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 3.0.0 and 3.0.1 [xx XXX xxxx]
35 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
37 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
42 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
43 S390X capability vector to zero. This simplifies testing of different code
44 paths on S390X architecture.
48 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
49 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
50 SP 800-38D". The communication will fail at this point.
54 * The EC_GROUP_clear_free() function is deprecated as there is nothing
55 confidential in EC_GROUP data.
59 * The byte order mark (BOM) character is ignored if encountered at the
60 beginning of a PEM-formatted file.
64 * Added CMS support for the Russian GOST algorithms.
68 * Due to move of the implementation of cryptographic operations
69 to the providers, validation of various operation parameters can
70 be postponed until the actual operation is executed where previously
71 it happened immediately when an operation parameter was set.
73 For example when setting an unsupported curve with
74 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
75 fail but later keygen operations with the EVP_PKEY_CTX will fail.
77 *OpenSSL team members and many third party contributors*
79 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
80 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
81 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
82 instead to retrieve these algorithms from a provider.
86 * On build targets where the multilib postfix is set in the build
87 configuration the libdir directory was changing based on whether
88 the lib directory with the multilib postfix exists on the system
89 or not. This unpredictable behavior was removed and eventual
90 multilib postfix is now always added to the default libdir. Use
91 `--libdir=lib` to override the libdir if adding the postfix is
96 * The triple DES key wrap functionality now conforms to RFC 3217 but is
97 no longer interoperable with OpenSSL 1.1.1.
101 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
102 function codes, this function can only cause problems for calling
107 * Add a configurable flag to output date formats as ISO 8601. Does not
108 change the default date format.
112 * Version of MSVC earlier than 1300 could get link warnings, which could
113 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
114 Support for this flag has been removed.
118 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
119 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
120 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
121 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
122 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
126 * The signatures of the functions to get and set options on SSL and
127 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
128 Some source code changes may be required.
132 * The public definitions of conf_method_st and conf_st have been
133 deprecated. They will be made opaque in a future release.
135 *Rich Salz and Tomáš Mráz*
137 * Client-initiated renegotiation is disabled by default. To allow it, use
138 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
139 flag, or the "ClientRenegotiation" config parameter as appropriate.
143 * Add "abspath" and "includedir" pragma's to config files, to prevent,
144 or modify relative pathname inclusion.
148 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
149 validated. Please consult the README-FIPS and
150 README-PROVIDERS files, as well as the migration guide.
152 *OpenSSL team members and many third party contributors*
154 * For the key types DH and DHX the allowed settable parameters are now different.
158 * The openssl commands that read keys, certificates, and CRLs now
159 automatically detect the PEM or DER format of the input files.
161 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
163 * Added enhanced PKCS#12 APIs which accept a library context.
167 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
171 * Added support for Kernel TLS (KTLS).
173 *Boris Pismenny, John Baldwin and Andrew Gallatin*
175 * Support for RFC 5746 secure renegotiation is now required by default for
176 SSL or TLS connections to succeed.
180 * The signature of the `copy` functional parameter of the
181 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
182 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
183 the signature of the `pub_decode` functional parameter of the
184 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
185 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
189 * The error return values from some control calls (ctrl) have changed.
193 * A public key check is now performed during EVP_PKEY_derive_set_peer().
197 * Many functions in the EVP_ namespace that are getters of values from
198 implementations or contexts were renamed to include get or get0 in their
199 names. Old names are provided as macro aliases for compatibility and
204 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
205 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
206 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
211 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
214 * The output from the command line applications may have minor
219 * The output from numerous "printing" may have minor changes.
223 * Windows thread synchronization uses read/write primitives (SRWLock) when
224 supported by the OS, otherwise CriticalSection continues to be used.
228 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
229 work on read only BIO source/sinks that do not support these functions.
230 This allows piping or redirection of a file BIO using stdin to be buffered
231 into memory. This is used internally in OSSL_DECODER_from_bio().
235 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
236 this function would return one of the values OSSL_STORE_INFO_NAME,
237 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
238 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
239 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
240 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
241 using this function should be amended to handle the changed return value.
245 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
246 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
247 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
248 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
249 contains more than one certificate identifier: This means that all
250 certificates referenced there MUST be part of the validation chain.
254 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
255 RC5, DESX and DES have been moved to the legacy provider.
259 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
260 RIPEMD-160 have been moved to the legacy provider.
264 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
269 * The deprecated functions EVP_PKEY_get0_RSA(),
270 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
271 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
272 well as the similarly named "get1" functions behave differently in
277 * A number of functions handling low-level keys or engines were deprecated
278 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
279 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
280 EVP_PKEY_get0_siphash().
284 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
285 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
286 will need to load the legacy crypto provider. This includes these PBE
287 algorithms which use this KDF:
288 - NID_pbeWithMD2AndDES_CBC
289 - NID_pbeWithMD5AndDES_CBC
290 - NID_pbeWithSHA1AndRC2_CBC
291 - NID_pbeWithMD2AndRC2_CBC
292 - NID_pbeWithMD5AndRC2_CBC
293 - NID_pbeWithSHA1AndDES_CBC
297 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
298 BIO_debug_callback() functions.
302 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
303 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
307 * The RAND_METHOD APIs have been deprecated.
311 * The SRP APIs have been deprecated.
315 * Add a compile time option to prevent the caching of provider fetched
316 algorithms. This is enabled by including the no-cached-fetch option
317 at configuration time.
321 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
322 count of PKCS12_DEFAULT_ITER.
324 *Tomáš Mráz and Sahana Prasad*
326 * The openssl speed command does not use low-level API calls anymore.
330 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
333 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
335 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
339 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
340 providers may supply their own group implementations (using either the "key
341 exchange" or the "key encapsulation" methods) which will automatically be
342 detected and used by libssl.
344 *Matt Caswell, Nicola Tuveri*
346 * The undocumented function X509_certificate_type() has been deprecated;
350 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
354 * Removed RSA padding mode for SSLv23 (which was only used for
355 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
356 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
361 * Deprecated the obsolete X9.31 RSA key generation related functions.
363 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
364 is not allowed to return a value > 1, this is no more taken as failure.
366 *Viktor Dukhovni and David von Oheimb*
368 * Deprecated the obsolete X9.31 RSA key generation related functions
369 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
370 BN_X931_generate_prime_ex().
374 * The default key generation method for the regular 2-prime RSA keys was
375 changed to the FIPS 186-4 B.3.6 method.
379 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
383 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
387 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
388 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
390 *Rich Salz, Richard Levitte, and David von Oheimb*
392 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
396 * Deprecated `OCSP_parse_url()`.
400 * Validation of SM2 keys has been separated from the validation of regular EC
405 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
406 switches: a validation failure triggers an early exit, returning a failure
407 exit status to the parent process.
411 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
412 to ignore unknown ciphers.
416 * The `-cipher-commands` and `-digest-commands` options
417 of the command line utility `list` have been deprecated.
418 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
422 * Added convenience functions for generating asymmetric key pairs:
423 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
424 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
428 * All of the low level EC_KEY functions have been deprecated.
430 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
432 * Deprecated all the libcrypto and libssl error string loading
437 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
438 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
443 * The `-crypt` option to the `passwd` command line tool has been removed.
447 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
452 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
456 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
457 EVP_PKEY_get1_tls_encodedpoint().
461 * The security callback, which can be customised by application code, supports
462 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
463 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
467 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
468 interface. Their functionality remains unchanged.
472 * Added new option for 'openssl list', '-providers', which will display the
473 list of loaded providers, their names, version and status. It optionally
474 displays their gettable parameters.
478 * Removed EVP_PKEY_set_alias_type().
482 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
483 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
487 * Changed all "STACK" functions to be macros instead of inline functions. Macro
488 parameters are still checked for type safety at compile time via helper
493 * Remove the RAND_DRBG API
495 *Paul Dale and Matthias St. Pierre*
497 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
498 as well as actual hostnames.
502 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
503 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
504 conversely, silently ignore DTLS protocol version bounds when configuring
505 TLS-based contexts. The commands can be repeated to set bounds of both
506 types. The same applies with the corresponding "min_protocol" and
507 "max_protocol" command-line switches, in case some application uses both TLS
510 SSL_CTX instances that are created for a fixed protocol version (e.g.
511 `TLSv1_server_method()`) also silently ignore version bounds. Previously
512 attempts to apply bounds to these protocol versions would result in an
513 error. Now only the "version-flexible" SSL_CTX instances are subject to
514 limits in configuration files in command-line options.
518 * Deprecated the `ENGINE` API. Engines should be replaced with providers
523 * Reworked the recorded ERR codes to make better space for system errors.
524 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
525 given code is a system error (true) or an OpenSSL error (false).
529 * Reworked the test perl framework to better allow parallel testing.
531 *Nicola Tuveri and David von Oheimb*
533 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
534 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
538 * 'Configure' has been changed to figure out the configuration target if
539 none is given on the command line. Consequently, the 'config' script is
540 now only a mere wrapper. All documentation is changed to only mention
543 *Rich Salz and Richard Levitte*
545 * Added a library context `OSSL_LIB_CTX` that applications as well as
546 other libraries can use to form a separate context within which
547 libcrypto operations are performed.
551 * Added various `_ex` functions to the OpenSSL API that support using
552 a non-default `OSSL_LIB_CTX`.
556 * Handshake now fails if Extended Master Secret extension is dropped
561 * Dropped interactive mode from the `openssl` program.
565 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
567 *David von Oheimb and Shane Lontis*
569 * Deprecated `EC_METHOD_get_field_type()`.
573 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
574 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
575 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
579 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
583 * Add CAdES-BES signature verification support, mostly derived
584 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
586 *Filipe Raimundo da Silva*
588 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
592 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
593 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
597 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
601 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
602 EC_KEY_precompute_mult().
606 * Deprecated EC_POINTs_mul().
610 * Removed FIPS_mode() and FIPS_mode_set().
614 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
618 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
619 EC_POINT_get_Jprojective_coordinates_GFp().
623 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
624 arrays to be more easily constructed via a series of utility functions.
625 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
626 the various push functions and finally convert to a passable OSSL_PARAM
627 array using OSSL_PARAM_BLD_to_param().
631 * The security strength of SHA1 and MD5 based signatures in TLS has been
636 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
637 contain a provider side internal key.
641 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
645 * Project text documents not yet having a proper file name extension
646 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
647 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
651 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
652 have been converted to Markdown with the goal to produce documents
653 which not only look pretty when viewed online in the browser, but
654 remain well readable inside a plain text editor.
656 To achieve this goal, a 'minimalistic' Markdown style has been applied
657 which avoids formatting elements that interfere too much with the
658 reading flow in the text file. For example, it
660 * avoids [ATX headings][] and uses [setext headings][] instead
661 (which works for `<h1>` and `<h2>` headings only).
662 * avoids [inline links][] and uses [reference links][] instead.
663 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
665 [ATX headings]: https://github.github.com/gfm/#atx-headings
666 [setext headings]: https://github.github.com/gfm/#setext-headings
667 [inline links]: https://github.github.com/gfm/#inline-link
668 [reference links]: https://github.github.com/gfm/#reference-link
669 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
670 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
672 *Matthias St. Pierre*
674 * The test suite is changed to preserve results of each test recipe.
675 A new directory test-runs/ with subdirectories named like the
676 test recipes are created in the build tree for this purpose.
680 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
681 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
682 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
684 *David von Oheimb, Martin Peylo*
686 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
687 It supports arbitrary request and response content types, GET redirection,
688 TLS, connections via HTTP(S) proxies, connections and exchange via
689 user-defined BIOs (allowing implicit connections), persistent connections,
690 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
691 The legacy OCSP-focused (and only partly documented) API
692 is retained for backward compatibility, while most of it is deprecated.
696 * Added `util/check-format.pl`, a tool for checking adherence to the
697 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
698 The checks performed are incomplete and yield some false positives.
699 Still the tool should be useful for detecting most typical glitches.
703 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
704 If domain name resolution yields multiple IP addresses all of them are tried
705 after `connect()` failures.
709 * All of the low level RSA functions have been deprecated.
713 * X509 certificates signed using SHA1 are no longer allowed at security
718 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
719 modified to use PKEY APIs. These commands are now in maintenance mode
720 and no new features will be added to them.
724 * The command line utility rsautl has been deprecated.
728 * The command line utilities genrsa and rsa have been modified to use PKEY
729 APIs. They now write PKCS#8 keys by default. These commands are now in
730 maintenance mode and no new features will be added to them.
734 * All of the low level DH functions have been deprecated.
736 *Paul Dale and Matt Caswell*
738 * All of the low level DSA functions have been deprecated.
742 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
743 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
747 * Deprecated low level ECDH and ECDSA functions.
751 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
755 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
756 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
757 a new formulation to include all the things it can be used for,
758 as well as words of caution.
762 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
766 * All of the low level HMAC functions have been deprecated.
768 *Paul Dale and David von Oheimb*
770 * Over two thousand fixes were made to the documentation, including:
771 - Common options (such as -rand/-writerand, TLS version control, etc)
772 were refactored and point to newly-enhanced descriptions in openssl.pod.
773 - Added style conformance for all options (with help from Richard Levitte),
774 documented all reported missing options, added a CI build to check
775 that all options are documented and that no unimplemented options
777 - Documented some internals, such as all use of environment variables.
778 - Addressed all internal broken L<> references.
782 * All of the low level CMAC functions have been deprecated.
786 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
787 functions have been deprecated.
789 *Paul Dale and David von Oheimb*
791 * Corrected the documentation of the return values from the `EVP_DigestSign*`
792 set of functions. The documentation mentioned negative values for some
793 errors, but this was never the case, so the mention of negative values
796 Code that followed the documentation and thereby check with something
797 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
801 * All of the low level cipher functions have been deprecated.
803 *Matt Caswell and Paul Dale*
805 * Removed include/openssl/opensslconf.h.in and replaced it with
806 include/openssl/configuration.h.in, which differs in not including
807 <openssl/macros.h>. A short header include/openssl/opensslconf.h
808 was added to include both.
810 This allows internal hacks where one might need to modify the set
811 of configured macros, for example this if deprecated symbols are
812 still supposed to be available internally:
814 #include <openssl/configuration.h>
816 #undef OPENSSL_NO_DEPRECATED
817 #define OPENSSL_SUPPRESS_DEPRECATED
819 #include <openssl/macros.h>
821 This should not be used by applications that use the exported
822 symbols, as that will lead to linking errors.
826 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
827 used in exponentiation with 512-bit moduli. No EC algorithms are
828 affected. Analysis suggests that attacks against 2-prime RSA1024,
829 3-prime RSA1536, and DSA1024 as a result of this defect would be very
830 difficult to perform and are not believed likely. Attacks against DH512
831 are considered just feasible. However, for an attack the target would
832 have to re-use the DH512 private key, which is not recommended anyway.
833 Also applications directly using the low-level API BN_mod_exp may be
834 affected if they use BN_FLG_CONSTTIME.
839 * Most memory-debug features have been deprecated, and the functionality
840 replaced with no-ops.
844 * Added documentation for the STACK API.
848 * Introduced a new method type and API, OSSL_ENCODER, to represent
849 generic encoders. These do the same sort of job that PEM writers
850 and d2i functions do, but with support for methods supplied by
851 providers, and the possibility for providers to support other
856 * Introduced a new method type and API, OSSL_DECODER, to represent
857 generic decoders. These do the same sort of job that PEM readers
858 and i2d functions do, but with support for methods supplied by
859 providers, and the possibility for providers to support other
864 * Added a .pragma directive to the syntax of configuration files, to
865 allow varying behavior in a supported and predictable manner.
866 Currently added pragma:
870 This allows dollar signs to be a keyword character unless it's
871 followed by a opening brace or parenthesis. This is useful for
872 platforms where dollar signs are commonly used in names, such as
873 volume names and system directory names on VMS.
877 * Added functionality to create an EVP_PKEY from user data.
881 * Change the interpretation of the '--api' configuration option to
882 mean that this is a desired API compatibility level with no
883 further meaning. The previous interpretation, that this would
884 also mean to remove all deprecated symbols up to and including
885 the given version, no requires that 'no-deprecated' is also used
886 in the configuration.
888 When building applications, the desired API compatibility level
889 can be set with the OPENSSL_API_COMPAT macro like before. For
890 API compatibility version below 3.0, the old style numerical
891 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
892 For version 3.0 and on, the value is expected to be the decimal
893 value calculated from the major and minor version like this:
895 MAJOR * 10000 + MINOR * 100
899 -DOPENSSL_API_COMPAT=30000 For 3.0
900 -DOPENSSL_API_COMPAT=30200 For 3.2
902 To hide declarations that are deprecated up to and including the
903 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
904 given when building the application as well.
908 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
909 access to certificate and CRL stores via URIs and OSSL_STORE
912 This adds the following functions:
914 - X509_LOOKUP_store()
915 - X509_STORE_load_file()
916 - X509_STORE_load_path()
917 - X509_STORE_load_store()
918 - SSL_add_store_cert_subjects_to_stack()
919 - SSL_CTX_set_default_verify_store()
920 - SSL_CTX_load_verify_file()
921 - SSL_CTX_load_verify_dir()
922 - SSL_CTX_load_verify_store()
926 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
927 The presence of this system service is determined at run-time.
931 * Added functionality to create an EVP_PKEY context based on data
932 for methods from providers. This takes an algorithm name and a
933 property query string and simply stores them, with the intent
934 that any operation that uses this context will use those strings
935 to fetch the needed methods implicitly, thereby making the port
936 of application written for pre-3.0 OpenSSL easier.
940 * The undocumented function NCONF_WIN32() has been deprecated; for
941 conversion details see the HISTORY section of doc/man5/config.pod
945 * Introduced the new functions EVP_DigestSignInit_ex() and
946 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
947 EVP_DigestVerifyUpdate() have been converted to functions. See the man
948 pages for further details.
952 * Over two thousand fixes were made to the documentation, including:
953 adding missing command flags, better style conformance, documentation
956 *Rich Salz, Richard Levitte*
958 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
959 X25519, X448, Ed25519 and Ed448.
963 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
968 * Deprecated the public definition of `ERR_STATE` as well as the function
969 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
974 * Added ERR functionality to give callers access to the stored function
975 names that have replaced the older function code based functions.
977 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
978 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
979 ERR_peek_error_all() and ERR_peek_last_error_all().
981 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
982 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
983 ERR_func_error_string().
987 * Extended testing to be verbose for failing tests only. The make variables
988 VERBOSE_FAILURE or VF can be used to enable this:
990 $ make VF=1 test # Unix
991 $ mms /macro=(VF=1) test ! OpenVMS
992 $ nmake VF=1 test # Windows
996 * Added the `-copy_extensions` option to the `x509` command for use with
997 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
998 all extensions in the request are copied to the certificate or vice versa.
1000 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1002 * Added the `-copy_extensions` option to the `req` command for use with
1003 `-x509`. When given with the `copy` or `copyall` argument,
1004 all extensions in the certification request are copied to the certificate.
1008 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1009 they generate are by default RFC 5280 compliant in the following sense:
1010 There is a subjectKeyIdentifier extension with a hash value of the public key
1011 and for not self-signed certs there is an authorityKeyIdentifier extension
1012 with a keyIdentifier field or issuer information identifying the signing key.
1013 This is done unless some configuration overrides the new default behavior,
1014 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1018 * Added several checks to `X509_verify_cert()` according to requirements in
1019 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1020 (which may be done by using the CLI option `-x509_strict`):
1021 * The basicConstraints of CA certificates must be marked critical.
1022 * CA certificates must explicitly include the keyUsage extension.
1023 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1024 * The issuer name of any certificate must not be empty.
1025 * The subject name of CA certs, certs with keyUsage crlSign,
1026 and certs without subjectAlternativeName must not be empty.
1027 * If a subjectAlternativeName extension is given it must not be empty.
1028 * The signatureAlgorithm field and the cert signature must be consistent.
1029 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1030 must not be marked critical.
1031 * The authorityKeyIdentifier must be given for X.509v3 certs
1032 unless they are self-signed.
1033 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1037 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1038 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1042 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1043 used even when parsing explicit parameters, when loading a encoded key
1044 or calling `EC_GROUP_new_from_ecpkparameters()`/
1045 `EC_GROUP_new_from_ecparameters()`.
1046 This prevents bypass of security hardening and performance gains,
1047 especially for curves with specialized EC_METHODs.
1048 By default, if a key encoded with explicit parameters is loaded and later
1049 encoded, the output is still encoded with explicit parameters, even if
1050 internally a "named" EC_GROUP is used for computation.
1054 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1055 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1056 NULL. After this change, only the cofactor parameter can be NULL. It also
1057 does some minimal sanity checks on the passed order.
1062 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1063 An attack is simple, if the first CMS_recipientInfo is valid but the
1064 second CMS_recipientInfo is chosen ciphertext. If the second
1065 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1066 encryption key will be replaced by garbage, and the message cannot be
1067 decoded, but if the RSA decryption fails, the correct encryption key is
1068 used and the recipient will not notice the attack.
1069 As a work around for this potential attack the length of the decrypted
1070 key must be equal to the cipher default key length, in case the
1071 certifiate is not given and all recipientInfo are tried out.
1072 The old behaviour can be re-enabled in the CMS code by setting the
1073 CMS_DEBUG_DECRYPT flag.
1077 * Early start up entropy quality from the DEVRANDOM seed source has been
1078 improved for older Linux systems. The RAND subsystem will wait for
1079 /dev/random to be producing output before seeding from /dev/urandom.
1080 The seeded state is stored for future library initialisations using
1081 a system global shared memory segment. The shared memory identifier
1082 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1083 the desired value. The default identifier is 114.
1087 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1088 when primes for RSA keys are computed.
1089 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1090 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1091 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1092 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1093 This avoids possible fingerprinting of newly generated RSA modules.
1097 * Correct the extended master secret constant on EBCDIC systems. Without this
1098 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1099 negotiate EMS will fail. Unfortunately this also means that TLS connections
1100 between EBCDIC systems with this fix, and EBCDIC systems without this
1101 fix will fail if they negotiate EMS.
1105 * Changed the library initialisation so that the config file is now loaded
1106 by default. This was already the case for libssl. It now occurs for both
1107 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1108 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1112 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1113 where the former acts as a replacement for `ERR_put_error()`, and the
1114 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1115 `ERR_raise_data()` adds more flexibility by taking a format string and
1116 an arbitrary number of arguments following it, to be processed with
1121 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1122 to check if a named provider is loaded and available. When called, it
1123 will also activate all fallback providers if such are still present.
1127 * Enforce a minimum DH modulus size of 512 bits.
1131 * Changed DH parameters to generate the order q subgroup instead of 2q.
1132 Previously generated DH parameters are still accepted by DH_check
1133 but DH_generate_key works around that by clearing bit 0 of the
1134 private key for those. This avoids leaking bit 0 of the private key.
1138 * Significantly reduce secure memory usage by the randomness pools.
1142 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1147 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1148 algorithms. An implementation of a key exchange algorithm can be obtained
1149 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1150 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1151 the older EVP_PKEY_derive_init() function. See the man pages for the new
1152 functions for further details.
1156 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1160 * Removed the function names from error messages and deprecated the
1165 * Removed NextStep support and the macro OPENSSL_UNISTD
1169 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1170 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1171 Also removed "export var as function" capability; we do not export
1172 variables, only functions.
1176 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1177 an error and 1 indicating success. In previous versions of OpenSSL this
1178 was a void type. If a key was set longer than the maximum possible this
1183 * Support SM2 signing and verification schemes with X509 certificate.
1187 * Use SHA256 as the default digest for TS query in the `ts` app.
1191 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1195 * Default cipher lists/suites are now available via a function, the
1196 #defines are deprecated.
1200 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1201 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1202 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1206 * Join the directories crypto/x509 and crypto/x509v3
1210 * Added command 'openssl kdf' that uses the EVP_KDF API.
1214 * Added command 'openssl mac' that uses the EVP_MAC API.
1218 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1219 as default directories. Also added the command 'openssl info'
1220 for scripting purposes.
1224 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1229 * Add prediction resistance to the DRBG reseeding process.
1233 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1234 mandated by IEEE Std 1619-2018.
1238 * Added newline escaping functionality to a filename when using openssl dgst.
1239 This output format is to replicate the output format found in the `*sum`
1240 checksum programs. This aims to preserve backward compatibility.
1242 *Matt Eaton, Richard Levitte, and Paul Dale*
1244 * Removed the heartbeat message in DTLS feature, as it has very
1245 little usage and doesn't seem to fulfill a valuable purpose.
1246 The configuration option is now deprecated.
1250 * Changed the output of 'openssl {digestname} < file' to display the
1251 digest name in its output.
1255 * Added a new generic trace API which provides support for enabling
1256 instrumentation through trace output.
1258 *Richard Levitte & Matthias St. Pierre*
1260 * Added build tests for C++. These are generated files that only do one
1261 thing, to include one public OpenSSL head file each. This tests that
1262 the public header files can be usefully included in a C++ application.
1264 This test isn't enabled by default. It can be enabled with the option
1265 'enable-buildtest-c++'.
1269 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1273 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1277 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1281 * Added KMAC to EVP_MAC.
1285 * Added property based algorithm implementation selection framework to
1290 * Added SCA hardening for modular field inversion in EC_GROUP through
1291 a new dedicated field_inv() pointer in EC_METHOD.
1292 This also addresses a leakage affecting conversions from projective
1293 to affine coordinates.
1295 *Billy Bob Brumley, Nicola Tuveri*
1297 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1298 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1299 those algorithms that were already supported through the EVP_PKEY API
1300 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1301 and scrypt are now wrappers that call EVP_KDF.
1305 * Build devcrypto engine as a dynamic engine.
1307 *Eneas U de Queiroz*
1309 * Add keyed BLAKE2 to EVP_MAC.
1313 * Fix a bug in the computation of the endpoint-pair shared secret used
1314 by DTLS over SCTP. This breaks interoperability with older versions
1315 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1316 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1317 interoperability with such broken implementations. However, enabling
1318 this switch breaks interoperability with correct implementations.
1320 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1321 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1325 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1329 * Changed the license to the Apache License v2.0.
1333 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1335 - Major releases (indicated by incrementing the MAJOR release number)
1336 may introduce incompatible API/ABI changes.
1337 - Minor releases (indicated by incrementing the MINOR release number)
1338 may introduce new features but retain API/ABI compatibility.
1339 - Patch releases (indicated by incrementing the PATCH number)
1340 are intended for bug fixes and other improvements of existing
1341 features only (like improving performance or adding documentation)
1342 and retain API/ABI compatibility.
1346 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1350 * Remove the 'dist' target and add a tarball building script. The
1351 'dist' target has fallen out of use, and it shouldn't be
1352 necessary to configure just to create a source distribution.
1356 * Recreate the OS390-Unix config target. It no longer relies on a
1357 special script like it did for OpenSSL pre-1.1.0.
1361 * Instead of having the source directories listed in Configure, add
1362 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1367 * Add GMAC to EVP_MAC.
1371 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1375 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1376 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1377 to facilitate the continued use of MACs through raw private keys in
1378 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1382 * Deprecate ECDH_KDF_X9_62().
1386 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1387 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1388 are retained for backwards compatibility.
1392 * AES-XTS mode now enforces that its two keys are different to mitigate
1393 the attacked described in "Efficient Instantiations of Tweakable
1394 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1395 Details of this attack can be obtained from:
1396 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1400 * Rename the object files, i.e. give them other names than in previous
1401 versions. Their names now include the name of the final product, as
1402 well as its type mnemonic (bin, lib, shlib).
1406 * Added new option for 'openssl list', '-objects', which will display the
1407 list of built in objects, i.e. OIDs with names.
1411 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1412 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1417 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1418 improves application performance by removing data copies and providing
1419 applications with zero-copy system calls such as sendfile and splice.
1423 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1427 * `PKCS12_parse` now maintains the order of the parsed certificates
1428 when outputting them via `*ca` (rather than reversing it).
1432 * Deprecated pthread fork support methods.
1436 * Added support for FFDHE key exchange in TLS 1.3.
1440 * Added a new concept for OpenSSL plugability: providers. This
1441 functionality is designed to replace the ENGINE API and ENGINE
1442 implementations, and to be much more dynamic, allowing provider
1443 authors to introduce new algorithms among other things, as long as
1444 there's an API that supports the algorithm type.
1446 With this concept comes a new core API for interaction between
1447 libcrypto and provider implementations. Public libcrypto functions
1448 that want to use providers do so through this core API.
1450 The main documentation for this core API is found in
1451 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1452 refer to other manuals describing the API specific for supported
1453 algorithm types (also called operations).
1460 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1462 * Fixed an SM2 Decryption Buffer Overflow.
1464 In order to decrypt SM2 encrypted data an application is expected to
1465 call the API function EVP_PKEY_decrypt(). Typically an application will
1466 call this function twice. The first time, on entry, the "out" parameter
1467 can be NULL and, on exit, the "outlen" parameter is populated with the
1468 buffer size required to hold the decrypted plaintext. The application
1469 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1470 again, but this time passing a non-NULL value for the "out" parameter.
1472 A bug in the implementation of the SM2 decryption code means that the
1473 calculation of the buffer size required to hold the plaintext returned
1474 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1475 size required by the second call. This can lead to a buffer overflow
1476 when EVP_PKEY_decrypt() is called by the application a second time with
1477 a buffer that is too small.
1479 A malicious attacker who is able present SM2 content for decryption to
1480 an application could cause attacker chosen data to overflow the buffer
1481 by up to a maximum of 62 bytes altering the contents of other data held
1482 after the buffer, possibly changing application behaviour or causing
1483 the application to crash. The location of the buffer is application
1484 dependent but is typically heap allocated.
1489 * Fixed various read buffer overruns processing ASN.1 strings
1491 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1492 structure which contains a buffer holding the string data and a field
1493 holding the buffer length. This contrasts with normal C strings which
1494 are repesented as a buffer for the string data which is terminated
1495 with a NUL (0) byte.
1497 Although not a strict requirement, ASN.1 strings that are parsed using
1498 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1499 well as any string whose value has been set with the ASN1_STRING_set()
1500 function will additionally NUL terminate the byte array in the
1501 ASN1_STRING structure.
1503 However, it is possible for applications to directly construct valid
1504 ASN1_STRING structures which do not NUL terminate the byte array by
1505 directly setting the "data" and "length" fields in the ASN1_STRING
1506 array. This can also happen by using the ASN1_STRING_set0() function.
1508 Numerous OpenSSL functions that print ASN.1 data have been found to
1509 assume that the ASN1_STRING byte array will be NUL terminated, even
1510 though this is not guaranteed for strings that have been directly
1511 constructed. Where an application requests an ASN.1 structure to be
1512 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1513 been directly constructed by the application without NUL terminating
1514 the "data" field, then a read buffer overrun can occur.
1516 The same thing can also occur during name constraints processing
1517 of certificates (for example if a certificate has been directly
1518 constructed by the application instead of loading it via the OpenSSL
1519 parsing functions, and the certificate contains non NUL terminated
1520 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1521 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1523 If a malicious actor can cause an application to directly construct an
1524 ASN1_STRING and then process it through one of the affected OpenSSL
1525 functions then this issue could be hit. This might result in a crash
1526 (causing a Denial of Service attack). It could also result in the
1527 disclosure of private memory contents (such as private keys, or
1528 sensitive plaintext).
1533 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1535 * Fixed a problem with verifying a certificate chain when using the
1536 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1537 the certificates present in a certificate chain. It is not set by default.
1539 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1540 the chain that have explicitly encoded elliptic curve parameters was added
1541 as an additional strict check.
1543 An error in the implementation of this check meant that the result of a
1544 previous check to confirm that certificates in the chain are valid CA
1545 certificates was overwritten. This effectively bypasses the check
1546 that non-CA certificates must not be able to issue other certificates.
1548 If a "purpose" has been configured then there is a subsequent opportunity
1549 for checks that the certificate is a valid CA. All of the named "purpose"
1550 values implemented in libcrypto perform this check. Therefore, where
1551 a purpose is set the certificate chain will still be rejected even when the
1552 strict flag has been used. A purpose is set by default in libssl client and
1553 server certificate verification routines, but it can be overridden or
1554 removed by an application.
1556 In order to be affected, an application must explicitly set the
1557 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1558 for the certificate verification or, in the case of TLS client or server
1559 applications, override the default purpose.
1564 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1565 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1566 renegotiation ClientHello omits the signature_algorithms extension (where it
1567 was present in the initial ClientHello), but includes a
1568 signature_algorithms_cert extension then a NULL pointer dereference will
1569 result, leading to a crash and a denial of service attack.
1571 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1572 (which is the default configuration). OpenSSL TLS clients are not impacted by
1576 *Peter Kästle and Samuel Sapalski*
1578 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1580 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1581 create a unique hash value based on the issuer and serial number data
1582 contained within an X509 certificate. However it was failing to correctly
1583 handle any errors that may occur while parsing the issuer field (which might
1584 occur if the issuer field is maliciously constructed). This may subsequently
1585 result in a NULL pointer deref and a crash leading to a potential denial of
1591 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1592 padding mode to correctly check for rollback attacks. This is considered a
1593 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1598 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1599 functions. Previously they could overflow the output length argument in some
1600 cases where the input length is close to the maximum permissable length for
1601 an integer on the platform. In such cases the return value from the function
1602 call would be 1 (indicating success), but the output length value would be
1603 negative. This could cause applications to behave incorrectly or crash.
1608 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1609 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1610 could be exploited in a side channel attack to recover the password. Since
1611 the attack is local host only this is outside of the current OpenSSL
1612 threat model and therefore no CVE is assigned.
1614 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1619 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1621 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1622 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1623 If an attacker can control both items being compared then this could lead
1624 to a possible denial of service attack. OpenSSL itself uses the
1625 GENERAL_NAME_cmp function for two purposes:
1626 1) Comparing CRL distribution point names between an available CRL and a
1627 CRL distribution point embedded in an X509 certificate
1628 2) When verifying that a timestamp response token signer matches the
1629 timestamp authority name (exposed via the API functions
1630 TS_RESP_verify_response and TS_RESP_verify_token)
1635 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1637 * Certificates with explicit curve parameters are now disallowed in
1638 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1642 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1643 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1644 conversely, silently ignore DTLS protocol version bounds when configuring
1645 TLS-based contexts. The commands can be repeated to set bounds of both
1646 types. The same applies with the corresponding "min_protocol" and
1647 "max_protocol" command-line switches, in case some application uses both TLS
1650 SSL_CTX instances that are created for a fixed protocol version (e.g.
1651 TLSv1_server_method()) also silently ignore version bounds. Previously
1652 attempts to apply bounds to these protocol versions would result in an
1653 error. Now only the "version-flexible" SSL_CTX instances are subject to
1654 limits in configuration files in command-line options.
1658 * Handshake now fails if Extended Master Secret extension is dropped
1663 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1665 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1667 * Fixed segmentation fault in SSL_check_chain()
1668 Server or client applications that call the SSL_check_chain() function
1669 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1670 dereference as a result of incorrect handling of the
1671 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1672 or unrecognised signature algorithm is received from the peer. This could
1673 be exploited by a malicious peer in a Denial of Service attack.
1678 * Added AES consttime code for no-asm configurations
1679 an optional constant time support for AES was added
1680 when building openssl for no-asm.
1681 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1682 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1683 At this time this feature is by default disabled.
1684 It will be enabled by default in 3.0.
1688 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1690 * Revert the change of EOF detection while reading in libssl to avoid
1691 regressions in applications depending on the current way of reporting
1692 the EOF. As the existing method is not fully accurate the change to
1693 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1694 branch and will be present in the 3.0 release.
1698 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1699 when primes for RSA keys are computed.
1700 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1701 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1702 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1703 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1704 This avoids possible fingerprinting of newly generated RSA modules.
1708 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1710 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1711 while reading in libssl then we would report an error back to the
1712 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1713 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1714 therefore give a hint as to what went wrong.
1718 * Check that ed25519 and ed448 are allowed by the security level. Previously
1719 signature algorithms not using an MD were not being checked that they were
1720 allowed by the security level.
1724 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1725 was not quite right. The behaviour was not consistent between resumption
1726 and normal handshakes, and also not quite consistent with historical
1727 behaviour. The behaviour in various scenarios has been clarified and
1728 it has been updated to make it match historical behaviour as closely as
1733 * *[VMS only]* The header files that the VMS compilers include automatically,
1734 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1735 that the C++ compiler doesn't understand. This is a shortcoming in the
1736 compiler, but can be worked around with `__cplusplus` guards.
1738 C++ applications that use OpenSSL libraries must be compiled using the
1739 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1740 functions. Otherwise, only functions with symbols of less than 31
1741 characters can be used, as the linker will not be able to successfully
1742 resolve symbols with longer names.
1746 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1747 The presence of this system service is determined at run-time.
1751 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1756 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1758 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1759 number generator (RNG). This was intended to include protection in the
1760 event of a fork() system call in order to ensure that the parent and child
1761 processes did not share the same RNG state. However this protection was not
1762 being used in the default case.
1764 A partial mitigation for this issue is that the output from a high
1765 precision timer is mixed into the RNG state so the likelihood of a parent
1766 and child process sharing state is significantly reduced.
1768 If an application already calls OPENSSL_init_crypto() explicitly using
1769 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1772 *Matthias St. Pierre*
1774 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1775 used even when parsing explicit parameters, when loading a encoded key
1776 or calling `EC_GROUP_new_from_ecpkparameters()`/
1777 `EC_GROUP_new_from_ecparameters()`.
1778 This prevents bypass of security hardening and performance gains,
1779 especially for curves with specialized EC_METHODs.
1780 By default, if a key encoded with explicit parameters is loaded and later
1781 encoded, the output is still encoded with explicit parameters, even if
1782 internally a "named" EC_GROUP is used for computation.
1786 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1787 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1788 NULL. After this change, only the cofactor parameter can be NULL. It also
1789 does some minimal sanity checks on the passed order.
1794 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1795 An attack is simple, if the first CMS_recipientInfo is valid but the
1796 second CMS_recipientInfo is chosen ciphertext. If the second
1797 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1798 encryption key will be replaced by garbage, and the message cannot be
1799 decoded, but if the RSA decryption fails, the correct encryption key is
1800 used and the recipient will not notice the attack.
1801 As a work around for this potential attack the length of the decrypted
1802 key must be equal to the cipher default key length, in case the
1803 certifiate is not given and all recipientInfo are tried out.
1804 The old behaviour can be re-enabled in the CMS code by setting the
1805 CMS_DEBUG_DECRYPT flag.
1810 * Early start up entropy quality from the DEVRANDOM seed source has been
1811 improved for older Linux systems. The RAND subsystem will wait for
1812 /dev/random to be producing output before seeding from /dev/urandom.
1813 The seeded state is stored for future library initialisations using
1814 a system global shared memory segment. The shared memory identifier
1815 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1816 the desired value. The default identifier is 114.
1820 * Correct the extended master secret constant on EBCDIC systems. Without this
1821 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1822 negotiate EMS will fail. Unfortunately this also means that TLS connections
1823 between EBCDIC systems with this fix, and EBCDIC systems without this
1824 fix will fail if they negotiate EMS.
1828 * Use Windows installation paths in the mingw builds
1830 Mingw isn't a POSIX environment per se, which means that Windows
1831 paths should be used for installation.
1836 * Changed DH_check to accept parameters with order q and 2q subgroups.
1837 With order 2q subgroups the bit 0 of the private key is not secret
1838 but DH_generate_key works around that by clearing bit 0 of the
1839 private key for those. This avoids leaking bit 0 of the private key.
1843 * Significantly reduce secure memory usage by the randomness pools.
1847 * Revert the DEVRANDOM_WAIT feature for Linux systems
1849 The DEVRANDOM_WAIT feature added a select() call to wait for the
1850 /dev/random device to become readable before reading from the
1851 /dev/urandom device.
1853 It turned out that this change had negative side effects on
1854 performance which were not acceptable. After some discussion it
1855 was decided to revert this feature and leave it up to the OS
1856 resp. the platform maintainer to ensure a proper initialization
1857 during early boot time.
1859 *Matthias St. Pierre*
1861 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1863 * Add build tests for C++. These are generated files that only do one
1864 thing, to include one public OpenSSL head file each. This tests that
1865 the public header files can be usefully included in a C++ application.
