long lines in MIME
11378 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11380 * Countermeasure against the Klima-Pokorny-Rosa extension of
11381 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11382 a protocol version number mismatch like a decryption error
11383 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11387 * Turn on RSA blinding by default in the default implementation
11388 to avoid a timing attack. Applications that don't want it can call
11389 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11390 They would be ill-advised to do so in most cases.
11392 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11394 * Change RSA blinding code so that it works when the PRNG is not
11395 seeded (in this case, the secret RSA exponent is abused as
11396 an unpredictable seed -- if it is not unpredictable, there
11397 is no point in blinding anyway). Make RSA blinding thread-safe
11398 by remembering the creator's thread ID in rsa->blinding and
11399 having all other threads use local one-time blinding factors
11400 (this requires more computation than sharing rsa->blinding, but
11401 avoids excessive locking; and if an RSA object is not shared
11402 between threads, blinding will still be very fast).
11406 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11407 ENGINE as defaults for all supported algorithms irrespective of
11408 the 'flags' parameter. 'flags' is now honoured, so applications
11409 should make sure they are passing it correctly.
11413 * Target "mingw" now allows native Windows code to be generated in
11414 the Cygwin environment as well as with the MinGW compiler.
11418 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11420 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11421 via timing by performing a MAC computation even if incorrect
11422 block cipher padding has been found. This is a countermeasure
11423 against active attacks where the attacker has to distinguish
11424 between bad padding and a MAC verification error. ([CVE-2003-0078])
11426 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11427 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11428 Martin Vuagnoux (EPFL, Ilion)*
11430 * Make the no-err option work as intended. The intention with no-err
11431 is not to have the whole error stack handling routines removed from
11432 libcrypto, it's only intended to remove all the function name and
11433 reason texts, thereby removing some of the footprint that may not
11434 be interesting if those errors aren't displayed anyway.
11436 NOTE: it's still possible for any application or module to have its
11437 own set of error texts inserted. The routines are there, just not
11438 used by default when no-err is given.
11442 * Add support for FreeBSD on IA64.
11444 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11446 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11447 Kerberos function mit_des_cbc_cksum(). Before this change,
11448 the value returned by DES_cbc_cksum() was like the one from
11449 mit_des_cbc_cksum(), except the bytes were swapped.
11451 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11453 * Allow an application to disable the automatic SSL chain building.
11454 Before this a rather primitive chain build was always performed in
11455 ssl3_output_cert_chain(): an application had no way to send the
11456 correct chain if the automatic operation produced an incorrect result.
11458 Now the chain builder is disabled if either:
11460 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11462 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11464 The reasoning behind this is that an application would not want the
11465 auto chain building to take place if extra chain certificates are
11466 present and it might also want a means of sending no additional
11467 certificates (for example the chain has two certificates and the
11472 * Add the possibility to build without the ENGINE framework.
11474 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11476 * Under Win32 gmtime() can return NULL: check return value in
11477 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11481 * DSA routines: under certain error conditions uninitialized BN objects
11482 could be freed. Solution: make sure initialization is performed early
11483 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11484 Nils Larsch <nla@trustcenter.de> via PR#459)
11488 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11489 checked on reconnect on the client side, therefore session resumption
11490 could still fail with a "ssl session id is different" error. This
11491 behaviour is masked when SSL_OP_ALL is used due to
11492 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11493 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11494 followup to PR #377.
11498 * IA-32 assembler support enhancements: unified ELF targets, support
11499 for SCO/Caldera platforms, fix for Cygwin shared build.
11503 * Add support for FreeBSD on sparc64. As a consequence, support for
11504 FreeBSD on non-x86 processors is separate from x86 processors on
11505 the config script, much like the NetBSD support.
11507 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11509 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11511 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11514 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11515 code (06) was taken as the first octet of the session ID and the last
11516 octet was ignored consequently. As a result SSLv2 client side session
11517 caching could not have worked due to the session ID mismatch between
11519 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11524 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11525 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11530 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11531 seems that in spite of existing for more than a year, many application
11532 author have done nothing to provide the necessary callbacks, which
11533 means that this particular engine will not work properly anywhere.
11534 This is a very unfortunate situation which forces us, in the name
11535 of usability, to give the hw_ncipher.c a static lock, which is part
11537 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11538 appear in 0.9.8 or later. We EXPECT application authors to have
11539 dealt properly with this when 0.9.8 is released (unless we actually
11540 make such changes in the libcrypto locking code that changes will
11541 have to be made anyway).
11545 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11546 octets have been read, EOF or an error occurs. Without this change
11547 some truncated ASN1 structures will not produce an error.
11551 * Disable Heimdal support, since it hasn't been fully implemented.
11552 Still give the possibility to force the use of Heimdal, but with
11553 warnings and a request that patches get sent to openssl-dev.
11557 * Add the VC-CE target, introduce the WINCE sysname, and add
11558 INSTALL.WCE and appropriate conditionals to make it build.
11560 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11562 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11563 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11564 edit numbers of the version.
11566 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11568 * Introduce safe string copy and catenation functions
11569 (BUF_strlcpy() and BUF_strlcat()).
11571 *Ben Laurie (CHATS) and Richard Levitte*
11573 * Avoid using fixed-size buffers for one-line DNs.
11575 *Ben Laurie (CHATS)*
11577 * Add BUF_MEM_grow_clean() to avoid information leakage when
11578 resizing buffers containing secrets, and use where appropriate.
11580 *Ben Laurie (CHATS)*
11582 * Avoid using fixed size buffers for configuration file location.
11584 *Ben Laurie (CHATS)*
11586 * Avoid filename truncation for various CA files.
11588 *Ben Laurie (CHATS)*
11590 * Use sizeof in preference to magic numbers.
11592 *Ben Laurie (CHATS)*
11594 * Avoid filename truncation in cert requests.
11596 *Ben Laurie (CHATS)*
11598 * Add assertions to check for (supposedly impossible) buffer
11601 *Ben Laurie (CHATS)*
11603 * Don't cache truncated DNS entries in the local cache (this could
11604 potentially lead to a spoofing attack).
11606 *Ben Laurie (CHATS)*
11608 * Fix various buffers to be large enough for hex/decimal
11609 representations in a platform independent manner.
11611 *Ben Laurie (CHATS)*
11613 * Add CRYPTO_realloc_clean() to avoid information leakage when
11614 resizing buffers containing secrets, and use where appropriate.
11616 *Ben Laurie (CHATS)*
11618 * Add BIO_indent() to avoid much slightly worrying code to do
11621 *Ben Laurie (CHATS)*
11623 * Convert sprintf()/BIO_puts() to BIO_printf().
11625 *Ben Laurie (CHATS)*
11627 * buffer_gets() could terminate with the buffer only half
11630 *Ben Laurie (CHATS)*
11632 * Add assertions to prevent user-supplied crypto functions from
11633 overflowing internal buffers by having large block sizes, etc.
11635 *Ben Laurie (CHATS)*
11637 * New OPENSSL_assert() macro (similar to assert(), but enabled
11640 *Ben Laurie (CHATS)*
11642 * Eliminate unused copy of key in RC4.
11644 *Ben Laurie (CHATS)*
11646 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11648 *Ben Laurie (CHATS)*
11650 * Fix off-by-one error in EGD path.
11652 *Ben Laurie (CHATS)*
11654 * If RANDFILE path is too long, ignore instead of truncating.
11656 *Ben Laurie (CHATS)*
11658 * Eliminate unused and incorrectly sized X.509 structure
11661 *Ben Laurie (CHATS)*
11663 * Eliminate unused and dangerous function knumber().
11665 *Ben Laurie (CHATS)*
11667 * Eliminate unused and dangerous structure, KSSL_ERR.
11669 *Ben Laurie (CHATS)*
11671 * Protect against overlong session ID context length in an encoded
11672 session object. Since these are local, this does not appear to be
11675 *Ben Laurie (CHATS)*
11677 * Change from security patch (see 0.9.6e below) that did not affect
11678 the 0.9.6 release series:
11680 Remote buffer overflow in SSL3 protocol - an attacker could
11681 supply an oversized master key in Kerberos-enabled versions.
11684 *Ben Laurie (CHATS)*
11686 * Change the SSL kerb5 codes to match RFC 2712.
11690 * Make -nameopt work fully for req and add -reqopt switch.
11692 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11694 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11696 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11698 * Make sure tests can be performed even if the corresponding algorithms
11699 have been removed entirely. This was also the last step to make
11700 OpenSSL compilable with DJGPP under all reasonable conditions.
11702 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11704 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11705 to allow version independent disabling of normally unselected ciphers,
11706 which may be activated as a side-effect of selecting a single cipher.
11708 (E.g., cipher list string "RSA" enables ciphersuites that are left
11709 out of "ALL" because they do not provide symmetric encryption.
11710 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11712 *Lutz Jaenicke, Bodo Moeller*
11714 * Add appropriate support for separate platform-dependent build
11715 directories. The recommended way to make a platform-dependent
11716 build directory is the following (tested on Linux), maybe with
11719 # Place yourself outside of the OpenSSL source tree. In
11720 # this example, the environment variable OPENSSL_SOURCE
11721 # is assumed to contain the absolute OpenSSL source directory.
11722 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11723 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11724 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11725 mkdir -p `dirname $F`
11726 ln -s $OPENSSL_SOURCE/$F $F
11729 To be absolutely sure not to disturb the source tree, a "make clean"
11730 is a good thing. If it isn't successful, don't worry about it,
11731 it probably means the source directory is very clean.
11735 * Make sure any ENGINE control commands make local copies of string
11736 pointers passed to them whenever necessary. Otherwise it is possible
11737 the caller may have overwritten (or deallocated) the original string
11738 data when a later ENGINE operation tries to use the stored values.
11740 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11742 * Improve diagnostics in file reading and command-line digests.
11744 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11746 * Add AES modes CFB and OFB to the object database. Correct an
11747 error in AES-CFB decryption.
11751 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11752 allows existing EVP_CIPHER_CTX structures to be reused after
11753 calling `EVP_*Final()`. This behaviour is used by encryption
11754 BIOs and some applications. This has the side effect that
11755 applications must explicitly clean up cipher contexts with
11756 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11760 * Check the values of dna and dnb in bn_mul_recursive before calling
11761 bn_mul_comba (a non zero value means the a or b arrays do not contain
11762 n2 elements) and fallback to bn_mul_normal if either is not zero.
11766 * Fix escaping of non-ASCII characters when using the -subj option
11767 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11771 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11772 form for "surname", serialNumber has no short form.
11773 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11774 therefore remove "mail" short name for "internet 7".
11775 The OID for unique identifiers in X509 certificates is
11776 x500UniqueIdentifier, not uniqueIdentifier.
11777 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11781 * Add an "init" command to the ENGINE config module and auto initialize
11782 ENGINEs. Without any "init" command the ENGINE will be initialized
11783 after all ctrl commands have been executed on it. If init=1 the
11784 ENGINE is initialized at that point (ctrls before that point are run
11785 on the uninitialized ENGINE and after on the initialized one). If
11786 init=0 then the ENGINE will not be initialized at all.
11790 * Fix the 'app_verify_callback' interface so that the user-defined
11791 argument is actually passed to the callback: In the
11792 SSL_CTX_set_cert_verify_callback() prototype, the callback
11793 declaration has been changed from
11796 int (*cb)(X509_STORE_CTX *,void *);
11797 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11798 i=s->ctx->app_verify_callback(&ctx)
11799 has been changed into
11800 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11802 To update applications using SSL_CTX_set_cert_verify_callback(),
11803 a dummy argument can be added to their callback functions.
11805 *D. K. Smetters <smetters@parc.xerox.com>*
11807 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11809 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11811 * Add and OPENSSL_LOAD_CONF define which will cause
11812 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11813 This allows older applications to transparently support certain
11814 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11815 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11816 load the config file and OPENSSL_add_all_algorithms_conf() which will
11817 always load it have also been added.
11821 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11822 Adjust NIDs and EVP layer.
11824 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11826 * Config modules support in openssl utility.
11828 Most commands now load modules from the config file,
11829 though in a few (such as version) this isn't done
11830 because it couldn't be used for anything.
11832 In the case of ca and req the config file used is
11833 the same as the utility itself: that is the -config
11834 command line option can be used to specify an
11839 * Move default behaviour from OPENSSL_config(). If appname is NULL
11840 use "openssl_conf" if filename is NULL use default openssl config file.
11844 * Add an argument to OPENSSL_config() to allow the use of an alternative
11845 config section name. Add a new flag to tolerate a missing config file
11846 and move code to CONF_modules_load_file().
11850 * Support for crypto accelerator cards from Accelerated Encryption
11851 Processing, www.aep.ie. (Use engine 'aep')
11852 The support was copied from 0.9.6c [engine] and adapted/corrected
11853 to work with the new engine framework.
11855 *AEP Inc. and Richard Levitte*
11857 * Support for SureWare crypto accelerator cards from Baltimore
11858 Technologies. (Use engine 'sureware')
11859 The support was copied from 0.9.6c [engine] and adapted
11860 to work with the new engine framework.
11864 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11865 make the newer ENGINE framework commands for the CHIL engine work.
11867 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11869 * Make it possible to produce shared libraries on ReliantUNIX.
11871 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11873 * Add the configuration target debug-linux-ppro.
11874 Make 'openssl rsa' use the general key loading routines
11875 implemented in `apps.c`, and make those routines able to
11876 handle the key format FORMAT_NETSCAPE and the variant
11879 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11881 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11883 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11885 * Add -keyform to rsautl, and document -engine.
11887 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11889 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11890 BIO_R_NO_SUCH_FILE error code rather than the generic
11891 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11895 * Add new functions
11896 ERR_peek_last_error
11897 ERR_peek_last_error_line
11898 ERR_peek_last_error_line_data.
11899 These are similar to
11901 ERR_peek_error_line
11902 ERR_peek_error_line_data,
11903 but report on the latest error recorded rather than the first one
11904 still in the error queue.
11906 *Ben Laurie, Bodo Moeller*
11908 * default_algorithms option in ENGINE config module. This allows things
11910 default_algorithms = ALL
11911 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11915 * Preliminary ENGINE config module.
11919 * New experimental application configuration code.
11923 * Change the AES code to follow the same name structure as all other
11924 symmetric ciphers, and behave the same way. Move everything to
11925 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11927 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11929 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11931 *Ben Laurie and Theo de Raadt*
11933 * Add option to output public keys in req command.
11935 *Massimiliano Pala madwolf@openca.org*
11937 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11938 (up to about 10% better than before for P-192 and P-224).
11942 * New functions/macros
11944 SSL_CTX_set_msg_callback(ctx, cb)
11945 SSL_CTX_set_msg_callback_arg(ctx, arg)
11946 SSL_set_msg_callback(ssl, cb)
11947 SSL_set_msg_callback_arg(ssl, arg)
11949 to request calling a callback function
11951 void cb(int write_p, int version, int content_type,
11952 const void *buf, size_t len, SSL *ssl, void *arg)
11954 whenever a protocol message has been completely received
11955 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11956 protocol version according to which the SSL library interprets
11957 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11958 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11959 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11960 specification (change_cipher_spec(20), alert(21), handshake(22)).
11961 'buf' and 'len' point to the actual message, 'ssl' to the
11962 SSL object, and 'arg' is the application-defined value set by
11963 SSL[_CTX]_set_msg_callback_arg().
11965 'openssl s_client' and 'openssl s_server' have new '-msg' options
11966 to enable a callback that displays all protocol messages.
11970 * Change the shared library support so shared libraries are built as
11971 soon as the corresponding static library is finished, and thereby get
11972 openssl and the test programs linked against the shared library.
11973 This still only happens when the keyword "shard" has been given to
11974 the configuration scripts.
11976 NOTE: shared library support is still an experimental thing, and
11977 backward binary compatibility is still not guaranteed.
11979 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11981 * Add support for Subject Information Access extension.
11983 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11985 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11986 additional bytes when new memory had to be allocated, not just
11987 when reusing an existing buffer.
11991 * New command line and configuration option 'utf8' for the req command.
11992 This allows field values to be specified as UTF8 strings.
11996 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11997 runs for the former and machine-readable output for the latter.
12001 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12002 of the e-mail address in the DN (i.e., it will go into a certificate
12003 extension only). The new configuration file option 'email_in_dn = no'
12004 has the same effect.
12006 *Massimiliano Pala madwolf@openca.org*
12008 * Change all functions with names starting with `des_` to be starting
12009 with `DES_` instead. Add wrappers that are compatible with libdes,
12010 but are named `_ossl_old_des_*`. Finally, add macros that map the
12011 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12012 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12013 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12016 Since we provide two compatibility mappings, the user needs to
12017 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12018 compatibility is desired. The default (i.e., when that macro
12019 isn't defined) is OpenSSL 0.9.6c compatibility.
12021 There are also macros that enable and disable the support of old
12022 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12023 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12024 are defined, the default will apply: to support the old des routines.
12026 In either case, one must include openssl/des.h to get the correct
12027 definitions. Do not try to just include openssl/des_old.h, that
12030 NOTE: This is a major break of an old API into a new one. Software
12031 authors are encouraged to switch to the `DES_` style functions. Some
12032 time in the future, des_old.h and the libdes compatibility functions
12033 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12034 default), and then completely removed.
12038 * Test for certificates which contain unsupported critical extensions.
12039 If such a certificate is found during a verify operation it is
12040 rejected by default: this behaviour can be overridden by either
12041 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12042 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12043 X509_supported_extension() has also been added which returns 1 if a
12044 particular extension is supported.
12048 * Modify the behaviour of EVP cipher functions in similar way to digests
12049 to retain compatibility with existing code.
12053 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12054 compatibility with existing code. In particular the 'ctx' parameter does
12055 not have to be to be initialized before the call to EVP_DigestInit() and
12056 it is tidied up after a call to EVP_DigestFinal(). New function
12057 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12058 EVP_MD_CTX_copy() changed to not require the destination to be
12059 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12060 requires the destination to be valid.
12062 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12063 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12067 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12068 so that complete 'Handshake' protocol structures are kept in memory
12069 instead of overwriting 'msg_type' and 'length' with 'body' data.
12073 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12075 *Massimo Santin via Richard Levitte*
12077 * Major restructuring to the underlying ENGINE code. This includes
12078 reduction of linker bloat, separation of pure "ENGINE" manipulation
12079 (initialisation, etc) from functionality dealing with implementations
12080 of specific crypto interfaces. This change also introduces integrated
12081 support for symmetric ciphers and digest implementations - so ENGINEs
12082 can now accelerate these by providing EVP_CIPHER and EVP_MD
12083 implementations of their own. This is detailed in
12084 [crypto/engine/README.md](crypto/engine/README.md)
12085 as it couldn't be adequately described here. However, there are a few
12086 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12087 were changed in the original introduction of ENGINE code have now
12088 reverted back - the hooking from this code to ENGINE is now a good
12089 deal more passive and at run-time, operations deal directly with
12090 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12091 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12092 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12093 they were not being used by the framework as there is no concept of a
12094 BIGNUM_METHOD and they could not be generalised to the new
12095 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12096 ENGINE_cpy() has been removed as it cannot be consistently defined in
12101 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12105 * Change mkdef.pl to sort symbols that get the same entry number,
12106 and make sure the automatically generated functions `ERR_load_*`
12107 become part of libeay.num as well.
12111 * New function SSL_renegotiate_pending(). This returns true once
12112 renegotiation has been requested (either SSL_renegotiate() call
12113 or HelloRequest/ClientHello received from the peer) and becomes
12114 false once a handshake has been completed.
12115 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12116 sends a HelloRequest, but does not ensure that a handshake takes
12117 place. SSL_renegotiate_pending() is useful for checking if the
12118 client has followed the request.)
12122 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12123 By default, clients may request session resumption even during
12124 renegotiation (if session ID contexts permit); with this option,
12125 session resumption is possible only in the first handshake.
12127 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12128 more bits available for options that should not be part of
12129 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12133 * Add some demos for certificate and certificate request creation.
12137 * Make maximum certificate chain size accepted from the peer application
12138 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12139 "Douglas E. Engert" <deengert@anl.gov>.
12143 * Add support for shared libraries for Unixware-7
12144 (Boyd Lynn Gerber <gerberb@zenez.com>).
12148 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12149 be done prior to destruction. Use this to unload error strings from
12150 ENGINEs that load their own error strings. NB: This adds two new API
12151 functions to "get" and "set" this destroy handler in an ENGINE.
12155 * Alter all existing ENGINE implementations (except "openssl" and
12156 "openbsd") to dynamically instantiate their own error strings. This
12157 makes them more flexible to be built both as statically-linked ENGINEs
12158 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12159 Also, add stub code to each that makes building them as self-contained
12160 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12164 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12165 implementations into applications that are completely implemented in
12166 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12167 commands that can be used to configure what shared-library to load and
12168 to control aspects of the way it is handled. Also, made an update to
12169 the [README-Engine.md](README-Engine.md) file
12170 that brings its information up-to-date and
12171 provides some information and instructions on the "dynamic" ENGINE
12172 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12176 * Make it possible to unload ranges of ERR strings with a new
12177 "ERR_unload_strings" function.
12181 * Add a copy() function to EVP_MD.
12185 * Make EVP_MD routines take a context pointer instead of just the
12186 md_data void pointer.
12190 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12191 that the digest can only process a single chunk of data
12192 (typically because it is provided by a piece of
12193 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12194 is only going to provide a single chunk of data, and hence the
12195 framework needn't accumulate the data for oneshot drivers.
12199 * As with "ERR", make it possible to replace the underlying "ex_data"
12200 functions. This change also alters the storage and management of global
12201 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12202 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12203 index counters. The API functions that use this state have been changed
12204 to take a "class_index" rather than pointers to the class's local STACK
12205 and counter, and there is now an API function to dynamically create new
12206 classes. This centralisation allows us to (a) plug a lot of the
12207 thread-safety problems that existed, and (b) makes it possible to clean
12208 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12209 such data would previously have always leaked in application code and
12210 workarounds were in place to make the memory debugging turn a blind eye
12211 to it. Application code that doesn't use this new function will still
12212 leak as before, but their memory debugging output will announce it now
12213 rather than letting it slide.
12215 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12216 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12217 has a return value to indicate success or failure.
12221 * Make it possible to replace the underlying "ERR" functions such that the
12222 global state (2 LHASH tables and 2 locks) is only used by the "default"
12223 implementation. This change also adds two functions to "get" and "set"
12224 the implementation prior to it being automatically set the first time
12225 any other ERR function takes place. Ie. an application can call "get",
12226 pass the return value to a module it has just loaded, and that module
12227 can call its own "set" function using that value. This means the
12228 module's "ERR" operations will use (and modify) the error state in the
12229 application and not in its own statically linked copy of OpenSSL code.
12233 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12234 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12235 the operation, and provides a more encapsulated way for external code
12236 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12237 to use these functions rather than manually incrementing the counts.
12239 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12243 * Add EVP test program.
12247 * Add symmetric cipher support to ENGINE. Expect the API to change!
12251 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12252 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12253 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12254 These allow a CRL to be built without having to access X509_CRL fields
12255 directly. Modify 'ca' application to use new functions.
12259 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12260 bug workarounds. Rollback attack detection is a security feature.
12261 The problem will only arise on OpenSSL servers when TLSv1 is not
12262 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12263 Software authors not wanting to support TLSv1 will have special reasons
12264 for their choice and can explicitly enable this option.