1867 This test isn't enabled by default. It can be enabled with the option
1868 'enable-buildtest-c++'.
1872 * Enable SHA3 pre-hashing for ECDSA and DSA.
1876 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1877 This changes the size when using the `genpkey` command when no size is given.
1878 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1879 generation commands to use 2048 bits by default.
1883 * Reorganize the manual pages to consistently have RETURN VALUES,
1884 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1885 util/fix-doc-nits accordingly.
1887 *Paul Yang, Joshua Lock*
1889 * Add the missing accessor EVP_PKEY_get0_engine()
1893 * Have commands like `s_client` and `s_server` output the signature scheme
1894 along with other cipher suite parameters when debugging.
1898 * Make OPENSSL_config() error agnostic again.
1902 * Do the error handling in RSA decryption constant time.
1906 * Prevent over long nonces in ChaCha20-Poly1305.
1908 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1909 for every encryption operation. RFC 7539 specifies that the nonce value
1910 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1911 and front pads the nonce with 0 bytes if it is less than 12
1912 bytes. However it also incorrectly allows a nonce to be set of up to 16
1913 bytes. In this case only the last 12 bytes are significant and any
1914 additional leading bytes are ignored.
1916 It is a requirement of using this cipher that nonce values are
1917 unique. Messages encrypted using a reused nonce value are susceptible to
1918 serious confidentiality and integrity attacks. If an application changes
1919 the default nonce length to be longer than 12 bytes and then makes a
1920 change to the leading bytes of the nonce expecting the new value to be a
1921 new unique nonce then such an application could inadvertently encrypt
1922 messages with a reused nonce.
1924 Additionally the ignored bytes in a long nonce are not covered by the
1925 integrity guarantee of this cipher. Any application that relies on the
1926 integrity of these ignored leading bytes of a long nonce may be further
1927 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1928 is safe because no such use sets such a long nonce value. However user
1929 applications that use this cipher directly and set a non-default nonce
1930 length to be longer than 12 bytes may be vulnerable.
1932 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1938 * Add DEVRANDOM_WAIT feature for Linux systems
1940 On older Linux systems where the getrandom() system call is not available,
1941 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1942 Contrary to getrandom(), the /dev/urandom device will not block during
1943 early boot when the kernel CSPRNG has not been seeded yet.
1945 To mitigate this known weakness, use select() to wait for /dev/random to
1946 become readable before reading from /dev/urandom.
1948 * Ensure that SM2 only uses SM3 as digest algorithm
1952 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1954 * Change the info callback signals for the start and end of a post-handshake
1955 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1956 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1957 confused by this and assume that a TLSv1.2 renegotiation has started. This
1958 can break KeyUpdate handling. Instead we no longer signal the start and end
1959 of a post handshake message exchange (although the messages themselves are
1960 still signalled). This could break some applications that were expecting
1961 the old signals. However without this KeyUpdate is not usable for many
1966 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1968 * Timing vulnerability in DSA signature generation
1970 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1971 timing side channel attack. An attacker could use variations in the signing
1972 algorithm to recover the private key.
1974 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1979 * Timing vulnerability in ECDSA signature generation
1981 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1982 timing side channel attack. An attacker could use variations in the signing
1983 algorithm to recover the private key.
1985 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1990 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1991 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1992 of two gigabytes and the error handling improved.
1994 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1995 categorized as a normal bug, not a security issue, because the DRBG reseeds
1996 automatically and is fully functional even without additional randomness
1997 provided by the application.
1999 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2001 * Add a new ClientHello callback. Provides a callback interface that gives
2002 the application the ability to adjust the nascent SSL object at the
2003 earliest stage of ClientHello processing, immediately after extensions have
2004 been collected but before they have been processed. In particular, this
2005 callback can adjust the supported TLS versions in response to the contents
2010 * Add SM2 base algorithm support.
2014 * s390x assembly pack: add (improved) hardware-support for the following
2015 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2016 aes-cfb/cfb8, aes-ecb.
2020 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2021 parameter is no longer accepted, as it leads to a corrupt table. NULL
2022 pem_str is reserved for alias entries only.
2026 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2027 step for prime curves. The new implementation is based on formulae from
2028 differential addition-and-doubling in homogeneous projective coordinates
2029 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2030 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2031 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2032 to work in projective coordinates.
2034 *Billy Bob Brumley, Nicola Tuveri*
2036 * Change generating and checking of primes so that the error rate of not
2037 being prime depends on the intended use based on the size of the input.
2038 For larger primes this will result in more rounds of Miller-Rabin.
2039 The maximal error rate for primes with more than 1080 bits is lowered
2042 *Kurt Roeckx, Annie Yousar*
2044 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2048 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2049 moving between systems, and to avoid confusion when a Windows build is
2050 done with mingw vs with MSVC. For POSIX installs, there's still a
2051 symlink or copy named 'tsget' to avoid that confusion as well.
2055 * Revert blinding in ECDSA sign and instead make problematic addition
2056 length-invariant. Switch even to fixed-length Montgomery multiplication.
2060 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2061 step for binary curves. The new implementation is based on formulae from
2062 differential addition-and-doubling in mixed Lopez-Dahab projective
2063 coordinates, modified to independently blind the operands.
2065 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2067 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2068 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2069 EC_METHODs to implement their own specialized "ladder step", to take
2070 advantage of more favorable coordinate systems or more efficient
2071 differential addition-and-doubling algorithms.
2073 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2075 * Modified the random device based seed sources to keep the relevant
2076 file descriptors open rather than reopening them on each access.
2077 This allows such sources to operate in a chroot() jail without
2078 the associated device nodes being available. This behaviour can be
2079 controlled using RAND_keep_random_devices_open().
2083 * Numerous side-channel attack mitigations have been applied. This may have
2084 performance impacts for some algorithms for the benefit of improved
2085 security. Specific changes are noted in this change log by their respective
2090 * AIX shared library support overhaul. Switch to AIX "natural" way of
2091 handling shared libraries, which means collecting shared objects of
2092 different versions and bitnesses in one common archive. This allows to
2093 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2094 doesn't affect the way 3rd party applications are linked, only how
2095 multi-version installation is managed.
2099 * Make ec_group_do_inverse_ord() more robust and available to other
2100 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2101 mitigations are applied to the fallback BN_mod_inverse().
2102 When using this function rather than BN_mod_inverse() directly, new
2103 EC cryptosystem implementations are then safer-by-default.
2107 * Add coordinate blinding for EC_POINT and implement projective
2108 coordinate blinding for generic prime curves as a countermeasure to
2109 chosen point SCA attacks.
2111 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2113 * Add blinding to ECDSA and DSA signatures to protect against side channel
2114 attacks discovered by Keegan Ryan (NCC Group).
2118 * Enforce checking in the `pkeyutl` command to ensure that the input
2119 length does not exceed the maximum supported digest length when performing
2120 a sign, verify or verifyrecover operation.
2124 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2125 I/O in combination with something like select() or poll() will hang. This
2126 can be turned off again using SSL_CTX_clear_mode().
2127 Many applications do not properly handle non-application data records, and
2128 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2129 around the problems in those applications, but can also break some.
2130 It's recommended to read the manpages about SSL_read(), SSL_write(),
2131 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2132 SSL_CTX_set_read_ahead() again.
2136 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2137 now allow empty (zero character) pass phrases.
2141 * Apply blinding to binary field modular inversion and remove patent
2142 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2146 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2147 binary and prime elliptic curves.
2151 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2152 constant time fixed point multiplication.
2156 * Revise elliptic curve scalar multiplication with timing attack
2157 defenses: ec_wNAF_mul redirects to a constant time implementation
2158 when computing fixed point and variable point multiplication (which
2159 in OpenSSL are mostly used with secret scalars in keygen, sign,
2160 ECDH derive operations).
2161 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2164 * Updated CONTRIBUTING
2168 * Updated DRBG / RAND to request nonce and additional low entropy
2169 randomness from the system.
2171 *Matthias St. Pierre*
2173 * Updated 'openssl rehash' to use OpenSSL consistent default.
2177 * Moved the load of the ssl_conf module to libcrypto, which helps
2178 loading engines that libssl uses before libssl is initialised.
2182 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2186 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2188 *Ingo Schwarze, Rich Salz*
2190 * Added output of accepting IP address and port for 'openssl s_server'
2194 * Added a new API for TLSv1.3 ciphersuites:
2195 SSL_CTX_set_ciphersuites()
2196 SSL_set_ciphersuites()
2200 * Memory allocation failures consistently add an error to the error
2205 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2206 in libcrypto when run as setuid/setgid.
2210 * Load any config file by default when libssl is used.
2214 * Added new public header file <openssl/rand_drbg.h> and documentation
2215 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2217 *Matthias St. Pierre*
2219 * QNX support removed (cannot find contributors to get their approval
2220 for the license change).
2224 * TLSv1.3 replay protection for early data has been implemented. See the
2225 SSL_read_early_data() man page for further details.
2229 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2230 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2231 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2232 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2233 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2234 configuration has been separated out. See the ciphers man page or the
2235 SSL_CTX_set_ciphersuites() man page for more information.
2239 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2240 in responder mode now supports the new "-multi" option, which
2241 spawns the specified number of child processes to handle OCSP
2242 requests. The "-timeout" option now also limits the OCSP
2243 responder's patience to wait to receive the full client request
2244 on a newly accepted connection. Child processes are respawned
2245 as needed, and the CA index file is automatically reloaded
2246 when changed. This makes it possible to run the "ocsp" responder
2247 as a long-running service, making the OpenSSL CA somewhat more
2248 feature-complete. In this mode, most diagnostic messages logged
2249 after entering the event loop are logged via syslog(3) rather than
2254 * Added support for X448 and Ed448. Heavily based on original work by
2259 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2260 objects loaded. This adds the functions OSSL_STORE_expect() and
2261 OSSL_STORE_find() as well as needed tools to construct searches and
2262 get the search data out of them.
2266 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2267 version of OpenSSL should review their configuration settings to ensure
2268 that they are still appropriate for TLSv1.3. For further information see:
2269 <https://wiki.openssl.org/index.php/TLS1.3>
2273 * Grand redesign of the OpenSSL random generator
2275 The default RAND method now utilizes an AES-CTR DRBG according to
2276 NIST standard SP 800-90Ar1. The new random generator is essentially
2277 a port of the default random generator from the OpenSSL FIPS 2.0
2278 object module. It is a hybrid deterministic random bit generator
2279 using an AES-CTR bit stream and which seeds and reseeds itself
2280 automatically using trusted system entropy sources.
2282 Some of its new features are:
2283 - Support for multiple DRBG instances with seed chaining.
2284 - The default RAND method makes use of a DRBG.
2285 - There is a public and private DRBG instance.
2286 - The DRBG instances are fork-safe.
2287 - Keep all global DRBG instances on the secure heap if it is enabled.
2288 - The public and private DRBG instance are per thread for lock free
2291 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2293 * Changed Configure so it only says what it does and doesn't dump
2294 so much data. Instead, ./configdata.pm should be used as a script
2295 to display all sorts of configuration data.
2299 * Added processing of "make variables" to Configure.
2303 * Added SHA512/224 and SHA512/256 algorithm support.
2307 * The last traces of Netware support, first removed in 1.1.0, have
2312 * Get rid of Makefile.shared, and in the process, make the processing
2313 of certain files (rc.obj, or the .def/.map/.opt files produced from
2314 the ordinal files) more visible and hopefully easier to trace and
2315 debug (or make silent).
2319 * Make it possible to have environment variable assignments as
2320 arguments to config / Configure.
2324 * Add multi-prime RSA (RFC 8017) support.
2328 * Add SM3 implemented according to GB/T 32905-2016
2329 *Jack Lloyd <jack.lloyd@ribose.com>,*
2330 *Ronald Tse <ronald.tse@ribose.com>,*
2331 *Erick Borsboom <erick.borsboom@ribose.com>*
2333 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2334 as documented in RFC6066.
2335 Based on a patch from Tomasz Moń
2337 *Filipe Raimundo da Silva*
2339 * Add SM4 implemented according to GB/T 32907-2016.
2340 *Jack Lloyd <jack.lloyd@ribose.com>,*
2341 *Ronald Tse <ronald.tse@ribose.com>,*
2342 *Erick Borsboom <erick.borsboom@ribose.com>*
2344 * Reimplement -newreq-nodes and ERR_error_string_n; the
2345 original author does not agree with the license change.
2349 * Add ARIA AEAD TLS support.
2353 * Some macro definitions to support VS6 have been removed. Visual
2354 Studio 6 has not worked since 1.1.0
2358 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2359 without clearing the errors.
2363 * Add "atfork" functions. If building on a system that without
2364 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2365 requirements. The RAND facility now uses/requires this.
2373 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2374 not possible to disable entirely. However, it's still possible to
2375 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2378 To disable, configure with 'no-ui-console'. 'no-ui' is still
2379 possible to use as an alias. Check at compile time with the
2380 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2381 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2385 * Add a STORE module, which implements a uniform and URI based reader of
2386 stores that can contain keys, certificates, CRLs and numerous other
2387 objects. The main API is loosely based on a few stdio functions,
2388 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2389 OSSL_STORE_error and OSSL_STORE_close.
2390 The implementation uses backends called "loaders" to implement arbitrary
2391 URI schemes. There is one built in "loader" for the 'file' scheme.
2395 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2396 then adjusted to work on FreeBSD 8.4 as well.
2397 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2398 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2402 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2403 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2404 error code calls like this:
2406 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2408 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2409 that can be encoded in C. For the foreseeable future, this will only
2412 *Richard Levitte and Tim Hudson*
2414 * Removed BSD cryptodev engine.
2418 * Add a build target 'build_all_generated', to build all generated files
2419 and only that. This can be used to prepare everything that requires
2420 things like perl for a system that lacks perl and then move everything
2421 to that system and do the rest of the build there.
2425 * In the UI interface, make it possible to duplicate the user data. This
2426 can be used by engines that need to retain the data for a longer time
2427 than just the call where this user data is passed.
2431 * Ignore the '-named_curve auto' value for compatibility of applications
2434 *Tomáš Mráz <tmraz@fedoraproject.org>*
2436 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2437 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2438 alerts across multiple records (some of which could be empty). In practice
2439 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2440 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2441 support this at all. Supporting it adds significant complexity to the
2442 record layer, and its removal is unlikely to cause interoperability
2447 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2448 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2449 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2454 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2455 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2457 *Richard Levitte, Andy Polyakov*
2459 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2464 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2465 platform rather than 'mingw'.
2469 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2470 success if they are asked to add an object which already exists
2471 in the store. This change cascades to other functions which load
2472 certificates and CRLs.
2476 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2477 facilitate stack unwinding even from assembly subroutines.
2481 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2482 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2486 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2487 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2488 which is the minimum version we support.
2492 * Certificate time validation (X509_cmp_time) enforces stricter
2493 compliance with RFC 5280. Fractional seconds and timezone offsets
2494 are no longer allowed.
2498 * Add support for ARIA
2502 * s_client will now send the Server Name Indication (SNI) extension by
2503 default unless the new "-noservername" option is used. The server name is
2504 based on the host provided to the "-connect" option unless overridden by
2505 using "-servername".
2509 * Add support for SipHash
2513 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2514 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2515 prevent issues where no progress is being made and the peer continually
2516 sends unrecognised record types, using up resources processing them.
2520 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2521 using the algorithm defined in
2522 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2526 * Heartbeat support has been removed; the ABI is changed for now.
2528 *Richard Levitte, Rich Salz*
2530 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2534 * The RSA "null" method, which was partially supported to avoid patent
2535 issues, has been replaced to always returns NULL.
2542 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2544 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2545 used even when parsing explicit parameters, when loading a encoded key
2546 or calling `EC_GROUP_new_from_ecpkparameters()`/
2547 `EC_GROUP_new_from_ecparameters()`.
2548 This prevents bypass of security hardening and performance gains,
2549 especially for curves with specialized EC_METHODs.
2550 By default, if a key encoded with explicit parameters is loaded and later
2551 encoded, the output is still encoded with explicit parameters, even if
2552 internally a "named" EC_GROUP is used for computation.
2556 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2557 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2558 NULL. After this change, only the cofactor parameter can be NULL. It also
2559 does some minimal sanity checks on the passed order.
2564 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2565 An attack is simple, if the first CMS_recipientInfo is valid but the
2566 second CMS_recipientInfo is chosen ciphertext. If the second
2567 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2568 encryption key will be replaced by garbage, and the message cannot be
2569 decoded, but if the RSA decryption fails, the correct encryption key is
2570 used and the recipient will not notice the attack.
2571 As a work around for this potential attack the length of the decrypted
2572 key must be equal to the cipher default key length, in case the
2573 certifiate is not given and all recipientInfo are tried out.
2574 The old behaviour can be re-enabled in the CMS code by setting the
2575 CMS_DEBUG_DECRYPT flag.
2580 * Use Windows installation paths in the mingw builds
2582 Mingw isn't a POSIX environment per se, which means that Windows
2583 paths should be used for installation.
2588 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2590 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2591 This changes the size when using the `genpkey` command when no size is given.
2592 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2593 generation commands to use 2048 bits by default.
2597 * Prevent over long nonces in ChaCha20-Poly1305.
2599 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2600 for every encryption operation. RFC 7539 specifies that the nonce value
2601 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2602 and front pads the nonce with 0 bytes if it is less than 12
2603 bytes. However it also incorrectly allows a nonce to be set of up to 16
2604 bytes. In this case only the last 12 bytes are significant and any
2605 additional leading bytes are ignored.
2607 It is a requirement of using this cipher that nonce values are
2608 unique. Messages encrypted using a reused nonce value are susceptible to
2609 serious confidentiality and integrity attacks. If an application changes
2610 the default nonce length to be longer than 12 bytes and then makes a
2611 change to the leading bytes of the nonce expecting the new value to be a
2612 new unique nonce then such an application could inadvertently encrypt
2613 messages with a reused nonce.
2615 Additionally the ignored bytes in a long nonce are not covered by the
2616 integrity guarantee of this cipher. Any application that relies on the
2617 integrity of these ignored leading bytes of a long nonce may be further
2618 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2619 is safe because no such use sets such a long nonce value. However user
2620 applications that use this cipher directly and set a non-default nonce
2621 length to be longer than 12 bytes may be vulnerable.
2623 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2629 * Added SCA hardening for modular field inversion in EC_GROUP through
2630 a new dedicated field_inv() pointer in EC_METHOD.
2631 This also addresses a leakage affecting conversions from projective
2632 to affine coordinates.
2634 *Billy Bob Brumley, Nicola Tuveri*
2636 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2637 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2641 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2645 * Remove the 'dist' target and add a tarball building script. The
2646 'dist' target has fallen out of use, and it shouldn't be
2647 necessary to configure just to create a source distribution.
2651 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2653 * Timing vulnerability in DSA signature generation
2655 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2656 timing side channel attack. An attacker could use variations in the signing
2657 algorithm to recover the private key.
2659 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2664 * Timing vulnerability in ECDSA signature generation
2666 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2667 timing side channel attack. An attacker could use variations in the signing
2668 algorithm to recover the private key.
2670 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2675 * Add coordinate blinding for EC_POINT and implement projective
2676 coordinate blinding for generic prime curves as a countermeasure to
2677 chosen point SCA attacks.
2679 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2681 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2683 * Client DoS due to large DH parameter
2685 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2686 malicious server can send a very large prime value to the client. This will
2687 cause the client to spend an unreasonably long period of time generating a
2688 key for this prime resulting in a hang until the client has finished. This
2689 could be exploited in a Denial Of Service attack.
2691 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2696 * Cache timing vulnerability in RSA Key Generation
2698 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2699 a cache timing side channel attack. An attacker with sufficient access to
2700 mount cache timing attacks during the RSA key generation process could
2701 recover the private key.
2703 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2704 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2709 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2710 parameter is no longer accepted, as it leads to a corrupt table. NULL
2711 pem_str is reserved for alias entries only.
2715 * Revert blinding in ECDSA sign and instead make problematic addition
2716 length-invariant. Switch even to fixed-length Montgomery multiplication.
2720 * Change generating and checking of primes so that the error rate of not
2721 being prime depends on the intended use based on the size of the input.
2722 For larger primes this will result in more rounds of Miller-Rabin.
2723 The maximal error rate for primes with more than 1080 bits is lowered
2726 *Kurt Roeckx, Annie Yousar*
2728 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2732 * Add blinding to ECDSA and DSA signatures to protect against side channel
2733 attacks discovered by Keegan Ryan (NCC Group).
2737 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2738 now allow empty (zero character) pass phrases.
2742 * Certificate time validation (X509_cmp_time) enforces stricter
2743 compliance with RFC 5280. Fractional seconds and timezone offsets
2744 are no longer allowed.
2748 * Fixed a text canonicalisation bug in CMS
2750 Where a CMS detached signature is used with text content the text goes
2751 through a canonicalisation process first prior to signing or verifying a
2752 signature. This process strips trailing space at the end of lines, converts
2753 line terminators to CRLF and removes additional trailing line terminators
2754 at the end of a file. A bug in the canonicalisation process meant that
2755 some characters, such as form-feed, were incorrectly treated as whitespace
2756 and removed. This is contrary to the specification (RFC5485). This fix
2757 could mean that detached text data signed with an earlier version of
2758 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2759 signed with a fixed OpenSSL may fail to verify with an earlier version of
2760 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2761 and use the "-binary" flag (for the "cms" command line application) or set
2762 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2766 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2768 * Constructed ASN.1 types with a recursive definition could exceed the stack
2770 Constructed ASN.1 types with a recursive definition (such as can be found
2771 in PKCS7) could eventually exceed the stack given malicious input with
2772 excessive recursion. This could result in a Denial Of Service attack. There
2773 are no such structures used within SSL/TLS that come from untrusted sources
2774 so this is considered safe.
2776 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2782 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2784 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2785 effectively reduced to only comparing the least significant bit of each
2786 byte. This allows an attacker to forge messages that would be considered as
2787 authenticated in an amount of tries lower than that guaranteed by the
2788 security claims of the scheme. The module can only be compiled by the
2789 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2791 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2797 * Add a build target 'build_all_generated', to build all generated files
2798 and only that. This can be used to prepare everything that requires
2799 things like perl for a system that lacks perl and then move everything
2800 to that system and do the rest of the build there.
2804 * Backport SSL_OP_NO_RENGOTIATION
2806 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2807 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2808 changes this is no longer possible in 1.1.0. Therefore the new
2809 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2810 1.1.0 to provide equivalent functionality.
2812 Note that if an application built against 1.1.0h headers (or above) is run
2813 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2814 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2818 * Removed the OS390-Unix config target. It relied on a script that doesn't
2823 * rsaz_1024_mul_avx2 overflow bug on x86_64
2825 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2826 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2827 Analysis suggests that attacks against RSA and DSA as a result of this
2828 defect would be very difficult to perform and are not believed likely.
2829 Attacks against DH1024 are considered just feasible, because most of the
2830 work necessary to deduce information about a private key may be performed
2831 offline. The amount of resources required for such an attack would be
2832 significant. However, for an attack on TLS to be meaningful, the server
2833 would have to share the DH1024 private key among multiple clients, which is
2834 no longer an option since CVE-2016-0701.
2836 This only affects processors that support the AVX2 but not ADX extensions
2837 like Intel Haswell (4th generation).
2839 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2840 was originally found via the OSS-Fuzz project.
2845 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2847 * bn_sqrx8x_internal carry bug on x86_64
2849 There is a carry propagating bug in the x86_64 Montgomery squaring
2850 procedure. No EC algorithms are affected. Analysis suggests that attacks
2851 against RSA and DSA as a result of this defect would be very difficult to
2852 perform and are not believed likely. Attacks against DH are considered just
2853 feasible (although very difficult) because most of the work necessary to
2854 deduce information about a private key may be performed offline. The amount
2855 of resources required for such an attack would be very significant and
2856 likely only accessible to a limited number of attackers. An attacker would
2857 additionally need online access to an unpatched system using the target
2858 private key in a scenario with persistent DH parameters and a private
2859 key that is shared between multiple clients.
2861 This only affects processors that support the BMI1, BMI2 and ADX extensions
2862 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2864 This issue was reported to OpenSSL by the OSS-Fuzz project.
2869 * Malformed X.509 IPAddressFamily could cause OOB read
2871 If an X.509 certificate has a malformed IPAddressFamily extension,
2872 OpenSSL could do a one-byte buffer overread. The most likely result
2873 would be an erroneous display of the certificate in text format.
2875 This issue was reported to OpenSSL by the OSS-Fuzz project.
2880 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2882 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2883 platform rather than 'mingw'.
2887 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2888 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2889 which is the minimum version we support.
2893 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2895 * Encrypt-Then-Mac renegotiation crash
2897 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2898 negotiated where it was not in the original handshake (or vice-versa) then
2899 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2900 and servers are affected.
2902 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2907 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2909 * Truncated packet could crash via OOB read
2911 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2912 cipher is being used, then a truncated packet can cause that host to
2913 perform an out-of-bounds read, usually resulting in a crash.
2915 This issue was reported to OpenSSL by Robert Święcki of Google.
2920 * Bad (EC)DHE parameters cause a client crash
2922 If a malicious server supplies bad parameters for a DHE or ECDHE key
2923 exchange then this can result in the client attempting to dereference a
2924 NULL pointer leading to a client crash. This could be exploited in a Denial
2927 This issue was reported to OpenSSL by Guido Vranken.
2932 * BN_mod_exp may produce incorrect results on x86_64
2934 There is a carry propagating bug in the x86_64 Montgomery squaring
2935 procedure. No EC algorithms are affected. Analysis suggests that attacks
2936 against RSA and DSA as a result of this defect would be very difficult to
2937 perform and are not believed likely. Attacks against DH are considered just
2938 feasible (although very difficult) because most of the work necessary to
2939 deduce information about a private key may be performed offline. The amount
2940 of resources required for such an attack would be very significant and
2941 likely only accessible to a limited number of attackers. An attacker would
2942 additionally need online access to an unpatched system using the target
2943 private key in a scenario with persistent DH parameters and a private
2944 key that is shared between multiple clients. For example this can occur by
2945 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2946 similar to CVE-2015-3193 but must be treated as a separate problem.
2948 This issue was reported to OpenSSL by the OSS-Fuzz project.
2953 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2955 * ChaCha20/Poly1305 heap-buffer-overflow
2957 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2958 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2959 crash. This issue is not considered to be exploitable beyond a DoS.
2961 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2966 * CMS Null dereference
2968 Applications parsing invalid CMS structures can crash with a NULL pointer
2969 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2970 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2971 structure callback if an attempt is made to free certain invalid encodings.
2972 Only CHOICE structures using a callback which do not handle NULL value are
2975 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2980 * Montgomery multiplication may produce incorrect results
2982 There is a carry propagating bug in the Broadwell-specific Montgomery
2983 multiplication procedure that handles input lengths divisible by, but
2984 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2985 and DH private keys are impossible. This is because the subroutine in
2986 question is not used in operations with the private key itself and an input
2987 of the attacker's direct choice. Otherwise the bug can manifest itself as
2988 transient authentication and key negotiation failures or reproducible
2989 erroneous outcome of public-key operations with specially crafted input.
2990 Among EC algorithms only Brainpool P-512 curves are affected and one
2991 presumably can attack ECDH key negotiation. Impact was not analyzed in
2992 detail, because pre-requisites for attack are considered unlikely. Namely
2993 multiple clients have to choose the curve in question and the server has to
2994 share the private key among them, neither of which is default behaviour.
2995 Even then only clients that chose the curve will be affected.
2997 This issue was publicly reported as transient failures and was not
2998 initially recognized as a security issue. Thanks to Richard Morgan for
2999 providing reproducible case.
3004 * Removed automatic addition of RPATH in shared libraries and executables,
3005 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3009 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3011 * Fix Use After Free for large message sizes
3013 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3014 message larger than approx 16k is received then the underlying buffer to
3015 store the incoming message is reallocated and moved. Unfortunately a
3016 dangling pointer to the old location is left which results in an attempt to
3017 write to the previously freed location. This is likely to result in a
3018 crash, however it could potentially lead to execution of arbitrary code.
3020 This issue only affects OpenSSL 1.1.0a.
3022 This issue was reported to OpenSSL by Robert Święcki.
3027 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3029 * OCSP Status Request extension unbounded memory growth
3031 A malicious client can send an excessively large OCSP Status Request
3032 extension. If that client continually requests renegotiation, sending a
3033 large OCSP Status Request extension each time, then there will be unbounded
3034 memory growth on the server. This will eventually lead to a Denial Of
3035 Service attack through memory exhaustion. Servers with a default
3036 configuration are vulnerable even if they do not support OCSP. Builds using
3037 the "no-ocsp" build time option are not affected.
3039 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3044 * SSL_peek() hang on empty record
3046 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3047 sends an empty record. This could be exploited by a malicious peer in a
3048 Denial Of Service attack.
3050 This issue was reported to OpenSSL by Alex Gaynor.
3055 * Excessive allocation of memory in tls_get_message_header() and
3056 dtls1_preprocess_fragment()
3058 A (D)TLS message includes 3 bytes for its length in the header for the
3059 message. This would allow for messages up to 16Mb in length. Messages of
3060 this length are excessive and OpenSSL includes a check to ensure that a
3061 peer is sending reasonably sized messages in order to avoid too much memory
3062 being consumed to service a connection. A flaw in the logic of version
3063 1.1.0 means that memory for the message is allocated too early, prior to
3064 the excessive message length check. Due to way memory is allocated in
3065 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3066 to service a connection. This could lead to a Denial of Service through
3067 memory exhaustion. However, the excessive message length check still takes
3068 place, and this would cause the connection to immediately fail. Assuming
3069 that the application calls SSL_free() on the failed connection in a timely
3070 manner then the 21Mb of allocated memory will then be immediately freed
3071 again. Therefore the excessive memory allocation will be transitory in
3072 nature. This then means that there is only a security impact if:
3074 1) The application does not call SSL_free() in a timely manner in the event
3075 that the connection fails
3077 2) The application is working in a constrained environment where there is
3078 very little free memory
3080 3) The attacker initiates multiple connection attempts such that there are
3081 multiple connections in a state where memory has been allocated for the
3082 connection; SSL_free() has not yet been called; and there is insufficient
3083 memory to service the multiple requests.
3085 Except in the instance of (1) above any Denial Of Service is likely to be
3086 transitory because as soon as the connection fails the memory is
3087 subsequently freed again in the SSL_free() call. However there is an
3088 increased risk during this period of application crashes due to the lack of
3089 memory - which would then mean a more serious Denial of Service.
3091 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3092 (CVE-2016-6307 and CVE-2016-6308)
3096 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3097 had to be removed. Primary reason is that vendor assembler can't
3098 assemble our modules with -KPIC flag. As result it, assembly
3099 support, was not even available as option. But its lack means
3100 lack of side-channel resistant code, which is incompatible with
3101 security by todays standards. Fortunately gcc is readily available
3102 prepackaged option, which we firmly point at...
3106 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3108 * Windows command-line tool supports UTF-8 opt-in option for arguments
3109 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3110 (to any value) allows Windows user to access PKCS#12 file generated
3111 with Windows CryptoAPI and protected with non-ASCII password, as well
3112 as files generated under UTF-8 locale on Linux also protected with
3117 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3118 have been disabled by default and removed from DEFAULT, just like RC4.
3119 See the RC4 item below to re-enable both.
3123 * The method for finding the storage location for the Windows RAND seed file
3124 has changed. First we check %RANDFILE%. If that is not set then we check
3125 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3126 all else fails we fall back to C:\.
3130 * The EVP_EncryptUpdate() function has had its return type changed from void
3131 to int. A return of 0 indicates and error while a return of 1 indicates
3136 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3137 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3138 off the constant time implementation for RSA, DSA and DH have been made
3139 no-ops and deprecated.
3143 * Windows RAND implementation was simplified to only get entropy by
3144 calling CryptGenRandom(). Various other RAND-related tickets
3147 *Joseph Wylie Yandle, Rich Salz*
3149 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3150 and `OPENSSL_LH_`, respectively. The old names are available
3151 with API compatibility. They new names are now completely documented.
3155 * Unify TYPE_up_ref(obj) methods signature.
3156 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3157 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3158 int (instead of void) like all others TYPE_up_ref() methods.
3159 So now these methods also check the return value of CRYPTO_atomic_add(),
3160 and the validity of object reference counter.
3162 *fdasilvayy@gmail.com*
3164 * With Windows Visual Studio builds, the .pdb files are installed
3165 alongside the installed libraries and executables. For a static
3166 library installation, ossl_static.pdb is the associate compiler
3167 generated .pdb file to be used when linking programs.
3171 * Remove openssl.spec. Packaging files belong with the packagers.
3175 * Automatic Darwin/OSX configuration has had a refresh, it will now
3176 recognise x86_64 architectures automatically. You can still decide
3177 to build for a different bitness with the environment variable
3178 KERNEL_BITS (can be 32 or 64), for example:
3180 KERNEL_BITS=32 ./config
3184 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3185 256 bit AES and HMAC with SHA256.
3189 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3193 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3197 * To enable users to have their own config files and build file templates,
3198 Configure looks in the directory indicated by the environment variable
3199 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3200 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3201 name and is used as is.
3205 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3206 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3207 X509_CERT_FILE_CTX was removed.
3211 * "shared" builds are now the default. To create only static libraries use
3212 the "no-shared" Configure option.
3216 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3217 All of these option have not worked for some while and are fundamental
3222 * Make various cleanup routines no-ops and mark them as deprecated. Most
3223 global cleanup functions are no longer required because they are handled
3224 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3225 Explicitly de-initing can cause problems (e.g. where a library that uses
3226 OpenSSL de-inits, but an application is still using it). The affected
3227 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3228 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3229 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3230 COMP_zlib_cleanup().
3234 * --strict-warnings no longer enables runtime debugging options
3235 such as REF_DEBUG. Instead, debug options are automatically
3236 enabled with '--debug' builds.
3238 *Andy Polyakov, Emilia Käsper*
3240 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3241 have been moved out of the public header files. New functions for managing
3242 these have been added.
3246 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3247 objects have been moved out of the public header files. New
3248 functions for managing these have been added.
3252 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3253 have been moved out of the public header files. New functions for managing
3254 these have been added.
3258 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3259 moved out of the public header files. New functions for managing these
3264 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3268 * Removed the mk1mf build scripts.
3272 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3273 it is always safe to #include a header now.
3277 * Removed the aged BC-32 config and all its supporting scripts
3281 * Removed support for Ultrix, Netware, and OS/2.
3285 * Add support for HKDF.
3287 *Alessandro Ghedini*
3289 * Add support for blake2b and blake2s
3293 * Added support for "pipelining". Ciphers that have the
3294 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3295 encryptions/decryptions simultaneously. There are currently no built-in
3296 ciphers with this property but the expectation is that engines will be able
3297 to offer it to significantly improve throughput. Support has been extended
3298 into libssl so that multiple records for a single connection can be
3299 processed in one go (for >=TLS 1.1).
3303 * Added the AFALG engine. This is an async capable engine which is able to
3304 offload work to the Linux kernel. In this initial version it only supports
3305 AES128-CBC. The kernel must be version 4.1.0 or greater.
3309 * OpenSSL now uses a new threading API. It is no longer necessary to
3310 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3311 are two supported threading models: pthreads and windows threads. It is
3312 also possible to configure OpenSSL at compile time for "no-threads". The
3313 old threading API should no longer be used. The functions have been
3314 replaced with "no-op" compatibility macros.
3316 *Alessandro Ghedini, Matt Caswell*
3318 * Modify behavior of ALPN to invoke callback after SNI/servername
3319 callback, such that updates to the SSL_CTX affect ALPN.
3323 * Add SSL_CIPHER queries for authentication and key-exchange.
3327 * Changes to the DEFAULT cipherlist:
3328 - Prefer (EC)DHE handshakes over plain RSA.
3329 - Prefer AEAD ciphers over legacy ciphers.
3330 - Prefer ECDSA over RSA when both certificates are available.
3331 - Prefer TLSv1.2 ciphers/PRF.
3332 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3337 * Change the ECC default curve list to be this, in order: x25519,
3338 secp256r1, secp521r1, secp384r1.
3342 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3343 disabled by default. They can be re-enabled using the
3344 enable-weak-ssl-ciphers option to Configure.
3348 * If the server has ALPN configured, but supports no protocols that the
3349 client advertises, send a fatal "no_application_protocol" alert.
3350 This behaviour is SHALL in RFC 7301, though it isn't universally
3351 implemented by other servers.
3355 * Add X25519 support.
3356 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3357 for public and private key encoding using the format documented in
3358 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3359 key generation and key derivation.
3361 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3366 * Deprecate SRP_VBASE_get_by_user.
3367 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3368 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3369 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3370 seed, even if the seed is configured.
3372 Users should use SRP_VBASE_get1_by_user instead. Note that in
3373 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3374 also that even though configuring the SRP seed attempts to hide
3375 invalid usernames by continuing the handshake with fake
3376 credentials, this behaviour is not constant time and no strong
3377 guarantees are made that the handshake is indistinguishable from
3378 that of a valid user.
3382 * Configuration change; it's now possible to build dynamic engines
3383 without having to build shared libraries and vice versa. This
3384 only applies to the engines in `engines/`, those in `crypto/engine/`
3385 will always be built into libcrypto (i.e. "static").
3387 Building dynamic engines is enabled by default; to disable, use
3388 the configuration option "disable-dynamic-engine".
3390 The only requirements for building dynamic engines are the
3391 presence of the DSO module and building with position independent
3392 code, so they will also automatically be disabled if configuring
3393 with "disable-dso" or "disable-pic".
3395 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3396 are also taken away from openssl/opensslconf.h, as they are
3401 * Configuration change; if there is a known flag to compile
3402 position independent code, it will always be applied on the
3403 libcrypto and libssl object files, and never on the application
3404 object files. This means other libraries that use routines from
3405 libcrypto / libssl can be made into shared libraries regardless
3406 of how OpenSSL was configured.
3408 If this isn't desirable, the configuration options "disable-pic"
3409 or "no-pic" can be used to disable the use of PIC. This will
3410 also disable building shared libraries and dynamic engines.
3414 * Removed JPAKE code. It was experimental and has no wide use.
3418 * The INSTALL_PREFIX Makefile variable has been renamed to
3419 DESTDIR. That makes for less confusion on what this variable
3420 is for. Also, the configuration option --install_prefix is
3425 * Heartbeat for TLS has been removed and is disabled by default
3426 for DTLS; configure with enable-heartbeats. Code that uses the
3427 old #define's might need to be updated.
3429 *Emilia Käsper, Rich Salz*
3431 * Rename REF_CHECK to REF_DEBUG.
3435 * New "unified" build system
3437 The "unified" build system is aimed to be a common system for all
3438 platforms we support. With it comes new support for VMS.
3440 This system builds supports building in a different directory tree
3441 than the source tree. It produces one Makefile (for unix family
3442 or lookalikes), or one descrip.mms (for VMS).
3444 The source of information to make the Makefile / descrip.mms is
3445 small files called 'build.info', holding the necessary
3446 information for each directory with source to compile, and a
3447 template in Configurations, like unix-Makefile.tmpl or
3450 With this change, the library names were also renamed on Windows
3451 and on VMS. They now have names that are closer to the standard
3452 on Unix, and include the major version number, and in certain
3453 cases, the architecture they are built for. See "Notes on shared
3454 libraries" in INSTALL.
3456 We rely heavily on the perl module Text::Template.
3460 * Added support for auto-initialisation and de-initialisation of the library.
3461 OpenSSL no longer requires explicit init or deinit routines to be called,
3462 except in certain circumstances. See the OPENSSL_init_crypto() and
3463 OPENSSL_init_ssl() man pages for further information.
3467 * The arguments to the DTLSv1_listen function have changed. Specifically the
3468 "peer" argument is now expected to be a BIO_ADDR object.
3470 * Rewrite of BIO networking library. The BIO library lacked consistent
3471 support of IPv6, and adding it required some more extensive
3472 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3473 which hold all types of addresses and chains of address information.
3474 It also introduces a new API, with functions like BIO_socket,
3475 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3476 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3477 have been adapted accordingly.
3481 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3486 * CRIME protection: disable compression by default, even if OpenSSL is
3487 compiled with zlib enabled. Applications can still enable compression
3488 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3489 using the SSL_CONF library to configure compression.
3493 * The signature of the session callback configured with
3494 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3495 was explicitly marked as `const unsigned char*` instead of
3500 * Always DPURIFY. Remove the use of uninitialized memory in the
3501 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3505 * Removed many obsolete configuration items, including
3506 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3507 MD2_CHAR, MD2_INT, MD2_LONG
3509 IDEA_SHORT, IDEA_LONG
3510 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3512 *Rich Salz, with advice from Andy Polyakov*
3514 * Many BN internals have been moved to an internal header file.
3516 *Rich Salz with help from Andy Polyakov*
3518 * Configuration and writing out the results from it has changed.
3519 Files such as Makefile include/openssl/opensslconf.h and are now
3520 produced through general templates, such as Makefile.in and
3521 crypto/opensslconf.h.in and some help from the perl module
3524 Also, the center of configuration information is no longer
3525 Makefile. Instead, Configure produces a perl module in
3526 configdata.pm which holds most of the config data (in the hash
3527 table %config), the target data that comes from the target
3528 configuration in one of the `Configurations/*.conf` files (in
3533 * To clarify their intended purposes, the Configure options
3534 --prefix and --openssldir change their semantics, and become more
3535 straightforward and less interdependent.
3537 --prefix shall be used exclusively to give the location INSTALLTOP
3538 where programs, scripts, libraries, include files and manuals are
3539 going to be installed. The default is now /usr/local.
3541 --openssldir shall be used exclusively to give the default
3542 location OPENSSLDIR where certificates, private keys, CRLs are
3543 managed. This is also where the default openssl.cnf gets
3545 If the directory given with this option is a relative path, the
3546 values of both the --prefix value and the --openssldir value will
3547 be combined to become OPENSSLDIR.
3548 The default for --openssldir is INSTALLTOP/ssl.
3550 Anyone who uses --openssldir to specify where OpenSSL is to be
3551 installed MUST change to use --prefix instead.
3555 * The GOST engine was out of date and therefore it has been removed. An up
3556 to date GOST engine is now being maintained in an external repository.
3557 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3558 support for GOST ciphersuites (these are only activated if a GOST engine
3563 * EGD is no longer supported by default; use enable-egd when
3566 *Ben Kaduk and Rich Salz*
3568 * The distribution now has Makefile.in files, which are used to
3569 create Makefile's when Configure is run. *Configure must be run
3570 before trying to build now.*
3574 * The return value for SSL_CIPHER_description() for error conditions
3579 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3581 Obtaining and performing DNSSEC validation of TLSA records is
3582 the application's responsibility. The application provides
3583 the TLSA records of its choice to OpenSSL, and these are then
3584 used to authenticate the peer.
3586 The TLSA records need not even come from DNS. They can, for
3587 example, be used to implement local end-entity certificate or
3588 trust-anchor "pinning", where the "pin" data takes the form
3589 of TLSA records, which can augment or replace verification
3590 based on the usual WebPKI public certification authorities.
3594 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3595 continues to support deprecated interfaces in default builds.
3596 However, applications are strongly advised to compile their
3597 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3598 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3599 or the 1.1.0 releases.
3601 In environments in which all applications have been ported to
3602 not use any deprecated interfaces OpenSSL's Configure script
3603 should be used with the --api=1.1.0 option to entirely remove
3604 support for the deprecated features from the library and
3605 unconditionally disable them in the installed headers.
3606 Essentially the same effect can be achieved with the "no-deprecated"
3607 argument to Configure, except that this will always restrict
3608 the build to just the latest API, rather than a fixed API
3611 As applications are ported to future revisions of the API,
3612 they should update their compile-time OPENSSL_API_COMPAT define
3613 accordingly, but in most cases should be able to continue to
3614 compile with later releases.
3616 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3617 0x10000000L and 0x00908000L, respectively. However those
3618 versions did not support the OPENSSL_API_COMPAT feature, and
3619 so applications are not typically tested for explicit support
3620 of just the undeprecated features of either release.
3624 * Add support for setting the minimum and maximum supported protocol.
3625 It can bet set via the SSL_set_min_proto_version() and
3626 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3627 MaxProtocol. It's recommended to use the new APIs to disable
3628 protocols instead of disabling individual protocols using
3629 SSL_set_options() or SSL_CONF's Protocol. This change also
3630 removes support for disabling TLS 1.2 in the OpenSSL TLS
3631 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3635 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3639 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3640 and integrates ECDSA and ECDH functionality into EC. Implementations can
3641 now redirect key generation and no longer need to convert to or from
3644 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3645 include the ec.h header file instead.
3649 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3650 ciphers who are no longer supported and drops support the ephemeral RSA key
3651 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3655 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3656 opaque. For HMAC_CTX, the following constructors and destructors
3659 HMAC_CTX *HMAC_CTX_new(void);
3660 void HMAC_CTX_free(HMAC_CTX *ctx);
3662 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3663 destroy such methods has been added. See EVP_MD_meth_new(3) and
3664 EVP_CIPHER_meth_new(3) for documentation.
3667 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3668 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3669 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3670 an already created structure.
3671 2) For consistency with the majority of our object creators and
3672 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3673 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3674 for deprecated builds.
3678 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3679 cryptographic operations to be performed asynchronously as long as an
3680 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3681 further details. Libssl has also had this capability integrated with the
3682 introduction of the new mode SSL_MODE_ASYNC and associated error
3683 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3684 pages. This work was developed in partnership with Intel Corp.
3688 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3689 always enabled now. If you want to disable the support you should
3690 exclude it using the list of supported ciphers. This also means that the
3691 "-no_ecdhe" option has been removed from s_server.
3695 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3696 SSL_{CTX_}set1_curves() which can set a list.
3700 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3701 curve you want to support using SSL_{CTX_}set1_curves().
3705 * State machine rewrite. The state machine code has been significantly
3706 refactored in order to remove much duplication of code and solve issues
3707 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3708 further details). This change does have some associated API changes.
3709 Notably the SSL_state() function has been removed and replaced by
3710 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3711 SSL_set_state() has been removed altogether. The previous handshake states
3712 defined in ssl.h and ssl3.h have also been removed.
3716 * All instances of the string "ssleay" in the public API were replaced
3717 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3718 Some error codes related to internal RSA_eay API's were renamed.
3722 * The demo files in crypto/threads were moved to demo/threads.
3726 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3729 *Matt Caswell, Rich Salz*
3731 * New ASN.1 embed macro.
3733 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3734 structure is not allocated: it is part of the parent. That is instead of
3742 This reduces memory fragmentation and make it impossible to accidentally
3743 set a mandatory field to NULL.
3745 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3746 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3747 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3752 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3756 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3757 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3758 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3759 DES and RC4 ciphersuites.
3763 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3764 This changes the decoding behaviour for some invalid messages,
3765 though the change is mostly in the more lenient direction, and
3766 legacy behaviour is preserved as much as possible.
3770 * Fix no-stdio build.
3771 *David Woodhouse <David.Woodhouse@intel.com> and also*
3772 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3774 * New testing framework
3775 The testing framework has been largely rewritten and is now using
3776 perl and the perl modules Test::Harness and an extended variant of
3777 Test::More called OpenSSL::Test to do its work. All test scripts in
3778 test/ have been rewritten into test recipes, and all direct calls to
3779 executables in test/Makefile have become individual recipes using the
3780 simplified testing OpenSSL::Test::Simple.
3782 For documentation on our testing modules, do:
3784 perldoc test/testlib/OpenSSL/Test/Simple.pm
3785 perldoc test/testlib/OpenSSL/Test.pm
3789 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3790 are used; the latter aborts on memory leaks (usually checked on exit).
3791 Some undocumented "set malloc, etc., hooks" functions were removed
3792 and others were changed. All are now documented.
3796 * In DSA_generate_parameters_ex, if the provided seed is too short,
3799 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3801 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3802 from RFC4279, RFC4785, RFC5487, RFC5489.
3804 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3805 original RSA_PSK patch.
3809 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3810 era flag was never set throughout the codebase (only read). Also removed
3811 SSL3_FLAGS_POP_BUFFER which was only used if
3812 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3816 * Changed the default name options in the "ca", "crl", "req" and "x509"
3817 to be "oneline" instead of "compat".
3821 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3822 not aware of clients that still exhibit this bug, and the workaround
3823 hasn't been working properly for a while.
3827 * The return type of BIO_number_read() and BIO_number_written() as well as
3828 the corresponding num_read and num_write members in the BIO structure has
3829 changed from unsigned long to uint64_t. On platforms where an unsigned
3830 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3835 * Given the pervasive nature of TLS extensions it is inadvisable to run
3836 OpenSSL without support for them. It also means that maintaining
3837 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3838 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3842 * Removed support for the two export grade static DH ciphersuites
3843 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3844 were newly added (along with a number of other static DH ciphersuites) to
3845 1.0.2. However the two export ones have *never* worked since they were
3846 introduced. It seems strange in any case to be adding new export
3847 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3851 * Version negotiation has been rewritten. In particular SSLv23_method(),
3852 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3853 and turned into macros which simply call the new preferred function names
3854 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3855 should use the new names instead. Also as part of this change the ssl23.h
3856 header file has been removed.
3860 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3861 code and the associated standard is no longer considered fit-for-purpose.
3865 * RT2547 was closed. When generating a private key, try to make the
3866 output file readable only by the owner. This behavior change might
3867 be noticeable when interacting with other software.
3869 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3874 * Added HTTP GET support to the ocsp command.
3878 * Changed default digest for the dgst and enc commands from MD5 to
3883 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3887 * Added support for TLS extended master secret from
3888 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3889 initial patch which was a great help during development.
3893 * All libssl internal structures have been removed from the public header
3894 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3895 now redundant). Users should not attempt to access internal structures
3896 directly. Instead they should use the provided API functions.
3900 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3901 Access to deprecated functions can be re-enabled by running config with
3902 "enable-deprecated". In addition applications wishing to use deprecated
3903 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3904 will, by default, disable some transitive includes that previously existed
3905 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3909 * Added support for OCB mode. OpenSSL has been granted a patent license
3910 compatible with the OpenSSL license for use of OCB. Details are available
3911 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3912 for OCB can be removed by calling config with no-ocb.
3916 * SSLv2 support has been removed. It still supports receiving a SSLv2
3917 compatible client hello.
3921 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3922 done while fixing the error code for the key-too-small case.
3924 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3926 * CA.sh has been removed; use CA.pl instead.
3930 * Removed old DES API.
3934 * Remove various unsupported platforms:
3940 Sinix/ReliantUNIX RM400
3945 16-bit platforms such as WIN16
3949 * Clean up OPENSSL_NO_xxx #define's
3950 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3951 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3952 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3953 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3954 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3955 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3956 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3957 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3958 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3959 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3963 * Cleaned up dead code
3964 Remove all but one '#ifdef undef' which is to be looked at.
3968 * Clean up calling of xxx_free routines.
3969 Just like free(), fix most of the xxx_free routines to accept
3970 NULL. Remove the non-null checks from callers. Save much code.
3974 * Add secure heap for storage of private keys (when possible).
3975 Add BIO_s_secmem(), CBIGNUM, etc.
3976 Contributed by Akamai Technologies under our Corporate CLA.
3980 * Experimental support for a new, fast, unbiased prime candidate generator,
3981 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3983 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3985 * New output format NSS in the sess_id command line tool. This allows
3986 exporting the session id and the master key in NSS keylog format.
3988 *Martin Kaiser <martin@kaiser.cx>*
3990 * Harmonize version and its documentation. -f flag is used to display
3993 *mancha <mancha1@zoho.com>*
3995 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3996 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3998 *mancha <mancha1@zoho.com>*
4000 * Fix some double frees. These are not thought to be exploitable.
4002 *mancha <mancha1@zoho.com>*
4004 * A missing bounds check in the handling of the TLS heartbeat extension
4005 can be used to reveal up to 64k of memory to a connected client or
4008 Thanks for Neel Mehta of Google Security for discovering this bug and to
4009 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4010 preparing the fix ([CVE-2014-0160])
4012 *Adam Langley, Bodo Moeller*
4014 * Fix for the attack described in the paper "Recovering OpenSSL
4015 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4016 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4017 <http://eprint.iacr.org/2014/140>
4019 Thanks to Yuval Yarom and Naomi Benger for discovering this
4020 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4022 *Yuval Yarom and Naomi Benger*
4024 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4025 this fixes a limitation in previous versions of OpenSSL.
4029 * Experimental encrypt-then-mac support.
4031 Experimental support for encrypt then mac from
4032 draft-gutmann-tls-encrypt-then-mac-02.txt
4034 To enable it set the appropriate extension number (0x42 for the test
4035 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4037 For non-compliant peers (i.e. just about everything) this should have no
4040 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4044 * Add EVP support for key wrapping algorithms, to avoid problems with
4045 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4046 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4047 algorithms and include tests cases.
4051 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4056 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4057 MGF1 digest and OAEP label.
4061 * Make openssl verify return errors.
4063 *Chris Palmer <palmer@google.com> and Ben Laurie*
4065 * New function ASN1_TIME_diff to calculate the difference between two
4066 ASN1_TIME structures or one structure and the current time.
4070 * Update fips_test_suite to support multiple command line options. New
4071 test to induce all self test errors in sequence and check expected
4076 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4077 sign or verify all in one operation.
4081 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4082 test programs and fips_test_suite. Includes functionality to parse
4083 the minimal script output of fipsalgest.pl directly.
4087 * Add authorisation parameter to FIPS_module_mode_set().
4091 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4095 * Use separate DRBG fields for internal and external flags. New function
4096 FIPS_drbg_health_check() to perform on demand health checking. Add
4097 generation tests to fips_test_suite with reduced health check interval to
4098 demonstrate periodic health checking. Add "nodh" option to
4099 fips_test_suite to skip very slow DH test.
4103 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4108 * More extensive health check for DRBG checking many more failure modes.
4109 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4110 combination: call this in fips_test_suite.
4114 * Add support for canonical generation of DSA parameter 'g'. See
4117 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4118 POST to handle HMAC cases.
4122 * Add functions FIPS_module_version() and FIPS_module_version_text()
4123 to return numerical and string versions of the FIPS module number.
4127 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4128 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4129 outside the validated module in the FIPS capable OpenSSL.
4133 * Minor change to DRBG entropy callback semantics. In some cases
4134 there is no multiple of the block length between min_len and
4135 max_len. Allow the callback to return more than max_len bytes
4136 of entropy but discard any extra: it is the callback's responsibility
4137 to ensure that the extra data discarded does not impact the
4138 requested amount of entropy.
4142 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4143 information in FIPS186-3, SP800-57 and SP800-131A.
4147 * CCM support via EVP. Interface is very similar to GCM case except we
4148 must supply all data in one chunk (i.e. no update, final) and the
4149 message length must be supplied if AAD is used. Add algorithm test
4154 * Initial version of POST overhaul. Add POST callback to allow the status
4155 of POST to be monitored and/or failures induced. Modify fips_test_suite
4156 to use callback. Always run all selftests even if one fails.
4160 * XTS support including algorithm test driver in the fips_gcmtest program.
4161 Note: this does increase the maximum key length from 32 to 64 bytes but
4162 there should be no binary compatibility issues as existing applications
4163 will never use XTS mode.
4167 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4168 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4169 performs algorithm blocking for unapproved PRNG types. Also do not
4170 set PRNG type in FIPS_mode_set(): leave this to the application.
4171 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4172 the standard OpenSSL PRNG: set additional data to a date time vector.
4176 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4177 This shouldn't present any incompatibility problems because applications
4178 shouldn't be using these directly and any that are will need to rethink
4179 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4183 * Extensive self tests and health checking required by SP800-90 DRBG.
4184 Remove strength parameter from FIPS_drbg_instantiate and always
4185 instantiate at maximum supported strength.
4189 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4193 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4197 * New function DH_compute_key_padded() to compute a DH key and pad with
4198 leading zeroes if needed: this complies with SP800-56A et al.
4202 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4203 anything, incomplete, subject to change and largely untested at present.
4207 * Modify fipscanisteronly build option to only build the necessary object
4208 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4212 * Add experimental option FIPSSYMS to give all symbols in
4213 fipscanister.o and FIPS or fips prefix. This will avoid
4214 conflicts with future versions of OpenSSL. Add perl script
4215 util/fipsas.pl to preprocess assembly language source files
4216 and rename any affected symbols.
4220 * Add selftest checks and algorithm block of non-fips algorithms in
4221 FIPS mode. Remove DES2 from selftests.
4225 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4226 return internal method without any ENGINE dependencies. Add new
4227 tiny fips sign and verify functions.
4231 * New build option no-ec2m to disable characteristic 2 code.
4235 * New build option "fipscanisteronly". This only builds fipscanister.o
4236 and (currently) associated fips utilities. Uses the file Makefile.fips
4237 instead of Makefile.org as the prototype.
4241 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4242 Update fips_gcmtest to use IV generator.
4246 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4247 setting output buffer to NULL. The `*Final` function must be
4248 called although it will not retrieve any additional data. The tag
4249 can be set or retrieved with a ctrl. The IV length is by default 12
4250 bytes (96 bits) but can be set to an alternative value. If the IV
4251 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4256 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4257 underlying do_cipher function handles all cipher semantics itself
4258 including padding and finalisation. This is useful if (for example)
4259 an ENGINE cipher handles block padding itself. The behaviour of
4260 do_cipher is subtly changed if this flag is set: the return value
4261 is the number of characters written to the output buffer (zero is
4262 no longer an error code) or a negative error code. Also if the
4263 input buffer is NULL and length 0 finalisation should be performed.
4267 * If a candidate issuer certificate is already part of the constructed
4268 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4272 * Improve forward-security support: add functions
4274 void SSL_CTX_set_not_resumable_session_callback(
4275 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4276 void SSL_set_not_resumable_session_callback(
4277 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4279 for use by SSL/TLS servers; the callback function will be called whenever a
4280 new session is created, and gets to decide whether the session may be
4281 cached to make it resumable (return 0) or not (return 1). (As by the
4282 SSL/TLS protocol specifications, the session_id sent by the server will be
4283 empty to indicate that the session is not resumable; also, the server will
4284 not generate RFC 4507 (RFC 5077) session tickets.)
4286 A simple reasonable callback implementation is to return is_forward_secure.
4287 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4288 by the SSL/TLS server library, indicating whether it can provide forward
4291 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4293 * New -verify_name option in command line utilities to set verification
4298 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4299 Add CMAC pkey methods.
4303 * Experimental renegotiation in s_server -www mode. If the client
4304 browses /reneg connection is renegotiated. If /renegcert it is
4305 renegotiated requesting a certificate.
4309 * Add an "external" session cache for debugging purposes to s_server. This
4310 should help trace issues which normally are only apparent in deployed
4311 multi-process servers.
4315 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4316 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4317 BIO_set_cipher() and some obscure PEM functions were changed so they
4318 can now return an error. The RAND changes required a change to the
4319 RAND_METHOD structure.
4323 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4324 a gcc attribute to warn if the result of a function is ignored. This
4325 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4326 whose return value is often ignored.
4330 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4331 These allow SCTs (signed certificate timestamps) to be requested and
4332 validated when establishing a connection.
4334 *Rob Percival <robpercival@google.com>*
4339 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4341 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4342 used even when parsing explicit parameters, when loading a encoded key
4343 or calling `EC_GROUP_new_from_ecpkparameters()`/
4344 `EC_GROUP_new_from_ecparameters()`.
4345 This prevents bypass of security hardening and performance gains,
4346 especially for curves with specialized EC_METHODs.
4347 By default, if a key encoded with explicit parameters is loaded and later
4348 encoded, the output is still encoded with explicit parameters, even if
4349 internally a "named" EC_GROUP is used for computation.
4353 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4354 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4355 NULL. After this change, only the cofactor parameter can be NULL. It also
4356 does some minimal sanity checks on the passed order.
4361 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4362 An attack is simple, if the first CMS_recipientInfo is valid but the
4363 second CMS_recipientInfo is chosen ciphertext. If the second
4364 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4365 encryption key will be replaced by garbage, and the message cannot be
4366 decoded, but if the RSA decryption fails, the correct encryption key is
4367 used and the recipient will not notice the attack.
4368 As a work around for this potential attack the length of the decrypted
4369 key must be equal to the cipher default key length, in case the
4370 certifiate is not given and all recipientInfo are tried out.
4371 The old behaviour can be re-enabled in the CMS code by setting the
4372 CMS_DEBUG_DECRYPT flag.
4377 * Document issue with installation paths in diverse Windows builds
4379 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4380 binaries and run-time config file.
4385 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4387 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4388 This changes the size when using the `genpkey` command when no size is given.
4389 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4390 generation commands to use 2048 bits by default.
4394 * Add FIPS support for Android Arm 64-bit
4396 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4397 Module in Version 2.0.10. For some reason, the corresponding target
4398 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4399 built with FIPS support on Android Arm 64-bit. This omission has been
4402 *Matthias St. Pierre*
4404 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4406 * 0-byte record padding oracle
4408 If an application encounters a fatal protocol error and then calls
4409 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4410 then OpenSSL can respond differently to the calling application if a 0 byte
4411 record is received with invalid padding compared to if a 0 byte record is
4412 received with an invalid MAC. If the application then behaves differently
4413 based on that in a way that is detectable to the remote peer, then this
4414 amounts to a padding oracle that could be used to decrypt data.
4416 In order for this to be exploitable "non-stitched" ciphersuites must be in
4417 use. Stitched ciphersuites are optimised implementations of certain
4418 commonly used ciphersuites. Also the application must call SSL_shutdown()
4419 twice even if a protocol error has occurred (applications should not do
4420 this but some do anyway).
4422 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4423 Aviram, with additional investigation by Steven Collison and Andrew
4424 Hourselt. It was reported to OpenSSL on 10th December 2018.
4429 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4433 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4435 * Microarchitecture timing vulnerability in ECC scalar multiplication
4437 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4438 shown to be vulnerable to a microarchitecture timing side channel attack.
4439 An attacker with sufficient access to mount local timing attacks during
4440 ECDSA signature generation could recover the private key.
4442 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4443 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4449 * Timing vulnerability in DSA signature generation
4451 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4452 timing side channel attack. An attacker could use variations in the signing
4453 algorithm to recover the private key.
4455 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4460 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4461 Module, accidentally introduced while backporting security fixes from the
4462 development branch and hindering the use of ECC in FIPS mode.
4466 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4468 * Client DoS due to large DH parameter
4470 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4471 malicious server can send a very large prime value to the client. This will
4472 cause the client to spend an unreasonably long period of time generating a
4473 key for this prime resulting in a hang until the client has finished. This
4474 could be exploited in a Denial Of Service attack.
4476 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4481 * Cache timing vulnerability in RSA Key Generation
4483 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4484 a cache timing side channel attack. An attacker with sufficient access to
4485 mount cache timing attacks during the RSA key generation process could
4486 recover the private key.
4488 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4489 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4494 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4495 parameter is no longer accepted, as it leads to a corrupt table. NULL
4496 pem_str is reserved for alias entries only.
4500 * Revert blinding in ECDSA sign and instead make problematic addition
4501 length-invariant. Switch even to fixed-length Montgomery multiplication.
4505 * Change generating and checking of primes so that the error rate of not
4506 being prime depends on the intended use based on the size of the input.
4507 For larger primes this will result in more rounds of Miller-Rabin.
4508 The maximal error rate for primes with more than 1080 bits is lowered
4511 *Kurt Roeckx, Annie Yousar*
4513 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4517 * Add blinding to ECDSA and DSA signatures to protect against side channel
4518 attacks discovered by Keegan Ryan (NCC Group).
4522 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4523 now allow empty (zero character) pass phrases.
4527 * Certificate time validation (X509_cmp_time) enforces stricter
4528 compliance with RFC 5280. Fractional seconds and timezone offsets
4529 are no longer allowed.
4533 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4535 * Constructed ASN.1 types with a recursive definition could exceed the stack
4537 Constructed ASN.1 types with a recursive definition (such as can be found
4538 in PKCS7) could eventually exceed the stack given malicious input with
4539 excessive recursion. This could result in a Denial Of Service attack. There
4540 are no such structures used within SSL/TLS that come from untrusted sources
4541 so this is considered safe.
4543 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4549 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4551 * Read/write after SSL object in error state
4553 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4554 mechanism. The intent was that if a fatal error occurred during a handshake
4555 then OpenSSL would move into the error state and would immediately fail if
4556 you attempted to continue the handshake. This works as designed for the
4557 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4558 SSL_connect()), however due to a bug it does not work correctly if
4559 SSL_read() or SSL_write() is called directly. In that scenario, if the
4560 handshake fails then a fatal error will be returned in the initial function
4561 call. If SSL_read()/SSL_write() is subsequently called by the application
4562 for the same SSL object then it will succeed and the data is passed without
4563 being decrypted/encrypted directly from the SSL/TLS record layer.
4565 In order to exploit this issue an application bug would have to be present
4566 that resulted in a call to SSL_read()/SSL_write() being issued after having
4567 already received a fatal error.
4569 This issue was reported to OpenSSL by David Benjamin (Google).
4574 * rsaz_1024_mul_avx2 overflow bug on x86_64
4576 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4577 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4578 Analysis suggests that attacks against RSA and DSA as a result of this
4579 defect would be very difficult to perform and are not believed likely.
4580 Attacks against DH1024 are considered just feasible, because most of the
4581 work necessary to deduce information about a private key may be performed
4582 offline. The amount of resources required for such an attack would be
4583 significant. However, for an attack on TLS to be meaningful, the server
4584 would have to share the DH1024 private key among multiple clients, which is
4585 no longer an option since CVE-2016-0701.
4587 This only affects processors that support the AVX2 but not ADX extensions
4588 like Intel Haswell (4th generation).
4590 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4591 was originally found via the OSS-Fuzz project.
4596 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4598 * bn_sqrx8x_internal carry bug on x86_64
4600 There is a carry propagating bug in the x86_64 Montgomery squaring
4601 procedure. No EC algorithms are affected. Analysis suggests that attacks
4602 against RSA and DSA as a result of this defect would be very difficult to
4603 perform and are not believed likely. Attacks against DH are considered just
4604 feasible (although very difficult) because most of the work necessary to
4605 deduce information about a private key may be performed offline. The amount
4606 of resources required for such an attack would be very significant and
4607 likely only accessible to a limited number of attackers. An attacker would
4608 additionally need online access to an unpatched system using the target
4609 private key in a scenario with persistent DH parameters and a private
4610 key that is shared between multiple clients.
4612 This only affects processors that support the BMI1, BMI2 and ADX extensions
4613 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4615 This issue was reported to OpenSSL by the OSS-Fuzz project.
4620 * Malformed X.509 IPAddressFamily could cause OOB read
4622 If an X.509 certificate has a malformed IPAddressFamily extension,
4623 OpenSSL could do a one-byte buffer overread. The most likely result
4624 would be an erroneous display of the certificate in text format.
4626 This issue was reported to OpenSSL by the OSS-Fuzz project.
4630 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4632 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4633 platform rather than 'mingw'.
4637 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4639 * Truncated packet could crash via OOB read
4641 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4642 cipher is being used, then a truncated packet can cause that host to
4643 perform an out-of-bounds read, usually resulting in a crash.
4645 This issue was reported to OpenSSL by Robert Święcki of Google.
4650 * BN_mod_exp may produce incorrect results on x86_64
4652 There is a carry propagating bug in the x86_64 Montgomery squaring
4653 procedure. No EC algorithms are affected. Analysis suggests that attacks
4654 against RSA and DSA as a result of this defect would be very difficult to
4655 perform and are not believed likely. Attacks against DH are considered just
4656 feasible (although very difficult) because most of the work necessary to
4657 deduce information about a private key may be performed offline. The amount
4658 of resources required for such an attack would be very significant and
4659 likely only accessible to a limited number of attackers. An attacker would
4660 additionally need online access to an unpatched system using the target
4661 private key in a scenario with persistent DH parameters and a private
4662 key that is shared between multiple clients. For example this can occur by
4663 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4664 similar to CVE-2015-3193 but must be treated as a separate problem.
4666 This issue was reported to OpenSSL by the OSS-Fuzz project.
4671 * Montgomery multiplication may produce incorrect results
4673 There is a carry propagating bug in the Broadwell-specific Montgomery
4674 multiplication procedure that handles input lengths divisible by, but
4675 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4676 and DH private keys are impossible. This is because the subroutine in
4677 question is not used in operations with the private key itself and an input
4678 of the attacker's direct choice. Otherwise the bug can manifest itself as
4679 transient authentication and key negotiation failures or reproducible
4680 erroneous outcome of public-key operations with specially crafted input.
4681 Among EC algorithms only Brainpool P-512 curves are affected and one
4682 presumably can attack ECDH key negotiation. Impact was not analyzed in
4683 detail, because pre-requisites for attack are considered unlikely. Namely
4684 multiple clients have to choose the curve in question and the server has to
4685 share the private key among them, neither of which is default behaviour.
4686 Even then only clients that chose the curve will be affected.
4688 This issue was publicly reported as transient failures and was not
4689 initially recognized as a security issue. Thanks to Richard Morgan for
4690 providing reproducible case.
4695 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4696 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4697 prevent issues where no progress is being made and the peer continually
4698 sends unrecognised record types, using up resources processing them.
4702 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4704 * Missing CRL sanity check
4706 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4707 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4708 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4710 This issue only affects the OpenSSL 1.0.2i
4715 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4717 * OCSP Status Request extension unbounded memory growth
4719 A malicious client can send an excessively large OCSP Status Request
4720 extension. If that client continually requests renegotiation, sending a
4721 large OCSP Status Request extension each time, then there will be unbounded
4722 memory growth on the server. This will eventually lead to a Denial Of
4723 Service attack through memory exhaustion. Servers with a default
4724 configuration are vulnerable even if they do not support OCSP. Builds using
4725 the "no-ocsp" build time option are not affected.
4727 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4732 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4735 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4741 * OOB write in MDC2_Update()
4743 An overflow can occur in MDC2_Update() either if called directly or
4744 through the EVP_DigestUpdate() function using MDC2. If an attacker
4745 is able to supply very large amounts of input data after a previous
4746 call to EVP_EncryptUpdate() with a partial block then a length check
4747 can overflow resulting in a heap corruption.
4749 The amount of data needed is comparable to SIZE_MAX which is impractical
4752 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4757 * Malformed SHA512 ticket DoS
4759 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4760 DoS attack where a malformed ticket will result in an OOB read which will
4763 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4764 a custom server callback and ticket lookup mechanism.
4766 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4771 * OOB write in BN_bn2dec()
4773 The function BN_bn2dec() does not check the return value of BN_div_word().
4774 This can cause an OOB write if an application uses this function with an
4775 overly large BIGNUM. This could be a problem if an overly large certificate
4776 or CRL is printed out from an untrusted source. TLS is not affected because
4777 record limits will reject an oversized certificate before it is parsed.
4779 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4784 * OOB read in TS_OBJ_print_bio()
4786 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4787 the total length the OID text representation would use and not the amount
4788 of data written. This will result in OOB reads when large OIDs are
4791 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4796 * Pointer arithmetic undefined behaviour
4798 Avoid some undefined pointer arithmetic
4800 A common idiom in the codebase is to check limits in the following manner:
4803 Where "p" points to some malloc'd data of SIZE bytes and
4806 "len" here could be from some externally supplied data (e.g. from a TLS
4809 The rules of C pointer arithmetic are such that "p + len" is only well
4810 defined where len <= SIZE. Therefore the above idiom is actually
4811 undefined behaviour.
4813 For example this could cause problems if some malloc implementation
4814 provides an address for "p" such that "p + len" actually overflows for
4815 values of len that are too big and therefore p + len < limit.
4817 This issue was reported to OpenSSL by Guido Vranken
4822 * Constant time flag not preserved in DSA signing
4824 Operations in the DSA signing algorithm should run in constant time in
4825 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4826 implementation means that a non-constant time codepath is followed for
4827 certain operations. This has been demonstrated through a cache-timing
4828 attack to be sufficient for an attacker to recover the private DSA key.
4830 This issue was reported by César Pereida (Aalto University), Billy Brumley
4831 (Tampere University of Technology), and Yuval Yarom (The University of
4832 Adelaide and NICTA).
4837 * DTLS buffered message DoS
4839 In a DTLS connection where handshake messages are delivered out-of-order
4840 those messages that OpenSSL is not yet ready to process will be buffered
4841 for later use. Under certain circumstances, a flaw in the logic means that
4842 those messages do not get removed from the buffer even though the handshake
4843 has been completed. An attacker could force up to approx. 15 messages to
4844 remain in the buffer when they are no longer required. These messages will
4845 be cleared when the DTLS connection is closed. The default maximum size for
4846 a message is 100k. Therefore the attacker could force an additional 1500k
4847 to be consumed per connection. By opening many simulataneous connections an
4848 attacker could cause a DoS attack through memory exhaustion.
4850 This issue was reported to OpenSSL by Quan Luo.
4855 * DTLS replay protection DoS
4857 A flaw in the DTLS replay attack protection mechanism means that records
4858 that arrive for future epochs update the replay protection "window" before
4859 the MAC for the record has been validated. This could be exploited by an
4860 attacker by sending a record for the next epoch (which does not have to
4861 decrypt or have a valid MAC), with a very large sequence number. This means
4862 that all subsequent legitimate packets are dropped causing a denial of
4863 service for a specific DTLS connection.
4865 This issue was reported to OpenSSL by the OCAP audit team.
4870 * Certificate message OOB reads
4872 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4873 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4874 theoretical DoS risk but this has not been observed in practice on common
4877 The messages affected are client certificate, client certificate request
4878 and server certificate. As a result the attack can only be performed
4879 against a client or a server which enables client authentication.
4881 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4886 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4888 * Prevent padding oracle in AES-NI CBC MAC check
4890 A MITM attacker can use a padding oracle attack to decrypt traffic
4891 when the connection uses an AES CBC cipher and the server support
4894 This issue was introduced as part of the fix for Lucky 13 padding
4895 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4896 constant time by making sure that always the same bytes are read and
4897 compared against either the MAC or padding bytes. But it no longer
4898 checked that there was enough data to have both the MAC and padding
4901 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4905 * Fix EVP_EncodeUpdate overflow
4907 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4908 Base64 encoding of binary data. If an attacker is able to supply very large
4909 amounts of input data then a length check can overflow resulting in a heap
4912 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4913 the `PEM_write_bio*` family of functions. These are mainly used within the
4914 OpenSSL command line applications, so any application which processes data
4915 from an untrusted source and outputs it as a PEM file should be considered
4916 vulnerable to this issue. User applications that call these APIs directly
4917 with large amounts of untrusted data may also be vulnerable.
4919 This issue was reported by Guido Vranken.
4924 * Fix EVP_EncryptUpdate overflow
4926 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4927 is able to supply very large amounts of input data after a previous call to
4928 EVP_EncryptUpdate() with a partial block then a length check can overflow
4929 resulting in a heap corruption. Following an analysis of all OpenSSL
4930 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4931 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4932 the first called function after an EVP_EncryptInit(), and therefore that
4933 specific call must be safe. The second form is where the length passed to
4934 EVP_EncryptUpdate() can be seen from the code to be some small value and
4935 therefore there is no possibility of an overflow. Since all instances are
4936 one of these two forms, it is believed that there can be no overflows in
4937 internal code due to this problem. It should be noted that
4938 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4939 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4940 of these calls have also been analysed too and it is believed there are no
4941 instances in internal usage where an overflow could occur.
4943 This issue was reported by Guido Vranken.
4948 * Prevent ASN.1 BIO excessive memory allocation
4950 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4951 a short invalid encoding can cause allocation of large amounts of memory
4952 potentially consuming excessive resources or exhausting memory.
4954 Any application parsing untrusted data through d2i BIO functions is
4955 affected. The memory based functions such as d2i_X509() are *not* affected.
4956 Since the memory based functions are used by the TLS library, TLS
4957 applications are not affected.
4959 This issue was reported by Brian Carpenter.
4966 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4967 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4968 in arbitrary stack data being returned in the buffer.
4970 This issue was reported by Guido Vranken.
4975 * Modify behavior of ALPN to invoke callback after SNI/servername
4976 callback, such that updates to the SSL_CTX affect ALPN.
4980 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4985 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4986 methods are enabled and ssl2 is disabled the methods return NULL.
4990 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4992 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4993 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4994 provide any "EXPORT" or "LOW" strength ciphers.
4998 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4999 is by default disabled at build-time. Builds that are not configured with
5000 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5001 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5002 will need to explicitly call either of:
5004 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5006 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5008 as appropriate. Even if either of those is used, or the application
5009 explicitly uses the version-specific SSLv2_method() or its client and
5010 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5011 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5012 ciphers, and SSLv2 56-bit DES are no longer available.
5017 * Fix a double-free in DSA code
5019 A double free bug was discovered when OpenSSL parses malformed DSA private
5020 keys and could lead to a DoS attack or memory corruption for applications
5021 that receive DSA private keys from untrusted sources. This scenario is
5024 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5030 * Disable SRP fake user seed to address a server memory leak.
5032 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5034 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5035 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5036 was changed to ignore the "fake user" SRP seed, even if the seed
5039 Users should use SRP_VBASE_get1_by_user instead. Note that in
5040 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5041 also that even though configuring the SRP seed attempts to hide
5042 invalid usernames by continuing the handshake with fake
5043 credentials, this behaviour is not constant time and no strong
5044 guarantees are made that the handshake is indistinguishable from
5045 that of a valid user.
5050 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5052 In the BN_hex2bn function the number of hex digits is calculated using an
5053 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5054 large values of `i` this can result in `bn_expand` not allocating any
5055 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5056 field as NULL leading to a subsequent NULL ptr deref. For very large values
5057 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5058 In this case memory is allocated to the internal BIGNUM data field, but it
5059 is insufficiently sized leading to heap corruption. A similar issue exists
5060 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5061 is ever called by user applications with very large untrusted hex/dec data.
5062 This is anticipated to be a rare occurrence.
5064 All OpenSSL internal usage of these functions use data that is not expected
5065 to be untrusted, e.g. config file data or application command line
5066 arguments. If user developed applications generate config file data based
5067 on untrusted data then it is possible that this could also lead to security
5068 consequences. This is also anticipated to be rare.
5070 This issue was reported to OpenSSL by Guido Vranken.
5075 * Fix memory issues in `BIO_*printf` functions
5077 The internal `fmtstr` function used in processing a "%s" format string in
5078 the `BIO_*printf` functions could overflow while calculating the length of a
5079 string and cause an OOB read when printing very long strings.
5081 Additionally the internal `doapr_outch` function can attempt to write to an
5082 OOB memory location (at an offset from the NULL pointer) in the event of a
5083 memory allocation failure. In 1.0.2 and below this could be caused where
5084 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5085 could be in processing a very long "%s" format string. Memory leaks can
5088 The first issue may mask the second issue dependent on compiler behaviour.
5089 These problems could enable attacks where large amounts of untrusted data
5090 is passed to the `BIO_*printf` functions. If applications use these functions
5091 in this way then they could be vulnerable. OpenSSL itself uses these
5092 functions when printing out human-readable dumps of ASN.1 data. Therefore
5093 applications that print this data could be vulnerable if the data is from
5094 untrusted sources. OpenSSL command line applications could also be
5095 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5096 as command line arguments.
5098 Libssl is not considered directly vulnerable. Additionally certificates etc
5099 received via remote connections via libssl are also unlikely to be able to
5100 trigger these issues because of message size limits enforced within libssl.
5102 This issue was reported to OpenSSL Guido Vranken.
5107 * Side channel attack on modular exponentiation
5109 A side-channel attack was found which makes use of cache-bank conflicts on
5110 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5111 of RSA keys. The ability to exploit this issue is limited as it relies on
5112 an attacker who has control of code in a thread running on the same
5113 hyper-threaded core as the victim thread which is performing decryptions.
5115 This issue was reported to OpenSSL by Yuval Yarom, The University of
5116 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5117 Nadia Heninger, University of Pennsylvania with more information at
5118 <http://cachebleed.info>.
5123 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5124 if no keysize is specified with default_bits. This fixes an
5125 omission in an earlier change that changed all RSA/DSA key generation
5126 commands to use 2048 bits by default.