12266 *Bodo Moeller, Lutz Jaenicke*
12268 * Rationalise EVP so it can be extended: don't include a union of
12269 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12270 (similar to those existing for EVP_CIPHER_CTX).
12275 EVP_MD_CTX_init(&md); /* new function call */
12276 EVP_DigestInit(&md, EVP_sha1());
12277 EVP_DigestUpdate(&md, in, len);
12278 EVP_DigestFinal(&md, out, NULL);
12279 EVP_MD_CTX_cleanup(&md); /* new function call */
12283 * Make DES key schedule conform to the usual scheme, as well as
12284 correcting its structure. This means that calls to DES functions
12285 now have to pass a pointer to a des_key_schedule instead of a
12286 plain des_key_schedule (which was actually always a pointer
12289 des_key_schedule ks;
12291 des_set_key_checked(..., &ks);
12292 des_ncbc_encrypt(..., &ks, ...);
12294 (Note that a later change renames 'des_...' into 'DES_...'.)
12298 * Initial reduction of linker bloat: the use of some functions, such as
12299 PEM causes large amounts of unused functions to be linked in due to
12300 poor organisation. For example pem_all.c contains every PEM function
12301 which has a knock on effect of linking in large amounts of (unused)
12302 ASN1 code. Grouping together similar functions and splitting unrelated
12303 functions prevents this.
12307 * Cleanup of EVP macros.
12311 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12312 correct `_ecb suffix`.
12316 * Add initial OCSP responder support to ocsp application. The
12317 revocation information is handled using the text based index
12318 use by the ca application. The responder can either handle
12319 requests generated internally, supplied in files (for example
12320 via a CGI script) or using an internal minimal server.
12324 * Add configuration choices to get zlib compression for TLS.
12328 * Changes to Kerberos SSL for RFC 2712 compliance:
12329 1. Implemented real KerberosWrapper, instead of just using
12330 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12331 2. Implemented optional authenticator field of KerberosWrapper.
12333 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12334 and authenticator structs; see crypto/krb5/.
12336 Generalized Kerberos calls to support multiple Kerberos libraries.
12337 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12338 via Richard Levitte*
12340 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12341 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12342 values for each of the key sizes rather than having just
12343 parameters (and 'speed' generating keys each time).
12347 * Speed up EVP routines.
12350 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12351 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12352 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12353 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12355 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12356 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12357 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12360 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12362 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12366 * Added the OS2-EMX target.
12368 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12370 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12371 New functions to support `NCONF` routines in extension code.
12372 New function `CONF_set_nconf()`
12373 to allow functions which take an `NCONF` to also handle the old `LHASH`
12374 structure: this means that the old `CONF` compatible routines can be
12375 retained (in particular w.rt. extensions) without having to duplicate the
12376 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12380 * Enhance the general user interface with mechanisms for inner control
12381 and with possibilities to have yes/no kind of prompts.
12385 * Change all calls to low-level digest routines in the library and
12386 applications to use EVP. Add missing calls to HMAC_cleanup() and
12387 don't assume HMAC_CTX can be copied using memcpy().
12389 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12391 * Add the possibility to control engines through control names but with
12392 arbitrary arguments instead of just a string.
12393 Change the key loaders to take a UI_METHOD instead of a callback
12394 function pointer. NOTE: this breaks binary compatibility with earlier
12395 versions of OpenSSL [engine].
12396 Adapt the nCipher code for these new conditions and add a card insertion
12401 * Enhance the general user interface with mechanisms to better support
12402 dialog box interfaces, application-defined prompts, the possibility
12403 to use defaults (for example default passwords from somewhere else)
12404 and interrupts/cancellations.
12408 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12409 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12413 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12414 tidy up some unnecessarily weird code in 'sk_new()').
12416 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12418 * Change the key loading routines for ENGINEs to use the same kind
12419 callback (pem_password_cb) as all other routines that need this
12424 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12425 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12426 than this minimum value is recommended.
12430 * New random seeder for OpenVMS, using the system process statistics
12431 that are easily reachable.
12435 * Windows apparently can't transparently handle global
12436 variables defined in DLLs. Initialisations such as:
12438 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12440 won't compile. This is used by the any applications that need to
12441 declare their own ASN1 modules. This was fixed by adding the option
12442 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12443 needed for static libraries under Win32.
12447 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12448 setting of purpose and trust fields. New X509_STORE trust and
12449 purpose functions and tidy up setting in other SSL functions.
12453 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12454 structure. These are inherited by X509_STORE_CTX when it is
12455 initialised. This allows various defaults to be set in the
12456 X509_STORE structure (such as flags for CRL checking and custom
12457 purpose or trust settings) for functions which only use X509_STORE_CTX
12458 internally such as S/MIME.
12460 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12461 trust settings if they are not set in X509_STORE. This allows X509_STORE
12462 purposes and trust (in S/MIME for example) to override any set by default.
12464 Add command line options for CRL checking to smime, s_client and s_server
12469 * Initial CRL based revocation checking. If the CRL checking flag(s)
12470 are set then the CRL is looked up in the X509_STORE structure and
12471 its validity and signature checked, then if the certificate is found
12472 in the CRL the verify fails with a revoked error.
12474 Various new CRL related callbacks added to X509_STORE_CTX structure.
12476 Command line options added to 'verify' application to support this.
12478 This needs some additional work, such as being able to handle multiple
12479 CRLs with different times, extension based lookup (rather than just
12480 by subject name) and ultimately more complete V2 CRL extension
12485 * Add a general user interface API (crypto/ui/). This is designed
12486 to replace things like des_read_password and friends (backward
12487 compatibility functions using this new API are provided).
12488 The purpose is to remove prompting functions from the DES code
12489 section as well as provide for prompting through dialog boxes in
12490 a window system and the like.
12494 * Add "ex_data" support to ENGINE so implementations can add state at a
12495 per-structure level rather than having to store it globally.
12499 * Make it possible for ENGINE structures to be copied when retrieved by
12500 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12501 This causes the "original" ENGINE structure to act like a template,
12502 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12503 operational state can be localised to each ENGINE structure, despite the
12504 fact they all share the same "methods". New ENGINE structures returned in
12505 this case have no functional references and the return value is the single
12506 structural reference. This matches the single structural reference returned
12507 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12512 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12513 needs to match any other type at all we need to manually clear the
12518 * Changes to the "openssl engine" utility to include;
12519 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12520 about an ENGINE's available control commands.
12521 - executing control commands from command line arguments using the
12522 '-pre' and '-post' switches. '-post' is only used if '-t' is
12523 specified and the ENGINE is successfully initialised. The syntax for
12524 the individual commands are colon-separated, for example;
12525 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12529 * New dynamic control command support for ENGINEs. ENGINEs can now
12530 declare their own commands (numbers), names (strings), descriptions,
12531 and input types for run-time discovery by calling applications. A
12532 subset of these commands are implicitly classed as "executable"
12533 depending on their input type, and only these can be invoked through
12534 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12535 can be based on user input, config files, etc). The distinction is
12536 that "executable" commands cannot return anything other than a boolean
12537 result and can only support numeric or string input, whereas some
12538 discoverable commands may only be for direct use through
12539 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12540 pointers, or other custom uses. The "executable" commands are to
12541 support parameterisations of ENGINE behaviour that can be
12542 unambiguously defined by ENGINEs and used consistently across any
12543 OpenSSL-based application. Commands have been added to all the
12544 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12545 control over shared-library paths without source code alterations.
12549 * Changed all ENGINE implementations to dynamically allocate their
12550 ENGINEs rather than declaring them statically. Apart from this being
12551 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12552 this also allows the implementations to compile without using the
12553 internal engine_int.h header.
12557 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12558 'const' value. Any code that should be able to modify a RAND_METHOD
12559 should already have non-const pointers to it (ie. they should only
12560 modify their own ones).
12564 * Made a variety of little tweaks to the ENGINE code.
12565 - "atalla" and "ubsec" string definitions were moved from header files
12566 to C code. "nuron" string definitions were placed in variables
12567 rather than hard-coded - allowing parameterisation of these values
12568 later on via ctrl() commands.
12569 - Removed unused "#if 0"'d code.
12570 - Fixed engine list iteration code so it uses ENGINE_free() to release
12571 structural references.
12572 - Constified the RAND_METHOD element of ENGINE structures.
12573 - Constified various get/set functions as appropriate and added
12574 missing functions (including a catch-all ENGINE_cpy that duplicates
12575 all ENGINE values onto a new ENGINE except reference counts/state).
12576 - Removed NULL parameter checks in get/set functions. Setting a method
12577 or function to NULL is a way of cancelling out a previously set
12578 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12579 and doesn't justify the extra error symbols and code.
12580 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12581 flags from engine_int.h to engine.h.
12582 - Changed prototypes for ENGINE handler functions (init(), finish(),
12583 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12587 * Implement binary inversion algorithm for BN_mod_inverse in addition
12588 to the algorithm using long division. The binary algorithm can be
12589 used only if the modulus is odd. On 32-bit systems, it is faster
12590 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12591 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12592 up to 450 bits. In 64-bit environments, the binary algorithm
12593 appears to be advantageous for much longer moduli; here we use it
12594 for moduli up to 2048 bits.
12598 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12599 could not support the combine flag in choice fields.
12603 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12604 extensions from a certificate request to the certificate.
12608 * Allow multiple 'certopt' and 'nameopt' options to be separated
12609 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12610 file: this allows the display of the certificate about to be
12611 signed to be customised, to allow certain fields to be included
12612 or excluded and extension details. The old system didn't display
12613 multicharacter strings properly, omitted fields not in the policy
12614 and couldn't display additional details such as extensions.
12618 * Function EC_POINTs_mul for multiple scalar multiplication
12619 of an arbitrary number of elliptic curve points
12620 \sum scalars[i]*points[i],
12621 optionally including the generator defined for the EC_GROUP:
12622 scalar*generator + \sum scalars[i]*points[i].
12624 EC_POINT_mul is a simple wrapper function for the typical case
12625 that the point list has just one item (besides the optional
12630 * First EC_METHODs for curves over GF(p):
12632 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12633 operations and provides various method functions that can also
12634 operate with faster implementations of modular arithmetic.
12636 EC_GFp_mont_method() reuses most functions that are part of
12637 EC_GFp_simple_method, but uses Montgomery arithmetic.
12639 *Bodo Moeller; point addition and point doubling
12640 implementation directly derived from source code provided by
12641 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12643 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12644 crypto/ec/ec_lib.c):
12646 Curves are EC_GROUP objects (with an optional group generator)
12647 based on EC_METHODs that are built into the library.
12649 Points are EC_POINT objects based on EC_GROUP objects.
12651 Most of the framework would be able to handle curves over arbitrary
12652 finite fields, but as there are no obvious types for fields other
12653 than GF(p), some functions are limited to that for now.
12657 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12658 that the file contains a complete HTTP response.
12662 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12663 change the def and num file printf format specifier from "%-40sXXX"
12664 to "%-39s XXX". The latter will always guarantee a space after the
12665 field while the former will cause them to run together if the field
12666 is 40 of more characters long.
12670 * Constify the cipher and digest 'method' functions and structures
12671 and modify related functions to take constant EVP_MD and EVP_CIPHER
12676 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12677 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12681 * Modify `EVP_Digest*()` routines so they now return values. Although the
12682 internal software routines can never fail additional hardware versions
12687 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12689 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12690 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12693 ERR_R_NESTED_ASN1_ERROR
12695 ERR_R_MISSING_ASN1_EOS
12696 were 4 .. 9, conflicting with
12697 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12699 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12700 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12702 Add new error code 'ERR_R_INTERNAL_ERROR'.
12706 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12711 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12712 sets the subject name for a new request or supersedes the
12713 subject name in a given request. Formats that can be parsed are
12714 'CN=Some Name, OU=myOU, C=IT'
12716 'CN=Some Name/OU=myOU/C=IT'.
12718 Add options '-batch' and '-verbose' to 'openssl req'.
12720 *Massimiliano Pala <madwolf@hackmasters.net>*
12722 * Introduce the possibility to access global variables through
12723 functions on platform were that's the best way to handle exporting
12724 global variables in shared libraries. To enable this functionality,
12725 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12726 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12727 is normally done by Configure or something similar).
12729 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12730 in the source file (foo.c) like this:
12732 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12733 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12735 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12736 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12738 OPENSSL_DECLARE_GLOBAL(int,foo);
12739 #define foo OPENSSL_GLOBAL_REF(foo)
12740 OPENSSL_DECLARE_GLOBAL(double,bar);
12741 #define bar OPENSSL_GLOBAL_REF(bar)
12743 The #defines are very important, and therefore so is including the
12744 header file everywhere where the defined globals are used.
12746 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12747 of ASN.1 items, but that structure is a bit different.
12749 The largest change is in util/mkdef.pl which has been enhanced with
12750 better and easier to understand logic to choose which symbols should
12751 go into the Windows .def files as well as a number of fixes and code
12752 cleanup (among others, algorithm keywords are now sorted
12753 lexicographically to avoid constant rewrites).
12757 * In BN_div() keep a copy of the sign of 'num' before writing the
12758 result to 'rm' because if rm==num the value will be overwritten
12759 and produce the wrong result if 'num' is negative: this caused
12760 problems with BN_mod() and BN_nnmod().
12764 * Function OCSP_request_verify(). This checks the signature on an
12765 OCSP request and verifies the signer certificate. The signer
12766 certificate is just checked for a generic purpose and OCSP request
12771 * Add OCSP_check_validity() function to check the validity of OCSP
12772 responses. OCSP responses are prepared in real time and may only
12773 be a few seconds old. Simply checking that the current time lies
12774 between thisUpdate and nextUpdate max reject otherwise valid responses
12775 caused by either OCSP responder or client clock inaccuracy. Instead
12776 we allow thisUpdate and nextUpdate to fall within a certain period of
12777 the current time. The age of the response can also optionally be
12778 checked. Two new options -validity_period and -status_age added to
12783 * If signature or public key algorithm is unrecognized print out its
12784 OID rather that just UNKNOWN.
12788 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12789 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12790 ID to be generated from the issuer certificate alone which can then be
12791 passed to OCSP_id_issuer_cmp().
12795 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12796 ASN1 modules to export functions returning ASN1_ITEM pointers
12797 instead of the ASN1_ITEM structures themselves. This adds several
12798 new macros which allow the underlying ASN1 function/structure to
12799 be accessed transparently. As a result code should not use ASN1_ITEM
12800 references directly (such as &X509_it) but instead use the relevant
12801 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12802 use of the new ASN1 code on platforms where exporting structures
12803 is problematical (for example in shared libraries) but exporting
12804 functions returning pointers to structures is not.
12808 * Add support for overriding the generation of SSL/TLS session IDs.
12809 These callbacks can be registered either in an SSL_CTX or per SSL.
12810 The purpose of this is to allow applications to control, if they wish,
12811 the arbitrary values chosen for use as session IDs, particularly as it
12812 can be useful for session caching in multiple-server environments. A
12813 command-line switch for testing this (and any client code that wishes
12814 to use such a feature) has been added to "s_server".
12816 *Geoff Thorpe, Lutz Jaenicke*
12818 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12819 of the form `#if defined(...) || defined(...) || ...` and
12820 `#if !defined(...) && !defined(...) && ...`. This also avoids
12821 the growing number of special cases it was previously handling.
12825 * Make all configuration macros available for application by making
12826 sure they are available in opensslconf.h, by giving them names starting
12827 with `OPENSSL_` to avoid conflicts with other packages and by making
12828 sure e_os2.h will cover all platform-specific cases together with
12830 Additionally, it is now possible to define configuration/platform-
12831 specific names (called "system identities"). In the C code, these
12832 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12833 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12834 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12839 * New option -set_serial to 'req' and 'x509' this allows the serial
12840 number to use to be specified on the command line. Previously self
12841 signed certificates were hard coded with serial number 0 and the
12842 CA options of 'x509' had to use a serial number in a file which was
12847 * New options to 'ca' utility to support V2 CRL entry extensions.
12848 Currently CRL reason, invalidity date and hold instruction are
12849 supported. Add new CRL extensions to V3 code and some new objects.
12853 * New function EVP_CIPHER_CTX_set_padding() this is used to
12854 disable standard block padding (aka PKCS#5 padding) in the EVP
12855 API, which was previously mandatory. This means that the data is
12856 not padded in any way and so the total length much be a multiple
12857 of the block size, otherwise an error occurs.
12861 * Initial (incomplete) OCSP SSL support.
12865 * New function OCSP_parse_url(). This splits up a URL into its host,
12866 port and path components: primarily to parse OCSP URLs. New -url
12867 option to ocsp utility.
12871 * New nonce behavior. The return value of OCSP_check_nonce() now
12872 reflects the various checks performed. Applications can decide
12873 whether to tolerate certain situations such as an absent nonce
12874 in a response when one was present in a request: the ocsp application
12875 just prints out a warning. New function OCSP_add1_basic_nonce()
12876 this is to allow responders to include a nonce in a response even if
12877 the request is nonce-less.
12881 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12882 skipped when using openssl x509 multiple times on a single input file,
12883 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12887 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12888 set string type: to handle setting ASN1_TIME structures. Fix ca
12889 utility to correctly initialize revocation date of CRLs.
12893 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12894 the clients preferred ciphersuites and rather use its own preferences.
12895 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12896 Internet Explorer by ensuring unchanged hash method during stepup.
12897 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12901 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12902 to aes and add a new 'exist' option to print out symbols that don't
12907 * Additional options to ocsp utility to allow flags to be set and
12908 additional certificates supplied.
12912 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12913 OCSP client a number of certificate to only verify the response
12918 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12919 handle the new API. Currently only ECB, CBC modes supported. Add new
12922 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12923 Encryption Standard (AES) Ciphersuites for Transport Layer
12924 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12925 not enabled by default and were not part of the "ALL" ciphersuite
12926 alias because they were not yet official; they could be
12927 explicitly requested by specifying the "AESdraft" ciphersuite
12928 group alias. In the final release of OpenSSL 0.9.7, the group
12929 alias is called "AES" and is part of "ALL".)
12931 *Ben Laurie, Steve Henson, Bodo Moeller*
12933 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12934 request to response.
12938 * Functions for OCSP responders. OCSP_request_onereq_count(),
12939 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12940 extract information from a certificate request. OCSP_response_create()
12941 creates a response and optionally adds a basic response structure.
12942 OCSP_basic_add1_status() adds a complete single response to a basic
12943 response and returns the OCSP_SINGLERESP structure just added (to allow
12944 extensions to be included for example). OCSP_basic_add1_cert() adds a
12945 certificate to a basic response and OCSP_basic_sign() signs a basic
12946 response with various flags. New helper functions ASN1_TIME_check()
12947 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12948 (converts ASN1_TIME to GeneralizedTime).
12952 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12953 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12954 structure from a certificate. X509_pubkey_digest() digests the public_key
12955 contents: this is used in various key identifiers.
12959 * Make sk_sort() tolerate a NULL argument.
12961 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12963 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12964 passed by the function are trusted implicitly. If any of them signed the
12965 response then it is assumed to be valid and is not verified.
12969 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12970 to data. This was previously part of the PKCS7 ASN1 code. This
12971 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12972 *Steve Henson, reported by Kenneth R. Robinette
12973 <support@securenetterm.com>*
12975 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12976 routines: without these tracing memory leaks is very painful.
12977 Fix leaks in PKCS12 and PKCS7 routines.
12981 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12982 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12983 effectively meant GeneralizedTime would never be used. Now it
12984 is initialised to -1 but X509_time_adj() now has to check the value
12985 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12986 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12987 *Steve Henson, reported by Kenneth R. Robinette
12988 <support@securenetterm.com>*
12990 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12991 result in a zero length in the ASN1_INTEGER structure which was
12992 not consistent with the structure when d2i_ASN1_INTEGER() was used
12993 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12994 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12995 where it did not print out a minus for negative ASN1_INTEGER.
12999 * Add summary printout to ocsp utility. The various functions which
13000 convert status values to strings have been renamed to:
13001 OCSP_response_status_str(), OCSP_cert_status_str() and
13002 OCSP_crl_reason_str() and are no longer static. New options
13003 to verify nonce values and to disable verification. OCSP response
13004 printout format cleaned up.
13008 * Add additional OCSP certificate checks. These are those specified
13009 in RFC2560. This consists of two separate checks: the CA of the
13010 certificate being checked must either be the OCSP signer certificate
13011 or the issuer of the OCSP signer certificate. In the latter case the
13012 OCSP signer certificate must contain the OCSP signing extended key
13013 usage. This check is performed by attempting to match the OCSP
13014 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13015 in the OCSP_CERTID structures of the response.
13019 * Initial OCSP certificate verification added to OCSP_basic_verify()
13020 and related routines. This uses the standard OpenSSL certificate
13021 verify routines to perform initial checks (just CA validity) and
13022 to obtain the certificate chain. Then additional checks will be
13023 performed on the chain. Currently the root CA is checked to see
13024 if it is explicitly trusted for OCSP signing. This is used to set
13025 a root CA as a global signing root: that is any certificate that
13026 chains to that CA is an acceptable OCSP signing certificate.
13030 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13031 extensions from a separate configuration file.
13032 As when reading extensions from the main configuration file,
13033 the '-extensions ...' option may be used for specifying the
13036 *Massimiliano Pala <madwolf@comune.modena.it>*
13038 * New OCSP utility. Allows OCSP requests to be generated or
13039 read. The request can be sent to a responder and the output
13040 parsed, outputted or printed in text form. Not complete yet:
13041 still needs to check the OCSP response validity.
13045 * New subcommands for 'openssl ca':
13046 `openssl ca -status <serial>` prints the status of the cert with
13047 the given serial number (according to the index file).
13048 `openssl ca -updatedb` updates the expiry status of certificates
13051 *Massimiliano Pala <madwolf@comune.modena.it>*
13053 * New '-newreq-nodes' command option to CA.pl. This is like
13054 '-newreq', but calls 'openssl req' with the '-nodes' option
13055 so that the resulting key is not encrypted.
13057 *Damien Miller <djm@mindrot.org>*
13059 * New configuration for the GNU Hurd.
13061 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13063 * Initial code to implement OCSP basic response verify. This
13064 is currently incomplete. Currently just finds the signer's
13065 certificate and verifies the signature on the response.
13069 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13070 value of OPENSSLDIR. This is available via the new '-d' option
13071 to 'openssl version', and is also included in 'openssl version -a'.
13075 * Allowing defining memory allocation callbacks that will be given
13076 file name and line number information in additional arguments
13077 (a `const char*` and an int). The basic functionality remains, as
13078 well as the original possibility to just replace malloc(),
13079 realloc() and free() by functions that do not know about these
13080 additional arguments. To register and find out the current
13081 settings for extended allocation functions, the following
13082 functions are provided:
13084 CRYPTO_set_mem_ex_functions
13085 CRYPTO_set_locked_mem_ex_functions
13086 CRYPTO_get_mem_ex_functions
13087 CRYPTO_get_locked_mem_ex_functions
13089 These work the same way as CRYPTO_set_mem_functions and friends.
13090 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13091 extended allocation function is enabled.
13092 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13093 a conventional allocation function is enabled.
13095 *Richard Levitte, Bodo Moeller*
13097 * Finish off removing the remaining LHASH function pointer casts.