5130 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5132 * DH small subgroups
5134 Historically OpenSSL only ever generated DH parameters based on "safe"
5135 primes. More recently (in version 1.0.2) support was provided for
5136 generating X9.42 style parameter files such as those required for RFC 5114
5137 support. The primes used in such files may not be "safe". Where an
5138 application is using DH configured with parameters based on primes that are
5139 not "safe" then an attacker could use this fact to find a peer's private
5140 DH exponent. This attack requires that the attacker complete multiple
5141 handshakes in which the peer uses the same private DH exponent. For example
5142 this could be used to discover a TLS server's private DH exponent if it's
5143 reusing the private DH exponent or it's using a static DH ciphersuite.
5145 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5146 TLS. It is not on by default. If the option is not set then the server
5147 reuses the same private DH exponent for the life of the server process and
5148 would be vulnerable to this attack. It is believed that many popular
5149 applications do set this option and would therefore not be at risk.
5151 The fix for this issue adds an additional check where a "q" parameter is
5152 available (as is the case in X9.42 based parameters). This detects the
5153 only known attack, and is the only possible defense for static DH
5154 ciphersuites. This could have some performance impact.
5156 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5157 default and cannot be disabled. This could have some performance impact.
5159 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5164 * SSLv2 doesn't block disabled ciphers
5166 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5167 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5168 been disabled, provided that the SSLv2 protocol was not also disabled via
5171 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5172 and Sebastian Schinzel.
5177 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5179 * BN_mod_exp may produce incorrect results on x86_64
5181 There is a carry propagating bug in the x86_64 Montgomery squaring
5182 procedure. No EC algorithms are affected. Analysis suggests that attacks
5183 against RSA and DSA as a result of this defect would be very difficult to
5184 perform and are not believed likely. Attacks against DH are considered just
5185 feasible (although very difficult) because most of the work necessary to
5186 deduce information about a private key may be performed offline. The amount
5187 of resources required for such an attack would be very significant and
5188 likely only accessible to a limited number of attackers. An attacker would
5189 additionally need online access to an unpatched system using the target
5190 private key in a scenario with persistent DH parameters and a private
5191 key that is shared between multiple clients. For example this can occur by
5192 default in OpenSSL DHE based SSL/TLS ciphersuites.
5194 This issue was reported to OpenSSL by Hanno Böck.
5199 * Certificate verify crash with missing PSS parameter
5201 The signature verification routines will crash with a NULL pointer
5202 dereference if presented with an ASN.1 signature using the RSA PSS
5203 algorithm and absent mask generation function parameter. Since these
5204 routines are used to verify certificate signature algorithms this can be
5205 used to crash any certificate verification operation and exploited in a
5206 DoS attack. Any application which performs certificate verification is
5207 vulnerable including OpenSSL clients and servers which enable client
5210 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5215 * X509_ATTRIBUTE memory leak
5217 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5218 memory. This structure is used by the PKCS#7 and CMS routines so any
5219 application which reads PKCS#7 or CMS data from untrusted sources is
5220 affected. SSL/TLS is not affected.
5222 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5228 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5229 This changes the decoding behaviour for some invalid messages,
5230 though the change is mostly in the more lenient direction, and
5231 legacy behaviour is preserved as much as possible.
5235 * In DSA_generate_parameters_ex, if the provided seed is too short,
5238 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5240 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5242 * Alternate chains certificate forgery
5244 During certificate verification, OpenSSL will attempt to find an
5245 alternative certificate chain if the first attempt to build such a chain
5246 fails. An error in the implementation of this logic can mean that an
5247 attacker could cause certain checks on untrusted certificates to be
5248 bypassed, such as the CA flag, enabling them to use a valid leaf
5249 certificate to act as a CA and "issue" an invalid certificate.
5251 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5256 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5258 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5259 incompatibility in the handling of HMAC. The previous ABI has now been
5264 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5266 * Malformed ECParameters causes infinite loop
5268 When processing an ECParameters structure OpenSSL enters an infinite loop
5269 if the curve specified is over a specially malformed binary polynomial
5272 This can be used to perform denial of service against any
5273 system which processes public keys, certificate requests or
5274 certificates. This includes TLS clients and TLS servers with
5275 client authentication enabled.
5277 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5282 * Exploitable out-of-bounds read in X509_cmp_time
5284 X509_cmp_time does not properly check the length of the ASN1_TIME
5285 string and can read a few bytes out of bounds. In addition,
5286 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5289 An attacker can use this to craft malformed certificates and CRLs of
5290 various sizes and potentially cause a segmentation fault, resulting in
5291 a DoS on applications that verify certificates or CRLs. TLS clients
5292 that verify CRLs are affected. TLS clients and servers with client
5293 authentication enabled may be affected if they use custom verification
5296 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5297 independently by Hanno Böck.
5302 * PKCS7 crash with missing EnvelopedContent
5304 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5305 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5306 with missing content and trigger a NULL pointer dereference on parsing.
5308 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5309 structures from untrusted sources are affected. OpenSSL clients and
5310 servers are not affected.
5312 This issue was reported to OpenSSL by Michal Zalewski (Google).
5317 * CMS verify infinite loop with unknown hash function
5319 When verifying a signedData message the CMS code can enter an infinite loop
5320 if presented with an unknown hash function OID. This can be used to perform
5321 denial of service against any system which verifies signedData messages using
5323 This issue was reported to OpenSSL by Johannes Bauer.
5328 * Race condition handling NewSessionTicket
5330 If a NewSessionTicket is received by a multi-threaded client when attempting to
5331 reuse a previous ticket then a race condition can occur potentially leading to
5332 a double free of the ticket data.
5337 * Only support 256-bit or stronger elliptic curves with the
5338 'ecdh_auto' setting (server) or by default (client). Of supported
5339 curves, prefer P-256 (both).
5343 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5345 * ClientHello sigalgs DoS fix
5347 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5348 invalid signature algorithms extension a NULL pointer dereference will
5349 occur. This can be exploited in a DoS attack against the server.
5351 This issue was was reported to OpenSSL by David Ramos of Stanford
5355 *Stephen Henson and Matt Caswell*
5357 * Multiblock corrupted pointer fix
5359 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5360 feature only applies on 64 bit x86 architecture platforms that support AES
5361 NI instructions. A defect in the implementation of "multiblock" can cause
5362 OpenSSL's internal write buffer to become incorrectly set to NULL when
5363 using non-blocking IO. Typically, when the user application is using a
5364 socket BIO for writing, this will only result in a failed connection.
5365 However if some other BIO is used then it is likely that a segmentation
5366 fault will be triggered, thus enabling a potential DoS attack.
5368 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5373 * Segmentation fault in DTLSv1_listen fix
5375 The DTLSv1_listen function is intended to be stateless and processes the
5376 initial ClientHello from many peers. It is common for user code to loop
5377 over the call to DTLSv1_listen until a valid ClientHello is received with
5378 an associated cookie. A defect in the implementation of DTLSv1_listen means
5379 that state is preserved in the SSL object from one invocation to the next
5380 that can lead to a segmentation fault. Errors processing the initial
5381 ClientHello can trigger this scenario. An example of such an error could be
5382 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5385 This issue was reported to OpenSSL by Per Allansson.
5390 * Segmentation fault in ASN1_TYPE_cmp fix
5392 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5393 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5394 certificate signature algorithm consistency this can be used to crash any
5395 certificate verification operation and exploited in a DoS attack. Any
5396 application which performs certificate verification is vulnerable including
5397 OpenSSL clients and servers which enable client authentication.
5402 * Segmentation fault for invalid PSS parameters fix
5404 The signature verification routines will crash with a NULL pointer
5405 dereference if presented with an ASN.1 signature using the RSA PSS
5406 algorithm and invalid parameters. Since these routines are used to verify
5407 certificate signature algorithms this can be used to crash any
5408 certificate verification operation and exploited in a DoS attack. Any
5409 application which performs certificate verification is vulnerable including
5410 OpenSSL clients and servers which enable client authentication.
5412 This issue was was reported to OpenSSL by Brian Carpenter.
5417 * ASN.1 structure reuse memory corruption fix
5419 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5420 memory corruption via an invalid write. Such reuse is and has been
5421 strongly discouraged and is believed to be rare.
5423 Applications that parse structures containing CHOICE or ANY DEFINED BY
5424 components may be affected. Certificate parsing (d2i_X509 and related
5425 functions) are however not affected. OpenSSL clients and servers are
5431 * PKCS7 NULL pointer dereferences fix
5433 The PKCS#7 parsing code does not handle missing outer ContentInfo
5434 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5435 missing content and trigger a NULL pointer dereference on parsing.
5437 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5438 otherwise parse PKCS#7 structures from untrusted sources are
5439 affected. OpenSSL clients and servers are not affected.
5441 This issue was reported to OpenSSL by Michal Zalewski (Google).
5446 * DoS via reachable assert in SSLv2 servers fix
5448 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5449 servers that both support SSLv2 and enable export cipher suites by sending
5450 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5452 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5453 (OpenSSL development team).
5458 * Empty CKE with client auth and DHE fix
5460 If client auth is used then a server can seg fault in the event of a DHE
5461 ciphersuite being selected and a zero length ClientKeyExchange message
5462 being sent by the client. This could be exploited in a DoS attack.
5467 * Handshake with unseeded PRNG fix
5469 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5470 with an unseeded PRNG. The conditions are:
5471 - The client is on a platform where the PRNG has not been seeded
5472 automatically, and the user has not seeded manually
5473 - A protocol specific client method version has been used (i.e. not
5474 SSL_client_methodv23)
5475 - A ciphersuite is used that does not require additional random data from
5476 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5478 If the handshake succeeds then the client random that has been used will
5479 have been generated from a PRNG with insufficient entropy and therefore the
5480 output may be predictable.
5482 For example using the following command with an unseeded openssl will
5483 succeed on an unpatched platform:
5485 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5490 * Use After Free following d2i_ECPrivatekey error fix
5492 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5493 could cause a use after free condition. This, in turn, could cause a double
5494 free in several private key parsing functions (such as d2i_PrivateKey
5495 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5496 for applications that receive EC private keys from untrusted
5497 sources. This scenario is considered rare.
5499 This issue was discovered by the BoringSSL project and fixed in their
5505 * X509_to_X509_REQ NULL pointer deref fix
5507 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5508 the certificate key is invalid. This function is rarely used in practice.
5510 This issue was discovered by Brian Carpenter.
5515 * Removed the export ciphers from the DEFAULT ciphers
5519 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5521 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5522 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5523 So far those who have to target multiple platforms would compromise
5524 and argue that binary targeting say ARMv5 would still execute on
5525 ARMv8. "Universal" build resolves this compromise by providing
5526 near-optimal performance even on newer platforms.
5530 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5531 (other platforms pending).
5533 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5535 * Add support for the SignedCertificateTimestampList certificate and
5536 OCSP response extensions from RFC6962.
5540 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5541 for corner cases. (Certain input points at infinity could lead to
5542 bogus results, with non-infinity inputs mapped to infinity too.)
5546 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5547 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5548 common cases are optimized and there still is room for further
5549 improvements. Vector Permutation AES for Altivec is also added.
5553 * Add support for little-endian ppc64 Linux target.
5555 *Marcelo Cerri (IBM)*
5557 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5558 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5559 are optimized and there still is room for further improvements.
5560 Both 32- and 64-bit modes are supported.
5562 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5564 * Improved ARMv7 NEON support.
5568 * Support for SPARC Architecture 2011 crypto extensions, first
5569 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5570 SHA256/512, MD5, GHASH and modular exponentiation.
5572 *Andy Polyakov, David Miller*
5574 * Accelerated modular exponentiation for Intel processors, a.k.a.
5577 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5579 * Support for new and upcoming Intel processors, including AVX2,
5580 BMI and SHA ISA extensions. This includes additional "stitched"
5581 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5584 This work was sponsored by Intel Corp.
5588 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5589 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5590 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5594 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5595 this fixes a limitation in previous versions of OpenSSL.
5599 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5600 MGF1 digest and OAEP label.
5604 * Add EVP support for key wrapping algorithms, to avoid problems with
5605 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5606 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5607 algorithms and include tests cases.
5611 * Add functions to allocate and set the fields of an ECDSA_METHOD
5614 *Douglas E. Engert, Steve Henson*
5616 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5617 difference in days and seconds between two tm or ASN1_TIME structures.
5621 * Add -rev test option to s_server to just reverse order of characters
5622 received by client and send back to server. Also prints an abbreviated
5623 summary of the connection parameters.
5627 * New option -brief for s_client and s_server to print out a brief summary
5628 of connection parameters.
5632 * Add callbacks for arbitrary TLS extensions.
5634 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5636 * New option -crl_download in several openssl utilities to download CRLs
5637 from CRLDP extension in certificates.
5641 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5645 * New function X509_CRL_diff to generate a delta CRL from the difference
5646 of two full CRLs. Add support to "crl" utility.
5650 * New functions to set lookup_crls function and to retrieve
5651 X509_STORE from X509_STORE_CTX.
5655 * Print out deprecated issuer and subject unique ID fields in
5660 * Extend OCSP I/O functions so they can be used for simple general purpose
5661 HTTP as well as OCSP. New wrapper function which can be used to download
5662 CRLs using the OCSP API.
5666 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5670 * `SSL_CONF*` functions. These provide a common framework for application
5671 configuration using configuration files or command lines.
5675 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5676 message callback and prints the results. Needs compile time option
5677 "enable-ssl-trace". New options to s_client and s_server to enable
5682 * New ctrl and macro to retrieve supported points extensions.
5683 Print out extension in s_server and s_client.
5687 * New functions to retrieve certificate signature and signature
5692 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5697 * New Suite B modes for TLS code. These use and enforce the requirements
5698 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5699 only use Suite B curves. The Suite B modes can be set by using the
5700 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5704 * New chain verification flags for Suite B levels of security. Check
5705 algorithms are acceptable when flags are set in X509_verify_cert.
5709 * Make tls1_check_chain return a set of flags indicating checks passed
5710 by a certificate chain. Add additional tests to handle client
5711 certificates: checks for matching certificate type and issuer name
5716 * If an attempt is made to use a signature algorithm not in the peer
5717 preference list abort the handshake. If client has no suitable
5718 signature algorithms in response to a certificate request do not
5719 use the certificate.
5723 * If server EC tmp key is not in client preference list abort handshake.
5727 * Add support for certificate stores in CERT structure. This makes it
5728 possible to have different stores per SSL structure or one store in
5729 the parent SSL_CTX. Include distinct stores for certificate chain
5730 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5731 to build and store a certificate chain in CERT structure: returning
5732 an error if the chain cannot be built: this will allow applications
5733 to test if a chain is correctly configured.
5735 Note: if the CERT based stores are not set then the parent SSL_CTX
5736 store is used to retain compatibility with existing behaviour.
5740 * New function ssl_set_client_disabled to set a ciphersuite disabled
5741 mask based on the current session, check mask when sending client
5742 hello and checking the requested ciphersuite.
5746 * New ctrls to retrieve and set certificate types in a certificate
5747 request message. Print out received values in s_client. If certificate
5748 types is not set with custom values set sensible values based on
5749 supported signature algorithms.
5753 * Support for distinct client and server supported signature algorithms.
5757 * Add certificate callback. If set this is called whenever a certificate
5758 is required by client or server. An application can decide which
5759 certificate chain to present based on arbitrary criteria: for example
5760 supported signature algorithms. Add very simple example to s_server.
5761 This fixes many of the problems and restrictions of the existing client
5762 certificate callback: for example you can now clear an existing
5763 certificate and specify the whole chain.
5767 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5768 the certificate can be used for (if anything). Set valid_flags field
5769 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5770 to have similar checks in it.
5772 Add new "cert_flags" field to CERT structure and include a "strict mode".
5773 This enforces some TLS certificate requirements (such as only permitting
5774 certificate signature algorithms contained in the supported algorithms
5775 extension) which some implementations ignore: this option should be used
5776 with caution as it could cause interoperability issues.
5780 * Update and tidy signature algorithm extension processing. Work out
5781 shared signature algorithms based on preferences and peer algorithms
5782 and print them out in s_client and s_server. Abort handshake if no
5783 shared signature algorithms.
5787 * Add new functions to allow customised supported signature algorithms
5788 for SSL and SSL_CTX structures. Add options to s_client and s_server
5793 * New function SSL_certs_clear() to delete all references to certificates
5794 from an SSL structure. Before this once a certificate had been added
5795 it couldn't be removed.
5799 * Integrate hostname, email address and IP address checking with certificate
5800 verification. New verify options supporting checking in openssl utility.
5804 * Fixes and wildcard matching support to hostname and email checking
5805 functions. Add manual page.
5807 *Florian Weimer (Red Hat Product Security Team)*
5809 * New functions to check a hostname email or IP address against a
5810 certificate. Add options x509 utility to print results of checks against
5815 * Fix OCSP checking.
5817 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5819 * Initial experimental support for explicitly trusted non-root CAs.
5820 OpenSSL still tries to build a complete chain to a root but if an
5821 intermediate CA has a trust setting included that is used. The first
5822 setting is used: whether to trust (e.g., -addtrust option to the x509
5827 * Add -trusted_first option which attempts to find certificates in the
5828 trusted store even if an untrusted chain is also supplied.
5832 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5833 platform support for Linux and Android.
5837 * Support for linux-x32, ILP32 environment in x86_64 framework.
5841 * Experimental multi-implementation support for FIPS capable OpenSSL.
5842 When in FIPS mode the approved implementations are used as normal,
5843 when not in FIPS mode the internal unapproved versions are used instead.
5844 This means that the FIPS capable OpenSSL isn't forced to use the
5845 (often lower performance) FIPS implementations outside FIPS mode.
5849 * Transparently support X9.42 DH parameters when calling
5850 PEM_read_bio_DHparameters. This means existing applications can handle
5851 the new parameter format automatically.
5855 * Initial experimental support for X9.42 DH parameter format: mainly
5856 to support use of 'q' parameter for RFC5114 parameters.
5860 * Add DH parameters from RFC5114 including test data to dhtest.
5864 * Support for automatic EC temporary key parameter selection. If enabled
5865 the most preferred EC parameters are automatically used instead of
5866 hardcoded fixed parameters. Now a server just has to call:
5867 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5868 support ECDH and use the most appropriate parameters.
5872 * Enhance and tidy EC curve and point format TLS extension code. Use
5873 static structures instead of allocation if default values are used.
5874 New ctrls to set curves we wish to support and to retrieve shared curves.
5875 Print out shared curves in s_server. New options to s_server and s_client
5876 to set list of supported curves.
5880 * New ctrls to retrieve supported signature algorithms and
5881 supported curve values as an array of NIDs. Extend openssl utility
5882 to print out received values.
5886 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5887 between NIDs and the more common NIST names such as "P-256". Enhance
5888 ecparam utility and ECC method to recognise the NIST names for curves.
5892 * Enhance SSL/TLS certificate chain handling to support different
5893 chains for each certificate instead of one chain in the parent SSL_CTX.
5897 * Support for fixed DH ciphersuite client authentication: where both
5898 server and client use DH certificates with common parameters.
5902 * Support for fixed DH ciphersuites: those requiring DH server
5907 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5909 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5910 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5911 X509_CINF_get_signature were reverted post internal team review.
5916 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5918 * OCSP Status Request extension unbounded memory growth
5920 A malicious client can send an excessively large OCSP Status Request
5921 extension. If that client continually requests renegotiation, sending a
5922 large OCSP Status Request extension each time, then there will be unbounded
5923 memory growth on the server. This will eventually lead to a Denial Of
5924 Service attack through memory exhaustion. Servers with a default
5925 configuration are vulnerable even if they do not support OCSP. Builds using
5926 the "no-ocsp" build time option are not affected.
5928 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5933 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5936 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5942 * OOB write in MDC2_Update()
5944 An overflow can occur in MDC2_Update() either if called directly or
5945 through the EVP_DigestUpdate() function using MDC2. If an attacker
5946 is able to supply very large amounts of input data after a previous
5947 call to EVP_EncryptUpdate() with a partial block then a length check
5948 can overflow resulting in a heap corruption.
5950 The amount of data needed is comparable to SIZE_MAX which is impractical
5953 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5958 * Malformed SHA512 ticket DoS
5960 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5961 DoS attack where a malformed ticket will result in an OOB read which will
5964 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5965 a custom server callback and ticket lookup mechanism.
5967 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5972 * OOB write in BN_bn2dec()
5974 The function BN_bn2dec() does not check the return value of BN_div_word().
5975 This can cause an OOB write if an application uses this function with an
5976 overly large BIGNUM. This could be a problem if an overly large certificate
5977 or CRL is printed out from an untrusted source. TLS is not affected because
5978 record limits will reject an oversized certificate before it is parsed.
5980 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5985 * OOB read in TS_OBJ_print_bio()
5987 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5988 the total length the OID text representation would use and not the amount
5989 of data written. This will result in OOB reads when large OIDs are
5992 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5997 * Pointer arithmetic undefined behaviour
5999 Avoid some undefined pointer arithmetic
6001 A common idiom in the codebase is to check limits in the following manner:
6004 Where "p" points to some malloc'd data of SIZE bytes and
6007 "len" here could be from some externally supplied data (e.g. from a TLS
6010 The rules of C pointer arithmetic are such that "p + len" is only well
6011 defined where len <= SIZE. Therefore the above idiom is actually
6012 undefined behaviour.
6014 For example this could cause problems if some malloc implementation
6015 provides an address for "p" such that "p + len" actually overflows for
6016 values of len that are too big and therefore p + len < limit.
6018 This issue was reported to OpenSSL by Guido Vranken
6023 * Constant time flag not preserved in DSA signing
6025 Operations in the DSA signing algorithm should run in constant time in
6026 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6027 implementation means that a non-constant time codepath is followed for
6028 certain operations. This has been demonstrated through a cache-timing
6029 attack to be sufficient for an attacker to recover the private DSA key.
6031 This issue was reported by César Pereida (Aalto University), Billy Brumley
6032 (Tampere University of Technology), and Yuval Yarom (The University of
6033 Adelaide and NICTA).
6038 * DTLS buffered message DoS
6040 In a DTLS connection where handshake messages are delivered out-of-order
6041 those messages that OpenSSL is not yet ready to process will be buffered
6042 for later use. Under certain circumstances, a flaw in the logic means that
6043 those messages do not get removed from the buffer even though the handshake
6044 has been completed. An attacker could force up to approx. 15 messages to
6045 remain in the buffer when they are no longer required. These messages will
6046 be cleared when the DTLS connection is closed. The default maximum size for
6047 a message is 100k. Therefore the attacker could force an additional 1500k
6048 to be consumed per connection. By opening many simulataneous connections an
6049 attacker could cause a DoS attack through memory exhaustion.
6051 This issue was reported to OpenSSL by Quan Luo.
6056 * DTLS replay protection DoS
6058 A flaw in the DTLS replay attack protection mechanism means that records
6059 that arrive for future epochs update the replay protection "window" before
6060 the MAC for the record has been validated. This could be exploited by an
6061 attacker by sending a record for the next epoch (which does not have to
6062 decrypt or have a valid MAC), with a very large sequence number. This means
6063 that all subsequent legitimate packets are dropped causing a denial of
6064 service for a specific DTLS connection.
6066 This issue was reported to OpenSSL by the OCAP audit team.
6071 * Certificate message OOB reads
6073 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6074 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6075 theoretical DoS risk but this has not been observed in practice on common
6078 The messages affected are client certificate, client certificate request
6079 and server certificate. As a result the attack can only be performed
6080 against a client or a server which enables client authentication.
6082 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6087 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6089 * Prevent padding oracle in AES-NI CBC MAC check
6091 A MITM attacker can use a padding oracle attack to decrypt traffic
6092 when the connection uses an AES CBC cipher and the server support
6095 This issue was introduced as part of the fix for Lucky 13 padding
6096 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6097 constant time by making sure that always the same bytes are read and
6098 compared against either the MAC or padding bytes. But it no longer
6099 checked that there was enough data to have both the MAC and padding
6102 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6107 * Fix EVP_EncodeUpdate overflow
6109 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6110 Base64 encoding of binary data. If an attacker is able to supply very large
6111 amounts of input data then a length check can overflow resulting in a heap
6114 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6115 the `PEM_write_bio*` family of functions. These are mainly used within the
6116 OpenSSL command line applications, so any application which processes data
6117 from an untrusted source and outputs it as a PEM file should be considered
6118 vulnerable to this issue. User applications that call these APIs directly
6119 with large amounts of untrusted data may also be vulnerable.
6121 This issue was reported by Guido Vranken.
6126 * Fix EVP_EncryptUpdate overflow
6128 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6129 is able to supply very large amounts of input data after a previous call to
6130 EVP_EncryptUpdate() with a partial block then a length check can overflow
6131 resulting in a heap corruption. Following an analysis of all OpenSSL
6132 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6133 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6134 the first called function after an EVP_EncryptInit(), and therefore that
6135 specific call must be safe. The second form is where the length passed to
6136 EVP_EncryptUpdate() can be seen from the code to be some small value and
6137 therefore there is no possibility of an overflow. Since all instances are
6138 one of these two forms, it is believed that there can be no overflows in
6139 internal code due to this problem. It should be noted that
6140 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6141 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6142 of these calls have also been analysed too and it is believed there are no
6143 instances in internal usage where an overflow could occur.
6145 This issue was reported by Guido Vranken.
6150 * Prevent ASN.1 BIO excessive memory allocation
6152 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6153 a short invalid encoding can casuse allocation of large amounts of memory
6154 potentially consuming excessive resources or exhausting memory.
6156 Any application parsing untrusted data through d2i BIO functions is
6157 affected. The memory based functions such as d2i_X509() are *not* affected.
6158 Since the memory based functions are used by the TLS library, TLS
6159 applications are not affected.
6161 This issue was reported by Brian Carpenter.
6168 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6169 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6170 in arbitrary stack data being returned in the buffer.
6172 This issue was reported by Guido Vranken.
6177 * Modify behavior of ALPN to invoke callback after SNI/servername
6178 callback, such that updates to the SSL_CTX affect ALPN.
6182 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6187 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6188 methods are enabled and ssl2 is disabled the methods return NULL.
6192 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6194 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6195 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6196 provide any "EXPORT" or "LOW" strength ciphers.
6200 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6201 is by default disabled at build-time. Builds that are not configured with
6202 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6203 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6204 will need to explicitly call either of:
6206 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6208 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6210 as appropriate. Even if either of those is used, or the application
6211 explicitly uses the version-specific SSLv2_method() or its client and
6212 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6213 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6214 ciphers, and SSLv2 56-bit DES are no longer available.
6219 * Fix a double-free in DSA code
6221 A double free bug was discovered when OpenSSL parses malformed DSA private
6222 keys and could lead to a DoS attack or memory corruption for applications
6223 that receive DSA private keys from untrusted sources. This scenario is
6226 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6232 * Disable SRP fake user seed to address a server memory leak.
6234 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6236 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6237 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6238 was changed to ignore the "fake user" SRP seed, even if the seed
6241 Users should use SRP_VBASE_get1_by_user instead. Note that in
6242 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6243 also that even though configuring the SRP seed attempts to hide
6244 invalid usernames by continuing the handshake with fake
6245 credentials, this behaviour is not constant time and no strong
6246 guarantees are made that the handshake is indistinguishable from
6247 that of a valid user.
6252 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6254 In the BN_hex2bn function the number of hex digits is calculated using an
6255 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6256 large values of `i` this can result in `bn_expand` not allocating any
6257 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6258 field as NULL leading to a subsequent NULL ptr deref. For very large values
6259 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6260 In this case memory is allocated to the internal BIGNUM data field, but it
6261 is insufficiently sized leading to heap corruption. A similar issue exists
6262 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6263 is ever called by user applications with very large untrusted hex/dec data.
6264 This is anticipated to be a rare occurrence.
6266 All OpenSSL internal usage of these functions use data that is not expected
6267 to be untrusted, e.g. config file data or application command line
6268 arguments. If user developed applications generate config file data based
6269 on untrusted data then it is possible that this could also lead to security
6270 consequences. This is also anticipated to be rare.
6272 This issue was reported to OpenSSL by Guido Vranken.
6277 * Fix memory issues in `BIO_*printf` functions
6279 The internal `fmtstr` function used in processing a "%s" format string in
6280 the `BIO_*printf` functions could overflow while calculating the length of a
6281 string and cause an OOB read when printing very long strings.
6283 Additionally the internal `doapr_outch` function can attempt to write to an
6284 OOB memory location (at an offset from the NULL pointer) in the event of a
6285 memory allocation failure. In 1.0.2 and below this could be caused where
6286 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6287 could be in processing a very long "%s" format string. Memory leaks can
6290 The first issue may mask the second issue dependent on compiler behaviour.
6291 These problems could enable attacks where large amounts of untrusted data
6292 is passed to the `BIO_*printf` functions. If applications use these functions
6293 in this way then they could be vulnerable. OpenSSL itself uses these
6294 functions when printing out human-readable dumps of ASN.1 data. Therefore
6295 applications that print this data could be vulnerable if the data is from
6296 untrusted sources. OpenSSL command line applications could also be
6297 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6298 as command line arguments.
6300 Libssl is not considered directly vulnerable. Additionally certificates etc
6301 received via remote connections via libssl are also unlikely to be able to
6302 trigger these issues because of message size limits enforced within libssl.
6304 This issue was reported to OpenSSL Guido Vranken.
6309 * Side channel attack on modular exponentiation
6311 A side-channel attack was found which makes use of cache-bank conflicts on
6312 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6313 of RSA keys. The ability to exploit this issue is limited as it relies on
6314 an attacker who has control of code in a thread running on the same
6315 hyper-threaded core as the victim thread which is performing decryptions.
6317 This issue was reported to OpenSSL by Yuval Yarom, The University of
6318 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6319 Nadia Heninger, University of Pennsylvania with more information at
6320 <http://cachebleed.info>.
6325 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6326 if no keysize is specified with default_bits. This fixes an
6327 omission in an earlier change that changed all RSA/DSA key generation
6328 commands to use 2048 bits by default.
6332 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6334 * Protection for DH small subgroup attacks
6336 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6337 switched on by default and cannot be disabled. This could have some
6342 * SSLv2 doesn't block disabled ciphers
6344 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6345 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6346 been disabled, provided that the SSLv2 protocol was not also disabled via
6349 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6350 and Sebastian Schinzel.
6355 * Reject DH handshakes with parameters shorter than 1024 bits.
6359 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6361 * Certificate verify crash with missing PSS parameter
6363 The signature verification routines will crash with a NULL pointer
6364 dereference if presented with an ASN.1 signature using the RSA PSS
6365 algorithm and absent mask generation function parameter. Since these
6366 routines are used to verify certificate signature algorithms this can be
6367 used to crash any certificate verification operation and exploited in a
6368 DoS attack. Any application which performs certificate verification is
6369 vulnerable including OpenSSL clients and servers which enable client
6372 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6377 * X509_ATTRIBUTE memory leak
6379 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6380 memory. This structure is used by the PKCS#7 and CMS routines so any
6381 application which reads PKCS#7 or CMS data from untrusted sources is
6382 affected. SSL/TLS is not affected.
6384 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6390 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6391 This changes the decoding behaviour for some invalid messages,
6392 though the change is mostly in the more lenient direction, and
6393 legacy behaviour is preserved as much as possible.
6397 * In DSA_generate_parameters_ex, if the provided seed is too short,
6398 use a random seed, as already documented.
6400 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6402 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6404 * Alternate chains certificate forgery
6406 During certificate verfification, OpenSSL will attempt to find an
6407 alternative certificate chain if the first attempt to build such a chain
6408 fails. An error in the implementation of this logic can mean that an
6409 attacker could cause certain checks on untrusted certificates to be
6410 bypassed, such as the CA flag, enabling them to use a valid leaf
6411 certificate to act as a CA and "issue" an invalid certificate.
6413 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6419 * Race condition handling PSK identify hint
6421 If PSK identity hints are received by a multi-threaded client then
6422 the values are wrongly updated in the parent SSL_CTX structure. This can
6423 result in a race condition potentially leading to a double free of the
6429 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6431 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6432 incompatibility in the handling of HMAC. The previous ABI has now been
6435 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6437 * Malformed ECParameters causes infinite loop
6439 When processing an ECParameters structure OpenSSL enters an infinite loop
6440 if the curve specified is over a specially malformed binary polynomial
6443 This can be used to perform denial of service against any
6444 system which processes public keys, certificate requests or
6445 certificates. This includes TLS clients and TLS servers with
6446 client authentication enabled.
6448 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6453 * Exploitable out-of-bounds read in X509_cmp_time
6455 X509_cmp_time does not properly check the length of the ASN1_TIME
6456 string and can read a few bytes out of bounds. In addition,
6457 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6460 An attacker can use this to craft malformed certificates and CRLs of
6461 various sizes and potentially cause a segmentation fault, resulting in
6462 a DoS on applications that verify certificates or CRLs. TLS clients
6463 that verify CRLs are affected. TLS clients and servers with client
6464 authentication enabled may be affected if they use custom verification
6467 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6468 independently by Hanno Böck.
6473 * PKCS7 crash with missing EnvelopedContent
6475 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6476 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6477 with missing content and trigger a NULL pointer dereference on parsing.
6479 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6480 structures from untrusted sources are affected. OpenSSL clients and
6481 servers are not affected.
6483 This issue was reported to OpenSSL by Michal Zalewski (Google).
6488 * CMS verify infinite loop with unknown hash function
6490 When verifying a signedData message the CMS code can enter an infinite loop
6491 if presented with an unknown hash function OID. This can be used to perform
6492 denial of service against any system which verifies signedData messages using
6494 This issue was reported to OpenSSL by Johannes Bauer.
6499 * Race condition handling NewSessionTicket
6501 If a NewSessionTicket is received by a multi-threaded client when attempting to
6502 reuse a previous ticket then a race condition can occur potentially leading to
6503 a double free of the ticket data.
6508 * Reject DH handshakes with parameters shorter than 768 bits.
6510 *Kurt Roeckx and Emilia Kasper*
6512 * dhparam: generate 2048-bit parameters by default.
6514 *Kurt Roeckx and Emilia Kasper*
6516 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6518 * Segmentation fault in ASN1_TYPE_cmp fix
6520 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6521 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6522 certificate signature algorithm consistency this can be used to crash any
6523 certificate verification operation and exploited in a DoS attack. Any
6524 application which performs certificate verification is vulnerable including
6525 OpenSSL clients and servers which enable client authentication.
6530 * ASN.1 structure reuse memory corruption fix
6532 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6533 memory corruption via an invalid write. Such reuse is and has been
6534 strongly discouraged and is believed to be rare.
6536 Applications that parse structures containing CHOICE or ANY DEFINED BY
6537 components may be affected. Certificate parsing (d2i_X509 and related
6538 functions) are however not affected. OpenSSL clients and servers are
6544 * PKCS7 NULL pointer dereferences fix
6546 The PKCS#7 parsing code does not handle missing outer ContentInfo
6547 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6548 missing content and trigger a NULL pointer dereference on parsing.
6550 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6551 otherwise parse PKCS#7 structures from untrusted sources are
6552 affected. OpenSSL clients and servers are not affected.
6554 This issue was reported to OpenSSL by Michal Zalewski (Google).
6559 * DoS via reachable assert in SSLv2 servers fix
6561 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6562 servers that both support SSLv2 and enable export cipher suites by sending
6563 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6565 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6566 (OpenSSL development team).
6571 * Use After Free following d2i_ECPrivatekey error fix
6573 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6574 could cause a use after free condition. This, in turn, could cause a double
6575 free in several private key parsing functions (such as d2i_PrivateKey
6576 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6577 for applications that receive EC private keys from untrusted
6578 sources. This scenario is considered rare.
6580 This issue was discovered by the BoringSSL project and fixed in their
6586 * X509_to_X509_REQ NULL pointer deref fix
6588 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6589 the certificate key is invalid. This function is rarely used in practice.
6591 This issue was discovered by Brian Carpenter.
6596 * Removed the export ciphers from the DEFAULT ciphers
6600 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6602 * Build fixes for the Windows and OpenVMS platforms
6604 *Matt Caswell and Richard Levitte*
6606 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6608 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6609 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6610 dereference. This could lead to a Denial Of Service attack. Thanks to
6611 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6616 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6617 dtls1_buffer_record function under certain conditions. In particular this
6618 could occur if an attacker sent repeated DTLS records with the same
6619 sequence number but for the next epoch. The memory leak could be exploited
6620 by an attacker in a Denial of Service attack through memory exhaustion.
6621 Thanks to Chris Mueller for reporting this issue.
6626 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6627 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6628 method would be set to NULL which could later result in a NULL pointer
6629 dereference. Thanks to Frank Schmirler for reporting this issue.
6634 * Abort handshake if server key exchange message is omitted for ephemeral
6637 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6638 reporting this issue.
6643 * Remove non-export ephemeral RSA code on client and server. This code
6644 violated the TLS standard by allowing the use of temporary RSA keys in
6645 non-export ciphersuites and could be used by a server to effectively
6646 downgrade the RSA key length used to a value smaller than the server
6647 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6648 INRIA or reporting this issue.
6653 * Fixed issue where DH client certificates are accepted without verification.
6654 An OpenSSL server will accept a DH certificate for client authentication
6655 without the certificate verify message. This effectively allows a client to
6656 authenticate without the use of a private key. This only affects servers
6657 which trust a client certificate authority which issues certificates
6658 containing DH keys: these are extremely rare and hardly ever encountered.
6659 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6665 * Ensure that the session ID context of an SSL is updated when its
6666 SSL_CTX is updated via SSL_set_SSL_CTX.
6668 The session ID context is typically set from the parent SSL_CTX,
6669 and can vary with the CTX.
6673 * Fix various certificate fingerprint issues.
6675 By using non-DER or invalid encodings outside the signed portion of a
6676 certificate the fingerprint can be changed without breaking the signature.
6677 Although no details of the signed portion of the certificate can be changed
6678 this can cause problems with some applications: e.g. those using the
6679 certificate fingerprint for blacklists.
6681 1. Reject signatures with non zero unused bits.
6683 If the BIT STRING containing the signature has non zero unused bits reject
6684 the signature. All current signature algorithms require zero unused bits.
6686 2. Check certificate algorithm consistency.
6688 Check the AlgorithmIdentifier inside TBS matches the one in the
6689 certificate signature. NB: this will result in signature failure
6690 errors for some broken certificates.
6692 Thanks to Konrad Kraszewski from Google for reporting this issue.
6694 3. Check DSA/ECDSA signatures use DER.
6696 Re-encode DSA/ECDSA signatures and compare with the original received
6697 signature. Return an error if there is a mismatch.
6699 This will reject various cases including garbage after signature
6700 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6701 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6702 (negative or with leading zeroes).
6704 Further analysis was conducted and fixes were developed by Stephen Henson
6705 of the OpenSSL core team.
6711 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6712 results on some platforms, including x86_64. This bug occurs at random
6713 with a very low probability, and is not known to be exploitable in any
6714 way, though its exact impact is difficult to determine. Thanks to Pieter
6715 Wuille (Blockstream) who reported this issue and also suggested an initial
6716 fix. Further analysis was conducted by the OpenSSL development team and
6717 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6718 the OpenSSL core team.
6723 * Do not resume sessions on the server if the negotiated protocol
6724 version does not match the session's version. Resuming with a different
6725 version, while not strictly forbidden by the RFC, is of questionable
6726 sanity and breaks all known clients.
6728 *David Benjamin, Emilia Käsper*
6730 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6731 early CCS messages during renegotiation. (Note that because
6732 renegotiation is encrypted, this early CCS was not exploitable.)
6736 * Tighten client-side session ticket handling during renegotiation:
6737 ensure that the client only accepts a session ticket if the server sends
6738 the extension anew in the ServerHello. Previously, a TLS client would
6739 reuse the old extension state and thus accept a session ticket if one was
6740 announced in the initial ServerHello.
6742 Similarly, ensure that the client requires a session ticket if one
6743 was advertised in the ServerHello. Previously, a TLS client would
6744 ignore a missing NewSessionTicket message.
6748 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6752 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6753 sends a carefully crafted handshake message, to cause OpenSSL to fail
6754 to free up to 64k of memory causing a memory leak. This could be
6755 exploited in a Denial Of Service attack. This issue affects OpenSSL
6756 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6757 whether SRTP is used or configured. Implementations of OpenSSL that
6758 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6760 The fix was developed by the OpenSSL team.