13098 There should no longer be any prototype-casting required when using
13099 the LHASH abstraction, and any casts that remain are "bugs". See
13100 the callback types and macros at the head of lhash.h for details
13101 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13105 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13106 If /dev/[u]random devices are not available or do not return enough
13107 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13109 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13110 /etc/entropy will be queried once each in this sequence, querying stops
13111 when enough entropy was collected without querying more sockets.
13115 * Change the Unix RAND_poll() variant to be able to poll several
13116 random devices, as specified by DEVRANDOM, until a sufficient amount
13117 of data has been collected. We spend at most 10 ms on each file
13118 (select timeout) and read in non-blocking mode. DEVRANDOM now
13119 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13120 (previously it was just the string "/dev/urandom"), so on typical
13121 platforms the 10 ms delay will never occur.
13122 Also separate out the Unix variant to its own file, rand_unix.c.
13123 For VMS, there's a currently-empty rand_vms.c.
13127 * Move OCSP client related routines to ocsp_cl.c. These
13128 provide utility functions which an application needing
13129 to issue a request to an OCSP responder and analyse the
13130 response will typically need: as opposed to those which an
13131 OCSP responder itself would need which will be added later.
13133 OCSP_request_sign() signs an OCSP request with an API similar
13134 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13135 response. OCSP_response_get1_basic() extracts basic response
13136 from response. OCSP_resp_find_status(): finds and extracts status
13137 information from an OCSP_CERTID structure (which will be created
13138 when the request structure is built). These are built from lower
13139 level functions which work on OCSP_SINGLERESP structures but
13140 won't normally be used unless the application wishes to examine
13141 extensions in the OCSP response for example.
13143 Replace nonce routines with a pair of functions.
13144 OCSP_request_add1_nonce() adds a nonce value and optionally
13145 generates a random value. OCSP_check_nonce() checks the
13146 validity of the nonce in an OCSP response.
13150 * Change function OCSP_request_add() to OCSP_request_add0_id().
13151 This doesn't copy the supplied OCSP_CERTID and avoids the
13152 need to free up the newly created id. Change return type
13153 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13154 This can then be used to add extensions to the request.
13155 Deleted OCSP_request_new(), since most of its functionality
13156 is now in OCSP_REQUEST_new() (and the case insensitive name
13157 clash) apart from the ability to set the request name which
13158 will be added elsewhere.
13162 * Update OCSP API. Remove obsolete extensions argument from
13163 various functions. Extensions are now handled using the new
13164 OCSP extension code. New simple OCSP HTTP function which
13165 can be used to send requests and parse the response.
13169 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13170 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13171 uses the special reorder version of SET OF to sort the attributes
13172 and reorder them to match the encoded order. This resolves a long
13173 standing problem: a verify on a PKCS7 structure just after signing
13174 it used to fail because the attribute order did not match the
13175 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13176 it uses the received order. This is necessary to tolerate some broken
13177 software that does not order SET OF. This is handled by encoding
13178 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13179 to produce the required SET OF.
13183 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13184 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13185 files to get correct declarations of the ASN.1 item variables.
13189 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13190 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13191 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13192 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13193 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13194 ASN1_ITEM and no wrapper functions.
13198 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13199 replace the old function pointer based I/O routines. Change most of
13200 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13204 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13205 lines, recognize more "algorithms" that can be deselected, and make
13206 it complain about algorithm deselection that isn't recognised.
13210 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13211 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13212 to use new functions. Add NO_ASN1_OLD which can be set to remove
13213 some old style ASN1 functions: this can be used to determine if old
13214 code will still work when these eventually go away.
13218 * New extension functions for OCSP structures, these follow the
13219 same conventions as certificates and CRLs.
13223 * New function X509V3_add1_i2d(). This automatically encodes and
13224 adds an extension. Its behaviour can be customised with various
13225 flags to append, replace or delete. Various wrappers added for
13226 certificates and CRLs.
13230 * Fix to avoid calling the underlying ASN1 print routine when
13231 an extension cannot be parsed. Correct a typo in the
13232 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13236 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13237 entries for variables.
13241 * Add functionality to `apps/openssl.c` for detecting locking
13242 problems: As the program is single-threaded, all we have
13243 to do is register a locking callback using an array for
13244 storing which locks are currently held by the program.
13248 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13249 SSL_get_ex_data_X509_STORE_idx(), which is used in
13250 ssl_verify_cert_chain() and thus can be called at any time
13251 during TLS/SSL handshakes so that thread-safety is essential.
13252 Unfortunately, the ex_data design is not at all suited
13253 for multi-threaded use, so it probably should be abolished.
13257 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13259 *Broadcom, tweaked and integrated by Geoff Thorpe*
13261 * Move common extension printing code to new function
13262 X509V3_print_extensions(). Reorganise OCSP print routines and
13263 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13267 * New function X509_signature_print() to remove duplication in some
13272 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13273 set (this was treated exactly the same as SET OF previously). This
13274 is used to reorder the STACK representing the structure to match the
13275 encoding. This will be used to get round a problem where a PKCS7
13276 structure which was signed could not be verified because the STACK
13277 order did not reflect the encoded order.
13281 * Reimplement the OCSP ASN1 module using the new code.
13285 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13286 for its ASN1 operations. The old style function pointers still exist
13287 for now but they will eventually go away.
13291 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13292 completely replaces the old ASN1 functionality with a table driven
13293 encoder and decoder which interprets an ASN1_ITEM structure describing
13294 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13295 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13296 has also been converted to the new form.
13300 * Change BN_mod_exp_recp so that negative moduli are tolerated
13301 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13302 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13303 for negative moduli.
13307 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13308 of not touching the result's sign bit.
13312 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13317 * Changed the LHASH code to use prototypes for callbacks, and created
13318 macros to declare and implement thin (optionally static) functions
13319 that provide type-safety and avoid function pointer casting for the
13320 type-specific callbacks.
13324 * Added Kerberos Cipher Suites to be used with TLS, as written in
13326 *Veers Staats <staatsvr@asc.hpc.mil>,
13327 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13329 * Reformat the FAQ so the different questions and answers can be divided
13330 in sections depending on the subject.
13334 * Have the zlib compression code load ZLIB.DLL dynamically under
13339 * New function BN_mod_sqrt for computing square roots modulo a prime
13340 (using the probabilistic Tonelli-Shanks algorithm unless
13341 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13342 be handled deterministically).
13344 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13346 * Make BN_mod_inverse faster by explicitly handling small quotients
13347 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13348 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13352 * New function BN_kronecker.
13356 * Fix BN_gcd so that it works on negative inputs; the result is
13357 positive unless both parameters are zero.
13358 Previously something reasonably close to an infinite loop was
13359 possible because numbers could be growing instead of shrinking
13360 in the implementation of Euclid's algorithm.
13364 * Fix BN_is_word() and BN_is_one() macros to take into account the
13365 sign of the number in question.
13367 Fix BN_is_word(a,w) to work correctly for w == 0.
13369 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13370 because its test if the absolute value of 'a' equals 'w'.
13371 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13372 it exists mostly for use in the implementations of BN_is_zero(),
13373 BN_is_one(), and BN_is_word().
13377 * New function BN_swap.
13381 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13382 the exponentiation functions are more likely to produce reasonable
13383 results on negative inputs.
13387 * Change BN_mod_mul so that the result is always non-negative.
13388 Previously, it could be negative if one of the factors was negative;
13389 I don't think anyone really wanted that behaviour.
13393 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13394 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13395 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13396 and add new functions:
13405 BN_mod_lshift1_quick
13407 BN_mod_lshift_quick
13409 These functions always generate non-negative results.
13411 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13412 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13414 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13415 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13416 be reduced modulo `m`.
13418 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13421 The following entry accidentally appeared in the CHANGES file
13422 distributed with OpenSSL 0.9.7. The modifications described in
13423 it do *not* apply to OpenSSL 0.9.7.
13425 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13426 was actually never needed) and in BN_mul(). The removal in BN_mul()
13427 required a small change in bn_mul_part_recursive() and the addition
13428 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13429 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13430 bn_sub_words() and bn_add_words() except they take arrays with
13436 * In 'openssl passwd', verify passwords read from the terminal
13437 unless the '-salt' option is used (which usually means that
13438 verification would just waste user's time since the resulting
13439 hash is going to be compared with some given password hash)
13440 or the new '-noverify' option is used.
13442 This is an incompatible change, but it does not affect
13443 non-interactive use of 'openssl passwd' (passwords on the command
13444 line, '-stdin' option, '-in ...' option) and thus should not
13445 cause any problems.
13449 * Remove all references to RSAref, since there's no more need for it.
13453 * Make DSO load along a path given through an environment variable
13454 (SHLIB_PATH) with shl_load().
13458 * Constify the ENGINE code as a result of BIGNUM constification.
13459 Also constify the RSA code and most things related to it. In a
13460 few places, most notable in the depth of the ASN.1 code, ugly
13461 casts back to non-const were required (to be solved at a later
13466 * Make it so the openssl application has all engines loaded by default.
13470 * Constify the BIGNUM routines a little more.
13474 * Add the following functions:
13476 ENGINE_load_cswift()
13478 ENGINE_load_atalla()
13479 ENGINE_load_nuron()
13480 ENGINE_load_builtin_engines()
13482 That way, an application can itself choose if external engines that
13483 are built-in in OpenSSL shall ever be used or not. The benefit is
13484 that applications won't have to be linked with libdl or other dso
13485 libraries unless it's really needed.
13487 Changed 'openssl engine' to load all engines on demand.
13488 Changed the engine header files to avoid the duplication of some
13489 declarations (they differed!).
13493 * 'openssl engine' can now list capabilities.
13497 * Better error reporting in 'openssl engine'.
13501 * Never call load_dh_param(NULL) in s_server.
13505 * Add engine application. It can currently list engines by name and
13506 identity, and test if they are actually available.
13510 * Improve RPM specification file by forcing symbolic linking and making
13511 sure the installed documentation is also owned by root.root.
13513 *Damien Miller <djm@mindrot.org>*
13515 * Give the OpenSSL applications more possibilities to make use of
13516 keys (public as well as private) handled by engines.
13520 * Add OCSP code that comes from CertCo.
13524 * Add VMS support for the Rijndael code.
13528 * Added untested support for Nuron crypto accelerator.
13532 * Add support for external cryptographic devices. This code was
13533 previously distributed separately as the "engine" branch.
13535 *Geoff Thorpe, Richard Levitte*
13537 * Rework the filename-translation in the DSO code. It is now possible to
13538 have far greater control over how a "name" is turned into a filename
13539 depending on the operating environment and any oddities about the
13540 different shared library filenames on each system.
13544 * Support threads on FreeBSD-elf in Configure.
13548 * Fix for SHA1 assembly problem with MASM: it produces
13549 warnings about corrupt line number information when assembling
13550 with debugging information. This is caused by the overlapping
13553 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13556 NCONF_get_number() has no error checking at all. As a replacement,
13557 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13558 promoted strongly. The old NCONF_get_number is kept around for
13559 binary backward compatibility.
13560 Make it possible for methods to load from something other than a BIO,
13561 by providing a function pointer that is given a name instead of a BIO.
13562 For example, this could be used to load configuration data from an
13567 * Fix for non blocking accept BIOs. Added new I/O special reason
13568 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13569 with non blocking I/O was not possible because no retry code was
13570 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13575 * Added the beginnings of Rijndael support.
13579 * Fix for bug in DirectoryString mask setting. Add support for
13580 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13581 to allow certificate printing to more controllable, additional
13582 'certopt' option to 'x509' to allow new printing options to be
13587 * Clean old EAY MD5 hack from e_os.h.
13591 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13593 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13594 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13596 *Joe Orton, Steve Henson*
13598 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13600 * Fix additional bug revealed by the NISCC test suite:
13602 Stop bug triggering large recursion when presented with
13603 certain ASN.1 tags ([CVE-2003-0851])
13607 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13609 * Fix various bugs revealed by running the NISCC test suite:
13611 Stop out of bounds reads in the ASN1 code when presented with
13612 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13614 If verify callback ignores invalid public key errors don't try to check
13615 certificate signature with the NULL public key.
13619 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13620 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13625 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13626 extra data after the compression methods not only for TLS 1.0
13627 but also for SSL 3.0 (as required by the specification).
13629 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13631 * Change X509_certificate_type() to mark the key as exported/exportable
13632 when it's 512 *bits* long, not 512 bytes.
13636 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13638 * Countermeasure against the Klima-Pokorny-Rosa extension of
13639 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13640 a protocol version number mismatch like a decryption error
13641 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13645 * Turn on RSA blinding by default in the default implementation
13646 to avoid a timing attack. Applications that don't want it can call
13647 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13648 They would be ill-advised to do so in most cases.
13650 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13652 * Change RSA blinding code so that it works when the PRNG is not
13653 seeded (in this case, the secret RSA exponent is abused as
13654 an unpredictable seed -- if it is not unpredictable, there
13655 is no point in blinding anyway). Make RSA blinding thread-safe
13656 by remembering the creator's thread ID in rsa->blinding and
13657 having all other threads use local one-time blinding factors
13658 (this requires more computation than sharing rsa->blinding, but
13659 avoids excessive locking; and if an RSA object is not shared
13660 between threads, blinding will still be very fast).
13664 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13666 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13667 via timing by performing a MAC computation even if incorrect
13668 block cipher padding has been found. This is a countermeasure
13669 against active attacks where the attacker has to distinguish
13670 between bad padding and a MAC verification error. ([CVE-2003-0078])
13672 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13673 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13674 Martin Vuagnoux (EPFL, Ilion)*
13676 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13678 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13679 memory from its contents. This is done with a counter that will
13680 place alternating values in each byte. This can be used to solve
13681 two issues: 1) the removal of calls to memset() by highly optimizing
13682 compilers, and 2) cleansing with other values than 0, since those can
13683 be read through on certain media, for example a swap space on disk.
13687 * Bugfix: client side session caching did not work with external caching,
13688 because the session->cipher setting was not restored when reloading
13689 from the external cache. This problem was masked, when
13690 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13691 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13695 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13696 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13698 *Zeev Lieber <zeev-l@yahoo.com>*
13700 * Undo an undocumented change introduced in 0.9.6e which caused
13701 repeated calls to OpenSSL_add_all_ciphers() and
13702 OpenSSL_add_all_digests() to be ignored, even after calling
13707 * Change the default configuration reader to deal with last line not
13708 being properly terminated.
13712 * Change X509_NAME_cmp() so it applies the special rules on handling
13713 DN values that are of type PrintableString, as well as RDNs of type
13714 emailAddress where the value has the type ia5String.
13716 *stefank@valicert.com via Richard Levitte*
13718 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13719 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13720 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13721 the bitwise-OR of the two for use by the majority of applications
13722 wanting this behaviour, and update the docs. The documented
13723 behaviour and actual behaviour were inconsistent and had been
13724 changing anyway, so this is more a bug-fix than a behavioural
13727 *Geoff Thorpe, diagnosed by Nadav Har'El*
13729 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13730 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13734 * Fix initialization code race conditions in
13735 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13736 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13737 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13738 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13739 ssl2_get_cipher_by_char(),
13740 ssl3_get_cipher_by_char().
13742 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13744 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13745 the cached sessions are flushed, as the remove_cb() might use ex_data
13746 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13747 (see [openssl.org #212]).
13749 *Geoff Thorpe, Lutz Jaenicke*
13751 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13752 length, instead of the encoding length to d2i_ASN1_OBJECT.
13756 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13758 * [In 0.9.6g-engine release:]
13759 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13761 *Lynn Gazis <lgazis@rainbow.com>*
13763 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13765 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13766 and get fix the header length calculation.
13767 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13768 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13770 * Use proper error handling instead of 'assertions' in buffer
13771 overflow checks added in 0.9.6e. This prevents DoS (the
13772 assertions could call abort()).
13774 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13776 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13778 * Add various sanity checks to asn1_get_length() to reject
13779 the ASN1 length bytes if they exceed sizeof(long), will appear
13780 negative or the content length exceeds the length of the
13783 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13785 * Fix cipher selection routines: ciphers without encryption had no flags
13786 for the cipher strength set and where therefore not handled correctly
13787 by the selection routines (PR #130).
13791 * Fix EVP_dsa_sha macro.
13796 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13797 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13798 that was added in OpenSSL 0.9.6d.
13800 As the countermeasure turned out to be incompatible with some
13801 broken SSL implementations, the new option is part of SSL_OP_ALL.
13802 SSL_OP_ALL is usually employed when compatibility with weird SSL
13803 implementations is desired (e.g. '-bugs' option to 's_client' and
13804 's_server'), so the new option is automatically set in many
13809 * Changes in security patch:
13811 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13812 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13813 Air Force Materiel Command, USAF, under agreement number
13816 * Add various sanity checks to asn1_get_length() to reject
13817 the ASN1 length bytes if they exceed sizeof(long), will appear
13818 negative or the content length exceeds the length of the
13819 supplied buffer. ([CVE-2002-0659])
13821 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13823 * Assertions for various potential buffer overflows, not known to
13824 happen in practice.
13826 *Ben Laurie (CHATS)*
13828 * Various temporary buffers to hold ASCII versions of integers were
13829 too small for 64 bit platforms. ([CVE-2002-0655])
13830 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13832 * Remote buffer overflow in SSL3 protocol - an attacker could
13833 supply an oversized session ID to a client. ([CVE-2002-0656])
13835 *Ben Laurie (CHATS)*
13837 * Remote buffer overflow in SSL2 protocol - an attacker could
13838 supply an oversized client master key. ([CVE-2002-0656])
13840 *Ben Laurie (CHATS)*
13842 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13844 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13845 encoded as NULL) with id-dsa-with-sha1.
13847 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13849 * Check various `X509_...()` return values in `apps/req.c`.
13851 *Nils Larsch <nla@trustcenter.de>*
13853 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13854 an end-of-file condition would erroneously be flagged, when the CRLF
13855 was just at the end of a processed block. The bug was discovered when
13856 processing data through a buffering memory BIO handing the data to a
13857 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13858 <ptsekov@syntrex.com> and Nedelcho Stanev.
13862 * Implement a countermeasure against a vulnerability recently found
13863 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13864 before application data chunks to avoid the use of known IVs
13865 with data potentially chosen by the attacker.
13869 * Fix length checks in ssl3_get_client_hello().
13873 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13874 to prevent ssl3_read_internal() from incorrectly assuming that
13875 ssl3_read_bytes() found application data while handshake
13876 processing was enabled when in fact s->s3->in_read_app_data was
13877 merely automatically cleared during the initial handshake.
13879 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13881 * Fix object definitions for Private and Enterprise: they were not
13882 recognized in their shortname (=lowercase) representation. Extend
13883 obj_dat.pl to issue an error when using undefined keywords instead
13884 of silently ignoring the problem (Svenning Sorensen
13885 <sss@sss.dnsalias.net>).
13889 * Fix DH_generate_parameters() so that it works for 'non-standard'
13890 generators, i.e. generators other than 2 and 5. (Previously, the
13891 code did not properly initialise the 'add' and 'rem' values to
13892 BN_generate_prime().)
13894 In the new general case, we do not insist that 'generator' is
13895 actually a primitive root: This requirement is rather pointless;
13896 a generator of the order-q subgroup is just as good, if not
13901 * Map new X509 verification errors to alerts. Discovered and submitted by
13902 Tom Wu <tom@arcot.com>.
13906 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13907 returning non-zero before the data has been completely received
13908 when using non-blocking I/O.
13910 *Bodo Moeller; problem pointed out by John Hughes*
13912 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13914 *Ben Laurie, Lutz Jaenicke*
13916 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13917 Yoram Zahavi <YoramZ@gilian.com>).
13921 * Add information about CygWin 1.3 and on, and preserve proper
13922 configuration for the versions before that.
13924 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13926 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13927 check whether we deal with a copy of a session and do not delete from
13928 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13929 <izhar@checkpoint.com>.
13933 * Do not store session data into the internal session cache, if it
13934 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13935 flag is set). Proposed by Aslam <aslam@funk.com>.
13939 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13944 * [In 0.9.6d-engine release:]
13945 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13947 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13949 * Add the configuration target linux-s390x.
13951 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13953 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13954 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13955 variable as an indication that a ClientHello message has been
13956 received. As the flag value will be lost between multiple
13957 invocations of ssl3_accept when using non-blocking I/O, the
13958 function may not be aware that a handshake has actually taken
13959 place, thus preventing a new session from being added to the
13962 To avoid this problem, we now set s->new_session to 2 instead of
13963 using a local variable.
13965 *Lutz Jaenicke, Bodo Moeller*
13967 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13968 if the SSL_R_LENGTH_MISMATCH error is detected.
13970 *Geoff Thorpe, Bodo Moeller*
13972 * New 'shared_ldflag' column in Configure platform table.
13976 * Fix EVP_CIPHER_mode macro.
13978 *"Dan S. Camper" <dan@bti.net>*
13980 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13981 type, we must throw them away by setting rr->length to 0.
13983 *D P Chang <dpc@qualys.com>*
13985 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13987 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13988 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13989 worked incorrectly for those cases where range = `10..._2` and
13990 `3*range` is two bits longer than range.)
13994 * Only add signing time to PKCS7 structures if it is not already
13999 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14000 OBJ_ld_ce should be OBJ_id_ce.
14001 Also some ip-pda OIDs in crypto/objects/objects.txt were
14002 incorrect (cf. RFC 3039).
14004 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14006 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14007 returns early because it has nothing to do.
14009 *Andy Schneider <andy.schneider@bjss.co.uk>*
14011 * [In 0.9.6c-engine release:]
14012 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14014 *Andy Schneider <andy.schneider@bjss.co.uk>*
14016 * [In 0.9.6c-engine release:]
14017 Add support for Cryptographic Appliance's keyserver technology.
14018 (Use engine 'keyclient')
14020 *Cryptographic Appliances and Geoff Thorpe*
14022 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14023 is called via tools/c89.sh because arguments have to be
14024 rearranged (all '-L' options must appear before the first object
14027 *Richard Shapiro <rshapiro@abinitio.com>*
14029 * [In 0.9.6c-engine release:]
14030 Add support for Broadcom crypto accelerator cards, backported
14033 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14035 * [In 0.9.6c-engine release:]
14036 Add support for SureWare crypto accelerator cards from
14037 Baltimore Technologies. (Use engine 'sureware')
14039 *Baltimore Technologies and Mark Cox*
14041 * [In 0.9.6c-engine release:]
14042 Add support for crypto accelerator cards from Accelerated
14043 Encryption Processing, www.aep.ie. (Use engine 'aep')
14045 *AEP Inc. and Mark Cox*
14047 * Add a configuration entry for gcc on UnixWare.
14049 *Gary Benson <gbenson@redhat.com>*
14051 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14052 messages are stored in a single piece (fixed-length part and
14053 variable-length part combined) and fix various bugs found on the way.
14057 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14058 instead. BIO_gethostbyname() does not know what timeouts are
14059 appropriate, so entries would stay in cache even when they have
14061 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14063 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14064 faced with a pathologically small ClientHello fragment that does
14065 not contain client_version: Instead of aborting with an error,
14066 simply choose the highest available protocol version (i.e.,
14067 TLS 1.0 unless it is disabled). In practice, ClientHello
14068 messages are never sent like this, but this change gives us
14069 strictly correct behaviour at least for TLS.
14073 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14074 never resets s->method to s->ctx->method when called from within
14075 one of the SSL handshake functions.
14077 *Bodo Moeller; problem pointed out by Niko Baric*
14079 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14080 (sent using the client's version number) if client_version is
14081 smaller than the protocol version in use. Also change
14082 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14083 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14084 the client will at least see that alert.
14088 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14093 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14094 client receives HelloRequest while in a handshake.
14096 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14098 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14099 should end in 'break', not 'goto end' which circumvents various
14100 cleanups done in state SSL_ST_OK. But session related stuff
14101 must be disabled for SSL_ST_OK in the case that we just sent a
14104 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14105 before just sending a HelloRequest.
14107 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14109 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14110 reveal whether illegal block cipher padding was found or a MAC
14111 verification error occurred. (Neither SSLerr() codes nor alerts
14112 are directly visible to potential attackers, but the information
14113 may leak via logfiles.)
14115 Similar changes are not required for the SSL 2.0 implementation
14116 because the number of padding bytes is sent in clear for SSL 2.0,
14117 and the extra bytes are just ignored. However ssl/s2_pkt.c
14118 failed to verify that the purported number of padding bytes is in
14123 * Add OpenUNIX-8 support including shared libraries
14124 (Boyd Lynn Gerber <gerberb@zenez.com>).
14128 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14129 'wristwatch attack' using huge encoding parameters (cf.
14130 James H. Manger's CRYPTO 2001 paper). Note that the
14131 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14132 encoding parameters and hence was not vulnerable.
14136 * BN_sqr() bug fix.
14138 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14140 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14141 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14142 followed by modular reduction.
14144 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14146 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14147 equivalent based on BN_pseudo_rand() instead of BN_rand().
14151 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14152 This function was broken, as the check for a new client hello message
14153 to handle SGC did not allow these large messages.
14154 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14158 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14162 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14163 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14167 * Rework the configuration and shared library support for Tru64 Unix.
14168 The configuration part makes use of modern compiler features and
14169 still retains old compiler behavior for those that run older versions
14170 of the OS. The shared library support part includes a variant that
14171 uses the RPATH feature, and is available through the special
14172 configuration target "alpha-cc-rpath", which will never be selected
14175 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14177 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14178 with the same message size as in ssl3_get_certificate_request().
14179 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14180 messages might inadvertently be reject as too long.
14182 *Petr Lampa <lampa@fee.vutbr.cz>*
14184 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14188 * Modified SSL library such that the verify_callback that has been set
14189 specifically for an SSL object with SSL_set_verify() is actually being
14190 used. Before the change, a verify_callback set with this function was
14191 ignored and the verify_callback() set in the SSL_CTX at the time of
14192 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14193 to allow the necessary settings.
14197 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14198 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14199 done automatically (in contradiction to the requirements of the C
14200 standard). This made problems when used from OpenSSH.
14204 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14205 dh->length and always used
14207 BN_rand_range(priv_key, dh->p).
14209 BN_rand_range() is not necessary for Diffie-Hellman, and this
14210 specific range makes Diffie-Hellman unnecessarily inefficient if
14211 dh->length (recommended exponent length) is much smaller than the
14212 length of dh->p. We could use BN_rand_range() if the order of
14213 the subgroup was stored in the DH structure, but we only have
14218 BN_rand(priv_key, l, ...)
14220 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14227 RSA_eay_public_encrypt
14228 RSA_eay_private_decrypt
14229 RSA_eay_private_encrypt (signing)
14230 RSA_eay_public_decrypt (signature verification)
14232 (default implementations for RSA_public_encrypt,
14233 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14234 always reject numbers >= n.
14238 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14239 to synchronize access to 'locking_thread'. This is necessary on
14240 systems where access to 'locking_thread' (an 'unsigned long'
14241 variable) is not atomic.
14245 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14246 *before* setting the 'crypto_lock_rand' flag. The previous code had
14247 a race condition if 0 is a valid thread ID.
14249 *Travis Vitek <vitek@roguewave.com>*
14251 * Add support for shared libraries under Irix.
14253 *Albert Chin-A-Young <china@thewrittenword.com>*
14255 * Add configuration option to build on Linux on both big-endian and
14256 little-endian MIPS.
14258 *Ralf Baechle <ralf@uni-koblenz.de>*
14260 * Add the possibility to create shared libraries on HP-UX.
14264 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14266 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14267 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14268 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14269 PRNG state recovery was possible based on the output of
14270 one PRNG request appropriately sized to gain knowledge on
14271 'md' followed by enough consecutive 1-byte PRNG requests
14272 to traverse all of 'state'.
14274 1. When updating 'md_local' (the current thread's copy of 'md')
14275 during PRNG output generation, hash all of the previous
14276 'md_local' value, not just the half used for PRNG output.
14278 2. Make the number of bytes from 'state' included into the hash
14279 independent from the number of PRNG bytes requested.
14281 The first measure alone would be sufficient to avoid
14282 Markku-Juhani's attack. (Actually it had never occurred
14283 to me that the half of 'md_local' used for chaining was the
14284 half from which PRNG output bytes were taken -- I had always
14285 assumed that the secret half would be used.) The second
14286 measure makes sure that additional data from 'state' is never
14287 mixed into 'md_local' in small portions; this heuristically
14288 further strengthens the PRNG.
14292 * Fix crypto/bn/asm/mips3.s.
14296 * When only the key is given to "enc", the IV is undefined. Print out
14297 an error message in this case.
14301 * Handle special case when X509_NAME is empty in X509 printing routines.
14305 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14306 positive and less than q.
14310 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14311 used: it isn't thread safe and the add_lock_callback should handle
14314 *Paul Rose <Paul.Rose@bridge.com>*
14316 * Verify that incoming data obeys the block size in
14317 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14323 *Ulf Möller, Bodo Möller*
14325 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14326 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14327 when fixing the server behaviour for backwards-compatible 'client
14328 hello' messages. (Note that the attack is impractical against
14329 SSL 3.0 and TLS 1.0 anyway because length and version checking
14330 means that the probability of guessing a valid ciphertext is
14331 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14334 Before 0.9.5, the countermeasure (hide the error by generating a
14335 random 'decryption result') did not work properly because
14336 ERR_clear_error() was missing, meaning that SSL_get_error() would
14337 detect the supposedly ignored error.
14339 Both problems are now fixed.
14343 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14344 (previously it was 1024).
14348 * Fix for compatibility mode trust settings: ignore trust settings
14349 unless some valid trust or reject settings are present.
14353 * Fix for blowfish EVP: its a variable length cipher.
14357 * Fix various bugs related to DSA S/MIME verification. Handle missing
14358 parameters in DSA public key structures and return an error in the
14359 DSA routines if parameters are absent.
14363 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14364 in the current directory if neither $RANDFILE nor $HOME was set.
14365 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14366 caused some confusion to Windows users who haven't defined $HOME.
14367 Thus RAND_file_name() is changed again: e_os.h can define a
14368 DEFAULT_HOME, which will be used if $HOME is not set.
14369 For Windows, we use "C:"; on other platforms, we still require
14370 environment variables.
14372 * Move 'if (!initialized) RAND_poll()' into regions protected by
14373 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14374 having multiple threads call RAND_poll() concurrently.
14378 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14379 combination of a flag and a thread ID variable.
14380 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14381 flag), *other* threads can enter ssleay_add_bytes without obeying
14382 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14383 that they do not hold after the first thread unsets add_do_not_lock).
14387 * Change bctest again: '-x' expressions are not available in all
14388 versions of 'test'.
14392 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14394 * Fix a couple of memory leaks in PKCS7_dataDecode()
14396 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14398 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14399 the default extension for executables, if any. Also, make the perl
14400 scripts that use symlink() to test if it really exists and use "cp"
14401 if it doesn't. All this made OpenSSL compilable and installable in
14406 * Fix for asn1_GetSequence() for indefinite length constructed data.
14407 If SEQUENCE is length is indefinite just set c->slen to the total
14408 amount of data available.
14410 *Steve Henson, reported by shige@FreeBSD.org*
14412 *This change does not apply to 0.9.7.*
14414 * Change bctest to avoid here-documents inside command substitution
14415 (workaround for FreeBSD /bin/sh bug).
14416 For compatibility with Ultrix, avoid shell functions (introduced
14417 in the bctest version that searches along $PATH).
14421 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14422 with des_encrypt() defined on some operating systems, like Solaris
14427 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14428 On the Importance of Eliminating Errors in Cryptographic
14429 Computations, J. Cryptology 14 (2001) 2, 101-119,
14430 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14434 * MIPS assembler BIGNUM division bug fix.
14438 * Disabled incorrect Alpha assembler code.
14442 * Fix PKCS#7 decode routines so they correctly update the length
14443 after reading an EOC for the EXPLICIT tag.
14447 *This change does not apply to 0.9.7.*
14449 * Fix bug in PKCS#12 key generation routines. This was triggered
14450 if a 3DES key was generated with a 0 initial byte. Include
14451 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14452 (but broken) behaviour.
14456 * Enhance bctest to search for a working bc along $PATH and print
14459 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14461 * Fix memory leaks in err.c: free err_data string if necessary;
14462 don't write to the wrong index in ERR_set_error_data.
14466 * Implement ssl23_peek (analogous to ssl23_read), which previously
14471 * Replace rdtsc with `_emit` statements for VC++ version 5.
14473 *Jeremy Cooper <jeremy@baymoo.org>*
14475 * Make it possible to reuse SSLv2 sessions.
14479 * In copy_email() check for >= 0 as a return value for
14480 X509_NAME_get_index_by_NID() since 0 is a valid index.
14482 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14484 * Avoid coredump with unsupported or invalid public keys by checking if
14485 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14486 PKCS7_verify() fails with non detached data.
14490 * Don't use getenv in library functions when run as setuid/setgid.
14491 New function OPENSSL_issetugid().
14495 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14496 due to incorrect handling of multi-threading:
14498 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14500 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14502 3. Count how many times MemCheck_off() has been called so that
14503 nested use can be treated correctly. This also avoids
14504 inband-signalling in the previous code (which relied on the
14505 assumption that thread ID 0 is impossible).
14509 * Add "-rand" option also to s_client and s_server.
14513 * Fix CPU detection on Irix 6.x.
14514 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14515 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14517 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14522 *This change does not apply to 0.9.7.*
14524 * Use the cached encoding of an X509_NAME structure rather than
14525 copying it. This is apparently the reason for the libsafe "errors"
14526 but the code is actually correct.
14530 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14531 Bleichenbacher's DSA attack.
14532 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14533 to be set and top=0 forces the highest bit to be set; top=-1 is new
14534 and leaves the highest bit random.
14536 *Ulf Moeller, Bodo Moeller*
14538 * In the `NCONF_...`-based implementations for `CONF_...` queries
14539 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14540 a temporary CONF structure with the data component set to NULL
14541 (which gives segmentation faults in lh_retrieve).
14542 Instead, use NULL for the CONF pointer in CONF_get_string and
14543 CONF_get_number (which may use environment variables) and directly
14544 return NULL from CONF_get_section.
14548 * Fix potential buffer overrun for EBCDIC.
14552 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14553 keyUsage if basicConstraints absent for a CA.
14557 * Make SMIME_write_PKCS7() write mail header values with a format that
14558 is more generally accepted (no spaces before the semicolon), since
14559 some programs can't parse those values properly otherwise. Also make
14560 sure BIO's that break lines after each write do not create invalid
14565 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14566 macros previously used would not encode an empty SEQUENCE OF
14567 and break the signature.
14571 *This change does not apply to 0.9.7.*
14573 * Zero the premaster secret after deriving the master secret in
14578 * Add some EVP_add_digest_alias registrations (as found in
14579 OpenSSL_add_all_digests()) to SSL_library_init()
14580 aka OpenSSL_add_ssl_algorithms(). This provides improved
14581 compatibility with peers using X.509 certificates
14582 with unconventional AlgorithmIdentifier OIDs.
14586 * Fix for Irix with NO_ASM.
14588 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14590 * ./config script fixes.
14592 *Ulf Moeller, Richard Levitte*
14594 * Fix 'openssl passwd -1'.
14598 * Change PKCS12_key_gen_asc() so it can cope with non null
14599 terminated strings whose length is passed in the passlen
14600 parameter, for example from PEM callbacks. This was done
14601 by adding an extra length parameter to asc2uni().
14603 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14605 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14606 call failed, free the DSA structure.
14610 * Fix to uni2asc() to cope with zero length Unicode strings.
14611 These are present in some PKCS#12 files.
14615 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14616 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14617 when writing a 32767 byte record.
14619 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14621 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14622 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14624 (RSA objects have a reference count access to which is protected
14625 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14626 so they are meant to be shared between threads.)
14627 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14628 "Reddie, Steven" <Steven.Reddie@ca.com>*
14630 * Fix a deadlock in CRYPTO_mem_leaks().
14634 * Use better test patterns in bntest.
14638 * rand_win.c fix for Borland C.
14642 * BN_rshift bugfix for n == 0.
14646 * Add a 'bctest' script that checks for some known 'bc' bugs
14647 so that 'make test' does not abort just because 'bc' is broken.
14651 * Store verify_result within SSL_SESSION also for client side to
14652 avoid potential security hole. (Re-used sessions on the client side
14653 always resulted in verify_result==X509_V_OK, not using the original
14654 result of the server certificate verification.)
14658 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14659 SSL3_RT_APPLICATION_DATA, return 0.
14660 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14665 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14666 releases, have been re-implemented by renaming the previous
14667 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14668 and ssl3_read_internal, respectively, and adding 'peek' parameters
14669 to them. The new ssl[23]_{read,peek} functions are calls to
14670 ssl[23]_read_internal with the 'peek' flag set appropriately.
14671 A 'peek' parameter has also been added to ssl3_read_bytes, which
14672 does the actual work for ssl3_read_internal.
14676 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14677 the method-specific "init()" handler. Also clean up ex_data after
14678 calling the method-specific "finish()" handler. Previously, this was
14679 happening the other way round.
14683 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14684 The previous value, 12, was not always sufficient for BN_mod_exp().
14688 * Make sure that shared libraries get the internal name engine with
14689 the full version number and not just 0. This should mark the
14690 shared libraries as not backward compatible. Of course, this should
14691 be changed again when we can guarantee backward binary compatibility.
14695 * Fix typo in get_cert_by_subject() in by_dir.c
14697 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14699 * Rework the system to generate shared libraries:
14701 - Make note of the expected extension for the shared libraries and
14702 if there is a need for symbolic links from for example libcrypto.so.0
14703 to libcrypto.so.0.9.7. There is extended info in Configure for
14706 - Make as few rebuilds of the shared libraries as possible.
14708 - Still avoid linking the OpenSSL programs with the shared libraries.
14710 - When installing, install the shared libraries separately from the
14715 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14717 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14718 and not in SSL_clear because the latter is also used by the
14719 accept/connect functions; previously, the settings made by
14720 SSL_set_read_ahead would be lost during the handshake.
14722 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14724 * Correct util/mkdef.pl to be selective about disabled algorithms.
14725 Previously, it would create entries for disabled algorithms no
14730 * Added several new manual pages for SSL_* function.
14734 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14736 * In ssl23_get_client_hello, generate an error message when faced
14737 with an initial SSL 3.0/TLS record that is too small to contain the
14738 first two bytes of the ClientHello message, i.e. client_version.
14739 (Note that this is a pathologic case that probably has never happened
14740 in real life.) The previous approach was to use the version number
14741 from the record header as a substitute; but our protocol choice
14742 should not depend on that one because it is not authenticated
14743 by the Finished messages.
14747 * More robust randomness gathering functions for Windows.
14749 *Jeffrey Altman <jaltman@columbia.edu>*
14751 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14752 not set then we don't setup the error code for issuer check errors
14753 to avoid possibly overwriting other errors which the callback does
14754 handle. If an application does set the flag then we assume it knows
14755 what it is doing and can handle the new informational codes
14760 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14761 a general "ANY" type, as such it should be able to decode anything
14762 including tagged types. However it didn't check the class so it would
14763 wrongly interpret tagged types in the same way as their universal
14764 counterpart and unknown types were just rejected. Changed so that the
14765 tagged and unknown types are handled in the same way as a SEQUENCE:
14766 that is the encoding is stored intact. There is also a new type
14767 "V_ASN1_OTHER" which is used when the class is not universal, in this
14768 case we have no idea what the actual type is so we just lump them all
14773 * On VMS, stdout may very well lead to a file that is written to
14774 in a record-oriented fashion. That means that every write() will
14775 write a separate record, which will be read separately by the
14776 programs trying to read from it. This can be very confusing.
14778 The solution is to put a BIO filter in the way that will buffer
14779 text until a linefeed is reached, and then write everything a
14780 line at a time, so every record written will be an actual line,
14781 not chunks of lines and not (usually doesn't happen, but I've
14782 seen it once) several lines in one record. BIO_f_linebuffer() is
14785 Currently, it's a VMS-only method, because that's where it has
14786 been tested well enough.
14790 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14791 it can return incorrect results.
14792 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14793 but it was in 0.9.6-beta[12].)
14797 * Disable the check for content being present when verifying detached
14798 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14799 include zero length content when signing messages.
14803 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14804 BIO_ctrl (for BIO pairs).
14808 * Add DSO method for VMS.
14812 * Bug fix: Montgomery multiplication could produce results with the
14817 * Add RPM specification openssl.spec and modify it to build three
14818 packages. The default package contains applications, application
14819 documentation and run-time libraries. The devel package contains
14820 include files, static libraries and function documentation. The
14821 doc package contains the contents of the doc directory. The original
14822 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14826 * Add a large number of documentation files for many SSL routines.
14828 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14830 * Add a configuration entry for Sony News 4.
14832 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14834 * Don't set the two most significant bits to one when generating a
14835 random number < q in the DSA library.
14839 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14840 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14841 the underlying transport is blocking) if a handshake took place.
14842 (The default behaviour is needed by applications such as s_client
14843 and s_server that use select() to determine when to use SSL_read;
14844 but for applications that know in advance when to expect data, it
14845 just makes things more complicated.)
14849 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14854 * Add a few more EBCDIC conditionals that make `req` and `x509`
14855 work better on such systems.
14857 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14859 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14860 Update PKCS12_parse() so it copies the friendlyName and the
14861 keyid to the certificates aux info.
14865 * Fix bug in PKCS7_verify() which caused an infinite loop
14866 if there was more than one signature.
14868 *Sven Uszpelkat <su@celocom.de>*
14870 * Major change in util/mkdef.pl to include extra information
14871 about each symbol, as well as presenting variables as well
14872 as functions. This change means that there's n more need
14873 to rebuild the .num files when some algorithms are excluded.
14877 * Allow the verify time to be set by an application,
14878 rather than always using the current time.
14882 * Phase 2 verify code reorganisation. The certificate
14883 verify code now looks up an issuer certificate by a
14884 number of criteria: subject name, authority key id
14885 and key usage. It also verifies self signed certificates
14886 by the same criteria. The main comparison function is
14887 X509_check_issued() which performs these checks.
14889 Lot of changes were necessary in order to support this
14890 without completely rewriting the lookup code.
14892 Authority and subject key identifier are now cached.
14894 The LHASH 'certs' is X509_STORE has now been replaced
14895 by a STACK_OF(X509_OBJECT). This is mainly because an
14896 LHASH can't store or retrieve multiple objects with
14897 the same hash value.
14899 As a result various functions (which were all internal
14900 use only) have changed to handle the new X509_STORE
14901 structure. This will break anything that messed round
14902 with X509_STORE internally.
14904 The functions X509_STORE_add_cert() now checks for an
14905 exact match, rather than just subject name.
14907 The X509_STORE API doesn't directly support the retrieval
14908 of multiple certificates matching a given criteria, however
14909 this can be worked round by performing a lookup first
14910 (which will fill the cache with candidate certificates)
14911 and then examining the cache for matches. This is probably
14912 the best we can do without throwing out X509_LOOKUP
14913 entirely (maybe later...).
14915 The X509_VERIFY_CTX structure has been enhanced considerably.
14917 All certificate lookup operations now go via a get_issuer()
14918 callback. Although this currently uses an X509_STORE it
14919 can be replaced by custom lookups. This is a simple way
14920 to bypass the X509_STORE hackery necessary to make this
14921 work and makes it possible to use more efficient techniques
14922 in future. A very simple version which uses a simple
14923 STACK for its trusted certificate store is also provided
14924 using X509_STORE_CTX_trusted_stack().
14926 The verify_cb() and verify() callbacks now have equivalents
14927 in the X509_STORE_CTX structure.
14929 X509_STORE_CTX also has a 'flags' field which can be used
14930 to customise the verify behaviour.
14934 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14935 excludes S/MIME capabilities.
14939 * When a certificate request is read in keep a copy of the
14940 original encoding of the signed data and use it when outputting
14941 again. Signatures then use the original encoding rather than
14942 a decoded, encoded version which may cause problems if the
14943 request is improperly encoded.
14947 * For consistency with other BIO_puts implementations, call
14948 buffer_write(b, ...) directly in buffer_puts instead of calling
14951 In BIO_puts, increment b->num_write as in BIO_write.
14953 *Peter.Sylvester@EdelWeb.fr*
14955 * Fix BN_mul_word for the case where the word is 0. (We have to use
14956 BN_zero, we may not return a BIGNUM with an array consisting of
14957 words set to zero.)
14961 * Avoid calling abort() from within the library when problems are
14962 detected, except if preprocessor symbols have been defined
14963 (such as REF_CHECK, BN_DEBUG etc.).
14967 * New openssl application 'rsautl'. This utility can be
14968 used for low-level RSA operations. DER public key
14969 BIO/fp routines also added.
14973 * New Configure entry and patches for compiling on QNX 4.
14975 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14977 * A demo state-machine implementation was sponsored by
14978 Nuron (<http://www.nuron.com/>) and is now available in
14979 demos/state_machine.
14983 * New options added to the 'dgst' utility for signature
14984 generation and verification.
14988 * Unrecognized PKCS#7 content types are now handled via a
14989 catch all ASN1_TYPE structure. This allows unsupported
14990 types to be stored as a "blob" and an application can
14991 encode and decode it manually.
14995 * Fix various signed/unsigned issues to make a_strex.c
14996 compile under VC++.
14998 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15000 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15001 length if passed a buffer. ASN1_INTEGER_to_BN failed
15002 if passed a NULL BN and its argument was negative.
15004 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15006 * Modification to PKCS#7 encoding routines to output definite
15007 length encoding. Since currently the whole structures are in
15008 memory there's not real point in using indefinite length
15009 constructed encoding. However if OpenSSL is compiled with
15010 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15014 * Added BIO_vprintf() and BIO_vsnprintf().
15018 * Added more prefixes to parse for in the strings written
15019 through a logging bio, to cover all the levels that are available
15020 through syslog. The prefixes are now:
15022 PANIC, EMERG, EMR => LOG_EMERG
15023 ALERT, ALR => LOG_ALERT
15024 CRIT, CRI => LOG_CRIT
15025 ERROR, ERR => LOG_ERR
15026 WARNING, WARN, WAR => LOG_WARNING
15027 NOTICE, NOTE, NOT => LOG_NOTICE
15028 INFO, INF => LOG_INFO
15029 DEBUG, DBG => LOG_DEBUG
15031 and as before, if none of those prefixes are present at the
15032 beginning of the string, LOG_ERR is chosen.