6765 * Session Ticket Memory Leak.
6767 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6768 integrity of that ticket is first verified. In the event of a session
6769 ticket integrity check failing, OpenSSL will fail to free memory
6770 causing a memory leak. By sending a large number of invalid session
6771 tickets an attacker could exploit this issue in a Denial Of Service
6777 * Build option no-ssl3 is incomplete.
6779 When OpenSSL is configured with "no-ssl3" as a build option, servers
6780 could accept and complete a SSL 3.0 handshake, and clients could be
6781 configured to send them.
6784 *Akamai and the OpenSSL team*
6786 * Add support for TLS_FALLBACK_SCSV.
6787 Client applications doing fallback retries should call
6788 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6791 *Adam Langley, Bodo Moeller*
6793 * Add additional DigestInfo checks.
6795 Re-encode DigestInto in DER and check against the original when
6796 verifying RSA signature: this will reject any improperly encoded
6797 DigestInfo structures.
6799 Note: this is a precautionary measure and no attacks are currently known.
6803 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6805 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6806 SRP code can be overrun an internal buffer. Add sanity check that
6807 g, A, B < N to SRP code.
6809 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6810 Group for discovering this issue.
6815 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6816 TLS 1.0 instead of higher protocol versions when the ClientHello message
6817 is badly fragmented. This allows a man-in-the-middle attacker to force a
6818 downgrade to TLS 1.0 even if both the server and the client support a
6819 higher protocol version, by modifying the client's TLS records.
6821 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6822 researching this issue.
6827 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6828 to a denial of service attack. A malicious server can crash the client
6829 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6830 ciphersuite and sending carefully crafted handshake messages.
6832 Thanks to Felix Gröbert (Google) for discovering and researching this
6838 * By sending carefully crafted DTLS packets an attacker could cause openssl
6839 to leak memory. This can be exploited through a Denial of Service attack.
6840 Thanks to Adam Langley for discovering and researching this issue.
6845 * An attacker can force openssl to consume large amounts of memory whilst
6846 processing DTLS handshake messages. This can be exploited through a
6847 Denial of Service attack.
6848 Thanks to Adam Langley for discovering and researching this issue.
6853 * An attacker can force an error condition which causes openssl to crash
6854 whilst processing DTLS packets due to memory being freed twice. This
6855 can be exploited through a Denial of Service attack.
6856 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6862 * If a multithreaded client connects to a malicious server using a resumed
6863 session and the server sends an ec point format extension it could write
6864 up to 255 bytes to freed memory.
6866 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6872 * A malicious server can crash an OpenSSL client with a null pointer
6873 dereference (read) by specifying an SRP ciphersuite even though it was not
6874 properly negotiated with the client. This can be exploited through a
6875 Denial of Service attack.
6877 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6878 discovering and researching this issue.
6883 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6884 X509_name_oneline, X509_name_print_ex et al. to leak some information
6885 from the stack. Applications may be affected if they echo pretty printing
6886 output to the attacker.
6888 Thanks to Ivan Fratric (Google) for discovering this issue.
6891 *Emilia Käsper, and Steve Henson*
6893 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6894 for corner cases. (Certain input points at infinity could lead to
6895 bogus results, with non-infinity inputs mapped to infinity too.)
6899 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6901 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6902 handshake can force the use of weak keying material in OpenSSL
6903 SSL/TLS clients and servers.
6905 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6906 researching this issue. ([CVE-2014-0224])
6908 *KIKUCHI Masashi, Steve Henson*
6910 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6911 OpenSSL DTLS client the code can be made to recurse eventually crashing
6914 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6917 *Imre Rad, Steve Henson*
6919 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6920 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6921 client or server. This is potentially exploitable to run arbitrary
6922 code on a vulnerable client or server.
6924 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6926 *Jüri Aedla, Steve Henson*
6928 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6929 are subject to a denial of service attack.
6931 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6932 this issue. ([CVE-2014-3470])
6934 *Felix Gröbert, Ivan Fratric, Steve Henson*
6936 * Harmonize version and its documentation. -f flag is used to display
6939 *mancha <mancha1@zoho.com>*
6941 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6942 in i2d_ECPrivateKey.
6944 *mancha <mancha1@zoho.com>*
6946 * Fix some double frees. These are not thought to be exploitable.
6948 *mancha <mancha1@zoho.com>*
6950 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6952 * A missing bounds check in the handling of the TLS heartbeat extension
6953 can be used to reveal up to 64k of memory to a connected client or
6956 Thanks for Neel Mehta of Google Security for discovering this bug and to
6957 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6958 preparing the fix ([CVE-2014-0160])
6960 *Adam Langley, Bodo Moeller*
6962 * Fix for the attack described in the paper "Recovering OpenSSL
6963 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6964 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6965 <http://eprint.iacr.org/2014/140>
6967 Thanks to Yuval Yarom and Naomi Benger for discovering this
6968 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6970 *Yuval Yarom and Naomi Benger*
6972 * TLS pad extension: draft-agl-tls-padding-03
6974 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6975 TLS client Hello record length value would otherwise be > 255 and
6976 less that 512 pad with a dummy extension containing zeroes so it
6977 is at least 512 bytes long.
6979 *Adam Langley, Steve Henson*
6981 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6983 * Fix for TLS record tampering bug. A carefully crafted invalid
6984 handshake could crash OpenSSL with a NULL pointer exception.
6985 Thanks to Anton Johansson for reporting this issues.
6988 * Keep original DTLS digest and encryption contexts in retransmission
6989 structures so we can use the previous session parameters if they need
6990 to be resent. ([CVE-2013-6450])
6994 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6995 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6996 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6997 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6998 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6999 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7001 *Rob Stradling, Adam Langley*
7003 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7005 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7006 supporting platforms or when small records were transferred.
7008 *Andy Polyakov, Steve Henson*
7010 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7012 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7014 This addresses the flaw in CBC record processing discovered by
7015 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7016 at: <http://www.isg.rhul.ac.uk/tls/>
7018 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7019 Security Group at Royal Holloway, University of London
7020 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7021 Emilia Käsper for the initial patch.
7024 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7026 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7027 ciphersuites which can be exploited in a denial of service attack.
7028 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7029 and detecting this bug and to Wolfgang Ettlinger
7030 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7035 * Return an error when checking OCSP signatures when key is NULL.
7036 This fixes a DoS attack. ([CVE-2013-0166])
7040 * Make openssl verify return errors.
7042 *Chris Palmer <palmer@google.com> and Ben Laurie*
7044 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7045 the right response is stapled. Also change SSL_get_certificate()
7046 so it returns the certificate actually sent.
7047 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7049 *Rob Stradling <rob.stradling@comodo.com>*
7051 * Fix possible deadlock when decoding public keys.
7055 * Don't use TLS 1.0 record version number in initial client hello
7060 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7062 * Sanity check record length before skipping explicit IV in TLS
7063 1.2, 1.1 and DTLS to fix DoS attack.
7065 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7066 fuzzing as a service testing platform.
7071 * Initialise tkeylen properly when encrypting CMS messages.
7072 Thanks to Solar Designer of Openwall for reporting this issue.
7076 * In FIPS mode don't try to use composite ciphers as they are not
7081 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7083 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7084 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7085 mean any application compiled against OpenSSL 1.0.0 headers setting
7086 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7087 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7088 0x10000000L Any application which was previously compiled against
7089 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7090 will need to be recompiled as a result. Letting be results in
7091 inability to disable specifically TLS 1.1 and in client context,
7092 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7096 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7097 disable just protocol X, but all protocols above X *if* there are
7098 protocols *below* X still enabled. In more practical terms it means
7099 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7100 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7101 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7106 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7108 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7109 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7110 in CRYPTO_realloc_clean.
7112 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7113 issue and to Adam Langley <agl@chromium.org> for fixing it.
7116 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7118 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7122 * Workarounds for some broken servers that "hang" if a client hello
7123 record length exceeds 255 bytes.
7125 1. Do not use record version number > TLS 1.0 in initial client
7126 hello: some (but not all) hanging servers will now work.
7127 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7128 the number of ciphers sent in the client hello. This should be
7129 set to an even number, such as 50, for example by passing:
7130 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7131 Most broken servers should now work.
7132 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7133 TLS 1.2 client support entirely.
7137 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7141 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7143 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7144 STRING form instead of a DigestInfo.
7148 * The format used for MDC2 RSA signatures is inconsistent between EVP
7149 and the RSA_sign/RSA_verify functions. This was made more apparent when
7150 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7151 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7152 the correct format in RSA_verify so both forms transparently work.
7156 * Some servers which support TLS 1.0 can choke if we initially indicate
7157 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7158 encrypted premaster secret. As a workaround use the maximum permitted
7159 client version in client hello, this should keep such servers happy
7160 and still work with previous versions of OpenSSL.
7164 * Add support for TLS/DTLS heartbeats.
7166 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7168 * Add support for SCTP.
7170 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7172 * Improved PRNG seeding for VOS.
7174 *Paul Green <Paul.Green@stratus.com>*
7176 * Extensive assembler packs updates, most notably:
7178 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7179 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7180 - x86_64: bit-sliced AES implementation;
7181 - ARM: NEON support, contemporary platforms optimizations;
7182 - s390x: z196 support;
7183 - `*`: GHASH and GF(2^m) multiplication implementations;
7187 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7188 (removal of unnecessary code)
7190 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7192 * Add TLS key material exporter from RFC 5705.
7196 * Add DTLS-SRTP negotiation from RFC 5764.
7200 * Add Next Protocol Negotiation,
7201 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7202 disabled with a no-npn flag to config or Configure. Code donated
7205 *Adam Langley <agl@google.com> and Ben Laurie*
7207 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7208 NIST-P256, NIST-P521, with constant-time single point multiplication on
7209 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7210 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7211 Code made available under Apache License version 2.0.
7213 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7214 line to include this in your build of OpenSSL, and run "make depend" (or
7215 "make update"). This enables the following EC_METHODs:
7217 EC_GFp_nistp224_method()
7218 EC_GFp_nistp256_method()
7219 EC_GFp_nistp521_method()
7221 EC_GROUP_new_by_curve_name() will automatically use these (while
7222 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7225 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7227 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7228 all platforms. Move ssize_t definition from e_os.h to the public
7229 header file e_os2.h as it now appears in public header file cms.h
7233 * New -sigopt option to the ca, req and x509 utilities. Additional
7234 signature parameters can be passed using this option and in
7239 * Add RSA PSS signing function. This will generate and set the
7240 appropriate AlgorithmIdentifiers for PSS based on those in the
7241 corresponding EVP_MD_CTX structure. No application support yet.
7245 * Support for companion algorithm specific ASN1 signing routines.
7246 New function ASN1_item_sign_ctx() signs a pre-initialised
7247 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7248 the appropriate parameters.
7252 * Add new algorithm specific ASN1 verification initialisation function
7253 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7254 handling will be the same no matter what EVP_PKEY_METHOD is used.
7255 Add a PSS handler to support verification of PSS signatures: checked
7256 against a number of sample certificates.
7260 * Add signature printing for PSS. Add PSS OIDs.
7262 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7264 * Add algorithm specific signature printing. An individual ASN1 method
7265 can now print out signatures instead of the standard hex dump.
7267 More complex signatures (e.g. PSS) can print out more meaningful
7268 information. Include DSA version that prints out the signature
7273 * Password based recipient info support for CMS library: implementing
7278 * Split password based encryption into PBES2 and PBKDF2 functions. This
7279 neatly separates the code into cipher and PBE sections and is required
7280 for some algorithms that split PBES2 into separate pieces (such as
7281 password based CMS).
7285 * Session-handling fixes:
7286 - Fix handling of connections that are resuming with a session ID,
7287 but also support Session Tickets.
7288 - Fix a bug that suppressed issuing of a new ticket if the client
7289 presented a ticket with an expired session.
7290 - Try to set the ticket lifetime hint to something reasonable.
7291 - Make tickets shorter by excluding irrelevant information.
7292 - On the client side, don't ignore renewed tickets.
7294 *Adam Langley, Bodo Moeller (Google)*
7296 * Fix PSK session representation.
7300 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7302 This work was sponsored by Intel.
7306 * Add GCM support to TLS library. Some custom code is needed to split
7307 the IV between the fixed (from PRF) and explicit (from TLS record)
7308 portions. This adds all GCM ciphersuites supported by RFC5288 and
7309 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7310 add a special AESGCM string for GCM only.
7314 * Expand range of ctrls for AES GCM. Permit setting invocation
7315 field on decrypt and retrieval of invocation field only on encrypt.
7319 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7320 As required by RFC5289 these ciphersuites cannot be used if for
7321 versions of TLS earlier than 1.2.
7325 * For FIPS capable OpenSSL interpret a NULL default public key method
7326 as unset and return the appropriate default but do *not* set the default.
7327 This means we can return the appropriate method in applications that
7328 switch between FIPS and non-FIPS modes.
7332 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7333 ENGINE is used then we cannot handle that in the FIPS module so we
7334 keep original code iff non-FIPS operations are allowed.
7338 * Add -attime option to openssl utilities.
7340 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7342 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7346 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7347 FIPS EC methods unconditionally for now.
7351 * New build option no-ec2m to disable characteristic 2 code.
7355 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7356 all cases can be covered as some introduce binary incompatibilities.
7360 * Redirect RSA operations to FIPS module including keygen,
7361 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7365 * Add similar low-level API blocking to ciphers.
7369 * low-level digest APIs are not approved in FIPS mode: any attempt
7370 to use these will cause a fatal error. Applications that *really* want
7371 to use them can use the `private_*` version instead.
7375 * Redirect cipher operations to FIPS module for FIPS builds.
7379 * Redirect digest operations to FIPS module for FIPS builds.
7383 * Update build system to add "fips" flag which will link in fipscanister.o
7384 for static and shared library builds embedding a signature if needed.
7388 * Output TLS supported curves in preference order instead of numerical
7389 order. This is currently hardcoded for the highest order curves first.
7390 This should be configurable so applications can judge speed vs strength.
7394 * Add TLS v1.2 server support for client authentication.
7398 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7403 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7404 FIPS modules versions.
7408 * Add TLS v1.2 client side support for client authentication. Keep cache
7409 of handshake records longer as we don't know the hash algorithm to use
7410 until after the certificate request message is received.
7414 * Initial TLS v1.2 client support. Add a default signature algorithms
7415 extension including all the algorithms we support. Parse new signature
7416 format in client key exchange. Relax some ECC signing restrictions for
7417 TLS v1.2 as indicated in RFC5246.
7421 * Add server support for TLS v1.2 signature algorithms extension. Switch
7422 to new signature format when needed using client digest preference.
7423 All server ciphersuites should now work correctly in TLS v1.2. No client
7424 support yet and no support for client certificates.
7428 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7429 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7430 ciphersuites. At present only RSA key exchange ciphersuites work with
7431 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7432 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7433 and version checking.
7437 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7438 with this defined it will not be affected by any changes to ssl internal
7439 structures. Add several utility functions to allow openssl application
7440 to work with OPENSSL_NO_SSL_INTERN defined.
7444 * A long standing patch to add support for SRP from EdelWeb (Peter
7445 Sylvester and Christophe Renou) was integrated.
7446 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7447 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7450 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7454 * Permit abbreviated handshakes when renegotiating using the function
7455 SSL_renegotiate_abbreviated().
7457 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7459 * Add call to ENGINE_register_all_complete() to
7460 ENGINE_load_builtin_engines(), so some implementations get used
7461 automatically instead of needing explicit application support.
7465 * Add support for TLS key exporter as described in RFC5705.
7467 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7469 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7470 a few changes are required:
7472 Add SSL_OP_NO_TLSv1_1 flag.
7473 Add TLSv1_1 methods.
7474 Update version checking logic to handle version 1.1.
7475 Add explicit IV handling (ported from DTLS code).
7476 Add command line options to s_client/s_server.
7483 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7485 * X509_ATTRIBUTE memory leak
7487 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7488 memory. This structure is used by the PKCS#7 and CMS routines so any
7489 application which reads PKCS#7 or CMS data from untrusted sources is
7490 affected. SSL/TLS is not affected.
7492 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7498 * Race condition handling PSK identify hint
7500 If PSK identity hints are received by a multi-threaded client then
7501 the values are wrongly updated in the parent SSL_CTX structure. This can
7502 result in a race condition potentially leading to a double free of the
7508 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7510 * Malformed ECParameters causes infinite loop
7512 When processing an ECParameters structure OpenSSL enters an infinite loop
7513 if the curve specified is over a specially malformed binary polynomial
7516 This can be used to perform denial of service against any
7517 system which processes public keys, certificate requests or
7518 certificates. This includes TLS clients and TLS servers with
7519 client authentication enabled.
7521 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7526 * Exploitable out-of-bounds read in X509_cmp_time
7528 X509_cmp_time does not properly check the length of the ASN1_TIME
7529 string and can read a few bytes out of bounds. In addition,
7530 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7533 An attacker can use this to craft malformed certificates and CRLs of
7534 various sizes and potentially cause a segmentation fault, resulting in
7535 a DoS on applications that verify certificates or CRLs. TLS clients
7536 that verify CRLs are affected. TLS clients and servers with client
7537 authentication enabled may be affected if they use custom verification
7540 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7541 independently by Hanno Böck.
7546 * PKCS7 crash with missing EnvelopedContent
7548 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7549 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7550 with missing content and trigger a NULL pointer dereference on parsing.
7552 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7553 structures from untrusted sources are affected. OpenSSL clients and
7554 servers are not affected.
7556 This issue was reported to OpenSSL by Michal Zalewski (Google).
7561 * CMS verify infinite loop with unknown hash function
7563 When verifying a signedData message the CMS code can enter an infinite loop
7564 if presented with an unknown hash function OID. This can be used to perform
7565 denial of service against any system which verifies signedData messages using
7567 This issue was reported to OpenSSL by Johannes Bauer.
7572 * Race condition handling NewSessionTicket
7574 If a NewSessionTicket is received by a multi-threaded client when attempting to
7575 reuse a previous ticket then a race condition can occur potentially leading to
7576 a double free of the ticket data.
7581 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7583 * Segmentation fault in ASN1_TYPE_cmp fix
7585 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7586 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7587 certificate signature algorithm consistency this can be used to crash any
7588 certificate verification operation and exploited in a DoS attack. Any
7589 application which performs certificate verification is vulnerable including
7590 OpenSSL clients and servers which enable client authentication.
7595 * ASN.1 structure reuse memory corruption fix
7597 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7598 memory corruption via an invalid write. Such reuse is and has been
7599 strongly discouraged and is believed to be rare.
7601 Applications that parse structures containing CHOICE or ANY DEFINED BY
7602 components may be affected. Certificate parsing (d2i_X509 and related
7603 functions) are however not affected. OpenSSL clients and servers are
7609 * PKCS7 NULL pointer dereferences fix
7611 The PKCS#7 parsing code does not handle missing outer ContentInfo
7612 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7613 missing content and trigger a NULL pointer dereference on parsing.
7615 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7616 otherwise parse PKCS#7 structures from untrusted sources are
7617 affected. OpenSSL clients and servers are not affected.
7619 This issue was reported to OpenSSL by Michal Zalewski (Google).
7624 * DoS via reachable assert in SSLv2 servers fix
7626 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7627 servers that both support SSLv2 and enable export cipher suites by sending
7628 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7630 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7631 (OpenSSL development team).
7636 * Use After Free following d2i_ECPrivatekey error fix
7638 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7639 could cause a use after free condition. This, in turn, could cause a double
7640 free in several private key parsing functions (such as d2i_PrivateKey
7641 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7642 for applications that receive EC private keys from untrusted
7643 sources. This scenario is considered rare.
7645 This issue was discovered by the BoringSSL project and fixed in their
7651 * X509_to_X509_REQ NULL pointer deref fix
7653 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7654 the certificate key is invalid. This function is rarely used in practice.
7656 This issue was discovered by Brian Carpenter.
7661 * Removed the export ciphers from the DEFAULT ciphers
7665 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7667 * Build fixes for the Windows and OpenVMS platforms
7669 *Matt Caswell and Richard Levitte*
7671 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7673 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7674 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7675 dereference. This could lead to a Denial Of Service attack. Thanks to
7676 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7681 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7682 dtls1_buffer_record function under certain conditions. In particular this
7683 could occur if an attacker sent repeated DTLS records with the same
7684 sequence number but for the next epoch. The memory leak could be exploited
7685 by an attacker in a Denial of Service attack through memory exhaustion.
7686 Thanks to Chris Mueller for reporting this issue.
7691 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7692 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7693 method would be set to NULL which could later result in a NULL pointer
7694 dereference. Thanks to Frank Schmirler for reporting this issue.
7699 * Abort handshake if server key exchange message is omitted for ephemeral
7702 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7703 reporting this issue.
7708 * Remove non-export ephemeral RSA code on client and server. This code
7709 violated the TLS standard by allowing the use of temporary RSA keys in
7710 non-export ciphersuites and could be used by a server to effectively
7711 downgrade the RSA key length used to a value smaller than the server
7712 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7713 INRIA or reporting this issue.
7718 * Fixed issue where DH client certificates are accepted without verification.
7719 An OpenSSL server will accept a DH certificate for client authentication
7720 without the certificate verify message. This effectively allows a client to
7721 authenticate without the use of a private key. This only affects servers
7722 which trust a client certificate authority which issues certificates
7723 containing DH keys: these are extremely rare and hardly ever encountered.
7724 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7730 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7731 results on some platforms, including x86_64. This bug occurs at random
7732 with a very low probability, and is not known to be exploitable in any
7733 way, though its exact impact is difficult to determine. Thanks to Pieter
7734 Wuille (Blockstream) who reported this issue and also suggested an initial
7735 fix. Further analysis was conducted by the OpenSSL development team and
7736 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7737 the OpenSSL core team.
7742 * Fix various certificate fingerprint issues.
7744 By using non-DER or invalid encodings outside the signed portion of a
7745 certificate the fingerprint can be changed without breaking the signature.
7746 Although no details of the signed portion of the certificate can be changed
7747 this can cause problems with some applications: e.g. those using the
7748 certificate fingerprint for blacklists.
7750 1. Reject signatures with non zero unused bits.
7752 If the BIT STRING containing the signature has non zero unused bits reject
7753 the signature. All current signature algorithms require zero unused bits.
7755 2. Check certificate algorithm consistency.
7757 Check the AlgorithmIdentifier inside TBS matches the one in the
7758 certificate signature. NB: this will result in signature failure
7759 errors for some broken certificates.
7761 Thanks to Konrad Kraszewski from Google for reporting this issue.
7763 3. Check DSA/ECDSA signatures use DER.
7765 Reencode DSA/ECDSA signatures and compare with the original received
7766 signature. Return an error if there is a mismatch.
7768 This will reject various cases including garbage after signature
7769 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7770 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7771 (negative or with leading zeroes).
7773 Further analysis was conducted and fixes were developed by Stephen Henson
7774 of the OpenSSL core team.
7780 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7782 * Session Ticket Memory Leak.
7784 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7785 integrity of that ticket is first verified. In the event of a session
7786 ticket integrity check failing, OpenSSL will fail to free memory
7787 causing a memory leak. By sending a large number of invalid session
7788 tickets an attacker could exploit this issue in a Denial Of Service
7794 * Build option no-ssl3 is incomplete.
7796 When OpenSSL is configured with "no-ssl3" as a build option, servers
7797 could accept and complete a SSL 3.0 handshake, and clients could be
7798 configured to send them.
7801 *Akamai and the OpenSSL team*
7803 * Add support for TLS_FALLBACK_SCSV.
7804 Client applications doing fallback retries should call
7805 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7808 *Adam Langley, Bodo Moeller*
7810 * Add additional DigestInfo checks.
7812 Reencode DigestInto in DER and check against the original when
7813 verifying RSA signature: this will reject any improperly encoded
7814 DigestInfo structures.
7816 Note: this is a precautionary measure and no attacks are currently known.
7820 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7822 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7823 to a denial of service attack. A malicious server can crash the client
7824 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7825 ciphersuite and sending carefully crafted handshake messages.
7827 Thanks to Felix Gröbert (Google) for discovering and researching this
7833 * By sending carefully crafted DTLS packets an attacker could cause openssl
7834 to leak memory. This can be exploited through a Denial of Service attack.
7835 Thanks to Adam Langley for discovering and researching this issue.
7840 * An attacker can force openssl to consume large amounts of memory whilst
7841 processing DTLS handshake messages. This can be exploited through a
7842 Denial of Service attack.
7843 Thanks to Adam Langley for discovering and researching this issue.
7848 * An attacker can force an error condition which causes openssl to crash
7849 whilst processing DTLS packets due to memory being freed twice. This
7850 can be exploited through a Denial of Service attack.
7851 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7857 * If a multithreaded client connects to a malicious server using a resumed
7858 session and the server sends an ec point format extension it could write
7859 up to 255 bytes to freed memory.
7861 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7867 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7868 X509_name_oneline, X509_name_print_ex et al. to leak some information
7869 from the stack. Applications may be affected if they echo pretty printing
7870 output to the attacker.
7872 Thanks to Ivan Fratric (Google) for discovering this issue.
7875 *Emilia Käsper, and Steve Henson*
7877 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7878 for corner cases. (Certain input points at infinity could lead to
7879 bogus results, with non-infinity inputs mapped to infinity too.)
7883 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7885 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7886 handshake can force the use of weak keying material in OpenSSL
7887 SSL/TLS clients and servers.
7889 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7890 researching this issue. ([CVE-2014-0224])
7892 *KIKUCHI Masashi, Steve Henson*
7894 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7895 OpenSSL DTLS client the code can be made to recurse eventually crashing
7898 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7901 *Imre Rad, Steve Henson*
7903 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7904 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7905 client or server. This is potentially exploitable to run arbitrary
7906 code on a vulnerable client or server.
7908 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7910 *Jüri Aedla, Steve Henson*
7912 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7913 are subject to a denial of service attack.
7915 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7916 this issue. ([CVE-2014-3470])
7918 *Felix Gröbert, Ivan Fratric, Steve Henson*
7920 * Harmonize version and its documentation. -f flag is used to display
7923 *mancha <mancha1@zoho.com>*
7925 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7926 in i2d_ECPrivateKey.
7928 *mancha <mancha1@zoho.com>*
7930 * Fix some double frees. These are not thought to be exploitable.
7932 *mancha <mancha1@zoho.com>*
7934 * Fix for the attack described in the paper "Recovering OpenSSL
7935 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7936 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7937 <http://eprint.iacr.org/2014/140>
7939 Thanks to Yuval Yarom and Naomi Benger for discovering this
7940 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7942 *Yuval Yarom and Naomi Benger*
7944 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7946 * Keep original DTLS digest and encryption contexts in retransmission
7947 structures so we can use the previous session parameters if they need
7948 to be resent. ([CVE-2013-6450])
7952 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7953 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7954 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7955 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7956 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7957 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7959 *Rob Stradling, Adam Langley*
7961 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7963 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7965 This addresses the flaw in CBC record processing discovered by
7966 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7967 at: <http://www.isg.rhul.ac.uk/tls/>
7969 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7970 Security Group at Royal Holloway, University of London
7971 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7972 Emilia Käsper for the initial patch.
7975 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7977 * Return an error when checking OCSP signatures when key is NULL.
7978 This fixes a DoS attack. ([CVE-2013-0166])
7982 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7983 the right response is stapled. Also change SSL_get_certificate()
7984 so it returns the certificate actually sent.
7985 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7986 (This is a backport)
7988 *Rob Stradling <rob.stradling@comodo.com>*
7990 * Fix possible deadlock when decoding public keys.
7994 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7996 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7999 * Sanity check record length before skipping explicit IV in DTLS
8002 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8003 fuzzing as a service testing platform.
8008 * Initialise tkeylen properly when encrypting CMS messages.
8009 Thanks to Solar Designer of Openwall for reporting this issue.
8013 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8015 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8016 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8017 in CRYPTO_realloc_clean.
8019 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8020 issue and to Adam Langley <agl@chromium.org> for fixing it.
8023 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8025 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8027 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8028 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8029 content decryption and always return the same error. Note: this attack
8030 needs on average 2^20 messages so it only affects automated senders. The
8031 old behaviour can be re-enabled in the CMS code by setting the
8032 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8033 an MMA defence is not necessary.
8034 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8035 this issue. ([CVE-2012-0884])
8039 * Fix CVE-2011-4619: make sure we really are receiving a
8040 client hello before rejecting multiple SGC restarts. Thanks to
8041 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8045 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8047 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8048 Thanks to Antonio Martin, Enterprise Secure Access Research and
8049 Development, Cisco Systems, Inc. for discovering this bug and
8050 preparing a fix. ([CVE-2012-0050])
8054 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8056 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8057 of the Vaudenay padding oracle attack on CBC mode encryption
8058 which enables an efficient plaintext recovery attack against
8059 the OpenSSL implementation of DTLS. Their attack exploits timing
8060 differences arising during decryption processing. A research
8061 paper describing this attack can be found at:
8062 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8063 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8064 Security Group at Royal Holloway, University of London
8065 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8066 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8067 for preparing the fix. ([CVE-2011-4108])
8069 *Robin Seggelmann, Michael Tuexen*
8071 * Clear bytes used for block padding of SSL 3.0 records.
8074 *Adam Langley (Google)*
8076 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8077 Kadianakis <desnacked@gmail.com> for discovering this issue and
8078 Adam Langley for preparing the fix. ([CVE-2011-4619])
8080 *Adam Langley (Google)*
8082 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8084 *Andrey Kulikov <amdeich@gmail.com>*
8086 * Prevent malformed RFC3779 data triggering an assertion failure.
8087 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8088 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8090 *Rob Austein <sra@hactrn.net>*
8092 * Improved PRNG seeding for VOS.
8094 *Paul Green <Paul.Green@stratus.com>*
8096 * Fix ssl_ciph.c set-up race.
8098 *Adam Langley (Google)*
8100 * Fix spurious failures in ecdsatest.c.
8102 *Emilia Käsper (Google)*
8104 * Fix the BIO_f_buffer() implementation (which was mixing different
8105 interpretations of the `..._len` fields).
8107 *Adam Langley (Google)*
8109 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8110 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8111 threads won't reuse the same blinding coefficients.
8113 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8114 lock to call BN_BLINDING_invert_ex, and avoids one use of
8115 BN_BLINDING_update for each BN_BLINDING structure (previously,
8116 the last update always remained unused).
8118 *Emilia Käsper (Google)*
8120 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8122 *Bob Buckholz (Google)*
8124 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8126 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8127 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8129 *Kaspar Brand <ossl@velox.ch>*
8131 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8132 for multi-threaded use of ECDH. ([CVE-2011-3210])
8134 *Adam Langley (Google)*
8136 * Fix x509_name_ex_d2i memory leak on bad inputs.
8140 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8141 signature public key algorithm by using OID xref utilities instead.
8142 Before this you could only use some ECC ciphersuites with SHA1 only.
8146 * Add protection against ECDSA timing attacks as mentioned in the paper
8147 by Billy Bob Brumley and Nicola Tuveri, see:
8148 <http://eprint.iacr.org/2011/232.pdf>
8150 *Billy Bob Brumley and Nicola Tuveri*
8152 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8154 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8156 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8158 * Fix bug in string printing code: if *any* escaping is enabled we must
8159 escape the escape character (backslash) or the resulting string is
8164 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8166 * Disable code workaround for ancient and obsolete Netscape browsers
8167 and servers: an attacker can use it in a ciphersuite downgrade attack.
8168 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8172 * Fixed J-PAKE implementation error, originally discovered by
8173 Sebastien Martini, further info and confirmation from Stefan
8174 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8178 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8180 * Fix extension code to avoid race conditions which can result in a buffer
8181 overrun vulnerability: resumed sessions must not be modified as they can
8182 be shared by multiple threads. CVE-2010-3864
8186 * Fix WIN32 build system to correctly link an ENGINE directory into
8191 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8193 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8196 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8198 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8200 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8201 context. The operation can be customised via the ctrl mechanism in
8202 case ENGINEs want to include additional functionality.
8206 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8210 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8211 output hashes compatible with older versions of OpenSSL.
8213 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8215 * Fix compression algorithm handling: if resuming a session use the
8216 compression algorithm of the resumed session instead of determining
8217 it from client hello again. Don't allow server to change algorithm.
8221 * Add load_crls() function to commands tidying load_certs() too. Add option
8222 to verify utility to allow additional CRLs to be included.
8226 * Update OCSP request code to permit adding custom headers to the request:
8227 some responders need this.
8231 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8234 *Julia Lawall <julia@diku.dk>*
8236 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8237 needlessly dereferenced structures, used obsolete functions and
8238 didn't handle all updated verify codes correctly.
8242 * Disable MD2 in the default configuration.
8246 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8247 indicate the initial BIO being pushed or popped. This makes it possible
8248 to determine whether the BIO is the one explicitly called or as a result
8249 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8250 it handles reference counts correctly and doesn't zero out the I/O bio
8251 when it is not being explicitly popped. WARNING: applications which
8252 included workarounds for the old buggy behaviour will need to be modified
8253 or they could free up already freed BIOs.
8257 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8258 renaming to all platforms (within the 0.9.8 branch, this was
8259 done conditionally on Netware platforms to avoid a name clash).
8261 *Guenter <lists@gknw.net>*
8263 * Add ECDHE and PSK support to DTLS.
8265 *Michael Tuexen <tuexen@fh-muenster.de>*
8267 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8272 * Add "missing" function EVP_MD_flags() (without this the only way to
8273 retrieve a digest flags is by accessing the structure directly. Update
8274 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8275 or cipher is registered as in the "from" argument. Print out all
8276 registered digests in the dgst usage message instead of manually
8277 attempting to work them out.
8281 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8282 this allows the use of compression and extensions. Change default cipher
8283 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8284 by default unless an application cipher string requests it.
8288 * Alter match criteria in PKCS12_parse(). It used to try to use local
8289 key ids to find matching certificates and keys but some PKCS#12 files
8290 don't follow the (somewhat unwritten) rules and this strategy fails.
8291 Now just gather all certificates together and the first private key
8292 then look for the first certificate that matches the key.
8296 * Support use of registered digest and cipher names for dgst and cipher
8297 commands instead of having to add each one as a special case. So now
8304 openssl dgst -sha256 foo
8306 and this works for ENGINE based algorithms too.
8310 * Update Gost ENGINE to support parameter files.
8312 *Victor B. Wagner <vitus@cryptocom.ru>*
8314 * Support GeneralizedTime in ca utility.
8316 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8318 * Enhance the hash format used for certificate directory links. The new
8319 form uses the canonical encoding (meaning equivalent names will work
8320 even if they aren't identical) and uses SHA1 instead of MD5. This form
8321 is incompatible with the older format and as a result c_rehash should
8322 be used to rebuild symbolic links.
8326 * Make PKCS#8 the default write format for private keys, replacing the
8327 traditional format. This form is standardised, more secure and doesn't
8328 include an implicit MD5 dependency.
8332 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8333 committed to OpenSSL should pass this lot as a minimum.
8337 * Add session ticket override functionality for use by EAP-FAST.
8339 *Jouni Malinen <j@w1.fi>*
8341 * Modify HMAC functions to return a value. Since these can be implemented
8342 in an ENGINE errors can occur.
8346 * Type-checked OBJ_bsearch_ex.
8350 * Type-checked OBJ_bsearch. Also some constification necessitated
8351 by type-checking. Still to come: TXT_DB, bsearch(?),
8352 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8357 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8358 seconds to a tm structure directly, instead of going through OS
8359 specific date routines. This avoids any issues with OS routines such
8360 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8361 and X509_time_adj_ex() to cover the extended range. The existing
8362 X509_time_adj() is still usable and will no longer have any date issues.
8366 * Delta CRL support. New use deltas option which will attempt to locate
8367 and search any appropriate delta CRLs available.
8369 This work was sponsored by Google.
8373 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8374 code and add additional score elements. Validate alternate CRL paths
8375 as part of the CRL checking and indicate a new error "CRL path validation
8376 error" in this case. Applications wanting additional details can use
8377 the verify callback and check the new "parent" field. If this is not
8378 NULL CRL path validation is taking place. Existing applications won't
8379 see this because it requires extended CRL support which is off by
8382 This work was sponsored by Google.
8386 * Support for freshest CRL extension.
8388 This work was sponsored by Google.
8392 * Initial indirect CRL support. Currently only supported in the CRLs
8393 passed directly and not via lookup. Process certificate issuer
8394 CRL entry extension and lookup CRL entries by bother issuer name
8395 and serial number. Check and process CRL issuer entry in IDP extension.
8397 This work was sponsored by Google.
8401 * Add support for distinct certificate and CRL paths. The CRL issuer
8402 certificate is validated separately in this case. Only enabled if
8403 an extended CRL support flag is set: this flag will enable additional
8404 CRL functionality in future.
8406 This work was sponsored by Google.
8410 * Add support for policy mappings extension.
8412 This work was sponsored by Google.
8416 * Fixes to pathlength constraint, self issued certificate handling,
8417 policy processing to align with RFC3280 and PKITS tests.
8419 This work was sponsored by Google.
8423 * Support for name constraints certificate extension. DN, email, DNS
8424 and URI types are currently supported.
8426 This work was sponsored by Google.
8430 * To cater for systems that provide a pointer-based thread ID rather
8431 than numeric, deprecate the current numeric thread ID mechanism and
8432 replace it with a structure and associated callback type. This
8433 mechanism allows a numeric "hash" to be extracted from a thread ID in
8434 either case, and on platforms where pointers are larger than 'long',
8435 mixing is done to help ensure the numeric 'hash' is usable even if it
8436 can't be guaranteed unique. The default mechanism is to use "&errno"
8437 as a pointer-based thread ID to distinguish between threads.
8439 Applications that want to provide their own thread IDs should now use
8440 CRYPTO_THREADID_set_callback() to register a callback that will call
8441 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8443 Note that ERR_remove_state() is now deprecated, because it is tied
8444 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8445 to free the current thread's error state should be replaced by
8446 ERR_remove_thread_state(NULL).
8448 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8449 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8450 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8451 application was previously providing a numeric thread callback that
8452 was inappropriate for distinguishing threads, then uniqueness might
8453 have been obtained with &errno that happened immediately in the
8454 intermediate development versions of OpenSSL; this is no longer the
8455 case, the numeric thread callback will now override the automatic use
8458 *Geoff Thorpe, with help from Bodo Moeller*
8460 * Initial support for different CRL issuing certificates. This covers a
8461 simple case where the self issued certificates in the chain exist and
8462 the real CRL issuer is higher in the existing chain.
8464 This work was sponsored by Google.
8468 * Removed effectively defunct crypto/store from the build.
8472 * Revamp of STACK to provide stronger type-checking. Still to come:
8473 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8474 ASN1_STRING, CONF_VALUE.
8478 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8479 RAM on SSL connections. This option can save about 34k per idle SSL.
8483 * Revamp of LHASH to provide stronger type-checking. Still to come:
8484 STACK, TXT_DB, bsearch, qsort.
8488 * Initial support for Cryptographic Message Syntax (aka CMS) based
8489 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8490 support for data, signedData, compressedData, digestedData and
8491 encryptedData, envelopedData types included. Scripts to check against
8492 RFC4134 examples draft and interop and consistency checks of many
8493 content types and variants.
8497 * Add options to enc utility to support use of zlib compression BIO.
8501 * Extend mk1mf to support importing of options and assembly language
8502 files from Configure script, currently only included in VC-WIN32.
8503 The assembly language rules can now optionally generate the source
8504 files from the associated perl scripts.
8508 * Implement remaining functionality needed to support GOST ciphersuites.
8509 Interop testing has been performed using CryptoPro implementations.
8511 *Victor B. Wagner <vitus@cryptocom.ru>*
8513 * s390x assembler pack.
8517 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8522 * Implement Opaque PRF Input TLS extension as specified in
8523 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8524 official specification yet and no extension type assignment by
8525 IANA exists, this extension (for now) will have to be explicitly
8526 enabled when building OpenSSL by providing the extension number
8527 to use. For example, specify an option
8529 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8531 to the "config" or "Configure" script to enable the extension,
8532 assuming extension number 0x9527 (which is a completely arbitrary
8533 and unofficial assignment based on the MD5 hash of the Internet
8534 Draft). Note that by doing so, you potentially lose
8535 interoperability with other TLS implementations since these might
8536 be using the same extension number for other purposes.
8538 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8539 opaque PRF input value to use in the handshake. This will create
8540 an internal copy of the length-'len' string at 'src', and will
8541 return non-zero for success.
8543 To get more control and flexibility, provide a callback function
8546 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8547 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8551 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8554 Callback function 'cb' will be called in handshakes, and is
8555 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8556 Argument 'arg' is for application purposes (the value as given to
8557 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8558 be provided to the callback function). The callback function
8559 has to return non-zero to report success: usually 1 to use opaque
8560 PRF input just if possible, or 2 to enforce use of the opaque PRF
8561 input. In the latter case, the library will abort the handshake
8562 if opaque PRF input is not successfully negotiated.
8564 Arguments 'peerinput' and 'len' given to the callback function
8565 will always be NULL and 0 in the case of a client. A server will
8566 see the client's opaque PRF input through these variables if
8567 available (NULL and 0 otherwise). Note that if the server
8568 provides an opaque PRF input, the length must be the same as the
8569 length of the client's opaque PRF input.
8571 Note that the callback function will only be called when creating
8572 a new session (session resumption can resume whatever was
8573 previously negotiated), and will not be called in SSL 2.0
8574 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8575 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8576 for applications that need to enforce opaque PRF input.
8580 * Update ssl code to support digests other than SHA1+MD5 for handshake
8583 *Victor B. Wagner <vitus@cryptocom.ru>*
8585 * Add RFC4507 support to OpenSSL. This includes the corrections in
8586 RFC4507bis. The encrypted ticket format is an encrypted encoded
8587 SSL_SESSION structure, that way new session features are automatically
8590 If a client application caches session in an SSL_SESSION structure
8591 support is transparent because tickets are now stored in the encoded
8594 The SSL_CTX structure automatically generates keys for ticket
8595 protection in servers so again support should be possible
8596 with no application modification.
8598 If a client or server wishes to disable RFC4507 support then the option
8599 SSL_OP_NO_TICKET can be set.
8601 Add a TLS extension debugging callback to allow the contents of any client
8602 or server extensions to be examined.
8604 This work was sponsored by Google.
8608 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8609 OpenSSL should now compile cleanly on gcc 4.2
8611 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8613 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8614 support including streaming MAC support: this is required for GOST
8615 ciphersuite support.
8617 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8619 * Add option -stream to use PKCS#7 streaming in smime utility. New
8620 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8621 to output in BER and PEM format.
8625 * Experimental support for use of HMAC via EVP_PKEY interface. This
8626 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8627 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8628 ENGINE support for HMAC keys which are unextractable. New -mac and
8629 -macopt options to dgst utility.
8633 * New option -sigopt to dgst utility. Update dgst to use
8634 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8635 alternative signing parameters such as X9.31 or PSS in the dgst
8640 * Change ssl_cipher_apply_rule(), the internal function that does
8641 the work each time a ciphersuite string requests enabling
8642 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8643 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8644 the order of disabled ciphersuites such that those ciphersuites
8645 that most recently went from enabled to disabled not only stay
8646 in order with respect to each other, but also have higher priority
8647 than other disabled ciphersuites the next time ciphersuites are
8650 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8651 the same ciphersuites as with "HIGH" alone, but in a specific
8652 order where the PSK ciphersuites come first (since they are the
8653 most recently disabled ciphersuites when "HIGH" is parsed).
8655 Also, change ssl_create_cipher_list() (using this new
8656 functionality) such that between otherwise identical
8657 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8662 * Change ssl_create_cipher_list() so that it automatically
8663 arranges the ciphersuites in reasonable order before starting
8664 to process the rule string. Thus, the definition for "DEFAULT"
8665 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8666 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8667 This makes it much easier to arrive at a reasonable default order
8668 in applications for which anonymous ciphers are OK (meaning
8669 that you can't actually use DEFAULT).
8671 *Bodo Moeller; suggested by Victor Duchovni*
8673 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8674 processing) into multiple integers instead of setting
8675 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8676 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8677 (These masks as well as the individual bit definitions are hidden
8678 away into the non-exported interface ssl/ssl_locl.h, so this
8679 change to the definition of the SSL_CIPHER structure shouldn't
8680 affect applications.) This give us more bits for each of these
8681 categories, so there is no longer a need to coagulate AES128 and
8682 AES256 into a single algorithm bit, and to coagulate Camellia128
8683 and Camellia256 into a single algorithm bit, which has led to all
8686 Thus, among other things, the kludge introduced in 0.9.7m and
8687 0.9.8e for masking out AES256 independently of AES128 or masking
8688 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8690 With the change, we also introduce new ciphersuite aliases that
8691 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8696 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8697 Use the leftmost N bytes of the signature input if the input is
8698 larger than the prime q (with N being the size in bytes of q).
8702 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8703 it yet and it is largely untested.
8707 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8711 * Initial incomplete changes to avoid need for function casts in OpenSSL
8712 some compilers (gcc 4.2 and later) reject their use. Safestack is
8713 reimplemented. Update ASN1 to avoid use of legacy functions.
8717 * Win32/64 targets are linked with Winsock2.
8721 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8722 to external functions. This can be used to increase CRL handling
8723 efficiency especially when CRLs are very large by (for example) storing
8724 the CRL revoked certificates in a database.
8728 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8729 new CRLs added to a directory can be used. New command line option
8730 -verify_return_error to s_client and s_server. This causes real errors
8731 to be returned by the verify callback instead of carrying on no matter
8732 what. This reflects the way a "real world" verify callback would behave.
8736 * GOST engine, supporting several GOST algorithms and public key formats.
8737 Kindly donated by Cryptocom.
8741 * Partial support for Issuing Distribution Point CRL extension. CRLs
8742 partitioned by DP are handled but no indirect CRL or reason partitioning
8743 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8744 selected via a scoring technique which handles IDP and AKID in CRLs.
8748 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8749 will ultimately be used for all verify operations: this will remove the
8750 X509_STORE dependency on certificate verification and allow alternative
8751 lookup methods. X509_STORE based implementations of these two callbacks.
8755 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8756 Modify get_crl() to find a valid (unexpired) CRL if possible.
8760 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8761 this would be called X509_CRL_cmp() but that name is already used by
8762 a function that just compares CRL issuer names. Cache several CRL
8763 extensions in X509_CRL structure and cache CRLDP in X509.
8767 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8768 this maps equivalent X509_NAME structures into a consistent structure.
8769 Name comparison can then be performed rapidly using memcmp().
8773 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8778 * Allow digests to supply their own micalg string for S/MIME type using
8779 the ctrl EVP_MD_CTRL_MICALG.
8783 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8784 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8785 ctrl. It can then customise the structure before and/or after signing
8790 * New function OBJ_add_sigid() to allow application defined signature OIDs
8791 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8792 to free up any added signature OIDs.
8796 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8797 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8798 digest and cipher tables. New options added to openssl utility:
8799 list-message-digest-algorithms and list-cipher-algorithms.
8803 * Change the array representation of binary polynomials: the list
8804 of degrees of non-zero coefficients is now terminated with -1.
8805 Previously it was terminated with 0, which was also part of the
8806 value; thus, the array representation was not applicable to
8807 polynomials where t^0 has coefficient zero. This change makes
8808 the array representation useful in a more general context.
8812 * Various modifications and fixes to SSL/TLS cipher string
8813 handling. For ECC, the code now distinguishes between fixed ECDH
8814 with RSA certificates on the one hand and with ECDSA certificates
8815 on the other hand, since these are separate ciphersuites. The
8816 unused code for Fortezza ciphersuites has been removed.
8818 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8819 (not "ECDHE"). For consistency with the code for DH
8820 certificates, use of ECDH certificates is now considered ECDH
8821 authentication, not RSA or ECDSA authentication (the latter is
8822 merely the CA's signing algorithm and not actively used in the
8825 The temporary ciphersuite alias "ECCdraft" is no longer
8826 available, and ECC ciphersuites are no longer excluded from "ALL"
8827 and "DEFAULT". The following aliases now exist for RFC 4492
8828 ciphersuites, most of these by analogy with the DH case:
8830 kECDHr - ECDH cert, signed with RSA
8831 kECDHe - ECDH cert, signed with ECDSA
8832 kECDH - ECDH cert (signed with either RSA or ECDSA)
8833 kEECDH - ephemeral ECDH
8834 ECDH - ECDH cert or ephemeral ECDH
8840 AECDH - anonymous ECDH
8841 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8845 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8846 Use correct micalg parameters depending on digest(s) in signed message.
8850 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8851 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8855 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8856 an engine to register a method. Add ENGINE lookups for methods and
8857 functional reference processing.
8861 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8862 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8867 * New -resign option to smime utility. This adds one or more signers
8868 to an existing PKCS#7 signedData structure. Also -md option to use an
8869 alternative message digest algorithm for signing.
8873 * Tidy up PKCS#7 routines and add new functions to make it easier to
8874 create PKCS7 structures containing multiple signers. Update smime
8875 application to support multiple signers.
8879 * New -macalg option to pkcs12 utility to allow setting of an alternative
8884 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8885 Reorganize PBE internals to lookup from a static table using NIDs,
8886 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8887 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8888 PRF which will be automatically used with PBES2.
8892 * Replace the algorithm specific calls to generate keys in "req" with the
8897 * Update PKCS#7 enveloped data routines to use new API. This is now
8898 supported by any public key method supporting the encrypt operation. A
8899 ctrl is added to allow the public key algorithm to examine or modify
8900 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8905 * Add a ctrl to asn1 method to allow a public key algorithm to express
8906 a default digest type to use. In most cases this will be SHA1 but some
8907 algorithms (such as GOST) need to specify an alternative digest. The
8908 return value indicates how strong the preference is 1 means optional and
8909 2 is mandatory (that is it is the only supported type). Modify
8910 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8911 use the default md. Update openssl utilities to use the default digest
8912 type for signing if it is not explicitly indicated.
8916 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8917 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8918 signing method from the key type. This effectively removes the link
8919 between digests and public key types.
8923 * Add an OID cross reference table and utility functions. Its purpose is to
8924 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8925 rsaEncryption. This will allow some of the algorithm specific hackery
8926 needed to use the correct OID to be removed.
8930 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8931 structures for PKCS7_sign(). They are now set up by the relevant public
8936 * Add provisional EC pkey method with support for ECDSA and ECDH.
8940 * Add support for key derivation (agreement) in the API, DH method and
8945 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8946 public and private key formats. As a side effect these add additional
8947 command line functionality not previously available: DSA signatures can be
8948 generated and verified using pkeyutl and DH key support and generation in
8955 *Oliver Tappe <zooey@hirschkaefer.de>*
8957 * New make target "install_html_docs" installs HTML renditions of the
8960 *Oliver Tappe <zooey@hirschkaefer.de>*
8962 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8963 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8964 support key and parameter generation and add initial key generation
8965 functionality for RSA.
8969 * Add functions for main EVP_PKEY_method operations. The undocumented
8970 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8971 `EVP_PKEY_{encrypt,decrypt}_old`.
8975 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8976 key API, doesn't do much yet.
8980 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8981 public key algorithms. New option to openssl utility:
8982 "list-public-key-algorithms" to print out info.
8986 * Implement the Supported Elliptic Curves Extension for
8987 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8991 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8992 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8996 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8997 utilities such as rsa, dsa, dsaparam etc except they process any key
9002 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9003 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9004 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9009 * Initial support for pluggable public key ASN1.
9010 De-spaghettify the public key ASN1 handling. Move public and private
9011 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9012 algorithm specific handling to a single module within the relevant
9013 algorithm directory. Add functions to allow (near) opaque processing
9014 of public and private key structures.
9018 * Implement the Supported Point Formats Extension for
9019 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9023 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9024 for the psk identity [hint] and the psk callback functions to the
9025 SSL_SESSION, SSL and SSL_CTX structure.
9028 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9032 SSL_CTX_use_psk_identity_hint
9033 SSL_get_psk_identity_hint
9034 SSL_get_psk_identity
9035 SSL_use_psk_identity_hint
9037 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9039 * Add RFC 3161 compliant time stamp request creation, response generation
9040 and response verification functionality.
9042 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9044 * Add initial support for TLS extensions, specifically for the server_name
9045 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9046 have new members for a host name. The SSL data structure has an
9047 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9048 stored in that context to allow for session resumption, even after the
9049 SSL has been switched to a new SSL_CTX in reaction to a client's
9050 server_name extension.
9052 New functions (subject to change):
9054 SSL_get_servername()
9055 SSL_get_servername_type()
9058 New CTRL codes and macros (subject to change):
9060 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9061 - SSL_CTX_set_tlsext_servername_callback()
9062 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9063 - SSL_CTX_set_tlsext_servername_arg()
9064 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9066 openssl s_client has a new '-servername ...' option.
9068 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9069 '-key2 ...', '-servername_fatal' (subject to change). This allows
9070 testing the HostName extension for a specific single host name ('-cert'
9071 and '-key' remain fallbacks for handshakes without HostName
9072 negotiation). If the unrecognized_name alert has to be sent, this by
9073 default is a warning; it becomes fatal with the '-servername_fatal'
9076 *Peter Sylvester, Remy Allais, Christophe Renou*
9078 * Whirlpool hash implementation is added.
9082 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9083 bn(64,32). Because of instruction set limitations it doesn't have
9084 any negative impact on performance. This was done mostly in order
9085 to make it possible to share assembler modules, such as bn_mul_mont
9086 implementations, between 32- and 64-bit builds without hassle.
9090 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9091 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9096 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9097 dedicated Montgomery multiplication procedure, is introduced.
9098 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9099 "64-bit" performance on certain 32-bit targets.
9103 * New option SSL_OP_NO_COMP to disable use of compression selectively
9104 in SSL structures. New SSL ctrl to set maximum send fragment size.
9105 Save memory by setting the I/O buffer sizes dynamically instead of
9106 using the maximum available value.
9110 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9111 in addition to the text details.
9115 * Very, very preliminary EXPERIMENTAL support for printing of general
9116 ASN1 structures. This currently produces rather ugly output and doesn't
9117 handle several customised structures at all.
9121 * Integrated support for PVK file format and some related formats such
9122 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9123 these in the 'rsa' and 'dsa' utilities.
9127 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9131 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9132 place for the (very old) "NETSCAPE" format certificates which are now
9133 handled using new ASN1 code equivalents.
9137 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9138 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9139 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9143 * Modify CRL distribution points extension code to print out previously
9144 unsupported fields. Enhance extension setting code to allow setting of
9149 * Add print and set support for Issuing Distribution Point CRL extension.
9153 * Change 'Configure' script to enable Camellia by default.
9160 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9162 * When rejecting SSL/TLS records due to an incorrect version number, never
9163 update s->server with a new major version number. As of
9164 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9165 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9166 the previous behavior could result in a read attempt at NULL when
9167 receiving specific incorrect SSL/TLS records once record payload
9168 protection is active. ([CVE-2010-0740])
9170 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9172 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9173 could be crashed if the relevant tables were not present (e.g. chrooted).
9175 *Tomas Hoger <thoger@redhat.com>*
9177 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9179 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9181 *Martin Olsson, Neel Mehta*
9183 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9184 accommodate for stack sorting, always a write lock!).
9188 * On some versions of WIN32 Heap32Next is very slow. This can cause
9189 excessive delays in the RAND_poll(): over a minute. As a workaround
9190 include a time check in the inner Heap32Next loop too.
9194 * The code that handled flushing of data in SSL/TLS originally used the
9195 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9196 the problem outlined in PR#1949. The fix suggested there however can
9197 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9198 of Apache). So instead simplify the code to flush unconditionally.
9199 This should be fine since flushing with no data to flush is a no op.
9203 * Handle TLS versions 2.0 and later properly and correctly use the
9204 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9205 off ancient servers have a habit of sticking around for a while...
9209 * Modify compression code so it frees up structures without using the
9210 ex_data callbacks. This works around a problem where some applications
9211 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9212 restarting) then use compression (e.g. SSL with compression) later.
9213 This results in significant per-connection memory leaks and
9214 has caused some security issues including CVE-2008-1678 and
9219 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9220 change when encrypting or decrypting.
9224 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9225 connect and renegotiate with servers which do not support RI.
9226 Until RI is more widely deployed this option is enabled by default.
9230 * Add "missing" ssl ctrls to clear options and mode.
9234 * If client attempts to renegotiate and doesn't support RI respond with
9235 a no_renegotiation alert as required by RFC5746. Some renegotiating
9236 TLS clients will continue a connection gracefully when they receive
9237 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9238 waiting for a server hello which it will never receive. Now we treat a
9239 received no_renegotiation alert as a fatal error. This is because
9240 applications requesting a renegotiation might well expect it to succeed
9241 and would have no code in place to handle the server denying it so the
9242 only safe thing to do is to terminate the connection.
9246 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9247 peer supports secure renegotiation and 0 otherwise. Print out peer
9248 renegotiation support in s_client/s_server.
9252 * Replace the highly broken and deprecated SPKAC certification method with
9253 the updated NID creation version. This should correctly handle UTF8.
9257 * Implement RFC5746. Re-enable renegotiation but require the extension
9258 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9259 turns out to be a bad idea. It has been replaced by
9260 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9261 SSL_CTX_set_options(). This is really not recommended unless you
9262 know what you are doing.
9264 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9266 * Fixes to stateless session resumption handling. Use initial_ctx when
9267 issuing and attempting to decrypt tickets in case it has changed during
9268 servername handling. Use a non-zero length session ID when attempting
9269 stateless session resumption: this makes it possible to determine if
9270 a resumption has occurred immediately after receiving server hello
9271 (several places in OpenSSL subtly assume this) instead of later in
9276 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9277 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9278 fixes for a few places where the return code is not checked
9281 *Julia Lawall <julia@diku.dk>*
9283 * Add --strict-warnings option to Configure script to include devteam
9284 warnings in other configurations.
9288 * Add support for --libdir option and LIBDIR variable in makefiles. This
9289 makes it possible to install openssl libraries in locations which
9290 have names other than "lib", for example "/usr/lib64" which some
9293 *Steve Henson, based on patch from Jeremy Utley*
9295 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9296 X690 8.9.12 and can produce some misleading textual output of OIDs.
9298 *Steve Henson, reported by Dan Kaminsky*
9300 * Delete MD2 from algorithm tables. This follows the recommendation in
9301 several standards that it is not used in new applications due to
9302 several cryptographic weaknesses. For binary compatibility reasons
9303 the MD2 API is still compiled in by default.
9307 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9312 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9313 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9316 *Guenter <lists@gknw.net>*
9318 * Fix the server certificate chain building code to use X509_verify_cert(),
9319 it used to have an ad-hoc builder which was unable to cope with anything
9320 other than a simple chain.
9322 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9324 * Don't check self signed certificate signatures in X509_verify_cert()
9325 by default (a flag can override this): it just wastes time without
9326 adding any security. As a useful side effect self signed root CAs
9327 with non-FIPS digests are now usable in FIPS mode.
9331 * In dtls1_process_out_of_seq_message() the check if the current message
9332 is already buffered was missing. For every new message was memory
9333 allocated, allowing an attacker to perform an denial of service attack
9334 with sending out of seq handshake messages until there is no memory
9335 left. Additionally every future message was buffered, even if the
9336 sequence number made no sense and would be part of another handshake.
9337 So only messages with sequence numbers less than 10 in advance will be
9338 buffered. ([CVE-2009-1378])
9340 *Robin Seggelmann, discovered by Daniel Mentz*
9342 * Records are buffered if they arrive with a future epoch to be
9343 processed after finishing the corresponding handshake. There is
9344 currently no limitation to this buffer allowing an attacker to perform
9345 a DOS attack with sending records with future epochs until there is no
9346 memory left. This patch adds the pqueue_size() function to determine
9347 the size of a buffer and limits the record buffer to 100 entries.
9350 *Robin Seggelmann, discovered by Daniel Mentz*
9352 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9353 parent structure is freed. ([CVE-2009-1379])
9357 * Handle non-blocking I/O properly in SSL_shutdown() call.
9359 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9361 * Add `2.5.4.*` OIDs
9363 *Ilya O. <vrghost@gmail.com>*
9365 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9367 * Disable renegotiation completely - this fixes a severe security
9368 problem ([CVE-2009-3555]) at the cost of breaking all
9369 renegotiation. Renegotiation can be re-enabled by setting
9370 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9371 run-time. This is really not recommended unless you know what
9376 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9378 * Don't set val to NULL when freeing up structures, it is freed up by
9379 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9380 zeroing past the valid field. ([CVE-2009-0789])
9382 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9384 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9385 checked correctly. This would allow some invalid signed attributes to
9386 appear to verify correctly. ([CVE-2009-0591])
9388 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9390 * Reject UniversalString and BMPString types with invalid lengths. This
9391 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9392 a legal length. ([CVE-2009-0590])
9396 * Set S/MIME signing as the default purpose rather than setting it
9397 unconditionally. This allows applications to override it at the store
9402 * Permit restricted recursion of ASN1 strings. This is needed in practice
9403 to handle some structures.
9407 * Improve efficiency of mem_gets: don't search whole buffer each time
9410 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9412 * New -hex option for openssl rand.
9416 * Print out UTF8String and NumericString when parsing ASN1.
9420 * Support NumericString type for name components.
9424 * Allow CC in the environment to override the automatically chosen
9425 compiler. Note that nothing is done to ensure flags work with the
9430 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9432 * Properly check EVP_VerifyFinal() and similar return values
9435 *Ben Laurie, Bodo Moeller, Google Security Team*
9437 * Enable TLS extensions by default.
9441 * Allow the CHIL engine to be loaded, whether the application is
9442 multithreaded or not. (This does not release the developer from the
9443 obligation to set up the dynamic locking callbacks.)
9445 *Sander Temme <sander@temme.net>*
9447 * Use correct exit code if there is an error in dgst command.
9449 *Steve Henson; problem pointed out by Roland Dirlewanger*
9451 * Tweak Configure so that you need to say "experimental-jpake" to enable
9452 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9456 * Add experimental JPAKE support, including demo authentication in
9457 s_client and s_server.
9461 * Set the comparison function in v3_addr_canonize().
9463 *Rob Austein <sra@hactrn.net>*
9465 * Add support for XMPP STARTTLS in s_client.
9467 *Philip Paeps <philip@freebsd.org>*
9469 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9470 to ensure that even with this option, only ciphersuites in the
9471 server's preference list will be accepted. (Note that the option
9472 applies only when resuming a session, so the earlier behavior was
9473 just about the algorithm choice for symmetric cryptography.)
9477 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9479 * Fix NULL pointer dereference if a DTLS server received
9480 ChangeCipherSpec as first record ([CVE-2009-1386]).
9484 * Fix a state transition in s3_srvr.c and d1_srvr.c
9485 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9489 * The fix in 0.9.8c that supposedly got rid of unsafe
9490 double-checked locking was incomplete for RSA blinding,
9491 addressing just one layer of what turns out to have been
9492 doubly unsafe triple-checked locking.
9494 So now fix this for real by retiring the MONT_HELPER macro
9495 in crypto/rsa/rsa_eay.c.
9497 *Bodo Moeller; problem pointed out by Marius Schilder*
9499 * Various precautionary measures:
9501 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9503 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9504 (NB: This would require knowledge of the secret session ticket key
9505 to exploit, in which case you'd be SOL either way.)
9507 - Change bn_nist.c so that it will properly handle input BIGNUMs
9508 outside the expected range.
9510 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9513 *Neel Mehta, Bodo Moeller*
9515 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9516 the load fails. Useful for distros.
9518 *Ben Laurie and the FreeBSD team*
9520 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9524 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9528 * Expand ENGINE to support engine supplied SSL client certificate functions.
9530 This work was sponsored by Logica.
9534 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9535 keystores. Support for SSL/TLS client authentication too.
9536 Not compiled unless enable-capieng specified to Configure.
9538 This work was sponsored by Logica.
9542 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9543 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9544 attribute creation routines such as certificate requests and PKCS#12
9549 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9551 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9552 handshake which could lead to a client crash as found using the
9553 Codenomicon TLS test suite ([CVE-2008-1672])
9555 *Steve Henson, Mark Cox*
9557 * Fix double free in TLS server name extensions which could lead to
9558 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9562 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9564 Clear the error queue to ensure that error entries left from
9565 older function calls do not interfere with the correct operation.
9567 *Lutz Jaenicke, Erik de Castro Lopo*
9569 * Remove root CA certificates of commercial CAs:
9571 The OpenSSL project does not recommend any specific CA and does not
9572 have any policy with respect to including or excluding any CA.
9573 Therefore it does not make any sense to ship an arbitrary selection
9574 of root CA certificates with the OpenSSL software.
9578 * RSA OAEP patches to fix two separate invalid memory reads.
9579 The first one involves inputs when 'lzero' is greater than
9580 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9581 before the beginning of from). The second one involves inputs where
9582 the 'db' section contains nothing but zeroes (there is a one-byte
9583 invalid read after the end of 'db').
9585 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9587 * Partial backport from 0.9.9-dev:
9589 Introduce bn_mul_mont (dedicated Montgomery multiplication
9590 procedure) as a candidate for BIGNUM assembler implementation.
9591 While 0.9.9-dev uses assembler for various architectures, only
9592 x86_64 is available by default here in the 0.9.8 branch, and
9593 32-bit x86 is available through a compile-time setting.
9595 To try the 32-bit x86 assembler implementation, use Configure
9596 option "enable-montasm" (which exists only for this backport).
9598 As "enable-montasm" for 32-bit x86 disclaims code stability
9599 anyway, in this constellation we activate additional code
9600 backported from 0.9.9-dev for further performance improvements,
9601 namely BN_from_montgomery_word. (To enable this otherwise,
9602 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9604 *Andy Polyakov (backport partially by Bodo Moeller)*
9606 * Add TLS session ticket callback. This allows an application to set
9607 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9608 values. This is useful for key rollover for example where several key
9609 sets may exist with different names.
9613 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9614 This was broken until now in 0.9.8 releases, such that the only way
9615 a registered ENGINE could be used (assuming it initialises
9616 successfully on the host) was to explicitly set it as the default
9617 for the relevant algorithms. This is in contradiction with 0.9.7
9618 behaviour and the documentation. With this fix, when an ENGINE is
9619 registered into a given algorithm's table of implementations, the
9620 'uptodate' flag is reset so that auto-discovery will be used next
9621 time a new context for that algorithm attempts to select an
9624 *Ian Lister (tweaked by Geoff Thorpe)*
9626 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9627 implementation in the following ways:
9629 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9632 Lack of BER streaming support means one pass streaming processing is
9633 only supported if data is detached: setting the streaming flag is
9634 ignored for embedded content.
9636 CMS support is disabled by default and must be explicitly enabled
9637 with the enable-cms configuration option.
9641 * Update the GMP engine glue to do direct copies between BIGNUM and
9642 mpz_t when openssl and GMP use the same limb size. Otherwise the
9643 existing "conversion via a text string export" trick is still used.
9645 *Paul Sheer <paulsheer@gmail.com>*
9647 * Zlib compression BIO. This is a filter BIO which compressed and
9648 uncompresses any data passed through it.
9652 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9653 RFC3394 compatible AES key wrapping.
9657 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9658 sets string data without copying. X509_ALGOR_set0() and
9659 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9660 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9661 from an X509_ATTRIBUTE structure optionally checking it occurs only
9662 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9667 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9668 to get the expected BN_FLG_CONSTTIME behavior.
9670 *Bodo Moeller (Google)*
9674 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9675 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9676 - added some more tests to do_tests.pl
9677 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9678 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9679 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9680 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9681 - various changes to netware.pl to enable gcc-cross builds on Win32
9683 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9684 - various changes to fix missing prototype warnings
9685 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9686 - added AES, WHIRLPOOL and CPUID assembler code to build files
9687 - added missing AES assembler make rules to mk1mf.pl
9688 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9690 *Guenter Knauf <eflash@gmx.net>*
9692 * Implement certificate status request TLS extension defined in RFC3546.
9693 A client can set the appropriate parameters and receive the encoded
9694 OCSP response via a callback. A server can query the supplied parameters
9695 and set the encoded OCSP response in the callback. Add simplified examples
9696 to s_client and s_server.
9700 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9703 + Binary incompatibility of ssl_ctx_st structure
9704 + DTLS interoperation with non-compliant servers
9705 + Don't call get_session_cb() without proposed session
9706 + Fix ia64 assembler code
9708 *Andy Polyakov, Steve Henson*
9710 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9712 * DTLS Handshake overhaul. There were longstanding issues with
9713 OpenSSL DTLS implementation, which were making it impossible for
9714 RFC 4347 compliant client to communicate with OpenSSL server.
9715 Unfortunately just fixing these incompatibilities would "cut off"
9716 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9717 server keeps tolerating non RFC compliant syntax. The opposite is
9718 not true, 0.9.8f client can not communicate with earlier server.
9719 This update even addresses CVE-2007-4995.
9723 * Changes to avoid need for function casts in OpenSSL: some compilers
9724 (gcc 4.2 and later) reject their use.
9725 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9728 * Add RFC4507 support to OpenSSL. This includes the corrections in
9729 RFC4507bis. The encrypted ticket format is an encrypted encoded
9730 SSL_SESSION structure, that way new session features are automatically
9733 If a client application caches session in an SSL_SESSION structure
9734 support is transparent because tickets are now stored in the encoded
9737 The SSL_CTX structure automatically generates keys for ticket
9738 protection in servers so again support should be possible
9739 with no application modification.
9741 If a client or server wishes to disable RFC4507 support then the option
9742 SSL_OP_NO_TICKET can be set.
9744 Add a TLS extension debugging callback to allow the contents of any client
9745 or server extensions to be examined.
9747 This work was sponsored by Google.
9751 * Add initial support for TLS extensions, specifically for the server_name
9752 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9753 have new members for a host name. The SSL data structure has an
9754 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9755 stored in that context to allow for session resumption, even after the
9756 SSL has been switched to a new SSL_CTX in reaction to a client's
9757 server_name extension.
9759 New functions (subject to change):
9761 SSL_get_servername()
9762 SSL_get_servername_type()
9765 New CTRL codes and macros (subject to change):
9767 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9768 - SSL_CTX_set_tlsext_servername_callback()
9769 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9770 - SSL_CTX_set_tlsext_servername_arg()
9771 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9773 openssl s_client has a new '-servername ...' option.
9775 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9776 '-key2 ...', '-servername_fatal' (subject to change). This allows
9777 testing the HostName extension for a specific single host name ('-cert'
9778 and '-key' remain fallbacks for handshakes without HostName
9779 negotiation). If the unrecognized_name alert has to be sent, this by
9780 default is a warning; it becomes fatal with the '-servername_fatal'
9783 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9785 * Add AES and SSE2 assembly language support to VC++ build.
9789 * Mitigate attack on final subtraction in Montgomery reduction.
9793 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9794 (which previously caused an internal error).
9798 * Squeeze another 10% out of IGE mode when in != out.
9802 * AES IGE mode speedup.
9804 *Dean Gaudet (Google)*
9806 * Add the Korean symmetric 128-bit cipher SEED (see
9807 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9808 add SEED ciphersuites from RFC 4162:
9810 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9811 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9812 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9813 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9815 To minimize changes between patchlevels in the OpenSSL 0.9.8
9816 series, SEED remains excluded from compilation unless OpenSSL
9817 is configured with 'enable-seed'.
9819 *KISA, Bodo Moeller*
9821 * Mitigate branch prediction attacks, which can be practical if a
9822 single processor is shared, allowing a spy process to extract
9823 information. For detailed background information, see
9824 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9825 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9826 and Necessary Software Countermeasures"). The core of the change
9827 are new versions BN_div_no_branch() and
9828 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9829 respectively, which are slower, but avoid the security-relevant
9830 conditional branches. These are automatically called by BN_div()
9831 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9832 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9833 remove a conditional branch.
9835 BN_FLG_CONSTTIME is the new name for the previous
9836 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9837 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9838 in the exponent causes BN_mod_exp_mont() to use the alternative
9839 implementation in BN_mod_exp_mont_consttime().) The old name
9840 remains as a deprecated alias.
9842 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9843 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9844 constant-time implementations for more than just exponentiation.
9845 Here too the old name is kept as a deprecated alias.
9847 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9848 the BN_BLINDING structure gets an independent copy of the
9849 modulus. This means that the previous `BIGNUM *m` argument to
9850 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9851 essentially becomes `const BIGNUM *m`, although we can't actually
9852 change this in the header file before 0.9.9. It allows
9853 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9854 enable BN_FLG_CONSTTIME.
9856 *Matthew D Wood (Intel Corp)*
9858 * In the SSL/TLS server implementation, be strict about session ID
9859 context matching (which matters if an application uses a single
9860 external cache for different purposes). Previously,
9861 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9862 set. This did ensure strict client verification, but meant that,
9863 with applications using a single external cache for quite
9864 different requirements, clients could circumvent ciphersuite
9865 restrictions for a given session ID context by starting a session
9866 in a different context.
9870 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9871 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9872 authentication-only ciphersuites.
9876 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9877 not complete and could lead to a possible single byte overflow
9878 ([CVE-2007-5135]) [Ben Laurie]
9880 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9882 * Since AES128 and AES256 (and similarly Camellia128 and
9883 Camellia256) share a single mask bit in the logic of
9884 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9885 kludge to work properly if AES128 is available and AES256 isn't
9886 (or if Camellia128 is available and Camellia256 isn't).
9890 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9891 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9892 When a point or a seed is encoded in a BIT STRING, we need to
9893 prevent the removal of trailing zero bits to get the proper DER
9894 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9895 of a NamedBitList, for which trailing 0 bits need to be removed.)
9899 * Have SSL/TLS server implementation tolerate "mismatched" record
9900 protocol version while receiving ClientHello even if the
9901 ClientHello is fragmented. (The server can't insist on the
9902 particular protocol version it has chosen before the ServerHello
9903 message has informed the client about his choice.)
9907 * Add RFC 3779 support.
9909 *Rob Austein for ARIN, Ben Laurie*
9911 * Load error codes if they are not already present instead of using a
9912 static variable. This allows them to be cleanly unloaded and reloaded.
9913 Improve header file function name parsing.
9917 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9918 or CAPABILITY handshake as required by RFCs.
9922 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9924 * Introduce limits to prevent malicious keys being able to
9925 cause a denial of service. ([CVE-2006-2940])
9927 *Steve Henson, Bodo Moeller*
9929 * Fix ASN.1 parsing of certain invalid structures that can result
9930 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9932 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9933 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9935 * Fix SSL client code which could crash if connecting to a
9936 malicious SSLv2 server. ([CVE-2006-4343])
9938 *Tavis Ormandy and Will Drewry, Google Security Team*
9940 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9941 match only those. Before that, "AES256-SHA" would be interpreted
9942 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9943 the same strength classification in 0.9.7h) as we currently only
9944 have a single AES bit in the ciphersuite description bitmap.
9945 That change, however, also applied to ciphersuite strings such as
9946 "RC4-MD5" that intentionally matched multiple ciphersuites --
9947 namely, SSL 2.0 ciphersuites in addition to the more common ones
9948 from SSL 3.0/TLS 1.0.
9950 So we change the selection algorithm again: Naming an explicit
9951 ciphersuite selects this one ciphersuite, and any other similar
9952 ciphersuite (same bitmap) from *other* protocol versions.
9953 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9954 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9956 Since SSL 2.0 does not have any ciphersuites for which the
9957 128/256 bit distinction would be relevant, this works for now.
9958 The proper fix will be to use different bits for AES128 and
9959 AES256, which would have avoided the problems from the beginning;
9960 however, bits are scarce, so we can only do this in a new release
9961 (not just a patchlevel) when we can change the SSL_CIPHER
9962 definition to split the single 'unsigned long mask' bitmap into
9963 multiple values to extend the available space.
9967 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9969 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9970 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9972 * Add AES IGE and biIGE modes.
9976 * Change the Unix randomness entropy gathering to use poll() when
9977 possible instead of select(), since the latter has some
9978 undesirable limitations.
9980 *Darryl Miles via Richard Levitte and Bodo Moeller*
9982 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9983 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9984 cannot be implicitly activated as part of, e.g., the "AES" alias.
9985 However, please upgrade to OpenSSL 0.9.9[-dev] for
9986 non-experimental use of the ECC ciphersuites to get TLS extension
9987 support, which is required for curve and point format negotiation
9988 to avoid potential handshake problems.
9992 * Disable rogue ciphersuites:
9994 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9995 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9996 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9998 The latter two were purportedly from
9999 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10002 Also deactivate the remaining ciphersuites from
10003 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10004 unofficial, and the ID has long expired.
10008 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10009 dual-core machines) and other potential thread-safety issues.
10013 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10014 versions), which is now available for royalty-free use
10015 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10016 Also, add Camellia TLS ciphersuites from RFC 4132.
10018 To minimize changes between patchlevels in the OpenSSL 0.9.8
10019 series, Camellia remains excluded from compilation unless OpenSSL
10020 is configured with 'enable-camellia'.
10024 * Disable the padding bug check when compression is in use. The padding
10025 bug check assumes the first packet is of even length, this is not
10026 necessarily true if compression is enabled and can result in false
10027 positives causing handshake failure. The actual bug test is ancient
10028 code so it is hoped that implementations will either have fixed it by
10029 now or any which still have the bug do not support compression.
10033 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10035 * When applying a cipher rule check to see if string match is an explicit
10036 cipher suite and only match that one cipher suite if it is.
10040 * Link in manifests for VC++ if needed.
10042 *Austin Ziegler <halostatue@gmail.com>*
10044 * Update support for ECC-based TLS ciphersuites according to
10045 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10046 TLS extensions, which are supported starting with the 0.9.9
10047 branch, not in the OpenSSL 0.9.8 branch).
10051 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10052 opaque EVP_CIPHER_CTX handling.
10056 * Fixes and enhancements to zlib compression code. We now only use
10057 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10058 to conform with the standards mentioned here:
10059 <http://www.zlib.net/DLL_FAQ.txt>
10060 Static zlib linking now works on Windows and the new --with-zlib-include
10061 --with-zlib-lib options to Configure can be used to supply the location
10062 of the headers and library. Gracefully handle case where zlib library
10067 * Several fixes and enhancements to the OID generation code. The old code
10068 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10069 handle numbers larger than ULONG_MAX, truncated printing and had a
10070 non standard OBJ_obj2txt() behaviour.
10074 * Add support for building of engines under engine/ as shared libraries
10075 under VC++ build system.
10079 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10080 Hopefully, we will not see any false combination of paths any more.
10084 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10086 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10087 (part of SSL_OP_ALL). This option used to disable the
10088 countermeasure against man-in-the-middle protocol-version
10089 rollback in the SSL 2.0 server implementation, which is a bad
10090 idea. ([CVE-2005-2969])
10092 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10093 for Information Security, National Institute of Advanced Industrial
10094 Science and Technology [AIST], Japan)*
10096 * Add two function to clear and return the verify parameter flags.
10100 * Keep cipherlists sorted in the source instead of sorting them at
10101 runtime, thus removing the need for a lock.
10105 * Avoid some small subgroup attacks in Diffie-Hellman.
10107 *Nick Mathewson and Ben Laurie*
10109 * Add functions for well-known primes.
10113 * Extended Windows CE support.
10115 *Satoshi Nakamura and Andy Polyakov*
10117 * Initialize SSL_METHOD structures at compile time instead of during
10118 runtime, thus removing the need for a lock.
10122 * Make PKCS7_decrypt() work even if no certificate is supplied by
10123 attempting to decrypt each encrypted key in turn. Add support to
10128 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10130 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10133 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10137 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10138 key into the same file any more.
10142 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10146 * Add -utf8 command line and config file option to 'ca'.
10148 *Stefan <stf@udoma.org*
10150 * Removed the macro des_crypt(), as it seems to conflict with some
10151 libraries. Use DES_crypt().
10155 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10156 involves renaming the source and generated shared-libs for
10157 both. The engines will accept the corrected or legacy ids
10158 ('ncipher' and '4758_cca' respectively) when binding. NB,
10159 this only applies when building 'shared'.