15034 On Win32, the `LOG_*` levels are mapped according to this:
15036 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15037 LOG_WARNING => EVENTLOG_WARNING_TYPE
15038 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15042 * Made it possible to reconfigure with just the configuration
15043 argument "reconf" or "reconfigure". The command line arguments
15044 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15045 and are retrieved from there when reconfiguring.
15051 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15053 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15057 * The obj_dat.pl script was messing up the sorting of object
15058 names. The reason was that it compared the quoted version
15059 of strings as a result "OCSP" > "OCSP Signing" because
15060 " > SPACE. Changed script to store unquoted versions of
15061 names and add quotes on output. It was also omitting some
15062 names from the lookup table if they were given a default
15063 value (that is if SN is missing it is given the same
15064 value as LN and vice versa), these are now added on the
15065 grounds that if an object has a name we should be able to
15066 look it up. Finally added warning output when duplicate
15067 short or long names are found.
15071 * Changes needed for Tandem NSK.
15073 *Scott Uroff <scott@xypro.com>*
15075 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15076 RSA_padding_check_SSLv23(), special padding was never detected
15077 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15078 version rollback attacks was not effective.
15080 In s23_clnt.c, don't use special rollback-attack detection padding
15081 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15082 client; similarly, in s23_srvr.c, don't do the rollback check if
15083 SSL 2.0 is the only protocol enabled in the server.
15087 * Make it possible to get hexdumps of unprintable data with 'openssl
15088 asn1parse'. By implication, the functions ASN1_parse_dump() and
15089 BIO_dump_indent() are added.
15093 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15094 these print out strings and name structures based on various
15095 flags including RFC2253 support and proper handling of
15096 multibyte characters. Added options to the 'x509' utility
15097 to allow the various flags to be set.
15101 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15102 Also change the functions X509_cmp_current_time() and
15103 X509_gmtime_adj() work with an ASN1_TIME structure,
15104 this will enable certificates using GeneralizedTime in validity
15105 dates to be checked.
15109 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15110 negative public key encodings) on by default,
15111 NO_NEG_PUBKEY_BUG can be set to disable it.
15115 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15116 content octets. An i2c_ASN1_OBJECT is unnecessary because
15117 the encoding can be trivially obtained from the structure.
15121 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15122 not read locks (`CRYPTO_r_[un]lock`).
15126 * A first attempt at creating official support for shared
15127 libraries through configuration. I've kept it so the
15128 default is static libraries only, and the OpenSSL programs
15129 are always statically linked for now, but there are
15130 preparations for dynamic linking in place.
15131 This has been tested on Linux and Tru64.
15135 * Randomness polling function for Win9x, as described in:
15136 Peter Gutmann, Software Generation of Practically Strong
15141 * Fix so PRNG is seeded in req if using an already existing
15146 * New options to smime application. -inform and -outform
15147 allow alternative formats for the S/MIME message including
15148 PEM and DER. The -content option allows the content to be
15149 specified separately. This should allow things like Netscape
15150 form signing output easier to verify.
15154 * Fix the ASN1 encoding of tags using the 'long form'.
15158 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15159 STRING types. These convert content octets to and from the
15160 underlying type. The actual tag and length octets are
15161 already assumed to have been read in and checked. These
15162 are needed because all other string types have virtually
15163 identical handling apart from the tag. By having versions
15164 of the ASN1 functions that just operate on content octets
15165 IMPLICIT tagging can be handled properly. It also allows
15166 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15167 and ASN1_INTEGER are identical apart from the tag.
15171 * Change the handling of OID objects as follows:
15173 - New object identifiers are inserted in objects.txt, following
15174 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15175 - objects.pl is used to process obj_mac.num and create a new
15177 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15180 This is currently kind of a hack, and the perl code in objects.pl
15181 isn't very elegant, but it works as I intended. The simplest way
15182 to check that it worked correctly is to look in obj_dat.h and
15183 check the array nid_objs and make sure the objects haven't moved
15184 around (this is important!). Additions are OK, as well as
15185 consistent name changes.
15189 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15193 * Addition of the command line parameter '-rand file' to 'openssl req'.
15194 The given file adds to whatever has already been seeded into the
15195 random pool through the RANDFILE configuration file option or
15196 environment variable, or the default random state file.
15200 * mkstack.pl now sorts each macro group into lexical order.
15201 Previously the output order depended on the order the files
15202 appeared in the directory, resulting in needless rewriting
15207 * Patches to make OpenSSL compile under Win32 again. Mostly
15208 work arounds for the VC++ problem that it treats func() as
15209 func(void). Also stripped out the parts of mkdef.pl that
15210 added extra typesafe functions: these no longer exist.
15214 * Reorganisation of the stack code. The macros are now all
15215 collected in safestack.h . Each macro is defined in terms of
15216 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15217 DEBUG_SAFESTACK is now handled in terms of function casts,
15218 this has the advantage of retaining type safety without the
15219 use of additional functions. If DEBUG_SAFESTACK is not defined
15220 then the non typesafe macros are used instead. Also modified the
15221 mkstack.pl script to handle the new form. Needs testing to see
15222 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15223 the default if no major problems. Similar behaviour for ASN1_SET_OF
15224 and PKCS12_STACK_OF.
15228 * When some versions of IIS use the 'NET' form of private key the
15229 key derivation algorithm is different. Normally MD5(password) is
15230 used as a 128 bit RC4 key. In the modified case
15231 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15232 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15233 as the old Netscape_RSA functions except they have an additional
15234 'sgckey' parameter which uses the modified algorithm. Also added
15235 an -sgckey command line option to the rsa utility. Thanks to
15236 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15237 algorithm to openssl-dev.
15241 * The evp_local.h macros were using 'c.##kname' which resulted in
15242 invalid expansion on some systems (SCO 5.0.5 for example).
15243 Corrected to 'c.kname'.
15245 *Phillip Porch <root@theporch.com>*
15247 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15248 a STACK of email addresses from a certificate or request, these look
15249 in the subject name and the subject alternative name extensions and
15250 omit any duplicate addresses.
15254 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15255 This makes DSA verification about 2 % faster.
15259 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15260 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15261 plus overhead for 1024 bit moduli).
15262 This makes exponentiations about 0.5 % faster for 1024 bit
15263 exponents (as measured by "openssl speed rsa2048").
15267 * Rename memory handling macros to avoid conflicts with other
15269 Malloc => OPENSSL_malloc
15270 Malloc_locked => OPENSSL_malloc_locked
15271 Realloc => OPENSSL_realloc
15272 Free => OPENSSL_free
15276 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15277 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15281 * CygWin32 support.
15283 *John Jarvie <jjarvie@newsguy.com>*
15285 * The type-safe stack code has been rejigged. It is now only compiled
15286 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15287 by default all type-specific stack functions are "#define"d back to
15288 standard stack functions. This results in more streamlined output
15289 but retains the type-safety checking possibilities of the original
15294 * The STACK code has been cleaned up, and certain type declarations
15295 that didn't make a lot of sense have been brought in line. This has
15296 also involved a cleanup of sorts in safestack.h to more correctly
15297 map type-safe stack functions onto their plain stack counterparts.
15298 This work has also resulted in a variety of "const"ifications of
15299 lots of the code, especially `_cmp` operations which should normally
15300 be prototyped with "const" parameters anyway.
15304 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15305 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15306 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15307 where all of 'md' is used each time the PRNG is used, but 'state'
15308 is used only indexed by a cyclic counter. As entropy may not be
15309 well distributed from the beginning, 'md' is important as a
15310 chaining variable. However, the output function chains only half
15311 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15312 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15313 in all of 'state' being rewritten, with the new values depending
15314 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15318 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15319 the handshake is continued after ssl_verify_cert_chain();
15320 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15321 can lead to 'unexplainable' connection aborts later.
15323 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15325 * Major EVP API cipher revision.
15326 Add hooks for extra EVP features. This allows various cipher
15327 parameters to be set in the EVP interface. Support added for variable
15328 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15329 setting of RC2 and RC5 parameters.
15331 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15334 Remove lots of duplicated code from the EVP library. For example *every*
15335 cipher init() function handles the 'iv' in the same way according to the
15336 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15337 for CFB and OFB modes they zero ctx->num.
15339 New functionality allows removal of S/MIME code RC2 hack.
15341 Most of the routines have the same form and so can be declared in terms
15344 By shifting this to the top level EVP_CipherInit() it can be removed from
15345 all individual ciphers. If the cipher wants to handle IVs or keys
15346 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15349 Change lots of functions like EVP_EncryptUpdate() to now return a
15350 value: although software versions of the algorithms cannot fail
15351 any installed hardware versions can.
15355 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15356 this option is set, tolerate broken clients that send the negotiated
15357 protocol version number instead of the requested protocol version
15362 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15363 i.e. non-zero for export ciphersuites, zero otherwise.
15364 Previous versions had this flag inverted, inconsistent with
15365 rsa_tmp_cb (..._TMP_RSA_CB).
15367 *Bodo Moeller; problem reported by Amit Chopra*
15369 * Add missing DSA library text string. Work around for some IIS
15370 key files with invalid SEQUENCE encoding.
15374 * Add a document (doc/standards.txt) that list all kinds of standards
15375 and so on that are implemented in OpenSSL.
15379 * Enhance c_rehash script. Old version would mishandle certificates
15380 with the same subject name hash and wouldn't handle CRLs at all.
15381 Added -fingerprint option to crl utility, to support new c_rehash
15386 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15390 * Fix for SSL server purpose checking. Server checking was
15391 rejecting certificates which had extended key usage present
15392 but no ssl client purpose.
15394 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15396 * Make PKCS#12 code work with no password. The PKCS#12 spec
15397 is a little unclear about how a blank password is handled.
15398 Since the password in encoded as a BMPString with terminating
15399 double NULL a zero length password would end up as just the
15400 double NULL. However no password at all is different and is
15401 handled differently in the PKCS#12 key generation code. NS
15402 treats a blank password as zero length. MSIE treats it as no
15403 password on export: but it will try both on import. We now do
15404 the same: PKCS12_parse() tries zero length and no password if
15405 the password is set to "" or NULL (NULL is now a valid password:
15406 it wasn't before) as does the pkcs12 application.
15410 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15411 perror when PEM_read_bio_X509_REQ fails, the error message must
15412 be obtained from the error queue.
15416 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15417 it in ERR_remove_state if appropriate, and change ERR_get_state
15418 accordingly to avoid race conditions (this is necessary because
15419 thread_hash is no longer constant once set).
15423 * Bugfix for linux-elf makefile.one.
15427 * RSA_get_default_method() will now cause a default
15428 RSA_METHOD to be chosen if one doesn't exist already.
15429 Previously this was only set during a call to RSA_new()
15430 or RSA_new_method(NULL) meaning it was possible for
15431 RSA_get_default_method() to return NULL.
15435 * Added native name translation to the existing DSO code
15436 that will convert (if the flag to do so is set) filenames
15437 that are sufficiently small and have no path information
15438 into a canonical native form. Eg. "blah" converted to
15439 "libblah.so" or "blah.dll" etc.
15443 * New function ERR_error_string_n(e, buf, len) which is like
15444 ERR_error_string(e, buf), but writes at most 'len' bytes
15445 including the 0 terminator. For ERR_error_string_n, 'buf'
15448 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15450 * CONF library reworked to become more general. A new CONF
15451 configuration file reader "class" is implemented as well as a
15452 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15453 old `CONF_*` functions are still there, but are reimplemented to
15454 work in terms of the new functions. Also, a set of functions
15455 to handle the internal storage of the configuration data is
15456 provided to make it easier to write new configuration file
15457 reader "classes" (I can definitely see something reading a
15458 configuration file in XML format, for example), called `_CONF_*`,
15459 or "the configuration storage API"...
15461 The new configuration file reading functions are:
15463 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15464 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15466 NCONF_default, NCONF_WIN32
15468 NCONF_dump_fp, NCONF_dump_bio
15470 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15471 NCONF_new creates a new CONF object. This works in the same way
15472 as other interfaces in OpenSSL, like the BIO interface.
15473 `NCONF_dump_*` dump the internal storage of the configuration file,
15474 which is useful for debugging. All other functions take the same
15475 arguments as the old `CONF_*` functions with the exception of the
15476 first that must be a `CONF *` instead of a `LHASH *`.
15478 To make it easier to use the new classes with the old `CONF_*` functions,
15479 the function CONF_set_default_method is provided.
15483 * Add '-tls1' option to 'openssl ciphers', which was already
15484 mentioned in the documentation but had not been implemented.
15485 (This option is not yet really useful because even the additional
15486 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15490 * Initial DSO code added into libcrypto for letting OpenSSL (and
15491 OpenSSL-based applications) load shared libraries and bind to
15492 them in a portable way.
15494 *Geoff Thorpe, with contributions from Richard Levitte*
15496 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15498 * Make sure _lrotl and _lrotr are only used with MSVC.
15500 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15501 (the default implementation of RAND_status).
15503 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15504 to '-clrext' (= clear extensions), as intended and documented.
15505 *Bodo Moeller; inconsistency pointed out by Michael Attili
15506 <attili@amaxo.com>*
15508 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15509 was larger than the MD block size.
15511 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15513 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15514 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15515 using the passed key: if the passed key was a private key the result
15516 of X509_print(), for example, would be to print out all the private key
15521 * des_quad_cksum() byte order bug fix.
15522 *Ulf Möller, using the problem description in krb4-0.9.7, where
15523 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15525 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15528 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15530 * For easily testing in shell scripts whether some command
15531 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15532 returns with exit code 0 iff no command of the given name is available.
15533 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15534 the output goes to stdout and nothing is printed to stderr.
15535 Additional arguments are always ignored.
15537 Since for each cipher there is a command of the same name,
15538 the 'no-cipher' compilation switches can be tested this way.
15540 ('openssl no-XXX' is not able to detect pseudo-commands such
15541 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15545 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15549 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15550 is set; it will be thrown away anyway because each handshake creates
15552 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15553 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15554 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15555 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15559 * New s_client option -ign_eof: EOF at stdin is ignored, and
15560 'Q' and 'R' lose their special meanings (quit/renegotiate).
15561 This is part of what -quiet does; unlike -quiet, -ign_eof
15562 does not suppress any output.
15566 * Add compatibility options to the purpose and trust code. The
15567 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15568 accepts a certificate or CA, this was the previous behaviour,
15569 with all the associated security issues.
15571 X509_TRUST_COMPAT is the old trust behaviour: only and
15572 automatically trust self signed roots in certificate store. A
15573 new trust setting X509_TRUST_DEFAULT is used to specify that
15574 a purpose has no associated trust setting and it should instead
15575 use the value in the default purpose.
15579 * Fix the PKCS#8 DSA private key code so it decodes keys again
15580 and fix a memory leak.
15584 * In util/mkerr.pl (which implements 'make errors'), preserve
15585 reason strings from the previous version of the .c file, as
15586 the default to have only downcase letters (and digits) in
15587 automatically generated reasons codes is not always appropriate.
15591 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15592 using strerror. Previously, ERR_reason_error_string() returned
15593 library names as reason strings for SYSerr; but SYSerr is a special
15594 case where small numbers are errno values, not library numbers.
15598 * Add '-dsaparam' option to 'openssl dhparam' application. This
15599 converts DSA parameters into DH parameters. (When creating parameters,
15600 DSA_generate_parameters is used.)
15604 * Include 'length' (recommended exponent length) in C code generated
15605 by 'openssl dhparam -C'.
15609 * The second argument to set_label in perlasm was already being used
15610 so couldn't be used as a "file scope" flag. Moved to third argument
15615 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15616 instead of RAND_bytes for encryption IVs and salts.
15620 * Include RAND_status() into RAND_METHOD instead of implementing
15621 it only for md_rand.c Otherwise replacing the PRNG by calling
15622 RAND_set_rand_method would be impossible.
15626 * Don't let DSA_generate_key() enter an infinite loop if the random
15627 number generation fails.
15631 * New 'rand' application for creating pseudo-random output.
15635 * Added configuration support for Linux/IA64
15637 *Rolf Haberrecker <rolf@suse.de>*
15639 * Assembler module support for Mingw32.
15643 * Shared library support for HPUX (in shlib/).
15645 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15647 * Shared library support for Solaris gcc.
15649 *Lutz Behnke <behnke@trustcenter.de>*
15651 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15653 * PKCS7_encrypt() was adding text MIME headers twice because they
15654 were added manually and by SMIME_crlf_copy().
15658 * In bntest.c don't call BN_rand with zero bits argument.
15660 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15662 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15663 case was implemented. This caused BN_div_recp() to fail occasionally.
15667 * Add an optional second argument to the set_label() in the perl
15668 assembly language builder. If this argument exists and is set
15669 to 1 it signals that the assembler should use a symbol whose
15670 scope is the entire file, not just the current function. This
15671 is needed with MASM which uses the format label:: for this scope.
15673 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15675 * Change the ASN1 types so they are typedefs by default. Before
15676 almost all types were #define'd to ASN1_STRING which was causing
15677 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15682 * Change names of new functions to the new get1/get0 naming
15683 convention: After 'get1', the caller owns a reference count
15684 and has to call `..._free`; 'get0' returns a pointer to some
15685 data structure without incrementing reference counters.
15686 (Some of the existing 'get' functions increment a reference
15687 counter, some don't.)
15688 Similarly, 'set1' and 'add1' functions increase reference
15689 counters or duplicate objects.
15693 * Allow for the possibility of temp RSA key generation failure:
15694 the code used to assume it always worked and crashed on failure.
15698 * Fix potential buffer overrun problem in BIO_printf().
15699 *Ulf Möller, using public domain code by Patrick Powell; problem
15700 pointed out by David Sacerdote <das33@cornell.edu>*
15702 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15703 RAND_egd() and RAND_status(). In the command line application,
15704 the EGD socket can be specified like a seed file using RANDFILE
15709 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15710 Some CAs (e.g. Verisign) distribute certificates in this form.
15714 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15715 list to exclude them. This means that no special compilation option
15716 is needed to use anonymous DH: it just needs to be included in the
15721 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15722 EVP_MD_type. The old functionality is available in a new macro called
15723 EVP_MD_md(). Change code that uses it and update docs.
15727 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15728 where the `void *` argument is replaced by a function pointer argument.
15729 Previously `void *` was abused to point to functions, which works on
15730 many platforms, but is not correct. As these functions are usually
15731 called by macros defined in OpenSSL header files, most source code
15732 should work without changes.
15736 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15737 sections with information on -D... compiler switches used for
15738 compiling the library so that applications can see them. To enable
15739 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15740 must be defined. E.g.,
15741 #define OPENSSL_ALGORITHM_DEFINES
15742 #include <openssl/opensslconf.h>
15743 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15745 *Richard Levitte, Ulf and Bodo Möller*
15747 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15752 * Change the 'other' type in certificate aux info to a STACK_OF
15753 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15754 the required ASN1 format: arbitrary types determined by an OID.
15758 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15759 argument to 'req'. This is not because the function is newer or
15760 better than others it just uses the work 'NEW' in the certificate
15761 request header lines. Some software needs this.
15765 * Reorganise password command line arguments: now passwords can be
15766 obtained from various sources. Delete the PEM_cb function and make
15767 it the default behaviour: i.e. if the callback is NULL and the
15768 usrdata argument is not NULL interpret it as a null terminated pass
15769 phrase. If usrdata and the callback are NULL then the pass phrase
15770 is prompted for as usual.
15774 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15775 the support is automatically enabled. The resulting binaries will
15776 autodetect the card and use it if present.
15778 *Ben Laurie and Compaq Inc.*
15780 * Work around for Netscape hang bug. This sends certificate request
15781 and server done in one record. Since this is perfectly legal in the
15782 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15783 the bugs/SSLv3 entry for more info.
15787 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15791 * Add -rand argument to smime and pkcs12 applications and read/write
15796 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15800 * Add command line password options to the remaining applications.
15804 * Bug fix for BN_div_recp() for numerators with an even number of
15809 * More tests in bntest.c, and changed test_bn output.
15813 * ./config recognizes MacOS X now.
15817 * Bug fix for BN_div() when the first words of num and divisor are
15818 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15822 * Add support for various broken PKCS#8 formats, and command line
15823 options to produce them.
15827 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15828 get temporary BIGNUMs from a BN_CTX.
15832 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15837 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15838 include a #define from the old name to the new. The original intent
15839 was that statically linked binaries could for example just call
15840 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15841 link with digests. This never worked because SSLeay_add_all_digests()
15842 and SSLeay_add_all_ciphers() were in the same source file so calling
15843 one would link with the other. They are now in separate source files.
15847 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15851 * Use a less unusual form of the Miller-Rabin primality test (it used
15852 a binary algorithm for exponentiation integrated into the Miller-Rabin
15853 loop, our standard modexp algorithms are faster).
15857 * Support for the EBCDIC character set completed.
15859 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15861 * Source code cleanups: use const where appropriate, eliminate casts,
15862 use `void *` instead of `char *` in lhash.
15866 * Bugfix: ssl3_send_server_key_exchange was not restartable
15867 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15868 this the server could overwrite ephemeral keys that the client
15873 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15874 using 50 iterations of the Rabin-Miller test.
15876 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15877 iterations of the Rabin-Miller test as required by the appendix
15878 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15879 As BN_is_prime_fasttest includes trial division, DSA parameter
15880 generation becomes much faster.
15882 This implies a change for the callback functions in DSA_is_prime
15883 and DSA_generate_parameters: The callback function is called once
15884 for each positive witness in the Rabin-Miller test, not just
15885 occasionally in the inner loop; and the parameters to the
15886 callback function now provide an iteration count for the outer
15887 loop rather than for the current invocation of the inner loop.
15888 DSA_generate_parameters additionally can call the callback
15889 function with an 'iteration count' of -1, meaning that a
15890 candidate has passed the trial division test (when q is generated
15891 from an application-provided seed, trial division is skipped).
15895 * New function BN_is_prime_fasttest that optionally does trial
15896 division before starting the Rabin-Miller test and has
15897 an additional BN_CTX * argument (whereas BN_is_prime always
15898 has to allocate at least one BN_CTX).
15899 'callback(1, -1, cb_arg)' is called when a number has passed the
15900 trial division stage.
15904 * Fix for bug in CRL encoding. The validity dates weren't being handled
15909 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15913 * New function BN_pseudo_rand().
15917 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15918 bignum version of BN_from_montgomery() with the working code from
15919 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15924 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15925 made it impossible to use the same SSL_SESSION data structure in
15926 SSL2 clients in multiple threads.
15930 * The return value of RAND_load_file() no longer counts bytes obtained
15931 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15932 to seed the PRNG (previously an explicit byte count was required).
15934 *Ulf Möller, Bodo Möller*
15936 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15937 used `char *` instead of `void *` and had casts all over the place.
15941 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15945 * Retain source code compatibility for BN_prime_checks macro:
15946 BN_is_prime(..., BN_prime_checks, ...) now uses
15947 BN_prime_checks_for_size to determine the appropriate number of
15948 Rabin-Miller iterations.
15952 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15953 DH_CHECK_P_NOT_SAFE_PRIME.
15954 (Check if this is true? OpenPGP calls them "strong".)
15958 * Merge the functionality of "dh" and "gendh" programs into a new program
15959 "dhparam". The old programs are retained for now but will handle DH keys
15960 (instead of parameters) in future.
15964 * Make the ciphers, s_server and s_client programs check the return values
15965 when a new cipher list is set.
15969 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15970 ciphers. Before when the 56bit ciphers were enabled the sorting was
15973 The syntax for the cipher sorting has been extended to support sorting by
15974 cipher-strength (using the strength_bits hard coded in the tables).
15975 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
15977 Fix a bug in the cipher-command parser: when supplying a cipher command
15978 string with an "undefined" symbol (neither command nor alphanumeric
15979 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15980 an error is flagged.
15982 Due to the strength-sorting extension, the code of the
15983 ssl_create_cipher_list() function was completely rearranged. I hope that
15984 the readability was also increased :-)
15986 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15988 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15989 for the first serial number and places 2 in the serial number file. This
15990 avoids problems when the root CA is created with serial number zero and
15991 the first user certificate has the same issuer name and serial number
15996 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15997 the new code. Add documentation for this stuff.
16001 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16002 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16003 structures and behave in an analogous way to the X509v3 functions:
16004 they shouldn't be called directly but wrapper functions should be used
16007 So we also now have some wrapper functions that call the X509at functions
16008 when passed certificate requests. (TO DO: similar things can be done with
16009 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16010 things. Some of these need some d2i or i2d and print functionality
16011 because they handle more complex structures.)
16015 * Add missing #ifndefs that caused missing symbols when building libssl
16016 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16017 NO_RSA in `ssl/s2*.c`.
16019 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16021 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16022 has a return value which indicates the quality of the random data
16023 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16024 error queue. New function RAND_pseudo_bytes() generates output that is
16025 guaranteed to be unique but not unpredictable. RAND_add is like
16026 RAND_seed, but takes an extra argument for an entropy estimate
16027 (RAND_seed always assumes full entropy).
16031 * Do more iterations of Rabin-Miller probable prime test (specifically,
16032 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16033 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16034 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16035 false-positive rate of at most 2^-80 for random input.
16039 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16043 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16044 in the 0.9.5 release), this returns the chain
16045 from an X509_CTX structure with a dup of the stack and all
16046 the X509 reference counts upped: so the stack will exist
16047 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16050 Also make SSL_SESSION_print() print out the verify return
16055 * Add manpage for the pkcs12 command. Also change the default
16056 behaviour so MAC iteration counts are used unless the new
16057 -nomaciter option is used. This improves file security and
16058 only older versions of MSIE (4.0 for example) need it.
16062 * Honor the no-xxx Configure options when creating .DEF files.
16066 * Add PKCS#10 attributes to field table: challengePassword,
16067 unstructuredName and unstructuredAddress. These are taken from
16068 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16069 international characters are used.
16071 More changes to X509_ATTRIBUTE code: allow the setting of types
16072 based on strings. Remove the 'loc' parameter when adding
16073 attributes because these will be a SET OF encoding which is sorted
16078 * Initial changes to the 'req' utility to allow request generation
16079 automation. This will allow an application to just generate a template
16080 file containing all the field values and have req construct the
16083 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16084 used all over the place including certificate requests and PKCS#7
16085 structures. They are currently handled manually where necessary with
16086 some primitive wrappers for PKCS#7. The new functions behave in a
16087 manner analogous to the X509 extension functions: they allow
16088 attributes to be looked up by NID and added.
16090 Later something similar to the X509V3 code would be desirable to
16091 automatically handle the encoding, decoding and printing of the
16092 more complex types. The string types like challengePassword can
16093 be handled by the string table functions.
16095 Also modified the multi byte string table handling. Now there is
16096 a 'global mask' which masks out certain types. The table itself
16097 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16098 is useful when for example there is only one permissible type
16099 (as in countryName) and using the mask might result in no valid
16104 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16105 SSL_get_peer_finished to allow applications to obtain the latest
16106 Finished messages sent to the peer or expected from the peer,
16107 respectively. (SSL_get_peer_finished is usually the Finished message
16108 actually received from the peer, otherwise the protocol will be aborted.)
16110 As the Finished message are message digests of the complete handshake
16111 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16112 be used for external authentication procedures when the authentication
16113 provided by SSL/TLS is not desired or is not enough.
16117 * Enhanced support for Alpha Linux is added. Now ./config checks if
16118 the host supports BWX extension and if Compaq C is present on the
16119 $PATH. Just exploiting of the BWX extension results in 20-30%
16120 performance kick for some algorithms, e.g. DES and RC4 to mention
16121 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16126 * Add support for MS "fast SGC". This is arguably a violation of the
16127 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16128 weak crypto and after checking the certificate is SGC a second one
16129 with strong crypto. MS SGC stops the first handshake after receiving
16130 the server certificate message and sends a second client hello. Since
16131 a server will typically do all the time consuming operations before
16132 expecting any further messages from the client (server key exchange
16133 is the most expensive) there is little difference between the two.
16135 To get OpenSSL to support MS SGC we have to permit a second client
16136 hello message after we have sent server done. In addition we have to
16137 reset the MAC if we do get this second client hello.
16141 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16142 if a DER encoded private key is RSA or DSA traditional format. Changed
16143 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16144 format DER encoded private key. Newer code should use PKCS#8 format which
16145 has the key type encoded in the ASN1 structure. Added DER private key
16146 support to pkcs8 application.
16150 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16151 ciphersuites has been selected (as required by the SSL 3/TLS 1
16152 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16153 is set, we interpret this as a request to violate the specification
16154 (the worst that can happen is a handshake failure, and 'correct'
16155 behaviour would result in a handshake failure anyway).
16159 * In SSL_CTX_add_session, take into account that there might be multiple
16160 SSL_SESSION structures with the same session ID (e.g. when two threads
16161 concurrently obtain them from an external cache).
16162 The internal cache can handle only one SSL_SESSION with a given ID,
16163 so if there's a conflict, we now throw out the old one to achieve
16168 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16169 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16170 some routines that use cipher OIDs: some ciphers do not have OIDs
16171 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16176 * Simplify the trust setting structure and code. Now we just have
16177 two sequences of OIDs for trusted and rejected settings. These will
16178 typically have values the same as the extended key usage extension
16179 and any application specific purposes.
16181 The trust checking code now has a default behaviour: it will just
16182 check for an object with the same NID as the passed id. Functions can
16183 be provided to override either the default behaviour or the behaviour
16184 for a given id. SSL client, server and email already have functions
16185 in place for compatibility: they check the NID and also return "trusted"
16186 if the certificate is self signed.
16190 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16191 traditional format into an EVP_PKEY structure.
16195 * Add a password callback function PEM_cb() which either prompts for
16196 a password if usr_data is NULL or otherwise assumes it is a null
16197 terminated password. Allow passwords to be passed on command line
16198 environment or config files in a few more utilities.
16202 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16203 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16204 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16205 Update documentation.
16209 * Support for ASN1 "NULL" type. This could be handled before by using
16210 ASN1_TYPE but there wasn't any function that would try to read a NULL
16211 and produce an error if it couldn't. For compatibility we also have
16212 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16213 don't allocate anything because they don't need to.
16217 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16220 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16222 * Rebuild of the memory allocation routines used by OpenSSL code and
16223 possibly others as well. The purpose is to make an interface that
16224 provide hooks so anyone can build a separate set of allocation and
16225 deallocation routines to be used by OpenSSL, for example memory
16226 pool implementations, or something else, which was previously hard
16227 since Malloc(), Realloc() and Free() were defined as macros having
16228 the values malloc, realloc and free, respectively (except for Win32
16229 compilations). The same is provided for memory debugging code.
16230 OpenSSL already comes with functionality to find memory leaks, but
16231 this gives people a chance to debug other memory problems.
16233 With these changes, a new set of functions and macros have appeared:
16235 CRYPTO_set_mem_debug_functions() [F]
16236 CRYPTO_get_mem_debug_functions() [F]
16237 CRYPTO_dbg_set_options() [F]
16238 CRYPTO_dbg_get_options() [F]
16239 CRYPTO_malloc_debug_init() [M]
16241 The memory debug functions are NULL by default, unless the library
16242 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16243 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16244 gives the standard debugging functions that come with OpenSSL) or
16245 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16246 provided by the library user) must be used. When the standard
16247 debugging functions are used, CRYPTO_dbg_set_options can be used to
16248 request additional information:
16249 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16250 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16252 Also, things like CRYPTO_set_mem_functions will always give the
16253 expected result (the new set of functions is used for allocation
16254 and deallocation) at all times, regardless of platform and compiler
16257 To finish it up, some functions that were never use in any other
16258 way than through macros have a new API and new semantic:
16260 CRYPTO_dbg_malloc()
16261 CRYPTO_dbg_realloc()
16264 All macros of value have retained their old syntax.
16266 *Richard Levitte and Bodo Moeller*
16268 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16269 ordering of SMIMECapabilities wasn't in "strength order" and there
16270 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16275 * Some ASN1 types with illegal zero length encoding (INTEGER,
16276 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16278 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16280 * Merge in my S/MIME library for OpenSSL. This provides a simple
16281 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16282 functionality to handle multipart/signed properly) and a utility
16283 called 'smime' to call all this stuff. This is based on code I
16284 originally wrote for Celo who have kindly allowed it to be
16285 included in OpenSSL.
16289 * Add variants des_set_key_checked and des_set_key_unchecked of
16290 des_set_key (aka des_key_sched). Global variable des_check_key
16291 decides which of these is called by des_set_key; this way
16292 des_check_key behaves as it always did, but applications and
16293 the library itself, which was buggy for des_check_key == 1,
16294 have a cleaner way to pick the version they need.
16298 * New function PKCS12_newpass() which changes the password of a
16303 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16304 dynamic mix. In both cases the ids can be used as an index into the
16305 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16306 functions so they accept a list of the field values and the
16307 application doesn't need to directly manipulate the X509_TRUST
16312 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16317 * Modify the way the V3 extension code looks up extensions. This now
16318 works in a similar way to the object code: we have some "standard"
16319 extensions in a static table which is searched with OBJ_bsearch()
16320 and the application can add dynamic ones if needed. The file
16321 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16322 updated whenever a new extension is added to the core code and kept
16323 in ext_nid order. There is a simple program 'tabtest.c' which checks
16324 this. New extensions are not added too often so this file can readily
16325 be maintained manually.
16327 There are two big advantages in doing things this way. The extensions
16328 can be looked up immediately and no longer need to be "added" using
16329 X509V3_add_standard_extensions(): this function now does nothing.
16330 Side note: I get *lots* of email saying the extension code doesn't
16331 work because people forget to call this function.
16332 Also no dynamic allocation is done unless new extensions are added:
16333 so if we don't add custom extensions there is no need to call
16334 X509V3_EXT_cleanup().
16338 * Modify enc utility's salting as follows: make salting the default. Add a
16339 magic header, so unsalted files fail gracefully instead of just decrypting
16340 to garbage. This is because not salting is a big security hole, so people
16341 should be discouraged from doing it.
16345 * Fixes and enhancements to the 'x509' utility. It allowed a message
16346 digest to be passed on the command line but it only used this
16347 parameter when signing a certificate. Modified so all relevant
16348 operations are affected by the digest parameter including the
16349 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16350 DSA key was used because it didn't fix the digest.
16354 * Initial certificate chain verify code. Currently tests the untrusted
16355 certificates for consistency with the verify purpose (which is set
16356 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16358 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16359 this is because it will reject chains with invalid extensions whereas
16360 every previous version of OpenSSL and SSLeay made no checks at all.
16362 Trust code: checks the root CA for the relevant trust settings. Trust
16363 settings have an initial value consistent with the verify purpose: e.g.
16364 if the verify purpose is for SSL client use it expects the CA to be
16365 trusted for SSL client use. However the default value can be changed to
16366 permit custom trust settings: one example of this would be to only trust
16367 certificates from a specific "secure" set of CAs.
16369 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16370 which should be used for version portability: especially since the
16371 verify structure is likely to change more often now.
16373 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16374 to set them. If not set then assume SSL clients will verify SSL servers
16377 Two new options to the verify program: -untrusted allows a set of
16378 untrusted certificates to be passed in and -purpose which sets the
16379 intended purpose of the certificate. If a purpose is set then the
16380 new chain verify code is used to check extension consistency.
16384 * Support for the authority information access extension.
16388 * Modify RSA and DSA PEM read routines to transparently handle
16389 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16390 public keys in a format compatible with certificate
16391 SubjectPublicKeyInfo structures. Unfortunately there were already
16392 functions called *_PublicKey_* which used various odd formats so
16393 these are retained for compatibility: however the DSA variants were
16394 never in a public release so they have been deleted. Changed dsa/rsa
16395 utilities to handle the new format: note no releases ever handled public
16396 keys so we should be OK.
16398 The primary motivation for this change is to avoid the same fiasco
16399 that dogs private keys: there are several incompatible private key
16400 formats some of which are standard and some OpenSSL specific and
16401 require various evil hacks to allow partial transparent handling and
16402 even then it doesn't work with DER formats. Given the option anything
16403 other than PKCS#8 should be dumped: but the other formats have to
16404 stay in the name of compatibility.
16406 With public keys and the benefit of hindsight one standard format
16407 is used which works with EVP_PKEY, RSA or DSA structures: though
16408 it clearly returns an error if you try to read the wrong kind of key.
16410 Added a -pubkey option to the 'x509' utility to output the public key.
16411 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16412 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16413 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16414 that do the same as the `EVP_PKEY_assign_*()` except they up the
16415 reference count of the added key (they don't "swallow" the
16420 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16421 CRLs would fail if the file contained no certificates or no CRLs:
16422 added a new function to read in both types and return the number
16423 read: this means that if none are read it will be an error. The
16424 DER versions of the certificate and CRL reader would always fail
16425 because it isn't possible to mix certificates and CRLs in DER format
16426 without choking one or the other routine. Changed this to just read
16427 a certificate: this is the best we can do. Also modified the code
16428 in `apps/verify.c` to take notice of return codes: it was previously
16429 attempting to read in certificates from NULL pointers and ignoring
16430 any errors: this is one reason why the cert and CRL reader seemed
16431 to work. It doesn't check return codes from the default certificate
16432 routines: these may well fail if the certificates aren't installed.
16436 * Code to support otherName option in GeneralName.
16440 * First update to verify code. Change the verify utility
16441 so it warns if it is passed a self signed certificate:
16442 for consistency with the normal behaviour. X509_verify
16443 has been modified to it will now verify a self signed
16444 certificate if *exactly* the same certificate appears
16445 in the store: it was previously impossible to trust a
16446 single self signed certificate. This means that:
16447 openssl verify ss.pem
16448 now gives a warning about a self signed certificate but
16449 openssl verify -CAfile ss.pem ss.pem
16454 * For servers, store verify_result in SSL_SESSION data structure
16455 (and add it to external session representation).
16456 This is needed when client certificate verifications fails,
16457 but an application-provided verification callback (set by
16458 SSL_CTX_set_cert_verify_callback) allows accepting the session
16459 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16460 but returns 1): When the session is reused, we have to set
16461 ssl->verify_result to the appropriate error code to avoid
16464 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16466 * Fix a bug in the new PKCS#7 code: it didn't consider the
16467 case in PKCS7_dataInit() where the signed PKCS7 structure
16468 didn't contain any existing data because it was being created.
16470 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16472 * Add a salt to the key derivation routines in enc.c. This
16473 forms the first 8 bytes of the encrypted file. Also add a
16474 -S option to allow a salt to be input on the command line.
16478 * New function X509_cmp(). Oddly enough there wasn't a function
16479 to compare two certificates. We do this by working out the SHA1
16480 hash and comparing that. X509_cmp() will be needed by the trust
16485 * SSL_get1_session() is like SSL_get_session(), but increments
16486 the reference count in the SSL_SESSION returned.
16488 *Geoff Thorpe <geoff@eu.c2.net>*
16490 * Fix for 'req': it was adding a null to request attributes.
16491 Also change the X509_LOOKUP and X509_INFO code to handle
16492 certificate auxiliary information.
16496 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16501 * Add the possibility to add extra information to the memory leak
16502 detecting output, to form tracebacks, showing from where each
16503 allocation was originated: CRYPTO_push_info("constant string") adds
16504 the string plus current file name and line number to a per-thread
16505 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16506 is like calling CYRPTO_pop_info() until the stack is empty.
16507 Also updated memory leak detection code to be multi-thread-safe.
16511 * Add options -text and -noout to pkcs7 utility and delete the
16512 encryption options which never did anything. Update docs.
16516 * Add options to some of the utilities to allow the pass phrase
16517 to be included on either the command line (not recommended on
16518 OSes like Unix) or read from the environment. Update the
16519 manpages and fix a few bugs.
16523 * Add a few manpages for some of the openssl commands.
16527 * Fix the -revoke option in ca. It was freeing up memory twice,
16528 leaking and not finding already revoked certificates.
16532 * Extensive changes to support certificate auxiliary information.
16533 This involves the use of X509_CERT_AUX structure and X509_AUX
16534 functions. An X509_AUX function such as PEM_read_X509_AUX()
16535 can still read in a certificate file in the usual way but it
16536 will also read in any additional "auxiliary information". By
16537 doing things this way a fair degree of compatibility can be
16538 retained: existing certificates can have this information added
16539 using the new 'x509' options.
16541 Current auxiliary information includes an "alias" and some trust
16542 settings. The trust settings will ultimately be used in enhanced
16543 certificate chain verification routines: currently a certificate
16544 can only be trusted if it is self signed and then it is trusted
16549 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16550 The problem was that one of the replacement routines had not been working
16551 since SSLeay releases. For now the offending routine has been replaced
16552 with non-optimised assembler. Even so, this now gives around 95%
16553 performance improvement for 1024 bit RSA signs.
16557 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16558 handling. Most clients have the effective key size in bits equal to
16559 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16560 A few however don't do this and instead use the size of the decrypted key
16561 to determine the RC2 key length and the AlgorithmIdentifier to determine
16562 the effective key length. In this case the effective key length can still
16563 be 40 bits but the key length can be 168 bits for example. This is fixed
16564 by manually forcing an RC2 key into the EVP_PKEY structure because the
16565 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16566 the key length and effective key length are equal.
16570 * Add a bunch of functions that should simplify the creation of
16571 X509_NAME structures. Now you should be able to do:
16572 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16573 and have it automatically work out the correct field type and fill in
16574 the structures. The more adventurous can try:
16575 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16576 and it will (hopefully) work out the correct multibyte encoding.
16580 * Change the 'req' utility to use the new field handling and multibyte
16581 copy routines. Before the DN field creation was handled in an ad hoc
16582 way in req, ca, and x509 which was rather broken and didn't support
16583 BMPStrings or UTF8Strings. Since some software doesn't implement
16584 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16585 using the dirstring_type option. See the new comment in the default
16586 openssl.cnf for more info.
16590 * Make crypto/rand/md_rand.c more robust:
16591 - Assure unique random numbers after fork().
16592 - Make sure that concurrent threads access the global counter and
16593 md serializably so that we never lose entropy in them
16594 or use exactly the same state in multiple threads.
16595 Access to the large state is not always serializable because
16596 the additional locking could be a performance killer, and
16597 md should be large enough anyway.
16601 * New file `apps/app_rand.c` with commonly needed functionality
16602 for handling the random seed file.
16604 Use the random seed file in some applications that previously did not:
16606 dsaparam -genkey (which also ignored its '-rand' option),
16609 x509 (when signing).
16610 Except on systems with /dev/urandom, it is crucial to have a random
16611 seed file at least for key creation, DSA signing, and for DH exchanges;
16612 for RSA signatures we could do without one.
16614 gendh and gendsa (unlike genrsa) used to read only the first byte
16615 of each file listed in the '-rand' option. The function as previously
16616 found in genrsa is now in app_rand.c and is used by all programs
16617 that support '-rand'.
16621 * In RAND_write_file, use mode 0600 for creating files;
16622 don't just chmod when it may be too late.
16626 * Report an error from X509_STORE_load_locations
16627 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16631 * New function ASN1_mbstring_copy() this copies a string in either
16632 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16633 into an ASN1_STRING type. A mask of permissible types is passed
16634 and it chooses the "minimal" type to use or an error if not type
16639 * Add function equivalents to the various macros in asn1.h. The old
16640 macros are retained with an `M_` prefix. Code inside the library can
16641 use the `M_` macros. External code (including the openssl utility)
16642 should *NOT* in order to be "shared library friendly".
16646 * Add various functions that can check a certificate's extensions
16647 to see if it usable for various purposes such as SSL client,
16648 server or S/MIME and CAs of these types. This is currently
16649 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16650 verification. Also added a -purpose flag to x509 utility to
16651 print out all the purposes.
16655 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16660 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16661 for, obtain and decode and extension and obtain its critical flag.
16662 This allows all the necessary extension code to be handled in a
16663 single function call.
16667 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16668 platforms. See crypto/rc4/rc4_enc.c for further details.
16672 * New -noout option to asn1parse. This causes no output to be produced
16673 its main use is when combined with -strparse and -out to extract data
16674 from a file (which may not be in ASN.1 format).
16678 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16679 when producing the local key id.
16681 *Richard Levitte <levitte@stacken.kth.se>*
16683 * New option -dhparam in s_server. This allows a DH parameter file to be
16684 stated explicitly. If it is not stated then it tries the first server
16685 certificate file. The previous behaviour hard coded the filename
16690 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16691 a public key to be input or output. For example:
16692 openssl rsa -in key.pem -pubout -out pubkey.pem
16693 Also added necessary DSA public key functions to handle this.
16697 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16698 in the message. This was handled by allowing
16699 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16701 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16703 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16704 to the end of the strings whereas this didn't. This would cause problems
16705 if strings read with d2i_ASN1_bytes() were later modified.
16707 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16709 * Fix for base64 decode bug. When a base64 bio reads only one line of
16710 data and it contains EOF it will end up returning an error. This is
16711 caused by input 46 bytes long. The cause is due to the way base64
16712 BIOs find the start of base64 encoded data. They do this by trying a
16713 trial decode on each line until they find one that works. When they
16714 do a flag is set and it starts again knowing it can pass all the
16715 data directly through the decoder. Unfortunately it doesn't reset
16716 the context it uses. This means that if EOF is reached an attempt
16717 is made to pass two EOFs through the context and this causes the
16718 resulting error. This can also cause other problems as well. As is
16719 usual with these problems it takes *ages* to find and the fix is
16720 trivial: move one line.
16722 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16724 * Ugly workaround to get s_client and s_server working under Windows. The
16725 old code wouldn't work because it needed to select() on sockets and the
16726 tty (for keypresses and to see if data could be written). Win32 only
16727 supports select() on sockets so we select() with a 1s timeout on the
16728 sockets and then see if any characters are waiting to be read, if none
16729 are present then we retry, we also assume we can always write data to
16730 the tty. This isn't nice because the code then blocks until we've
16731 received a complete line of data and it is effectively polling the
16732 keyboard at 1s intervals: however it's quite a bit better than not
16733 working at all :-) A dedicated Windows application might handle this
16734 with an event loop for example.
16738 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16739 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16740 will be called when RSA_sign() and RSA_verify() are used. This is useful
16741 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16742 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16743 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16744 This necessitated the support of an extra signature type NID_md5_sha1
16745 for SSL signatures and modifications to the SSL library to use it instead
16746 of calling RSA_public_decrypt() and RSA_private_encrypt().