10161 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10163 * Add attribute functions to EVP_PKEY structure. Modify
10164 PKCS12_create() to recognize a CSP name attribute and
10165 use it. Make -CSP option work again in pkcs12 utility.
10169 * Add new functionality to the bn blinding code:
10170 - automatic re-creation of the BN_BLINDING parameters after
10171 a fixed number of uses (currently 32)
10172 - add new function for parameter creation
10173 - introduce flags to control the update behaviour of the
10174 BN_BLINDING parameters
10175 - hide BN_BLINDING structure
10176 Add a second BN_BLINDING slot to the RSA structure to improve
10177 performance when a single RSA object is shared among several
10182 * Add support for DTLS.
10184 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10186 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10187 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10191 * Remove buggy and incomplete DH cert support from
10192 ssl/ssl_rsa.c and ssl/s3_both.c
10196 * Use SHA-1 instead of MD5 as the default digest algorithm for
10197 the `apps/openssl` commands.
10201 * Compile clean with "-Wall -Wmissing-prototypes
10202 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10203 DEBUG_SAFESTACK must also be set.
10207 * Change ./Configure so that certain algorithms can be disabled by default.
10208 The new counterpiece to "no-xxx" is "enable-xxx".
10210 The patented RC5 and MDC2 algorithms will now be disabled unless
10211 "enable-rc5" and "enable-mdc2", respectively, are specified.
10213 (IDEA remains enabled despite being patented. This is because IDEA
10214 is frequently required for interoperability, and there is no license
10215 fee for non-commercial use. As before, "no-idea" can be used to
10216 avoid this algorithm.)
10220 * Add processing of proxy certificates (see RFC 3820). This work was
10221 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10222 EGEE (Enabling Grids for E-science in Europe).
10226 * RC4 performance overhaul on modern architectures/implementations, such
10227 as Intel P4, IA-64 and AMD64.
10231 * New utility extract-section.pl. This can be used specify an alternative
10232 section number in a pod file instead of having to treat each file as
10233 a separate case in Makefile. This can be done by adding two lines to the
10236 =for comment openssl_section:XXX
10238 The blank line is mandatory.
10242 * New arguments -certform, -keyform and -pass for s_client and s_server
10243 to allow alternative format key and certificate files and passphrase
10248 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10249 update associated structures and add various utility functions.
10251 Add new policy related verify parameters, include policy checking in
10252 standard verify code. Enhance 'smime' application with extra parameters
10253 to support policy checking and print out.
10257 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10258 Nehemiah processors. These extensions support AES encryption in hardware
10259 as well as RNG (though RNG support is currently disabled).
10261 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10263 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10267 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10269 *Andy Polyakov and a number of other people*
10271 * Improved PowerPC platform support. Most notably BIGNUM assembler
10272 implementation contributed by IBM.
10274 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10276 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10277 exponent rather than 'unsigned long'. There is a corresponding change to
10278 the new 'rsa_keygen' element of the RSA_METHOD structure.
10280 *Jelte Jansen, Geoff Thorpe*
10282 * Functionality for creating the initial serial number file is now
10283 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10285 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10286 number file to 1, which is bound to cause problems. To avoid
10287 the problems while respecting compatibility between different 0.9.7
10288 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10289 CA.pl for serial number initialization. With the new release 0.9.8,
10290 we can fix the problem directly in the 'ca' utility.)
10294 * Reduced header interdependencies by declaring more opaque objects in
10295 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10296 give fewer recursive includes, which could break lazy source code - so
10297 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10298 developers should define this symbol when building and using openssl to
10299 ensure they track the recommended behaviour, interfaces, [etc], but
10300 backwards-compatible behaviour prevails when this isn't defined.
10304 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10308 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10309 This will generate a random key of the appropriate length based on the
10310 cipher context. The EVP_CIPHER can provide its own random key generation
10311 routine to support keys of a specific form. This is used in the des and
10312 3des routines to generate a key of the correct parity. Update S/MIME
10313 code to use new functions and hence generate correct parity DES keys.
10314 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10315 valid (weak or incorrect parity).
10319 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10320 as looking them up. This is useful when the verified structure may contain
10321 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10322 present unless the new PKCS7_NO_CRL flag is asserted.
10326 * Extend ASN1 oid configuration module. It now additionally accepts the
10329 shortName = some long name, 1.2.3.4
10333 * Reimplemented the BN_CTX implementation. There is now no more static
10334 limitation on the number of variables it can handle nor the depth of the
10335 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10336 information can now expand as required, and rather than having a single
10337 static array of bignums, BN_CTX now uses a linked-list of such arrays
10338 allowing it to expand on demand whilst maintaining the usefulness of
10339 BN_CTX's "bundling".
10343 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10344 to allow all RSA operations to function using a single BN_CTX.
10348 * Preliminary support for certificate policy evaluation and checking. This
10349 is initially intended to pass the tests outlined in "Conformance Testing
10350 of Relying Party Client Certificate Path Processing Logic" v1.07.
10354 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10355 remained unused and not that useful. A variety of other little bignum
10356 tweaks and fixes have also been made continuing on from the audit (see
10361 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10362 associated ASN1, EVP and SSL functions and old ASN1 macros.
10366 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10367 and this should never fail. So the return value from the use of
10368 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10369 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10373 * BN_CTX_get() should return zero-valued bignums, providing the same
10374 initialised value as BN_new().
10376 *Geoff Thorpe, suggested by Ulf Möller*
10378 * Support for inhibitAnyPolicy certificate extension.
10382 * An audit of the BIGNUM code is underway, for which debugging code is
10383 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10384 is considered valid when processing BIGNUMs, and causes execution to
10385 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10386 further steps are taken to deliberately pollute unused data in BIGNUM
10387 structures to try and expose faulty code further on. For now, openssl will
10388 (in its default mode of operation) continue to tolerate the inconsistent
10389 forms that it has tolerated in the past, but authors and packagers should
10390 consider trying openssl and their own applications when compiled with
10391 these debugging symbols defined. It will help highlight potential bugs in
10392 their own code, and will improve the test coverage for OpenSSL itself. At
10393 some point, these tighter rules will become openssl's default to improve
10394 maintainability, though the assert()s and other overheads will remain only
10395 in debugging configurations. See bn.h for more details.
10397 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10399 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10400 that can only be obtained through BN_CTX_new() (which implicitly
10401 initialises it). The presence of this function only made it possible
10402 to overwrite an existing structure (and cause memory leaks).
10406 * Because of the callback-based approach for implementing LHASH as a
10407 template type, lh_insert() adds opaque objects to hash-tables and
10408 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10409 to clean up those corresponding objects before destroying the hash table
10410 (and losing the object pointers). So some over-zealous constifications in
10411 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10412 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10413 prototyped to have "const" restrictions on the object pointers they are
10414 given (and so aren't required to cast them away any more).
10418 * The tmdiff.h API was so ugly and minimal that our own timing utility
10419 (speed) prefers to use its own implementation. The two implementations
10420 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10421 its object type properly exposed (MS_TM) instead of casting to/from
10422 `char *`. This may still change yet if someone realises MS_TM and
10424 aren't necessarily the greatest nomenclatures - but this is what was used
10425 internally to the implementation so I've used that for now.
10429 * Ensure that deprecated functions do not get compiled when
10430 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10431 the self-tests were still using deprecated key-generation functions so
10432 these have been updated also.
10436 * Reorganise PKCS#7 code to separate the digest location functionality
10437 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10438 New function PKCS7_set_digest() to set the digest type for PKCS#7
10439 digestedData type. Add additional code to correctly generate the
10440 digestedData type and add support for this type in PKCS7 initialization
10445 * New function PKCS7_set0_type_other() this initializes a PKCS7
10446 structure of type "other".
10450 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10451 sure the loop does correctly stop and breaking ("division by zero")
10452 modulus operations are not performed. The (pre-generated) prime
10453 table crypto/bn/bn_prime.h was already correct, but it could not be
10454 re-generated on some platforms because of the "division by zero"
10455 situation in the script.
10457 *Ralf S. Engelschall*
10459 * Update support for ECC-based TLS ciphersuites according to
10460 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10461 SHA-1 now is only used for "small" curves (where the
10462 representation of a field element takes up to 24 bytes); for
10463 larger curves, the field element resulting from ECDH is directly
10464 used as premaster secret.
10466 *Douglas Stebila (Sun Microsystems Laboratories)*
10468 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10469 curve secp160r1 to the tests.
10471 *Douglas Stebila (Sun Microsystems Laboratories)*
10473 * Add the possibility to load symbols globally with DSO.
10475 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10477 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10478 control of the error stack.
10482 * Add support for STORE in ENGINE.
10486 * Add the STORE type. The intention is to provide a common interface
10487 to certificate and key stores, be they simple file-based stores, or
10488 HSM-type store, or LDAP stores, or...
10489 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10493 * Add a generic structure called OPENSSL_ITEM. This can be used to
10494 pass a list of arguments to any function as well as provide a way
10495 for a function to pass data back to the caller.
10499 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10500 works like BUF_strdup() but can be used to duplicate a portion of
10501 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10506 * Add the function sk_find_ex() which works like sk_find(), but will
10507 return an index to an element even if an exact match couldn't be
10508 found. The index is guaranteed to point at the element where the
10509 searched-for key would be inserted to preserve sorting order.
10513 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10514 takes an extra flags argument for optional functionality. Currently,
10515 the following flags are defined:
10517 OBJ_BSEARCH_VALUE_ON_NOMATCH
10518 This one gets OBJ_bsearch_ex() to return a pointer to the first
10519 element where the comparing function returns a negative or zero
10522 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10523 This one gets OBJ_bsearch_ex() to return a pointer to the first
10524 element where the comparing function returns zero. This is useful
10525 if there are more than one element where the comparing function
10530 * Make it possible to create self-signed certificates with 'openssl ca'
10531 in such a way that the self-signed certificate becomes part of the
10532 CA database and uses the same mechanisms for serial number generation
10533 as all other certificate signing. The new flag '-selfsign' enables
10534 this functionality. Adapt CA.sh and CA.pl.in.
10538 * Add functionality to check the public key of a certificate request
10539 against a given private. This is useful to check that a certificate
10540 request can be signed by that key (self-signing).
10544 * Make it possible to have multiple active certificates with the same
10545 subject in the CA index file. This is done only if the keyword
10546 'unique_subject' is set to 'no' in the main CA section (default
10547 if 'CA_default') of the configuration file. The value is saved
10548 with the database itself in a separate index attribute file,
10549 named like the index file with '.attr' appended to the name.
10553 * Generate multi-valued AVAs using '+' notation in config files for
10558 * Support for nameConstraints certificate extension.
10562 * Support for policyConstraints certificate extension.
10566 * Support for policyMappings certificate extension.
10570 * Make sure the default DSA_METHOD implementation only uses its
10571 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10572 and change its own handlers to be NULL so as to remove unnecessary
10573 indirection. This lets alternative implementations fallback to the
10574 default implementation more easily.
10578 * Support for directoryName in GeneralName related extensions
10583 * Make it possible to link applications using Makefile.shared.
10584 Make that possible even when linking against static libraries!
10588 * Support for single pass processing for S/MIME signing. This now
10589 means that S/MIME signing can be done from a pipe, in addition
10590 cleartext signing (multipart/signed type) is effectively streaming
10591 and the signed data does not need to be all held in memory.
10593 This is done with a new flag PKCS7_STREAM. When this flag is set
10594 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10595 is done after the data is output (and digests calculated) in
10596 SMIME_write_PKCS7().
10600 * Add full support for -rpath/-R, both in shared libraries and
10601 applications, at least on the platforms where it's known how
10606 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10607 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10608 will now compute a table of multiples of the generator that
10609 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10610 faster (notably in the case of a single point multiplication,
10611 scalar * generator).
10613 *Nils Larsch, Bodo Moeller*
10615 * IPv6 support for certificate extensions. The various extensions
10616 which use the IP:a.b.c.d can now take IPv6 addresses using the
10617 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10622 * Added an ENGINE that implements RSA by performing private key
10623 exponentiations with the GMP library. The conversions to and from
10624 GMP's mpz_t format aren't optimised nor are any montgomery forms
10625 cached, and on x86 it appears OpenSSL's own performance has caught up.
10626 However there are likely to be other architectures where GMP could
10627 provide a boost. This ENGINE is not built in by default, but it can be
10628 specified at Configure time and should be accompanied by the necessary
10629 linker additions, eg;
10630 ./config -DOPENSSL_USE_GMP -lgmp
10634 * "openssl engine" will not display ENGINE/DSO load failure errors when
10635 testing availability of engines with "-t" - the old behaviour is
10636 produced by increasing the feature's verbosity with "-tt".
10640 * ECDSA routines: under certain error conditions uninitialized BN objects
10641 could be freed. Solution: make sure initialization is performed early
10642 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10647 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10648 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10649 software implementations. For DSA and DH, parameter generation can
10650 also be overridden by providing the appropriate method callbacks.
10654 * Change the "progress" mechanism used in key-generation and
10655 primality testing to functions that take a new BN_GENCB pointer in
10656 place of callback/argument pairs. The new API functions have `_ex`
10657 postfixes and the older functions are reimplemented as wrappers for
10658 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10659 declarations of the old functions to help (graceful) attempts to
10660 migrate to the new functions. Also, the new key-generation API
10661 functions operate on a caller-supplied key-structure and return
10662 success/failure rather than returning a key or NULL - this is to
10663 help make "keygen" another member function of RSA_METHOD etc.
10665 Example for using the new callback interface:
10667 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10668 void *my_arg = ...;
10671 BN_GENCB_set(&my_cb, my_callback, my_arg);
10673 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10674 /* For the meaning of a, b in calls to my_callback(), see the
10675 * documentation of the function that calls the callback.
10676 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10677 * my_callback should return 1 if it wants BN_is_prime_ex()
10678 * to continue, or 0 to stop.
10683 * Change the ZLIB compression method to be stateful, and make it
10684 available to TLS with the number defined in
10685 draft-ietf-tls-compression-04.txt.
10689 * Add the ASN.1 structures and functions for CertificatePair, which
10690 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10692 CertificatePair ::= SEQUENCE {
10693 forward [0] Certificate OPTIONAL,
10694 reverse [1] Certificate OPTIONAL,
10695 -- at least one of the pair shall be present -- }
10697 Also implement the PEM functions to read and write certificate
10698 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10700 This needed to be defined, mostly for the sake of the LDAP
10701 attribute crossCertificatePair, but may prove useful elsewhere as
10706 * Make it possible to inhibit symlinking of shared libraries in
10707 Makefile.shared, for Cygwin's sake.
10711 * Extend the BIGNUM API by creating a function
10712 void BN_set_negative(BIGNUM *a, int neg);
10713 and a macro that behave like
10714 int BN_is_negative(const BIGNUM *a);
10716 to avoid the need to access 'a->neg' directly in applications.
10720 * Implement fast modular reduction for pseudo-Mersenne primes
10721 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10722 EC_GROUP_new_curve_GFp() will now automatically use this
10725 *Nils Larsch <nla@trustcenter.de>*
10727 * Add new lock type (CRYPTO_LOCK_BN).
10731 * Change the ENGINE framework to automatically load engines
10732 dynamically from specific directories unless they could be
10733 found to already be built in or loaded. Move all the
10734 current engines except for the cryptodev one to a new
10735 directory engines/.
10736 The engines in engines/ are built as shared libraries if
10737 the "shared" options was given to ./Configure or ./config.
10738 Otherwise, they are inserted in libcrypto.a.
10739 /usr/local/ssl/engines is the default directory for dynamic
10740 engines, but that can be overridden at configure time through
10741 the usual use of --prefix and/or --openssldir, and at run
10742 time with the environment variable OPENSSL_ENGINES.
10744 *Geoff Thorpe and Richard Levitte*
10746 * Add Makefile.shared, a helper makefile to build shared
10747 libraries. Adapt Makefile.org.
10751 * Add version info to Win32 DLLs.
10753 *Peter 'Luna' Runestig" <peter@runestig.com>*
10755 * Add new 'medium level' PKCS#12 API. Certificates and keys
10756 can be added using this API to created arbitrary PKCS#12
10757 files while avoiding the low-level API.
10759 New options to PKCS12_create(), key or cert can be NULL and
10760 will then be omitted from the output file. The encryption
10761 algorithm NIDs can be set to -1 for no encryption, the mac
10762 iteration count can be set to 0 to omit the mac.
10764 Enhance pkcs12 utility by making the -nokeys and -nocerts
10765 options work when creating a PKCS#12 file. New option -nomac
10766 to omit the mac, NONE can be set for an encryption algorithm.
10767 New code is modified to use the enhanced PKCS12_create()
10768 instead of the low-level API.
10772 * Extend ASN1 encoder to support indefinite length constructed
10773 encoding. This can output sequences tags and octet strings in
10774 this form. Modify pk7_asn1.c to support indefinite length
10775 encoding. This is experimental and needs additional code to
10776 be useful, such as an ASN1 bio and some enhanced streaming
10779 Extend template encode functionality so that tagging is passed
10780 down to the template encoder.
10784 * Let 'openssl req' fail if an argument to '-newkey' is not
10785 recognized instead of using RSA as a default.
10789 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10790 As these are not official, they are not included in "ALL";
10791 the "ECCdraft" ciphersuite group alias can be used to select them.
10793 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10795 * Add ECDH engine support.
10797 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10799 * Add ECDH in new directory crypto/ecdh/.
10801 *Douglas Stebila (Sun Microsystems Laboratories)*
10803 * Let BN_rand_range() abort with an error after 100 iterations
10804 without success (which indicates a broken PRNG).
10808 * Change BN_mod_sqrt() so that it verifies that the input value
10809 is really the square of the return value. (Previously,
10810 BN_mod_sqrt would show GIGO behaviour.)
10814 * Add named elliptic curves over binary fields from X9.62, SECG,
10815 and WAP/WTLS; add OIDs that were still missing.
10817 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10819 * Extend the EC library for elliptic curves over binary fields
10820 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10823 EC_GF2m_simple_method
10827 EC_GROUP_new_curve_GF2m
10828 EC_GROUP_set_curve_GF2m
10829 EC_GROUP_get_curve_GF2m
10830 EC_POINT_set_affine_coordinates_GF2m
10831 EC_POINT_get_affine_coordinates_GF2m
10832 EC_POINT_set_compressed_coordinates_GF2m
10834 Point compression for binary fields is disabled by default for
10835 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10838 As binary polynomials are represented as BIGNUMs, various members
10839 of the EC_GROUP and EC_POINT data structures can be shared
10840 between the implementations for prime fields and binary fields;
10841 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10842 are essentially identical to their `..._GFp` counterparts.
10843 (For simplicity, the `..._GFp` prefix has been dropped from
10844 various internal method names.)
10846 An internal 'field_div' method (similar to 'field_mul' and
10847 'field_sqr') has been added; this is used only for binary fields.
10849 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10851 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10852 through methods ('mul', 'precompute_mult').
10854 The generic implementations (now internally called 'ec_wNAF_mul'
10855 and 'ec_wNAF_precomputed_mult') remain the default if these
10856 methods are undefined.
10858 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10860 * New function EC_GROUP_get_degree, which is defined through
10861 EC_METHOD. For curves over prime fields, this returns the bit
10862 length of the modulus.
10864 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10866 * New functions EC_GROUP_dup, EC_POINT_dup.
10867 (These simply call ..._new and ..._copy).
10869 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10871 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10872 Polynomials are represented as BIGNUMs (where the sign bit is not
10873 used) in the following functions [macros]:
10876 BN_GF2m_sub [= BN_GF2m_add]
10877 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10878 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10879 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10881 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10882 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10883 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10884 BN_GF2m_cmp [= BN_ucmp]
10886 (Note that only the 'mod' functions are actually for fields GF(2^m).
10887 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10889 For some functions, an the irreducible polynomial defining a
10890 field can be given as an 'unsigned int[]' with strictly
10891 decreasing elements giving the indices of those bits that are set;
10892 i.e., p[] represents the polynomial
10893 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10895 p[0] > p[1] > ... > p[k] = 0.
10896 This applies to the following functions:
10899 BN_GF2m_mod_mul_arr
10900 BN_GF2m_mod_sqr_arr
10901 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10902 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10903 BN_GF2m_mod_exp_arr
10904 BN_GF2m_mod_sqrt_arr
10905 BN_GF2m_mod_solve_quad_arr
10909 Conversion can be performed by the following functions:
10914 bntest.c has additional tests for binary polynomial arithmetic.
10916 Two implementations for BN_GF2m_mod_div() are available.
10917 The default algorithm simply uses BN_GF2m_mod_inv() and
10918 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10919 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10920 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10922 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10924 * Add new error code 'ERR_R_DISABLED' that can be used when some
10925 functionality is disabled at compile-time.
10927 *Douglas Stebila <douglas.stebila@sun.com>*
10929 * Change default behaviour of 'openssl asn1parse' so that more
10930 information is visible when viewing, e.g., a certificate:
10932 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10933 mode the content of non-printable OCTET STRINGs is output in a
10934 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10935 avoid the appearance of a printable string.
10937 *Nils Larsch <nla@trustcenter.de>*
10939 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10941 EC_GROUP_set_asn1_flag()
10942 EC_GROUP_get_asn1_flag()
10943 EC_GROUP_set_point_conversion_form()
10944 EC_GROUP_get_point_conversion_form()
10945 These control ASN1 encoding details:
10946 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10947 has been set to OPENSSL_EC_NAMED_CURVE.
10948 - Points are encoded in uncompressed form by default; options for
10949 asn1_for are as for point2oct, namely
10950 POINT_CONVERSION_COMPRESSED
10951 POINT_CONVERSION_UNCOMPRESSED
10952 POINT_CONVERSION_HYBRID
10954 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10956 EC_GROUP_set_seed()
10957 EC_GROUP_get0_seed()
10958 EC_GROUP_get_seed_len()
10959 This is used only for ASN1 purposes (so far).
10961 *Nils Larsch <nla@trustcenter.de>*
10963 * Add 'field_type' member to EC_METHOD, which holds the NID
10964 of the appropriate field type OID. The new function
10965 EC_METHOD_get_field_type() returns this value.
10967 *Nils Larsch <nla@trustcenter.de>*
10970 EC_POINT_point2bn()
10971 EC_POINT_bn2point()
10972 EC_POINT_point2hex()
10973 EC_POINT_hex2point()
10974 providing useful interfaces to EC_POINT_point2oct() and
10975 EC_POINT_oct2point().
10977 *Nils Larsch <nla@trustcenter.de>*
10979 * Change internals of the EC library so that the functions
10980 EC_GROUP_set_generator()
10981 EC_GROUP_get_generator()
10982 EC_GROUP_get_order()
10983 EC_GROUP_get_cofactor()
10984 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10985 to methods, which would lead to unnecessary code duplication when
10986 adding different types of curves.
10988 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10990 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10991 arithmetic, and such that modified wNAFs are generated
10992 (which avoid length expansion in many cases).
10996 * Add a function EC_GROUP_check_discriminant() (defined via
10997 EC_METHOD) that verifies that the curve discriminant is non-zero.
10999 Add a function EC_GROUP_check() that makes some sanity tests
11000 on a EC_GROUP, its generator and order. This includes
11001 EC_GROUP_check_discriminant().
11003 *Nils Larsch <nla@trustcenter.de>*
11005 * Add ECDSA in new directory crypto/ecdsa/.
11007 Add applications 'openssl ecparam' and 'openssl ecdsa'
11008 (these are based on 'openssl dsaparam' and 'openssl dsa').
11010 ECDSA support is also included in various other files across the
11011 library. Most notably,
11012 - 'openssl req' now has a '-newkey ecdsa:file' option;
11013 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11014 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11015 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11016 them suitable for ECDSA where domain parameters must be
11017 extracted before the specific public key;
11018 - ECDSA engine support has been added.
11020 *Nils Larsch <nla@trustcenter.de>*
11022 * Include some named elliptic curves, and add OIDs from X9.62,
11023 SECG, and WAP/WTLS. Each curve can be obtained from the new
11025 EC_GROUP_new_by_curve_name(),
11026 and the list of available named curves can be obtained with
11027 EC_get_builtin_curves().
11028 Also add a 'curve_name' member to EC_GROUP objects, which can be
11030 EC_GROUP_set_curve_name()
11031 EC_GROUP_get_curve_name()
11033 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11035 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11036 was actually never needed) and in BN_mul(). The removal in BN_mul()
11037 required a small change in bn_mul_part_recursive() and the addition
11038 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11039 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11040 bn_sub_words() and bn_add_words() except they take arrays with
11045 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11047 * Cleanse PEM buffers before freeing them since they may contain
11050 *Benjamin Bennett <ben@psc.edu>*
11052 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11053 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11054 authentication-only ciphersuites.
11058 * Since AES128 and AES256 share a single mask bit in the logic of
11059 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11060 kludge to work properly if AES128 is available and AES256 isn't.
11064 * Expand security boundary to match 1.1.1 module.
11068 * Remove redundant features: hash file source, editing of test vectors
11069 modify fipsld to use external fips_premain.c signature.
11073 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11074 run algorithm test programs.
11078 * Make algorithm test programs more tolerant of whitespace.
11082 * Have SSL/TLS server implementation tolerate "mismatched" record
11083 protocol version while receiving ClientHello even if the
11084 ClientHello is fragmented. (The server can't insist on the
11085 particular protocol version it has chosen before the ServerHello
11086 message has informed the client about his choice.)
11090 * Load error codes if they are not already present instead of using a
11091 static variable. This allows them to be cleanly unloaded and reloaded.
11095 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11097 * Introduce limits to prevent malicious keys being able to
11098 cause a denial of service. ([CVE-2006-2940])
11100 *Steve Henson, Bodo Moeller*
11102 * Fix ASN.1 parsing of certain invalid structures that can result
11103 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11105 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11106 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11108 * Fix SSL client code which could crash if connecting to a
11109 malicious SSLv2 server. ([CVE-2006-4343])
11111 *Tavis Ormandy and Will Drewry, Google Security Team*
11113 * Change ciphersuite string processing so that an explicit
11114 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11115 will no longer include "AES128-SHA"), and any other similar
11116 ciphersuite (same bitmap) from *other* protocol versions (so that
11117 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11118 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11119 changes from 0.9.8b and 0.9.8d.
11123 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11125 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11126 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11128 * Change the Unix randomness entropy gathering to use poll() when
11129 possible instead of select(), since the latter has some
11130 undesirable limitations.
11132 *Darryl Miles via Richard Levitte and Bodo Moeller*
11134 * Disable rogue ciphersuites:
11136 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11137 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11138 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11140 The latter two were purportedly from
11141 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11144 Also deactivate the remaining ciphersuites from
11145 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11146 unofficial, and the ID has long expired.
11150 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11151 dual-core machines) and other potential thread-safety issues.
11155 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11157 * Adapt fipsld and the build system to link against the validated FIPS
11158 module in FIPS mode.
11162 * Fixes for VC++ 2005 build under Windows.
11166 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11167 from a Windows bash shell such as MSYS. It is autodetected from the
11168 "config" script when run from a VC++ environment. Modify standard VC++
11169 build to use fipscanister.o from the GNU make build.
11173 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11175 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11176 The value now differs depending on if you build for FIPS or not.
11177 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11178 safely run with a non-FIPSed libcrypto, as it may crash because of
11179 the difference induced by this change.
11183 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11185 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11186 (part of SSL_OP_ALL). This option used to disable the
11187 countermeasure against man-in-the-middle protocol-version
11188 rollback in the SSL 2.0 server implementation, which is a bad
11189 idea. ([CVE-2005-2969])
11191 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11192 for Information Security, National Institute of Advanced Industrial
11193 Science and Technology [AIST, Japan)]*
11195 * Minimal support for X9.31 signatures and PSS padding modes. This is
11196 mainly for FIPS compliance and not fully integrated at this stage.
11200 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11201 the exponentiation using a fixed-length exponent. (Otherwise,
11202 the information leaked through timing could expose the secret key
11203 after many signatures; cf. Bleichenbacher's attack on DSA with
11208 * Make a new fixed-window mod_exp implementation the default for
11209 RSA, DSA, and DH private-key operations so that the sequence of
11210 squares and multiplies and the memory access pattern are
11211 independent of the particular secret key. This will mitigate
11212 cache-timing and potential related attacks.
11214 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11215 and this is automatically used by BN_mod_exp_mont() if the new flag
11216 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11217 will use this BN flag for private exponents unless the flag
11218 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11219 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11221 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11223 * Change the client implementation for SSLv23_method() and
11224 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11225 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11226 (Previously, the SSL 2.0 backwards compatible Client Hello
11227 message format would be used even with SSL_OP_NO_SSLv2.)
11231 * Add support for smime-type MIME parameter in S/MIME messages which some
11236 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11237 a threadsafe manner. Modify rsa code to use new function and add calls
11238 to dsa and dh code (which had race conditions before).
11242 * Include the fixed error library code in the C error file definitions
11243 instead of fixing them up at runtime. This keeps the error code
11244 structures constant.
11248 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11250 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11253 * Fixes for newer kerberos headers. NB: the casts are needed because
11254 the 'length' field is signed on one version and unsigned on another
11255 with no (?) obvious way to tell the difference, without these VC++
11256 complains. Also the "definition" of FAR (blank) is no longer included
11257 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11258 some needed definitions.
11262 * Undo Cygwin change.
11266 * Added support for proxy certificates according to RFC 3820.
11267 Because they may be a security thread to unaware applications,
11268 they must be explicitly allowed in run-time. See
11269 docs/HOWTO/proxy_certificates.txt for further information.
11273 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11275 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11276 server and client random values. Previously
11277 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11278 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11280 This change has negligible security impact because:
11282 1. Server and client random values still have 24 bytes of pseudo random
11285 2. Server and client random values are sent in the clear in the initial
11288 3. The master secret is derived using the premaster secret (48 bytes in
11289 size for static RSA ciphersuites) as well as client server and random
11292 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11295 *Stephen Henson, reported by UK NISCC*
11297 * Use Windows randomness collection on Cygwin.
11301 * Fix hang in EGD/PRNGD query when communication socket is closed
11302 prematurely by EGD/PRNGD.
11304 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11306 * Prompt for pass phrases when appropriate for PKCS12 input format.
11310 * Back-port of selected performance improvements from development
11311 branch, as well as improved support for PowerPC platforms.
11315 * Add lots of checks for memory allocation failure, error codes to indicate
11316 failure and freeing up memory if a failure occurs.
11318 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11320 * Add new -passin argument to dgst.
11324 * Perform some character comparisons of different types in X509_NAME_cmp:
11325 this is needed for some certificates that re-encode DNs into UTF8Strings
11326 (in violation of RFC3280) and can't or won't issue name rollover
11331 * Make an explicit check during certificate validation to see that
11332 the CA setting in each certificate on the chain is correct. As a
11333 side effect always do the following basic checks on extensions,
11334 not just when there's an associated purpose to the check:
11336 - if there is an unhandled critical extension (unless the user
11337 has chosen to ignore this fault)
11338 - if the path length has been exceeded (if one is set at all)
11339 - that certain extensions fit the associated purpose (if one has
11344 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11346 * Avoid a race condition when CRLs are checked in a multi threaded
11347 environment. This would happen due to the reordering of the revoked
11348 entries during signature checking and serial number lookup. Now the
11349 encoding is cached and the serial number sort performed under a lock.
11350 Add new STACK function sk_is_sorted().
11354 * Add Delta CRL to the extension code.
11358 * Various fixes to s3_pkt.c so alerts are sent properly.
11360 *David Holmes <d.holmes@f5.com>*
11362 * Reduce the chances of duplicate issuer name and serial numbers (in
11363 violation of RFC3280) using the OpenSSL certificate creation utilities.
11364 This is done by creating a random 64 bit value for the initial serial
11365 number when a serial number file is created or when a self signed
11366 certificate is created using 'openssl req -x509'. The initial serial
11367 number file is created using 'openssl x509 -next_serial' in CA.pl
11368 rather than being initialized to 1.
11372 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11374 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11375 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11377 *Joe Orton, Steve Henson*
11379 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11382 *Joe Orton, Steve Henson*
11384 * Make it possible to have multiple active certificates with the same
11385 subject in the CA index file. This is done only if the keyword
11386 'unique_subject' is set to 'no' in the main CA section (default
11387 if 'CA_default') of the configuration file. The value is saved
11388 with the database itself in a separate index attribute file,
11389 named like the index file with '.attr' appended to the name.
11393 * X509 verify fixes. Disable broken certificate workarounds when
11394 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11395 keyUsage extension present. Don't accept CRLs with unhandled critical
11396 extensions: since verify currently doesn't process CRL extensions this
11397 rejects a CRL with *any* critical extensions. Add new verify error codes
11402 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11403 A clarification of RFC2560 will require the use of OCTET STRINGs and
11404 some implementations cannot handle the current raw format. Since OpenSSL
11405 copies and compares OCSP nonces as opaque blobs without any attempt at
11406 parsing them this should not create any compatibility issues.
11410 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11411 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11412 this HMAC (and other) operations are several times slower than OpenSSL
11417 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11419 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11421 * Use the correct content when signing type "other".
11425 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11427 * Fix various bugs revealed by running the NISCC test suite:
11429 Stop out of bounds reads in the ASN1 code when presented with
11430 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11432 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11434 If verify callback ignores invalid public key errors don't try to check
11435 certificate signature with the NULL public key.
11439 * New -ignore_err option in ocsp application to stop the server
11440 exiting on the first error in a request.
11444 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11445 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11450 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11451 extra data after the compression methods not only for TLS 1.0
11452 but also for SSL 3.0 (as required by the specification).
11454 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11456 * Change X509_certificate_type() to mark the key as exported/exportable
11457 when it's 512 *bits* long, not 512 bytes.
11461 * Change AES_cbc_encrypt() so it outputs exact multiple of
11462 blocks during encryption.
11466 * Various fixes to base64 BIO and non blocking I/O. On write
11467 flushes were not handled properly if the BIO retried. On read
11468 data was not being buffered properly and had various logic bugs.
11469 This also affects blocking I/O when the data being decoded is a
11474 * Various S/MIME bugfixes and compatibility changes:
11475 output correct application/pkcs7 MIME type if
11476 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11477 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11478 of files as .eml work). Correctly handle very long lines in MIME
11483 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11485 * Countermeasure against the Klima-Pokorny-Rosa extension of
11486 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11487 a protocol version number mismatch like a decryption error
11488 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11492 * Turn on RSA blinding by default in the default implementation
11493 to avoid a timing attack. Applications that don't want it can call
11494 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11495 They would be ill-advised to do so in most cases.
11497 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11499 * Change RSA blinding code so that it works when the PRNG is not
11500 seeded (in this case, the secret RSA exponent is abused as
11501 an unpredictable seed -- if it is not unpredictable, there
11502 is no point in blinding anyway). Make RSA blinding thread-safe
11503 by remembering the creator's thread ID in rsa->blinding and
11504 having all other threads use local one-time blinding factors
11505 (this requires more computation than sharing rsa->blinding, but
11506 avoids excessive locking; and if an RSA object is not shared
11507 between threads, blinding will still be very fast).
11511 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11512 ENGINE as defaults for all supported algorithms irrespective of
11513 the 'flags' parameter. 'flags' is now honoured, so applications
11514 should make sure they are passing it correctly.
11518 * Target "mingw" now allows native Windows code to be generated in
11519 the Cygwin environment as well as with the MinGW compiler.
11523 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11525 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11526 via timing by performing a MAC computation even if incorrect
11527 block cipher padding has been found. This is a countermeasure
11528 against active attacks where the attacker has to distinguish
11529 between bad padding and a MAC verification error. ([CVE-2003-0078])
11531 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11532 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11533 Martin Vuagnoux (EPFL, Ilion)*
11535 * Make the no-err option work as intended. The intention with no-err
11536 is not to have the whole error stack handling routines removed from
11537 libcrypto, it's only intended to remove all the function name and
11538 reason texts, thereby removing some of the footprint that may not
11539 be interesting if those errors aren't displayed anyway.
11541 NOTE: it's still possible for any application or module to have its
11542 own set of error texts inserted. The routines are there, just not
11543 used by default when no-err is given.
11547 * Add support for FreeBSD on IA64.
11549 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11551 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11552 Kerberos function mit_des_cbc_cksum(). Before this change,
11553 the value returned by DES_cbc_cksum() was like the one from
11554 mit_des_cbc_cksum(), except the bytes were swapped.
11556 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11558 * Allow an application to disable the automatic SSL chain building.
11559 Before this a rather primitive chain build was always performed in
11560 ssl3_output_cert_chain(): an application had no way to send the
11561 correct chain if the automatic operation produced an incorrect result.
11563 Now the chain builder is disabled if either:
11565 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11567 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11569 The reasoning behind this is that an application would not want the
11570 auto chain building to take place if extra chain certificates are
11571 present and it might also want a means of sending no additional
11572 certificates (for example the chain has two certificates and the
11577 * Add the possibility to build without the ENGINE framework.
11579 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11581 * Under Win32 gmtime() can return NULL: check return value in
11582 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11586 * DSA routines: under certain error conditions uninitialized BN objects
11587 could be freed. Solution: make sure initialization is performed early
11588 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11589 Nils Larsch <nla@trustcenter.de> via PR#459)
11593 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11594 checked on reconnect on the client side, therefore session resumption
11595 could still fail with a "ssl session id is different" error. This
11596 behaviour is masked when SSL_OP_ALL is used due to
11597 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11598 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11599 followup to PR #377.
11603 * IA-32 assembler support enhancements: unified ELF targets, support
11604 for SCO/Caldera platforms, fix for Cygwin shared build.
11608 * Add support for FreeBSD on sparc64. As a consequence, support for
11609 FreeBSD on non-x86 processors is separate from x86 processors on
11610 the config script, much like the NetBSD support.
11612 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11614 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11616 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11619 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11620 code (06) was taken as the first octet of the session ID and the last
11621 octet was ignored consequently. As a result SSLv2 client side session
11622 caching could not have worked due to the session ID mismatch between
11624 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11629 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11630 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11635 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11636 seems that in spite of existing for more than a year, many application
11637 author have done nothing to provide the necessary callbacks, which
11638 means that this particular engine will not work properly anywhere.
11639 This is a very unfortunate situation which forces us, in the name
11640 of usability, to give the hw_ncipher.c a static lock, which is part
11642 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11643 appear in 0.9.8 or later. We EXPECT application authors to have
11644 dealt properly with this when 0.9.8 is released (unless we actually
11645 make such changes in the libcrypto locking code that changes will
11646 have to be made anyway).
11650 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11651 octets have been read, EOF or an error occurs. Without this change
11652 some truncated ASN1 structures will not produce an error.
11656 * Disable Heimdal support, since it hasn't been fully implemented.
11657 Still give the possibility to force the use of Heimdal, but with
11658 warnings and a request that patches get sent to openssl-dev.
11662 * Add the VC-CE target, introduce the WINCE sysname, and add
11663 INSTALL.WCE and appropriate conditionals to make it build.
11665 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11667 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11668 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11669 edit numbers of the version.
11671 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11673 * Introduce safe string copy and catenation functions
11674 (BUF_strlcpy() and BUF_strlcat()).
11676 *Ben Laurie (CHATS) and Richard Levitte*
11678 * Avoid using fixed-size buffers for one-line DNs.
11680 *Ben Laurie (CHATS)*
11682 * Add BUF_MEM_grow_clean() to avoid information leakage when
11683 resizing buffers containing secrets, and use where appropriate.
11685 *Ben Laurie (CHATS)*
11687 * Avoid using fixed size buffers for configuration file location.
11689 *Ben Laurie (CHATS)*
11691 * Avoid filename truncation for various CA files.
11693 *Ben Laurie (CHATS)*
11695 * Use sizeof in preference to magic numbers.
11697 *Ben Laurie (CHATS)*
11699 * Avoid filename truncation in cert requests.
11701 *Ben Laurie (CHATS)*
11703 * Add assertions to check for (supposedly impossible) buffer
11706 *Ben Laurie (CHATS)*
11708 * Don't cache truncated DNS entries in the local cache (this could
11709 potentially lead to a spoofing attack).
11711 *Ben Laurie (CHATS)*
11713 * Fix various buffers to be large enough for hex/decimal
11714 representations in a platform independent manner.