16750 * Add new -verify -CAfile and -CApath options to the crl program, these
16751 will lookup a CRL issuers certificate and verify the signature in a
16752 similar way to the verify program. Tidy up the crl program so it
16753 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16754 less strict. It will now permit CRL extensions even if it is not
16755 a V2 CRL: this will allow it to tolerate some broken CRLs.
16759 * Initialize all non-automatic variables each time one of the openssl
16760 sub-programs is started (this is necessary as they may be started
16761 multiple times from the "OpenSSL>" prompt).
16763 *Lennart Bang, Bodo Moeller*
16765 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16766 removing all other RSA functionality (this is what NO_RSA does). This
16767 is so (for example) those in the US can disable those operations covered
16768 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16773 * Non-copying interface to BIO pairs.
16774 (still largely untested)
16778 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16779 ASCII string. This was handled independently in various places before.
16783 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16784 UTF8 strings a character at a time.
16788 * Use client_version from client hello to select the protocol
16789 (s23_srvr.c) and for RSA client key exchange verification
16790 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16794 * Add various utility functions to handle SPKACs, these were previously
16795 handled by poking round in the structure internals. Added new function
16796 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16797 print, verify and generate SPKACs. Based on an original idea from
16798 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16802 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16806 * Allow the config file extension section to be overwritten on the
16807 command line. Based on an original idea from Massimiliano Pala
16808 <madwolf@comune.modena.it>. The new option is called -extensions
16809 and can be applied to ca, req and x509. Also -reqexts to override
16810 the request extensions in req and -crlexts to override the crl extensions
16815 * Add new feature to the SPKAC handling in ca. Now you can include
16816 the same field multiple times by preceding it by "XXXX." for example:
16819 this is the same syntax as used in the req config file.
16823 * Allow certificate extensions to be added to certificate requests. These
16824 are specified in a 'req_extensions' option of the req section of the
16825 config file. They can be printed out with the -text option to req but
16826 are otherwise ignored at present.
16830 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16831 data read consists of only the final block it would not decrypted because
16832 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16833 A misplaced 'break' also meant the decrypted final block might not be
16834 copied until the next read.
16838 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16839 a few extra parameters to the DH structure: these will be useful if
16840 for example we want the value of 'q' or implement X9.42 DH.
16844 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16845 provides hooks that allow the default DSA functions or functions on a
16846 "per key" basis to be replaced. This allows hardware acceleration and
16847 hardware key storage to be handled without major modification to the
16848 library. Also added low-level modexp hooks and CRYPTO_EX structure and
16849 associated functions.
16853 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16854 as "read only": it can't be written to and the buffer it points to will
16855 not be freed. Reading from a read only BIO is much more efficient than
16856 a normal memory BIO. This was added because there are several times when
16857 an area of memory needs to be read from a BIO. The previous method was
16858 to create a memory BIO and write the data to it, this results in two
16859 copies of the data and an O(n^2) reading algorithm. There is a new
16860 function BIO_new_mem_buf() which creates a read only memory BIO from
16861 an area of memory. Also modified the PKCS#7 routines to use read only
16866 * Bugfix: ssl23_get_client_hello did not work properly when called in
16867 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16868 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16869 but a retry condition occurred while trying to read the rest.
16873 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16874 NID_pkcs7_encrypted by default: this was wrong since this should almost
16875 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16876 the encrypted data type: this is a more sensible place to put it and it
16877 allows the PKCS#12 code to be tidied up that duplicated this
16882 * Changed obj_dat.pl script so it takes its input and output files on
16883 the command line. This should avoid shell escape redirection problems
16888 * Initial support for certificate extension requests, these are included
16889 in things like Xenroll certificate requests. Included functions to allow
16890 extensions to be obtained and added.
16894 * -crlf option to s_client and s_server for sending newlines as
16895 CRLF (as required by many protocols).
16899 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16901 * Install libRSAglue.a when OpenSSL is built with RSAref.
16903 *Ralf S. Engelschall*
16905 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16907 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16909 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16914 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16915 DH parameters/keys (q is lost during that conversion, but the resulting
16916 DH parameters contain its length).
16918 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16919 much faster than DH_generate_parameters (which creates parameters
16920 where `p = 2*q + 1`), and also the smaller q makes DH computations
16921 much more efficient (160-bit exponentiation instead of 1024-bit
16922 exponentiation); so this provides a convenient way to support DHE
16923 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16924 utter importance to use
16925 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16927 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16928 when such DH parameters are used, because otherwise small subgroup
16929 attacks may become possible!
16933 * Avoid memory leak in i2d_DHparams.
16937 * Allow the -k option to be used more than once in the enc program:
16938 this allows the same encrypted message to be read by multiple recipients.
16942 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16943 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16944 it will always use the numerical form of the OID, even if it has a short
16949 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16950 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16951 otherwise bn_mod_exp was called. In the case of hardware keys for example
16952 no private key components need be present and it might store extra data
16953 in the RSA structure, which cannot be accessed from bn_mod_exp.
16954 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16955 private key operations.
16959 * Added support for SPARC Linux.
16963 * pem_password_cb function type incompatibly changed from
16964 typedef int pem_password_cb(char *buf, int size, int rwflag);
16966 ....(char *buf, int size, int rwflag, void *userdata);
16967 so that applications can pass data to their callbacks:
16968 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16969 additional void * argument, which is just handed through whenever
16970 the password callback is called.
16972 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16974 New function SSL_CTX_set_default_passwd_cb_userdata.
16976 Compatibility note: As many C implementations push function arguments
16977 onto the stack in reverse order, the new library version is likely to
16978 interoperate with programs that have been compiled with the old
16979 pem_password_cb definition (PEM_whatever takes some data that
16980 happens to be on the stack as its last argument, and the callback
16981 just ignores this garbage); but there is no guarantee whatsoever that
16984 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16985 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16986 problems not only on Windows, but also on some Unix platforms.
16987 To avoid problematic command lines, these definitions are now in an
16988 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16989 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16993 * MIPS III/IV assembler module is reimplemented.
16997 * More DES library cleanups: remove references to srand/rand and
16998 delete an unused file.
17002 * Add support for the free Netwide assembler (NASM) under Win32,
17003 since not many people have MASM (ml) and it can be hard to obtain.
17004 This is currently experimental but it seems to work OK and pass all
17005 the tests. Check out INSTALL.W32 for info.
17009 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17010 without temporary keys kept an extra copy of the server key,
17011 and connections with temporary keys did not free everything in case
17016 * New function RSA_check_key and new openssl rsa option -check
17017 for verifying the consistency of RSA keys.
17019 *Ulf Moeller, Bodo Moeller*
17021 * Various changes to make Win32 compile work:
17022 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17023 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17024 comparison" warnings.
17025 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17029 * Add a debugging option to PKCS#5 v2 key generation function: when
17030 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17031 derived keys are printed to stderr.
17035 * Copy the flags in ASN1_STRING_dup().
17037 *Roman E. Pavlov <pre@mo.msk.ru>*
17039 * The x509 application mishandled signing requests containing DSA
17040 keys when the signing key was also DSA and the parameters didn't match.
17042 It was supposed to omit the parameters when they matched the signing key:
17043 the verifying software was then supposed to automatically use the CA's
17044 parameters if they were absent from the end user certificate.
17046 Omitting parameters is no longer recommended. The test was also
17047 the wrong way round! This was probably due to unusual behaviour in
17048 EVP_cmp_parameters() which returns 1 if the parameters match.
17049 This meant that parameters were omitted when they *didn't* match and
17050 the certificate was useless. Certificates signed with 'ca' didn't have
17053 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17055 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17056 The interface is as follows:
17057 Applications can use
17058 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17059 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17060 "off" is now the default.
17061 The library internally uses
17062 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17063 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17064 to disable memory-checking temporarily.
17066 Some inconsistent states that previously were possible (and were
17067 even the default) are now avoided.
17069 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17070 with each memory chunk allocated; this is occasionally more helpful
17071 than just having a counter.
17073 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17075 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17080 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17081 which largely parallels "options", but is for changing API behaviour,
17082 whereas "options" are about protocol behaviour.
17083 Initial "mode" flags are:
17085 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17086 a single record has been written.
17087 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17088 retries use the same buffer location.
17089 (But all of the contents must be
17094 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17097 * Fix problems with no-hmac etc.
17099 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17101 * New functions RSA_get_default_method(), RSA_set_method() and
17102 RSA_get_method(). These allows replacement of RSA_METHODs without having
17103 to mess around with the internals of an RSA structure.
17107 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17108 Also really enable memory leak checks in openssl.c and in some
17111 *Chad C. Mulligan, Bodo Moeller*
17113 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17114 up the length of negative integers. This has now been simplified to just
17115 store the length when it is first determined and use it later, rather
17116 than trying to keep track of where data is copied and updating it to
17118 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17120 * Add a new function PKCS7_signatureVerify. This allows the verification
17121 of a PKCS#7 signature but with the signing certificate passed to the
17122 function itself. This contrasts with PKCS7_dataVerify which assumes the
17123 certificate is present in the PKCS#7 structure. This isn't always the
17124 case: certificates can be omitted from a PKCS#7 structure and be
17125 distributed by "out of band" means (such as a certificate database).
17129 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17130 function prototypes in pem.h, also change util/mkdef.pl to add the
17131 necessary function names.
17135 * mk1mf.pl (used by Windows builds) did not properly read the
17136 options set by Configure in the top level Makefile, and Configure
17137 was not even able to write more than one option correctly.
17138 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17142 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17143 file to be loaded from a BIO or FILE pointer. The BIO version will
17144 for example allow memory BIOs to contain config info.
17148 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17149 Whoever hopes to achieve shared-library compatibility across versions
17150 must use this, not the compile-time macro.
17151 (Exercise 0.9.4: Which is the minimum library version required by
17153 Note: All this applies only to multi-threaded programs, others don't
17158 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17159 through a BIO pair triggered the default case, i.e.
17160 SSLerr(...,SSL_R_UNKNOWN_STATE).
17164 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17165 can use the SSL library even if none of the specific BIOs is
17170 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17171 for the encoded length.
17173 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17175 * Add initial documentation of the X509V3 functions.
17179 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17180 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17181 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17182 secure PKCS#8 private key format with a high iteration count.
17186 * Fix determination of Perl interpreter: A perl or perl5
17187 *directory* in $PATH was also accepted as the interpreter.
17189 *Ralf S. Engelschall*
17191 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17192 wrong with it but it was very old and did things like calling
17193 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17194 unusual formatting.
17198 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17199 to use the new extension code.
17203 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17204 with macros. This should make it easier to change their form, add extra
17205 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17210 * Add to configuration table a new entry that can specify an alternative
17211 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17212 according to Mark Crispin <MRC@Panda.COM>.
17216 * DES CBC did not update the IV. Weird.
17220 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17221 Changing the behaviour of the former might break existing programs --
17222 where IV updating is needed, des_ncbc_encrypt can be used.
17225 * When bntest is run from "make test" it drives bc to check its
17226 calculations, as well as internally checking them. If an internal check
17227 fails, it needs to cause bc to give a non-zero result or make test carries
17228 on without noticing the failure. Fixed.
17232 * DES library cleanups.
17236 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17237 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17238 ciphers. NOTE: although the key derivation function has been verified
17239 against some published test vectors it has not been extensively tested
17240 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17245 * Instead of "mkdir -p", which is not fully portable, use new
17246 Perl script "util/mkdir-p.pl".
17250 * Rewrite the way password based encryption (PBE) is handled. It used to
17251 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17252 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17253 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17254 the 'parameter' field of the AlgorithmIdentifier is passed to the
17255 underlying key generation function so it must do its own ASN1 parsing.
17256 This has also changed the EVP_PBE_CipherInit() function which now has a
17257 'parameter' argument instead of literal salt and iteration count values
17258 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17262 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17263 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17264 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17265 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17266 value was just used as a "magic string" and not used directly its
17267 value doesn't matter.
17271 * Introduce some semblance of const correctness to BN. Shame C doesn't
17276 * "linux-sparc64" configuration (ultrapenguin).
17278 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17279 "linux-sparc" configuration.
17281 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17283 * config now generates no-xxx options for missing ciphers.
17287 * Support the EBCDIC character set (work in progress).
17288 File ebcdic.c not yet included because it has a different license.
17290 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17292 * Support BS2000/OSD-POSIX.
17294 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17296 * Make callbacks for key generation use `void *` instead of `char *`.
17300 * Make S/MIME samples compile (not yet tested).
17304 * Additional typesafe stacks.
17308 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17312 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17314 * New configuration variant "sco5-gcc".
17316 * Updated some demos.
17318 *Sean O Riordain, Wade Scholine*
17320 * Add missing BIO_free at exit of pkcs12 application.
17324 * Fix memory leak in conf.c.
17328 * Updates for Win32 to assembler version of MD5.
17332 * Set #! path to perl in `apps/der_chop` to where we found it
17333 instead of using a fixed path.
17337 * SHA library changes for irix64-mips4-cc.
17341 * Improvements for VMS support.
17345 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17347 * Bignum library bug fix. IRIX 6 passes "make test" now!
17348 This also avoids the problems with SC4.2 and unpatched SC5.
17350 *Andy Polyakov <appro@fy.chalmers.se>*
17352 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17353 These are required because of the typesafe stack would otherwise break
17354 existing code. If old code used a structure member which used to be STACK
17355 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17356 sk_num or sk_value it would produce an error because the num, data members
17357 are not present in STACK_OF. Now it just produces a warning. sk_set
17358 replaces the old method of assigning a value to sk_value
17359 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17360 that does this will no longer work (and should use sk_set instead) but
17361 this could be regarded as a "questionable" behaviour anyway.
17365 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17366 correctly handle encrypted S/MIME data.
17370 * Change type of various DES function arguments from des_cblock
17371 (which means, in function argument declarations, pointer to char)
17372 to des_cblock * (meaning pointer to array with 8 char elements),
17373 which allows the compiler to do more typechecking; it was like
17374 that back in SSLeay, but with lots of ugly casts.
17376 Introduce new type const_des_cblock.
17380 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17381 problems: find RecipientInfo structure that matches recipient certificate
17382 and initialise the ASN1 structures properly based on passed cipher.
17386 * Belatedly make the BN tests actually check the results.
17390 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17391 to and from BNs: it was completely broken. New compilation option
17392 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17393 key elements as negative integers.
17397 * Reorganize and speed up MD5.
17399 *Andy Polyakov <appro@fy.chalmers.se>*
17403 *Richard Levitte <richard@levitte.org>*
17405 * New option -out to asn1parse to allow the parsed structure to be
17406 output to a file. This is most useful when combined with the -strparse
17407 option to examine the output of things like OCTET STRINGS.
17411 * Make SSL library a little more fool-proof by not requiring any longer
17412 that `SSL_set_{accept,connect}_state` be called before
17413 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17414 in many applications because usually everything *appeared* to work as
17415 intended anyway -- now it really works as intended).
17419 * Move openssl.cnf out of lib/.
17423 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17424 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17425 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17427 *Ralf S. Engelschall*
17429 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17430 handle PKCS#7 enveloped data properly.
17432 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17434 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17435 copying pointers. The cert_st handling is changed by this in
17436 various ways (and thus what used to be known as ctx->default_cert
17437 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17438 any longer when s->cert does not give us what we need).
17439 ssl_cert_instantiate becomes obsolete by this change.
17440 As soon as we've got the new code right (possibly it already is?),
17441 we have solved a couple of bugs of the earlier code where s->cert
17442 was used as if it could not have been shared with other SSL structures.
17444 Note that using the SSL API in certain dirty ways now will result
17445 in different behaviour than observed with earlier library versions:
17446 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17447 does not influence s as it used to.
17449 In order to clean up things more thoroughly, inside SSL_SESSION
17450 we don't use CERT any longer, but a new structure SESS_CERT
17451 that holds per-session data (if available); currently, this is
17452 the peer's certificate chain and, for clients, the server's certificate
17453 and temporary key. CERT holds only those values that can have
17454 meaningful defaults in an SSL_CTX.
17458 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17459 from the internal representation. Various PKCS#7 fixes: remove some
17460 evil casts and set the enc_dig_alg field properly based on the signing
17465 * Allow PKCS#12 password to be set from the command line or the
17466 environment. Let 'ca' get its config file name from the environment
17467 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17472 * Allow certificate policies extension to use an IA5STRING for the
17473 organization field. This is contrary to the PKIX definition but
17474 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17479 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17480 without disallowing inline assembler and the like for non-pedantic builds.
17484 * Support Borland C++ builder.
17486 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17492 * SHA-1 cleanups and performance enhancements.
17494 *Andy Polyakov <appro@fy.chalmers.se>*
17496 * Sparc v8plus assembler for the bignum library.
17498 *Andy Polyakov <appro@fy.chalmers.se>*
17500 * Accept any -xxx and +xxx compiler options in Configure.
17504 * Update HPUX configuration.
17508 * Add missing `sk_<type>_unshift()` function to safestack.h
17510 *Ralf S. Engelschall*
17512 * New function SSL_CTX_use_certificate_chain_file that sets the
17513 "extra_cert"s in addition to the certificate. (This makes sense
17514 only for "PEM" format files, as chains as a whole are not
17519 * Support verify_depth from the SSL API.
17520 x509_vfy.c had what can be considered an off-by-one-error:
17521 Its depth (which was not part of the external interface)
17522 was actually counting the number of certificates in a chain;
17523 now it really counts the depth.
17527 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17528 instead of X509err, which often resulted in confusing error
17529 messages since the error codes are not globally unique
17530 (e.g. an alleged error in ssl3_accept when a certificate
17531 didn't match the private key).
17533 * New function SSL_CTX_set_session_id_context that allows to set a default
17534 value (so that you don't need SSL_set_session_id_context for each
17535 connection using the SSL_CTX).
17539 * OAEP decoding bug fix.
17543 * Support INSTALL_PREFIX for package builders, as proposed by
17548 * New Configure options "threads" and "no-threads". For systems
17549 where the proper compiler options are known (currently Solaris
17550 and Linux), "threads" is the default.
17554 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17558 * Install various scripts to $(OPENSSLDIR)/misc, not to
17559 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17560 such as /usr/local/bin.
17564 * "make linux-shared" to build shared libraries.
17566 *Niels Poppe <niels@netbox.org>*
17568 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17572 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17573 extension adding in x509 utility.
17577 * Remove NOPROTO sections and error code comments.
17581 * Partial rewrite of the DEF file generator to now parse the ANSI
17586 * New Configure options --prefix=DIR and --openssldir=DIR.
17590 * Complete rewrite of the error code script(s). It is all now handled
17591 by one script at the top level which handles error code gathering,
17592 header rewriting and C source file generation. It should be much better
17593 than the old method: it now uses a modified version of Ulf's parser to
17594 read the ANSI prototypes in all header files (thus the old K&R definitions
17595 aren't needed for error creation any more) and do a better job of
17596 translating function codes into names. The old 'ASN1 error code embedded
17597 in a comment' is no longer necessary and it doesn't use .err files which
17598 have now been deleted. Also the error code call doesn't have to appear all
17599 on one line (which resulted in some large lines...).
17603 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17607 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17608 0 (which usually indicates a closed connection), but continue reading.
17612 * Fix some race conditions.
17616 * Add support for CRL distribution points extension. Add Certificate
17617 Policies and CRL distribution points documentation.
17621 * Move the autogenerated header file parts to crypto/opensslconf.h.
17625 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17626 8 of keying material. Merlin has also confirmed interop with this fix
17627 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17629 *Merlin Hughes <merlin@baltimore.ie>*
17631 * Fix lots of warnings.
17633 *Richard Levitte <levitte@stacken.kth.se>*
17635 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17636 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17638 *Richard Levitte <levitte@stacken.kth.se>*
17640 * Fix problems with sizeof(long) == 8.
17642 *Andy Polyakov <appro@fy.chalmers.se>*
17644 * Change functions to ANSI C.
17648 * Fix typos in error codes.
17650 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17652 * Remove defunct assembler files from Configure.
17656 * SPARC v8 assembler BIGNUM implementation.
17658 *Andy Polyakov <appro@fy.chalmers.se>*
17660 * Support for Certificate Policies extension: both print and set.
17661 Various additions to support the r2i method this uses.
17665 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17666 return a const string when you are expecting an allocated buffer.
17670 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17671 types DirectoryString and DisplayText.
17675 * Add code to allow r2i extensions to access the configuration database,
17676 add an LHASH database driver and add several ctx helper functions.
17680 * Fix an evil bug in bn_expand2() which caused various BN functions to
17681 fail when they extended the size of a BIGNUM.
17685 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17686 support typesafe stack.
17690 * Fix typo in SSL_[gs]et_options().
17692 *Nils Frostberg <nils@medcom.se>*
17694 * Delete various functions and files that belonged to the (now obsolete)
17695 old X509V3 handling code.
17699 * New Configure option "rsaref".
17703 * Don't auto-generate pem.h.
17707 * Introduce type-safe ASN.1 SETs.
17711 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17713 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17715 * Introduce type-safe STACKs. This will almost certainly break lots of code
17716 that links with OpenSSL (well at least cause lots of warnings), but fear
17717 not: the conversion is trivial, and it eliminates loads of evil casts. A
17718 few STACKed things have been converted already. Feel free to convert more.
17719 In the fullness of time, I'll do away with the STACK type altogether.
17723 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17724 specified in `<certfile>` by updating the entry in the index.txt file.
17725 This way one no longer has to edit the index.txt file manually for
17726 revoking a certificate. The -revoke option does the gory details now.
17728 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17730 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17731 `-text` option at all and this way the `-noout -text` combination was
17732 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17734 *Ralf S. Engelschall*
17736 * Make sure a corresponding plain text error message exists for the
17737 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17738 verify callback function determined that a certificate was revoked.
17740 *Ralf S. Engelschall*
17742 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17743 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17744 all available ciphers including rc5, which was forgotten until now.
17745 In order to let the testing shell script know which algorithms
17746 are available, a new (up to now undocumented) command
17747 `openssl list-cipher-commands` is used.
17751 * Bugfix: s_client occasionally would sleep in select() when
17752 it should have checked SSL_pending() first.
17756 * New functions DSA_do_sign and DSA_do_verify to provide access to
17757 the raw DSA values prior to ASN.1 encoding.
17761 * Tweaks to Configure
17763 *Niels Poppe <niels@netbox.org>*
17765 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17770 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17774 * New config option to avoid instructions that are illegal on the 80386.
17775 The default code is faster, but requires at least a 486.
17779 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17780 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17781 same as SSL2_VERSION anyway.
17785 * New "-showcerts" option for s_client.
17789 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17790 application. Various cleanups and fixes.
17794 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17795 modify error routines to work internally. Add error codes and PBE init
17796 to library startup routines.
17800 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17801 packing functions to asn1 and evp. Changed function names and error
17802 codes along the way.
17806 * PKCS12 integration: and so it begins... First of several patches to
17807 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17808 objects to objects.h
17812 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17813 and display support for Thawte strong extranet extension.
17817 * Add LinuxPPC support.
17819 *Jeff Dubrule <igor@pobox.org>*
17821 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17822 bn_div_words in alpha.s.
17824 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17826 * Make sure the RSA OAEP test is skipped under -DRSAref because
17827 OAEP isn't supported when OpenSSL is built with RSAref.
17829 *Ulf Moeller <ulf@fitug.de>*
17831 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17832 so they no longer are missing under -DNOPROTO.
17834 *Soren S. Jorvang <soren@t.dk>*
17836 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17838 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17839 doesn't work when the session is reused. Coming soon!
17843 * Fix a security hole, that allows sessions to be reused in the wrong
17844 context thus bypassing client cert protection! All software that uses
17845 client certs and session caches in multiple contexts NEEDS PATCHING to
17846 allow session reuse! A fuller solution is in the works.
17848 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17850 * Some more source tree cleanups (removed obsolete files
17851 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17852 permission on "config" script to be executable) and a fix for the INSTALL
17855 *Ulf Moeller <ulf@fitug.de>*
17857 * Remove some legacy and erroneous uses of malloc, free instead of
17860 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17862 * Make rsa_oaep_test return non-zero on error.
17864 *Ulf Moeller <ulf@fitug.de>*
17866 * Add support for native Solaris shared libraries. Configure
17867 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17868 if someone would make that last step automatic.
17870 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17872 * ctx_size was not built with the right compiler during "make links". Fixed.
17876 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17877 except NULL ciphers". This means the default cipher list will no longer
17878 enable NULL ciphers. They need to be specifically enabled e.g. with
17879 the string "DEFAULT:eNULL".
17883 * Fix to RSA private encryption routines: if p < q then it would
17884 occasionally produce an invalid result. This will only happen with
17885 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17889 * Be less restrictive and allow also `perl util/perlpath.pl
17890 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17891 because this way one can also use an interpreter named `perl5` (which is
17892 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17893 installed as `perl`).
17895 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17897 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17899 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17901 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17902 advapi32.lib to Win32 build and change the pem test comparison
17903 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17904 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17905 and crypto/des/ede_cbcm_enc.c.
17909 * DES quad checksum was broken on big-endian architectures. Fixed.
17913 * Comment out two functions in bio.h that aren't implemented. Fix up the
17914 Win32 test batch file so it (might) work again. The Win32 test batch file
17915 is horrible: I feel ill....
17919 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17920 in e_os.h. Audit of header files to check ANSI and non ANSI
17921 sections: 10 functions were absent from non ANSI section and not exported
17922 from Windows DLLs. Fixed up libeay.num for new functions.
17926 * Make `openssl version` output lines consistent.
17928 *Ralf S. Engelschall*
17930 * Fix Win32 symbol export lists for BIO functions: Added
17931 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17932 to ms/libeay{16,32}.def.
17934 *Ralf S. Engelschall*
17936 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17937 fine under Unix and passes some trivial tests I've now added. But the
17938 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17939 added to make sure no one expects that this stuff really works in the
17940 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17941 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17944 *Ralf S. Engelschall*
17946 * Fix the generation of two part addresses in perl.
17948 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17950 * Add config entry for Linux on MIPS.
17952 *John Tobey <jtobey@channel1.com>*
17954 * Make links whenever Configure is run, unless we are on Windoze.
17958 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17959 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17964 * Add a useful kludge to allow package maintainers to specify compiler and
17965 other platforms details on the command line without having to patch the
17966 Configure script every time: One now can use
17967 `perl Configure <id>:<details>`,
17968 i.e. platform ids are allowed to have details appended
17969 to them (separated by colons). This is treated as there would be a static
17970 pre-configured entry in Configure's %table under key `<id>` with value
17971 `<details>` and `perl Configure <id>` is called. So, when you want to
17972 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17973 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17974 now, which overrides the FreeBSD-elf entry on-the-fly.
17976 *Ralf S. Engelschall*
17978 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17982 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17983 on the `perl Configure ...` command line. This way one can compile
17984 OpenSSL libraries with Position Independent Code (PIC) which is needed
17985 for linking it into DSOs.
17987 *Ralf S. Engelschall*
17989 * Remarkably, export ciphers were totally broken and no-one had noticed!
17994 * Cleaned up the LICENSE document: The official contact for any license
17995 questions now is the OpenSSL core team under openssl-core@openssl.org.
17996 And add a paragraph about the dual-license situation to make sure people
17997 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17998 to the OpenSSL toolkit.
18000 *Ralf S. Engelschall*
18002 * General source tree makefile cleanups: Made `making xxx in yyy...`
18003 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18004 Additionally cleaned up the `make links` target: Remove unnecessary
18005 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18006 to speed processing and no longer clutter the display with confusing
18007 stuff. Instead only the actually done links are displayed.
18009 *Ralf S. Engelschall*
18011 * Permit null encryption ciphersuites, used for authentication only. It used
18012 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18013 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18018 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18019 signed attributes when verifying signatures (this would break them),
18020 the detached data encoding was wrong and public keys obtained using
18021 X509_get_pubkey() weren't freed.
18025 * Add text documentation for the BUFFER functions. Also added a work around
18026 to a Win95 console bug. This was triggered by the password read stuff: the
18027 last character typed gets carried over to the next fread(). If you were
18028 generating a new cert request using 'req' for example then the last
18029 character of the passphrase would be CR which would then enter the first
18034 * Added the new 'Includes OpenSSL Cryptography Software' button as
18035 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18036 button and can be used by applications based on OpenSSL to show the
18037 relationship to the OpenSSL project.
18039 *Ralf S. Engelschall*
18041 * Remove confusing variables in function signatures in files
18042 ssl/ssl_lib.c and ssl/ssl.h.
18044 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18046 * Don't install bss_file.c under PREFIX/include/
18048 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18050 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18051 functions that return function pointers and has support for NT specific
18052 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18053 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18054 unsigned to signed types: this was killing the Win32 compile.
18058 * Add new certificate file to stack functions,
18059 SSL_add_dir_cert_subjects_to_stack() and
18060 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18061 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18062 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18063 This means that Apache-SSL and similar packages don't have to mess around
18064 to add as many CAs as they want to the preferred list.
18068 * Experiment with doxygen documentation. Currently only partially applied to
18070 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18071 openssl.doxy as the configuration file.
18075 * Get rid of remaining C++-style comments which strict C compilers hate.
18077 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18079 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18080 compiled in by default: it has problems with large keys.
18084 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18085 DH private keys and/or callback functions which directly correspond to
18086 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18087 is needed for applications which have to configure certificates on a
18088 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18090 For the RSA certificate situation is makes no difference, but
18091 for the DSA certificate situation this fixes the "no shared cipher"
18092 problem where the OpenSSL cipher selection procedure failed because the
18093 temporary keys were not overtaken from the context and the API provided
18094 no way to reconfigure them.
18095 The new functions now let applications reconfigure the stuff and they
18096 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18097 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18098 non-public-API function ssl_cert_instantiate() is used as a helper
18099 function and also to reduce code redundancy inside ssl_rsa.c.
18101 *Ralf S. Engelschall*
18103 * Move s_server -dcert and -dkey options out of the undocumented feature
18104 area because they are useful for the DSA situation and should be
18105 recognized by the users.
18107 *Ralf S. Engelschall*
18109 * Fix the cipher decision scheme for export ciphers: the export bits are
18110 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18111 SSL_EXP_MASK. So, the original variable has to be used instead of the
18112 already masked variable.
18114 *Richard Levitte <levitte@stacken.kth.se>*
18116 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18118 *Richard Levitte <levitte@stacken.kth.se>*
18120 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18121 from `int` to `unsigned int` because it is a length and initialized by
18122 EVP_DigestFinal() which expects an `unsigned int *`.
18124 *Richard Levitte <levitte@stacken.kth.se>*
18126 * Don't hard-code path to Perl interpreter on shebang line of Configure
18127 script. Instead use the usual Shell->Perl transition trick.
18129 *Ralf S. Engelschall*
18131 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18132 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18133 -noout -modulus` as it's already the case for `openssl rsa -noout
18134 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18135 currently the public key is printed (a decision which was already done by
18136 `openssl dsa -modulus` in the past) which serves a similar purpose.
18137 Additionally the NO_RSA no longer completely removes the whole -modulus
18138 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18141 *Ralf S. Engelschall*
18143 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18144 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18146 *Arne Ansper <arne@ats.cyber.ee>*
18148 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18149 to be added. Now both 'req' and 'ca' can use new objects defined in the
18154 * Add cool BIO that does syslog (or event log on NT).
18156 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18158 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18159 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18160 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18161 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18165 * Add preliminary config info for new extension code.
18169 * Make RSA_NO_PADDING really use no padding.
18171 *Ulf Moeller <ulf@fitug.de>*
18173 * Generate errors when private/public key check is done.
18177 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18178 for some CRL extensions and new objects added.
18182 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18183 key usage extension and fuller support for authority key id.
18187 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18188 padding method for RSA, which is recommended for new applications in PKCS
18189 #1 v2.0 (RFC 2437, October 1998).
18190 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18191 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18192 against Bleichbacher's attack on RSA.
18193 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18196 * Updates to the new SSL compression code
18198 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18200 * Fix so that the version number in the master secret, when passed
18201 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18202 (because the server will not accept higher), that the version number
18203 is 0x03,0x01, not 0x03,0x00
18205 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18207 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18208 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18209 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18213 * Support for RAW extensions where an arbitrary extension can be
18214 created by including its DER encoding. See `apps/openssl.cnf` for
18219 * Make sure latest Perl versions don't interpret some generated C array
18220 code as Perl array code in the crypto/err/err_genc.pl script.
18222 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18224 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18225 not many people have the assembler. Various Win32 compilation fixes and
18226 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18227 build instructions.
18231 * Modify configure script 'Configure' to automatically create crypto/date.h
18232 file under Win32 and also build pem.h from pem.org. New script
18233 util/mkfiles.pl to create the MINFO file on environments that can't do a
18234 'make files': perl util/mkfiles.pl >MINFO should work.
18238 * Major rework of DES function declarations, in the pursuit of correctness
18239 and purity. As a result, many evil casts evaporated, and some weirdness,
18240 too. You may find this causes warnings in your code. Zapping your evil
18241 casts will probably fix them. Mostly.
18245 * Fix for a typo in asn1.h. Bug fix to object creation script
18246 obj_dat.pl. It considered a zero in an object definition to mean
18247 "end of object": none of the objects in objects.h have any zeros
18248 so it wasn't spotted.
18250 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18252 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18253 Masking (CBCM). In the absence of test vectors, the best I have been able
18254 to do is check that the decrypt undoes the encrypt, so far. Send me test
18255 vectors if you have them.
18259 * Correct calculation of key length for export ciphers (too much space was
18260 allocated for null ciphers). This has not been tested!
18264 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18265 message is now correct (it understands "crypto" and "ssl" on its
18266 command line). There is also now an "update" option. This will update
18267 the util/ssleay.num and util/libeay.num files with any new functions.
18269 perl util/mkdef.pl crypto ssl update
18270 it will update them.
18274 * Overhauled the Perl interface:
18275 - ported BN stuff to OpenSSL's different BN library
18276 - made the perl/ source tree CVS-aware
18277 - renamed the package from SSLeay to OpenSSL (the files still contain
18278 their history because I've copied them in the repository)
18279 - removed obsolete files (the test scripts will be replaced
18280 by better Test::Harness variants in the future)
18282 *Ralf S. Engelschall*
18284 * First cut for a very conservative source tree cleanup:
18285 1. merge various obsolete readme texts into doc/ssleay.txt
18286 where we collect the old documents and readme texts.
18287 2. remove the first part of files where I'm already sure that we no
18288 longer need them because of three reasons: either they are just temporary
18289 files which were left by Eric or they are preserved original files where
18290 I've verified that the diff is also available in the CVS via "cvs diff
18291 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18292 the crypto/md/ stuff).
18294 *Ralf S. Engelschall*
18296 * More extension code. Incomplete support for subject and issuer alt
18297 name, issuer and authority key id. Change the i2v function parameters
18298 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18299 what that's for :-) Fix to ASN1 macro which messed up
18300 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18304 * Preliminary support for ENUMERATED type. This is largely copied from the
18309 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18311 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18313 * Make sure `make rehash` target really finds the `openssl` program.
18315 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18317 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18318 like to hear about it if this slows down other processors.
18322 * Add CygWin32 platform information to Configure script.
18324 *Alan Batie <batie@aahz.jf.intel.com>*
18326 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18328 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18330 * New program nseq to manipulate netscape certificate sequences
18334 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18339 * Fixes to BN code. Previously the default was to define BN_RECURSION
18340 but the BN code had some problems that would cause failures when
18341 doing certificate verification and some other functions.
18343 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18345 * Add ASN1 and PEM code to support netscape certificate sequences.
18349 * Add ASN1 and PEM code to support netscape certificate sequences.
18353 * Add several PKIX and private extended key usage OIDs.
18357 * Modify the 'ca' program to handle the new extension code. Modify
18358 openssl.cnf for new extension format, add comments.
18362 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18363 and add a sample to openssl.cnf so req -x509 now adds appropriate
18368 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18369 error code, add initial support to X509_print() and x509 application.
18373 * Takes a deep breath and start adding X509 V3 extension support code. Add
18374 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18375 stuff is currently isolated and isn't even compiled yet.
18379 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18380 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18381 Removed the versions check from X509 routines when loading extensions:
18382 this allows certain broken certificates that don't set the version
18383 properly to be processed.
18387 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18388 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18389 can still be regenerated with "make depend".
18393 * Spelling mistake in C version of CAST-128.
18395 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18397 * Changes to the error generation code. The perl script err-code.pl
18398 now reads in the old error codes and retains the old numbers, only
18399 adding new ones if necessary. It also only changes the .err files if new
18400 codes are added. The makefiles have been modified to only insert errors
18401 when needed (to avoid needlessly modifying header files). This is done
18402 by only inserting errors if the .err file is newer than the auto generated
18403 C file. To rebuild all the error codes from scratch (the old behaviour)
18404 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18405 or delete all the .err files.
18409 * CAST-128 was incorrectly implemented for short keys. The C version has
18410 been fixed, but is untested. The assembler versions are also fixed, but
18411 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18412 to regenerate it if needed.
18413 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18414 Hagino <itojun@kame.net>*
18416 * File was opened incorrectly in randfile.c.
18418 *Ulf Möller <ulf@fitug.de>*
18420 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18421 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18422 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18423 al: it's just almost always a UTCTime. Note this patch adds new error
18424 codes so do a "make errors" if there are problems.
18428 * Correct Linux 1 recognition in config.
18430 *Ulf Möller <ulf@fitug.de>*
18432 * Remove pointless MD5 hash when using DSA keys in ca.
18434 *Anonymous <nobody@replay.com>*
18436 * Generate an error if given an empty string as a cert directory. Also
18437 generate an error if handed NULL (previously returned 0 to indicate an
18438 error, but didn't set one).
18440 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18442 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18446 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18447 parameters. This was causing a warning which killed off the Win32 compile.
18451 * Remove C++ style comments from crypto/bn/bn_local.h.
18453 *Neil Costigan <neil.costigan@celocom.com>*
18455 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18456 based on a text string, looking up short and long names and finally
18457 "dot" format. The "dot" format stuff didn't work. Added new function
18458 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18459 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18460 OID is not part of the table.
18464 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18465 X509_LOOKUP_by_alias().
18469 * Sort openssl functions by name.
18473 * Get the `gendsa` command working and add it to the `list` command. Remove
18474 encryption from sample DSA keys (in case anyone is interested the password
18479 * Make *all* `*_free` functions accept a NULL pointer.
18481 *Frans Heymans <fheymans@isaserver.be>*
18483 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18486 *Anonymous <nobody@replay.com>*
18488 * s_server should send the CAfile as acceptable CAs, not its own cert.
18490 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18492 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18494 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18496 * Temp key "for export" tests were wrong in s3_srvr.c.
18498 *Anonymous <nobody@replay.com>*
18500 * Add prototype for temp key callback functions
18501 SSL_CTX_set_tmp_{rsa,dh}_callback().
18505 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18506 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18510 * X509_name_add_entry() freed the wrong thing after an error.
18512 *Arne Ansper <arne@ats.cyber.ee>*
18514 * rsa_eay.c would attempt to free a NULL context.
18516 *Arne Ansper <arne@ats.cyber.ee>*
18518 * BIO_s_socket() had a broken should_retry() on Windoze.
18520 *Arne Ansper <arne@ats.cyber.ee>*
18522 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18524 *Arne Ansper <arne@ats.cyber.ee>*
18526 * Make sure the already existing X509_STORE->depth variable is initialized
18527 in X509_STORE_new(), but document the fact that this variable is still
18528 unused in the certificate verification process.
18530 *Ralf S. Engelschall*
18532 * Fix the various library and `apps/` files to free up pkeys obtained from
18533 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18537 * Fix reference counting in X509_PUBKEY_get(). This makes
18538 demos/maurice/example2.c work, amongst others, probably.
18540 *Steve Henson and Ben Laurie*
18542 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18543 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18544 are no longer created. This way we have a single and consistent command
18545 line interface `openssl <command>`, similar to `cvs <command>`.
18547 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18549 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18550 BIT STRING wrapper always have zero unused bits.
18554 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18558 * Make the top-level INSTALL documentation easier to understand.
18562 * Makefiles updated to exit if an error occurs in a sub-directory
18563 make (including if user presses ^C) [Paul Sutton]
18565 * Make Montgomery context stuff explicit in RSA data structure.
18569 * Fix build order of pem and err to allow for generated pem.h.
18573 * Fix renumbering bug in X509_NAME_delete_entry().
18577 * Enhanced the err-ins.pl script so it makes the error library number
18578 global and can add a library name. This is needed for external ASN1 and
18579 other error libraries.
18583 * Fixed sk_insert which never worked properly.
18587 * Fix ASN1 macros so they can handle indefinite length constructed
18588 EXPLICIT tags. Some non standard certificates use these: they can now
18593 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18594 into a single doc/ssleay.txt bundle. This way the information is still
18595 preserved but no longer messes up this directory. Now it's new room for
18596 the new set of documentation files.
18598 *Ralf S. Engelschall*
18600 * SETs were incorrectly DER encoded. This was a major pain, because they
18601 shared code with SEQUENCEs, which aren't coded the same. This means that
18602 almost everything to do with SETs or SEQUENCEs has either changed name or
18603 number of arguments.
18605 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18607 * Fix test data to work with the above.
18611 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18612 was already fixed by Eric for 0.9.1 it seems.
18614 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18616 * Autodetect FreeBSD3.
18620 * Fix various bugs in Configure. This affects the following platforms:
18624 unixware-2.0-pentium
18629 * Eliminate generated files from CVS. Reorder tests to regenerate files
18630 before they are needed.
18634 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18638 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18640 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18641 changed SSLeay to OpenSSL in version strings.
18643 *Ralf S. Engelschall*
18645 * Some fixups to the top-level documents.
18649 * Fixed the nasty bug where rsaref.h was not found under compile-time
18650 because the symlink to include/ was missing.
18652 *Ralf S. Engelschall*
18654 * Incorporated the popular no-RSA/DSA-only patches
18655 which allow to compile a RSA-free SSLeay.
18657 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18659 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18660 when "ssleay" is still not found.
18662 *Ralf S. Engelschall*
18664 * Added more platforms to Configure: Cray T3E, HPUX 11,
18666 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18668 * Updated the README file.
18670 *Ralf S. Engelschall*
18672 * Added various .cvsignore files in the CVS repository subdirs
18673 to make a "cvs update" really silent.
18675 *Ralf S. Engelschall*
18677 * Recompiled the error-definition header files and added
18678 missing symbols to the Win32 linker tables.
18680 *Ralf S. Engelschall*
18682 * Cleaned up the top-level documents;
18683 o new files: CHANGES and LICENSE
18684 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18685 o merged COPYRIGHT into LICENSE
18686 o removed obsolete TODO file
18687 o renamed MICROSOFT to INSTALL.W32
18689 *Ralf S. Engelschall*
18691 * Removed dummy files from the 0.9.1b source tree:
18692 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18693 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18694 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18695 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18696 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18698 *Ralf S. Engelschall*
18700 * Added various platform portability fixes.
18704 * The Genesis of the OpenSSL rpject:
18705 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18706 Young and Tim J. Hudson created while they were working for C2Net until
18709 *The OpenSSL Project*
18711 ### Changes between 0.9.0b and 0.9.1b [not released]
18713 * Updated a few CA certificates under certs/
18717 * Changed some BIGNUM api stuff.
18721 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18722 DGUX x86, Linux Alpha, etc.
18726 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18727 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18732 * Add -strparse option to asn1pars program which parses nested
18735 *Dr Stephen Henson <shenson@bigfoot.com>*
18737 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18741 * DSA fix for "ca" program.
18745 * Added "-genkey" option to "dsaparam" program.
18749 * Added RIPE MD160 (rmd160) message digest.
18753 * Added -a (all) option to "ssleay version" command.
18757 * Added PLATFORM define which is the id given to Configure.
18761 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18765 * Extended the ASN.1 parser routines.
18769 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18773 * Added a BN_CTX to the BN library.
18777 * Fixed the weak key values in DES library
18781 * Changed API in EVP library for cipher aliases.
18785 * Added support for RC2/64bit cipher.
18789 * Converted the lhash library to the crypto/mem.c functions.
18793 * Added more recognized ASN.1 object ids.
18797 * Added more RSA padding checks for SSL/TLS.
18801 * Added BIO proxy/filter functionality.
18805 * Added extra_certs to SSL_CTX which can be used
18806 send extra CA certificates to the client in the CA cert chain sending
18807 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18811 * Now Fortezza is denied in the authentication phase because
18812 this is key exchange mechanism is not supported by SSLeay at all.
18816 * Additional PKCS1 checks.
18820 * Support the string "TLSv1" for all TLS v1 ciphers.
18824 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18825 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18829 * Fixed a few memory leaks.
18833 * Fixed various code and comment typos.
18837 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18838 bytes sent in the client random.
18840 *Edward Bishop <ebishop@spyglass.com>*
18844 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18845 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18846 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18847 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18848 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18849 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18850 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18851 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18852 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18853 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18854 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18855 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18856 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18857 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18858 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18859 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18860 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18861 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18862 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18863 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18864 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18865 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18866 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18867 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18868 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18869 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18870 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18871 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18872 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18873 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18874 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18875 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18876 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18877 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18878 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18879 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18880 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18881 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18882 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18883 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18884 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18885 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18886 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18887 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18888 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18889 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18890 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18891 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18892 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18893 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18894 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18895 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18896 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18897 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18898 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18899 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18900 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18901 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18902 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18903 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18904 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18905 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18906 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18907 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18908 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18909 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18910 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18911 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18912 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18913 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18914 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18915 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18916 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18917 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18918 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18919 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18920 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18921 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18922 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18923 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18924 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18925 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18926 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18927 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18928 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18929 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18930 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18931 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18932 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18933 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18934 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18935 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18936 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18937 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18938 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18939 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18940 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18941 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18942 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18943 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18944 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18945 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18946 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18947 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18948 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18949 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18950 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18951 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18952 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18953 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18954 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18955 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18956 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18957 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18958 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18959 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18960 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18961 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18962 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18963 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18964 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18965 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18966 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18967 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18968 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18969 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18970 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18971 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18972 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18973 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18974 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18975 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18976 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18977 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18978 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18979 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18980 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18981 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18982 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18983 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18984 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18985 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18986 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18987 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18988 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18989 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18990 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18991 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18992 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18993 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18994 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18995 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18996 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18997 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18998 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18999 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19000 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19001 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19002 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19003 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19004 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19005 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655