11716 *Ben Laurie (CHATS)*
11718 * Add CRYPTO_realloc_clean() to avoid information leakage when
11719 resizing buffers containing secrets, and use where appropriate.
11721 *Ben Laurie (CHATS)*
11723 * Add BIO_indent() to avoid much slightly worrying code to do
11726 *Ben Laurie (CHATS)*
11728 * Convert sprintf()/BIO_puts() to BIO_printf().
11730 *Ben Laurie (CHATS)*
11732 * buffer_gets() could terminate with the buffer only half
11735 *Ben Laurie (CHATS)*
11737 * Add assertions to prevent user-supplied crypto functions from
11738 overflowing internal buffers by having large block sizes, etc.
11740 *Ben Laurie (CHATS)*
11742 * New OPENSSL_assert() macro (similar to assert(), but enabled
11745 *Ben Laurie (CHATS)*
11747 * Eliminate unused copy of key in RC4.
11749 *Ben Laurie (CHATS)*
11751 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11753 *Ben Laurie (CHATS)*
11755 * Fix off-by-one error in EGD path.
11757 *Ben Laurie (CHATS)*
11759 * If RANDFILE path is too long, ignore instead of truncating.
11761 *Ben Laurie (CHATS)*
11763 * Eliminate unused and incorrectly sized X.509 structure
11766 *Ben Laurie (CHATS)*
11768 * Eliminate unused and dangerous function knumber().
11770 *Ben Laurie (CHATS)*
11772 * Eliminate unused and dangerous structure, KSSL_ERR.
11774 *Ben Laurie (CHATS)*
11776 * Protect against overlong session ID context length in an encoded
11777 session object. Since these are local, this does not appear to be
11780 *Ben Laurie (CHATS)*
11782 * Change from security patch (see 0.9.6e below) that did not affect
11783 the 0.9.6 release series:
11785 Remote buffer overflow in SSL3 protocol - an attacker could
11786 supply an oversized master key in Kerberos-enabled versions.
11789 *Ben Laurie (CHATS)*
11791 * Change the SSL kerb5 codes to match RFC 2712.
11795 * Make -nameopt work fully for req and add -reqopt switch.
11797 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11799 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11801 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11803 * Make sure tests can be performed even if the corresponding algorithms
11804 have been removed entirely. This was also the last step to make
11805 OpenSSL compilable with DJGPP under all reasonable conditions.
11807 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11809 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11810 to allow version independent disabling of normally unselected ciphers,
11811 which may be activated as a side-effect of selecting a single cipher.
11813 (E.g., cipher list string "RSA" enables ciphersuites that are left
11814 out of "ALL" because they do not provide symmetric encryption.
11815 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11817 *Lutz Jaenicke, Bodo Moeller*
11819 * Add appropriate support for separate platform-dependent build
11820 directories. The recommended way to make a platform-dependent
11821 build directory is the following (tested on Linux), maybe with
11824 # Place yourself outside of the OpenSSL source tree. In
11825 # this example, the environment variable OPENSSL_SOURCE
11826 # is assumed to contain the absolute OpenSSL source directory.
11827 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11828 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11829 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11830 mkdir -p `dirname $F`
11831 ln -s $OPENSSL_SOURCE/$F $F
11834 To be absolutely sure not to disturb the source tree, a "make clean"
11835 is a good thing. If it isn't successful, don't worry about it,
11836 it probably means the source directory is very clean.
11840 * Make sure any ENGINE control commands make local copies of string
11841 pointers passed to them whenever necessary. Otherwise it is possible
11842 the caller may have overwritten (or deallocated) the original string
11843 data when a later ENGINE operation tries to use the stored values.
11845 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11847 * Improve diagnostics in file reading and command-line digests.
11849 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11851 * Add AES modes CFB and OFB to the object database. Correct an
11852 error in AES-CFB decryption.
11856 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11857 allows existing EVP_CIPHER_CTX structures to be reused after
11858 calling `EVP_*Final()`. This behaviour is used by encryption
11859 BIOs and some applications. This has the side effect that
11860 applications must explicitly clean up cipher contexts with
11861 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11865 * Check the values of dna and dnb in bn_mul_recursive before calling
11866 bn_mul_comba (a non zero value means the a or b arrays do not contain
11867 n2 elements) and fallback to bn_mul_normal if either is not zero.
11871 * Fix escaping of non-ASCII characters when using the -subj option
11872 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11876 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11877 form for "surname", serialNumber has no short form.
11878 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11879 therefore remove "mail" short name for "internet 7".
11880 The OID for unique identifiers in X509 certificates is
11881 x500UniqueIdentifier, not uniqueIdentifier.
11882 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11886 * Add an "init" command to the ENGINE config module and auto initialize
11887 ENGINEs. Without any "init" command the ENGINE will be initialized
11888 after all ctrl commands have been executed on it. If init=1 the
11889 ENGINE is initialized at that point (ctrls before that point are run
11890 on the uninitialized ENGINE and after on the initialized one). If
11891 init=0 then the ENGINE will not be initialized at all.
11895 * Fix the 'app_verify_callback' interface so that the user-defined
11896 argument is actually passed to the callback: In the
11897 SSL_CTX_set_cert_verify_callback() prototype, the callback
11898 declaration has been changed from
11901 int (*cb)(X509_STORE_CTX *,void *);
11902 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11903 i=s->ctx->app_verify_callback(&ctx)
11904 has been changed into
11905 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11907 To update applications using SSL_CTX_set_cert_verify_callback(),
11908 a dummy argument can be added to their callback functions.
11910 *D. K. Smetters <smetters@parc.xerox.com>*
11912 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11914 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11916 * Add and OPENSSL_LOAD_CONF define which will cause
11917 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11918 This allows older applications to transparently support certain
11919 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11920 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11921 load the config file and OPENSSL_add_all_algorithms_conf() which will
11922 always load it have also been added.
11926 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11927 Adjust NIDs and EVP layer.
11929 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11931 * Config modules support in openssl utility.
11933 Most commands now load modules from the config file,
11934 though in a few (such as version) this isn't done
11935 because it couldn't be used for anything.
11937 In the case of ca and req the config file used is
11938 the same as the utility itself: that is the -config
11939 command line option can be used to specify an
11944 * Move default behaviour from OPENSSL_config(). If appname is NULL
11945 use "openssl_conf" if filename is NULL use default openssl config file.
11949 * Add an argument to OPENSSL_config() to allow the use of an alternative
11950 config section name. Add a new flag to tolerate a missing config file
11951 and move code to CONF_modules_load_file().
11955 * Support for crypto accelerator cards from Accelerated Encryption
11956 Processing, www.aep.ie. (Use engine 'aep')
11957 The support was copied from 0.9.6c [engine] and adapted/corrected
11958 to work with the new engine framework.
11960 *AEP Inc. and Richard Levitte*
11962 * Support for SureWare crypto accelerator cards from Baltimore
11963 Technologies. (Use engine 'sureware')
11964 The support was copied from 0.9.6c [engine] and adapted
11965 to work with the new engine framework.
11969 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11970 make the newer ENGINE framework commands for the CHIL engine work.
11972 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11974 * Make it possible to produce shared libraries on ReliantUNIX.
11976 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11978 * Add the configuration target debug-linux-ppro.
11979 Make 'openssl rsa' use the general key loading routines
11980 implemented in `apps.c`, and make those routines able to
11981 handle the key format FORMAT_NETSCAPE and the variant
11984 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11986 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11988 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11990 * Add -keyform to rsautl, and document -engine.
11992 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11994 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11995 BIO_R_NO_SUCH_FILE error code rather than the generic
11996 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12000 * Add new functions
12001 ERR_peek_last_error
12002 ERR_peek_last_error_line
12003 ERR_peek_last_error_line_data.
12004 These are similar to
12006 ERR_peek_error_line
12007 ERR_peek_error_line_data,
12008 but report on the latest error recorded rather than the first one
12009 still in the error queue.
12011 *Ben Laurie, Bodo Moeller*
12013 * default_algorithms option in ENGINE config module. This allows things
12015 default_algorithms = ALL
12016 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12020 * Preliminary ENGINE config module.
12024 * New experimental application configuration code.
12028 * Change the AES code to follow the same name structure as all other
12029 symmetric ciphers, and behave the same way. Move everything to
12030 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12032 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12034 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12036 *Ben Laurie and Theo de Raadt*
12038 * Add option to output public keys in req command.
12040 *Massimiliano Pala madwolf@openca.org*
12042 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12043 (up to about 10% better than before for P-192 and P-224).
12047 * New functions/macros
12049 SSL_CTX_set_msg_callback(ctx, cb)
12050 SSL_CTX_set_msg_callback_arg(ctx, arg)
12051 SSL_set_msg_callback(ssl, cb)
12052 SSL_set_msg_callback_arg(ssl, arg)
12054 to request calling a callback function
12056 void cb(int write_p, int version, int content_type,
12057 const void *buf, size_t len, SSL *ssl, void *arg)
12059 whenever a protocol message has been completely received
12060 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12061 protocol version according to which the SSL library interprets
12062 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12063 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12064 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12065 specification (change_cipher_spec(20), alert(21), handshake(22)).
12066 'buf' and 'len' point to the actual message, 'ssl' to the
12067 SSL object, and 'arg' is the application-defined value set by
12068 SSL[_CTX]_set_msg_callback_arg().
12070 'openssl s_client' and 'openssl s_server' have new '-msg' options
12071 to enable a callback that displays all protocol messages.
12075 * Change the shared library support so shared libraries are built as
12076 soon as the corresponding static library is finished, and thereby get
12077 openssl and the test programs linked against the shared library.
12078 This still only happens when the keyword "shard" has been given to
12079 the configuration scripts.
12081 NOTE: shared library support is still an experimental thing, and
12082 backward binary compatibility is still not guaranteed.
12084 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12086 * Add support for Subject Information Access extension.
12088 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12090 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12091 additional bytes when new memory had to be allocated, not just
12092 when reusing an existing buffer.
12096 * New command line and configuration option 'utf8' for the req command.
12097 This allows field values to be specified as UTF8 strings.
12101 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12102 runs for the former and machine-readable output for the latter.
12106 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12107 of the e-mail address in the DN (i.e., it will go into a certificate
12108 extension only). The new configuration file option 'email_in_dn = no'
12109 has the same effect.
12111 *Massimiliano Pala madwolf@openca.org*
12113 * Change all functions with names starting with `des_` to be starting
12114 with `DES_` instead. Add wrappers that are compatible with libdes,
12115 but are named `_ossl_old_des_*`. Finally, add macros that map the
12116 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12117 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12118 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12121 Since we provide two compatibility mappings, the user needs to
12122 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12123 compatibility is desired. The default (i.e., when that macro
12124 isn't defined) is OpenSSL 0.9.6c compatibility.
12126 There are also macros that enable and disable the support of old
12127 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12128 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12129 are defined, the default will apply: to support the old des routines.
12131 In either case, one must include openssl/des.h to get the correct
12132 definitions. Do not try to just include openssl/des_old.h, that
12135 NOTE: This is a major break of an old API into a new one. Software
12136 authors are encouraged to switch to the `DES_` style functions. Some
12137 time in the future, des_old.h and the libdes compatibility functions
12138 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12139 default), and then completely removed.
12143 * Test for certificates which contain unsupported critical extensions.
12144 If such a certificate is found during a verify operation it is
12145 rejected by default: this behaviour can be overridden by either
12146 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12147 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12148 X509_supported_extension() has also been added which returns 1 if a
12149 particular extension is supported.
12153 * Modify the behaviour of EVP cipher functions in similar way to digests
12154 to retain compatibility with existing code.
12158 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12159 compatibility with existing code. In particular the 'ctx' parameter does
12160 not have to be to be initialized before the call to EVP_DigestInit() and
12161 it is tidied up after a call to EVP_DigestFinal(). New function
12162 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12163 EVP_MD_CTX_copy() changed to not require the destination to be
12164 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12165 requires the destination to be valid.
12167 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12168 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12172 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12173 so that complete 'Handshake' protocol structures are kept in memory
12174 instead of overwriting 'msg_type' and 'length' with 'body' data.
12178 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12180 *Massimo Santin via Richard Levitte*
12182 * Major restructuring to the underlying ENGINE code. This includes
12183 reduction of linker bloat, separation of pure "ENGINE" manipulation
12184 (initialisation, etc) from functionality dealing with implementations
12185 of specific crypto interfaces. This change also introduces integrated
12186 support for symmetric ciphers and digest implementations - so ENGINEs
12187 can now accelerate these by providing EVP_CIPHER and EVP_MD
12188 implementations of their own. This is detailed in
12189 [crypto/engine/README.md](crypto/engine/README.md)
12190 as it couldn't be adequately described here. However, there are a few
12191 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12192 were changed in the original introduction of ENGINE code have now
12193 reverted back - the hooking from this code to ENGINE is now a good
12194 deal more passive and at run-time, operations deal directly with
12195 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12196 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12197 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12198 they were not being used by the framework as there is no concept of a
12199 BIGNUM_METHOD and they could not be generalised to the new
12200 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12201 ENGINE_cpy() has been removed as it cannot be consistently defined in
12206 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12210 * Change mkdef.pl to sort symbols that get the same entry number,
12211 and make sure the automatically generated functions `ERR_load_*`
12212 become part of libeay.num as well.
12216 * New function SSL_renegotiate_pending(). This returns true once
12217 renegotiation has been requested (either SSL_renegotiate() call
12218 or HelloRequest/ClientHello received from the peer) and becomes
12219 false once a handshake has been completed.
12220 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12221 sends a HelloRequest, but does not ensure that a handshake takes
12222 place. SSL_renegotiate_pending() is useful for checking if the
12223 client has followed the request.)
12227 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12228 By default, clients may request session resumption even during
12229 renegotiation (if session ID contexts permit); with this option,
12230 session resumption is possible only in the first handshake.
12232 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12233 more bits available for options that should not be part of
12234 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12238 * Add some demos for certificate and certificate request creation.
12242 * Make maximum certificate chain size accepted from the peer application
12243 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12244 "Douglas E. Engert" <deengert@anl.gov>.
12248 * Add support for shared libraries for Unixware-7
12249 (Boyd Lynn Gerber <gerberb@zenez.com>).
12253 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12254 be done prior to destruction. Use this to unload error strings from
12255 ENGINEs that load their own error strings. NB: This adds two new API
12256 functions to "get" and "set" this destroy handler in an ENGINE.
12260 * Alter all existing ENGINE implementations (except "openssl" and
12261 "openbsd") to dynamically instantiate their own error strings. This
12262 makes them more flexible to be built both as statically-linked ENGINEs
12263 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12264 Also, add stub code to each that makes building them as self-contained
12265 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12269 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12270 implementations into applications that are completely implemented in
12271 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12272 commands that can be used to configure what shared-library to load and
12273 to control aspects of the way it is handled. Also, made an update to
12274 the [README-Engine.md](README-Engine.md) file
12275 that brings its information up-to-date and
12276 provides some information and instructions on the "dynamic" ENGINE
12277 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12281 * Make it possible to unload ranges of ERR strings with a new
12282 "ERR_unload_strings" function.
12286 * Add a copy() function to EVP_MD.
12290 * Make EVP_MD routines take a context pointer instead of just the
12291 md_data void pointer.
12295 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12296 that the digest can only process a single chunk of data
12297 (typically because it is provided by a piece of
12298 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12299 is only going to provide a single chunk of data, and hence the
12300 framework needn't accumulate the data for oneshot drivers.
12304 * As with "ERR", make it possible to replace the underlying "ex_data"
12305 functions. This change also alters the storage and management of global
12306 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12307 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12308 index counters. The API functions that use this state have been changed
12309 to take a "class_index" rather than pointers to the class's local STACK
12310 and counter, and there is now an API function to dynamically create new
12311 classes. This centralisation allows us to (a) plug a lot of the
12312 thread-safety problems that existed, and (b) makes it possible to clean
12313 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12314 such data would previously have always leaked in application code and
12315 workarounds were in place to make the memory debugging turn a blind eye
12316 to it. Application code that doesn't use this new function will still
12317 leak as before, but their memory debugging output will announce it now
12318 rather than letting it slide.
12320 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12321 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12322 has a return value to indicate success or failure.
12326 * Make it possible to replace the underlying "ERR" functions such that the
12327 global state (2 LHASH tables and 2 locks) is only used by the "default"
12328 implementation. This change also adds two functions to "get" and "set"
12329 the implementation prior to it being automatically set the first time
12330 any other ERR function takes place. Ie. an application can call "get",
12331 pass the return value to a module it has just loaded, and that module
12332 can call its own "set" function using that value. This means the
12333 module's "ERR" operations will use (and modify) the error state in the
12334 application and not in its own statically linked copy of OpenSSL code.
12338 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12339 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12340 the operation, and provides a more encapsulated way for external code
12341 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12342 to use these functions rather than manually incrementing the counts.
12344 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12348 * Add EVP test program.
12352 * Add symmetric cipher support to ENGINE. Expect the API to change!
12356 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12357 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12358 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12359 These allow a CRL to be built without having to access X509_CRL fields
12360 directly. Modify 'ca' application to use new functions.
12364 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12365 bug workarounds. Rollback attack detection is a security feature.
12366 The problem will only arise on OpenSSL servers when TLSv1 is not
12367 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12368 Software authors not wanting to support TLSv1 will have special reasons
12369 for their choice and can explicitly enable this option.
12371 *Bodo Moeller, Lutz Jaenicke*
12373 * Rationalise EVP so it can be extended: don't include a union of
12374 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12375 (similar to those existing for EVP_CIPHER_CTX).
12380 EVP_MD_CTX_init(&md); /* new function call */
12381 EVP_DigestInit(&md, EVP_sha1());
12382 EVP_DigestUpdate(&md, in, len);
12383 EVP_DigestFinal(&md, out, NULL);
12384 EVP_MD_CTX_cleanup(&md); /* new function call */
12388 * Make DES key schedule conform to the usual scheme, as well as
12389 correcting its structure. This means that calls to DES functions
12390 now have to pass a pointer to a des_key_schedule instead of a
12391 plain des_key_schedule (which was actually always a pointer
12394 des_key_schedule ks;
12396 des_set_key_checked(..., &ks);
12397 des_ncbc_encrypt(..., &ks, ...);
12399 (Note that a later change renames 'des_...' into 'DES_...'.)
12403 * Initial reduction of linker bloat: the use of some functions, such as
12404 PEM causes large amounts of unused functions to be linked in due to
12405 poor organisation. For example pem_all.c contains every PEM function
12406 which has a knock on effect of linking in large amounts of (unused)
12407 ASN1 code. Grouping together similar functions and splitting unrelated
12408 functions prevents this.
12412 * Cleanup of EVP macros.
12416 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12417 correct `_ecb suffix`.
12421 * Add initial OCSP responder support to ocsp application. The
12422 revocation information is handled using the text based index
12423 use by the ca application. The responder can either handle
12424 requests generated internally, supplied in files (for example
12425 via a CGI script) or using an internal minimal server.
12429 * Add configuration choices to get zlib compression for TLS.
12433 * Changes to Kerberos SSL for RFC 2712 compliance:
12434 1. Implemented real KerberosWrapper, instead of just using
12435 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12436 2. Implemented optional authenticator field of KerberosWrapper.
12438 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12439 and authenticator structs; see crypto/krb5/.
12441 Generalized Kerberos calls to support multiple Kerberos libraries.
12442 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12443 via Richard Levitte*
12445 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12446 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12447 values for each of the key sizes rather than having just
12448 parameters (and 'speed' generating keys each time).
12452 * Speed up EVP routines.
12455 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12456 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12457 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12458 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12460 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12461 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12462 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12465 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12467 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12471 * Added the OS2-EMX target.
12473 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12475 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12476 New functions to support `NCONF` routines in extension code.
12477 New function `CONF_set_nconf()`
12478 to allow functions which take an `NCONF` to also handle the old `LHASH`
12479 structure: this means that the old `CONF` compatible routines can be
12480 retained (in particular w.rt. extensions) without having to duplicate the
12481 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12485 * Enhance the general user interface with mechanisms for inner control
12486 and with possibilities to have yes/no kind of prompts.
12490 * Change all calls to low-level digest routines in the library and
12491 applications to use EVP. Add missing calls to HMAC_cleanup() and
12492 don't assume HMAC_CTX can be copied using memcpy().
12494 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12496 * Add the possibility to control engines through control names but with
12497 arbitrary arguments instead of just a string.
12498 Change the key loaders to take a UI_METHOD instead of a callback
12499 function pointer. NOTE: this breaks binary compatibility with earlier
12500 versions of OpenSSL [engine].
12501 Adapt the nCipher code for these new conditions and add a card insertion
12506 * Enhance the general user interface with mechanisms to better support
12507 dialog box interfaces, application-defined prompts, the possibility
12508 to use defaults (for example default passwords from somewhere else)
12509 and interrupts/cancellations.
12513 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12514 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12518 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12519 tidy up some unnecessarily weird code in 'sk_new()').
12521 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12523 * Change the key loading routines for ENGINEs to use the same kind
12524 callback (pem_password_cb) as all other routines that need this
12529 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12530 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12531 than this minimum value is recommended.
12535 * New random seeder for OpenVMS, using the system process statistics
12536 that are easily reachable.
12540 * Windows apparently can't transparently handle global
12541 variables defined in DLLs. Initialisations such as:
12543 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12545 won't compile. This is used by the any applications that need to
12546 declare their own ASN1 modules. This was fixed by adding the option
12547 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12548 needed for static libraries under Win32.
12552 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12553 setting of purpose and trust fields. New X509_STORE trust and
12554 purpose functions and tidy up setting in other SSL functions.
12558 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12559 structure. These are inherited by X509_STORE_CTX when it is
12560 initialised. This allows various defaults to be set in the
12561 X509_STORE structure (such as flags for CRL checking and custom
12562 purpose or trust settings) for functions which only use X509_STORE_CTX
12563 internally such as S/MIME.
12565 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12566 trust settings if they are not set in X509_STORE. This allows X509_STORE
12567 purposes and trust (in S/MIME for example) to override any set by default.
12569 Add command line options for CRL checking to smime, s_client and s_server
12574 * Initial CRL based revocation checking. If the CRL checking flag(s)
12575 are set then the CRL is looked up in the X509_STORE structure and
12576 its validity and signature checked, then if the certificate is found
12577 in the CRL the verify fails with a revoked error.
12579 Various new CRL related callbacks added to X509_STORE_CTX structure.
12581 Command line options added to 'verify' application to support this.
12583 This needs some additional work, such as being able to handle multiple
12584 CRLs with different times, extension based lookup (rather than just
12585 by subject name) and ultimately more complete V2 CRL extension
12590 * Add a general user interface API (crypto/ui/). This is designed
12591 to replace things like des_read_password and friends (backward
12592 compatibility functions using this new API are provided).
12593 The purpose is to remove prompting functions from the DES code
12594 section as well as provide for prompting through dialog boxes in
12595 a window system and the like.
12599 * Add "ex_data" support to ENGINE so implementations can add state at a
12600 per-structure level rather than having to store it globally.
12604 * Make it possible for ENGINE structures to be copied when retrieved by
12605 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12606 This causes the "original" ENGINE structure to act like a template,
12607 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12608 operational state can be localised to each ENGINE structure, despite the
12609 fact they all share the same "methods". New ENGINE structures returned in
12610 this case have no functional references and the return value is the single
12611 structural reference. This matches the single structural reference returned
12612 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12617 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12618 needs to match any other type at all we need to manually clear the
12623 * Changes to the "openssl engine" utility to include;
12624 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12625 about an ENGINE's available control commands.
12626 - executing control commands from command line arguments using the
12627 '-pre' and '-post' switches. '-post' is only used if '-t' is
12628 specified and the ENGINE is successfully initialised. The syntax for
12629 the individual commands are colon-separated, for example;
12630 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12634 * New dynamic control command support for ENGINEs. ENGINEs can now
12635 declare their own commands (numbers), names (strings), descriptions,
12636 and input types for run-time discovery by calling applications. A
12637 subset of these commands are implicitly classed as "executable"
12638 depending on their input type, and only these can be invoked through
12639 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12640 can be based on user input, config files, etc). The distinction is
12641 that "executable" commands cannot return anything other than a boolean
12642 result and can only support numeric or string input, whereas some
12643 discoverable commands may only be for direct use through
12644 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12645 pointers, or other custom uses. The "executable" commands are to
12646 support parameterisations of ENGINE behaviour that can be
12647 unambiguously defined by ENGINEs and used consistently across any
12648 OpenSSL-based application. Commands have been added to all the
12649 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12650 control over shared-library paths without source code alterations.
12654 * Changed all ENGINE implementations to dynamically allocate their
12655 ENGINEs rather than declaring them statically. Apart from this being
12656 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12657 this also allows the implementations to compile without using the
12658 internal engine_int.h header.
12662 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12663 'const' value. Any code that should be able to modify a RAND_METHOD
12664 should already have non-const pointers to it (ie. they should only
12665 modify their own ones).
12669 * Made a variety of little tweaks to the ENGINE code.
12670 - "atalla" and "ubsec" string definitions were moved from header files
12671 to C code. "nuron" string definitions were placed in variables
12672 rather than hard-coded - allowing parameterisation of these values
12673 later on via ctrl() commands.
12674 - Removed unused "#if 0"'d code.
12675 - Fixed engine list iteration code so it uses ENGINE_free() to release
12676 structural references.
12677 - Constified the RAND_METHOD element of ENGINE structures.
12678 - Constified various get/set functions as appropriate and added
12679 missing functions (including a catch-all ENGINE_cpy that duplicates
12680 all ENGINE values onto a new ENGINE except reference counts/state).
12681 - Removed NULL parameter checks in get/set functions. Setting a method
12682 or function to NULL is a way of cancelling out a previously set
12683 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12684 and doesn't justify the extra error symbols and code.
12685 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12686 flags from engine_int.h to engine.h.
12687 - Changed prototypes for ENGINE handler functions (init(), finish(),
12688 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12692 * Implement binary inversion algorithm for BN_mod_inverse in addition
12693 to the algorithm using long division. The binary algorithm can be
12694 used only if the modulus is odd. On 32-bit systems, it is faster
12695 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12696 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12697 up to 450 bits. In 64-bit environments, the binary algorithm
12698 appears to be advantageous for much longer moduli; here we use it
12699 for moduli up to 2048 bits.
12703 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12704 could not support the combine flag in choice fields.
12708 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12709 extensions from a certificate request to the certificate.
12713 * Allow multiple 'certopt' and 'nameopt' options to be separated
12714 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12715 file: this allows the display of the certificate about to be
12716 signed to be customised, to allow certain fields to be included
12717 or excluded and extension details. The old system didn't display
12718 multicharacter strings properly, omitted fields not in the policy
12719 and couldn't display additional details such as extensions.
12723 * Function EC_POINTs_mul for multiple scalar multiplication
12724 of an arbitrary number of elliptic curve points
12725 \sum scalars[i]*points[i],
12726 optionally including the generator defined for the EC_GROUP:
12727 scalar*generator + \sum scalars[i]*points[i].
12729 EC_POINT_mul is a simple wrapper function for the typical case
12730 that the point list has just one item (besides the optional
12735 * First EC_METHODs for curves over GF(p):
12737 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12738 operations and provides various method functions that can also
12739 operate with faster implementations of modular arithmetic.
12741 EC_GFp_mont_method() reuses most functions that are part of
12742 EC_GFp_simple_method, but uses Montgomery arithmetic.
12744 *Bodo Moeller; point addition and point doubling
12745 implementation directly derived from source code provided by
12746 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12748 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12749 crypto/ec/ec_lib.c):
12751 Curves are EC_GROUP objects (with an optional group generator)
12752 based on EC_METHODs that are built into the library.
12754 Points are EC_POINT objects based on EC_GROUP objects.
12756 Most of the framework would be able to handle curves over arbitrary
12757 finite fields, but as there are no obvious types for fields other
12758 than GF(p), some functions are limited to that for now.
12762 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12763 that the file contains a complete HTTP response.
12767 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12768 change the def and num file printf format specifier from "%-40sXXX"
12769 to "%-39s XXX". The latter will always guarantee a space after the
12770 field while the former will cause them to run together if the field
12771 is 40 of more characters long.
12775 * Constify the cipher and digest 'method' functions and structures
12776 and modify related functions to take constant EVP_MD and EVP_CIPHER
12781 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12782 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12786 * Modify `EVP_Digest*()` routines so they now return values. Although the
12787 internal software routines can never fail additional hardware versions
12792 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12794 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12795 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12798 ERR_R_NESTED_ASN1_ERROR
12800 ERR_R_MISSING_ASN1_EOS
12801 were 4 .. 9, conflicting with
12802 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12804 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12805 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12807 Add new error code 'ERR_R_INTERNAL_ERROR'.
12811 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12816 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12817 sets the subject name for a new request or supersedes the
12818 subject name in a given request. Formats that can be parsed are
12819 'CN=Some Name, OU=myOU, C=IT'
12821 'CN=Some Name/OU=myOU/C=IT'.
12823 Add options '-batch' and '-verbose' to 'openssl req'.
12825 *Massimiliano Pala <madwolf@hackmasters.net>*
12827 * Introduce the possibility to access global variables through
12828 functions on platform were that's the best way to handle exporting
12829 global variables in shared libraries. To enable this functionality,
12830 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12831 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12832 is normally done by Configure or something similar).
12834 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12835 in the source file (foo.c) like this:
12837 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12838 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12840 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12841 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12843 OPENSSL_DECLARE_GLOBAL(int,foo);
12844 #define foo OPENSSL_GLOBAL_REF(foo)
12845 OPENSSL_DECLARE_GLOBAL(double,bar);
12846 #define bar OPENSSL_GLOBAL_REF(bar)
12848 The #defines are very important, and therefore so is including the
12849 header file everywhere where the defined globals are used.
12851 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12852 of ASN.1 items, but that structure is a bit different.
12854 The largest change is in util/mkdef.pl which has been enhanced with
12855 better and easier to understand logic to choose which symbols should
12856 go into the Windows .def files as well as a number of fixes and code
12857 cleanup (among others, algorithm keywords are now sorted
12858 lexicographically to avoid constant rewrites).
12862 * In BN_div() keep a copy of the sign of 'num' before writing the
12863 result to 'rm' because if rm==num the value will be overwritten
12864 and produce the wrong result if 'num' is negative: this caused
12865 problems with BN_mod() and BN_nnmod().
12869 * Function OCSP_request_verify(). This checks the signature on an
12870 OCSP request and verifies the signer certificate. The signer
12871 certificate is just checked for a generic purpose and OCSP request
12876 * Add OCSP_check_validity() function to check the validity of OCSP
12877 responses. OCSP responses are prepared in real time and may only
12878 be a few seconds old. Simply checking that the current time lies
12879 between thisUpdate and nextUpdate max reject otherwise valid responses
12880 caused by either OCSP responder or client clock inaccuracy. Instead
12881 we allow thisUpdate and nextUpdate to fall within a certain period of
12882 the current time. The age of the response can also optionally be
12883 checked. Two new options -validity_period and -status_age added to
12888 * If signature or public key algorithm is unrecognized print out its
12889 OID rather that just UNKNOWN.
12893 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12894 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12895 ID to be generated from the issuer certificate alone which can then be
12896 passed to OCSP_id_issuer_cmp().
12900 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12901 ASN1 modules to export functions returning ASN1_ITEM pointers
12902 instead of the ASN1_ITEM structures themselves. This adds several
12903 new macros which allow the underlying ASN1 function/structure to
12904 be accessed transparently. As a result code should not use ASN1_ITEM
12905 references directly (such as &X509_it) but instead use the relevant
12906 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12907 use of the new ASN1 code on platforms where exporting structures
12908 is problematical (for example in shared libraries) but exporting
12909 functions returning pointers to structures is not.
12913 * Add support for overriding the generation of SSL/TLS session IDs.
12914 These callbacks can be registered either in an SSL_CTX or per SSL.
12915 The purpose of this is to allow applications to control, if they wish,
12916 the arbitrary values chosen for use as session IDs, particularly as it
12917 can be useful for session caching in multiple-server environments. A
12918 command-line switch for testing this (and any client code that wishes
12919 to use such a feature) has been added to "s_server".
12921 *Geoff Thorpe, Lutz Jaenicke*
12923 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12924 of the form `#if defined(...) || defined(...) || ...` and
12925 `#if !defined(...) && !defined(...) && ...`. This also avoids
12926 the growing number of special cases it was previously handling.
12930 * Make all configuration macros available for application by making
12931 sure they are available in opensslconf.h, by giving them names starting
12932 with `OPENSSL_` to avoid conflicts with other packages and by making
12933 sure e_os2.h will cover all platform-specific cases together with
12935 Additionally, it is now possible to define configuration/platform-
12936 specific names (called "system identities"). In the C code, these
12937 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12938 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12939 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12944 * New option -set_serial to 'req' and 'x509' this allows the serial
12945 number to use to be specified on the command line. Previously self
12946 signed certificates were hard coded with serial number 0 and the
12947 CA options of 'x509' had to use a serial number in a file which was
12952 * New options to 'ca' utility to support V2 CRL entry extensions.
12953 Currently CRL reason, invalidity date and hold instruction are
12954 supported. Add new CRL extensions to V3 code and some new objects.
12958 * New function EVP_CIPHER_CTX_set_padding() this is used to
12959 disable standard block padding (aka PKCS#5 padding) in the EVP
12960 API, which was previously mandatory. This means that the data is
12961 not padded in any way and so the total length much be a multiple
12962 of the block size, otherwise an error occurs.
12966 * Initial (incomplete) OCSP SSL support.
12970 * New function OCSP_parse_url(). This splits up a URL into its host,
12971 port and path components: primarily to parse OCSP URLs. New -url
12972 option to ocsp utility.
12976 * New nonce behavior. The return value of OCSP_check_nonce() now
12977 reflects the various checks performed. Applications can decide
12978 whether to tolerate certain situations such as an absent nonce
12979 in a response when one was present in a request: the ocsp application
12980 just prints out a warning. New function OCSP_add1_basic_nonce()
12981 this is to allow responders to include a nonce in a response even if
12982 the request is nonce-less.
12986 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12987 skipped when using openssl x509 multiple times on a single input file,
12988 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12992 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12993 set string type: to handle setting ASN1_TIME structures. Fix ca
12994 utility to correctly initialize revocation date of CRLs.
12998 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12999 the clients preferred ciphersuites and rather use its own preferences.
13000 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13001 Internet Explorer by ensuring unchanged hash method during stepup.
13002 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13006 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13007 to aes and add a new 'exist' option to print out symbols that don't
13012 * Additional options to ocsp utility to allow flags to be set and
13013 additional certificates supplied.
13017 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13018 OCSP client a number of certificate to only verify the response
13023 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13024 handle the new API. Currently only ECB, CBC modes supported. Add new
13027 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13028 Encryption Standard (AES) Ciphersuites for Transport Layer
13029 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13030 not enabled by default and were not part of the "ALL" ciphersuite
13031 alias because they were not yet official; they could be
13032 explicitly requested by specifying the "AESdraft" ciphersuite
13033 group alias. In the final release of OpenSSL 0.9.7, the group
13034 alias is called "AES" and is part of "ALL".)
13036 *Ben Laurie, Steve Henson, Bodo Moeller*
13038 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13039 request to response.
13043 * Functions for OCSP responders. OCSP_request_onereq_count(),
13044 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13045 extract information from a certificate request. OCSP_response_create()
13046 creates a response and optionally adds a basic response structure.
13047 OCSP_basic_add1_status() adds a complete single response to a basic
13048 response and returns the OCSP_SINGLERESP structure just added (to allow
13049 extensions to be included for example). OCSP_basic_add1_cert() adds a
13050 certificate to a basic response and OCSP_basic_sign() signs a basic
13051 response with various flags. New helper functions ASN1_TIME_check()
13052 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13053 (converts ASN1_TIME to GeneralizedTime).
13057 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13058 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13059 structure from a certificate. X509_pubkey_digest() digests the public_key
13060 contents: this is used in various key identifiers.
13064 * Make sk_sort() tolerate a NULL argument.
13066 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13068 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13069 passed by the function are trusted implicitly. If any of them signed the
13070 response then it is assumed to be valid and is not verified.
13074 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13075 to data. This was previously part of the PKCS7 ASN1 code. This
13076 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13077 *Steve Henson, reported by Kenneth R. Robinette
13078 <support@securenetterm.com>*
13080 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13081 routines: without these tracing memory leaks is very painful.
13082 Fix leaks in PKCS12 and PKCS7 routines.
13086 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13087 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13088 effectively meant GeneralizedTime would never be used. Now it
13089 is initialised to -1 but X509_time_adj() now has to check the value
13090 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13091 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13092 *Steve Henson, reported by Kenneth R. Robinette
13093 <support@securenetterm.com>*
13095 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13096 result in a zero length in the ASN1_INTEGER structure which was
13097 not consistent with the structure when d2i_ASN1_INTEGER() was used
13098 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13099 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13100 where it did not print out a minus for negative ASN1_INTEGER.
13104 * Add summary printout to ocsp utility. The various functions which
13105 convert status values to strings have been renamed to:
13106 OCSP_response_status_str(), OCSP_cert_status_str() and
13107 OCSP_crl_reason_str() and are no longer static. New options
13108 to verify nonce values and to disable verification. OCSP response
13109 printout format cleaned up.
13113 * Add additional OCSP certificate checks. These are those specified
13114 in RFC2560. This consists of two separate checks: the CA of the
13115 certificate being checked must either be the OCSP signer certificate
13116 or the issuer of the OCSP signer certificate. In the latter case the
13117 OCSP signer certificate must contain the OCSP signing extended key
13118 usage. This check is performed by attempting to match the OCSP
13119 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13120 in the OCSP_CERTID structures of the response.
13124 * Initial OCSP certificate verification added to OCSP_basic_verify()
13125 and related routines. This uses the standard OpenSSL certificate
13126 verify routines to perform initial checks (just CA validity) and
13127 to obtain the certificate chain. Then additional checks will be
13128 performed on the chain. Currently the root CA is checked to see
13129 if it is explicitly trusted for OCSP signing. This is used to set
13130 a root CA as a global signing root: that is any certificate that
13131 chains to that CA is an acceptable OCSP signing certificate.
13135 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13136 extensions from a separate configuration file.
13137 As when reading extensions from the main configuration file,
13138 the '-extensions ...' option may be used for specifying the
13141 *Massimiliano Pala <madwolf@comune.modena.it>*
13143 * New OCSP utility. Allows OCSP requests to be generated or
13144 read. The request can be sent to a responder and the output
13145 parsed, outputted or printed in text form. Not complete yet:
13146 still needs to check the OCSP response validity.
13150 * New subcommands for 'openssl ca':
13151 `openssl ca -status <serial>` prints the status of the cert with
13152 the given serial number (according to the index file).
13153 `openssl ca -updatedb` updates the expiry status of certificates
13156 *Massimiliano Pala <madwolf@comune.modena.it>*
13158 * New '-newreq-nodes' command option to CA.pl. This is like
13159 '-newreq', but calls 'openssl req' with the '-nodes' option
13160 so that the resulting key is not encrypted.
13162 *Damien Miller <djm@mindrot.org>*
13164 * New configuration for the GNU Hurd.
13166 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13168 * Initial code to implement OCSP basic response verify. This
13169 is currently incomplete. Currently just finds the signer's
13170 certificate and verifies the signature on the response.
13174 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13175 value of OPENSSLDIR. This is available via the new '-d' option
13176 to 'openssl version', and is also included in 'openssl version -a'.
13180 * Allowing defining memory allocation callbacks that will be given
13181 file name and line number information in additional arguments
13182 (a `const char*` and an int). The basic functionality remains, as
13183 well as the original possibility to just replace malloc(),
13184 realloc() and free() by functions that do not know about these
13185 additional arguments. To register and find out the current
13186 settings for extended allocation functions, the following
13187 functions are provided:
13189 CRYPTO_set_mem_ex_functions
13190 CRYPTO_set_locked_mem_ex_functions
13191 CRYPTO_get_mem_ex_functions
13192 CRYPTO_get_locked_mem_ex_functions
13194 These work the same way as CRYPTO_set_mem_functions and friends.
13195 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13196 extended allocation function is enabled.
13197 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13198 a conventional allocation function is enabled.
13200 *Richard Levitte, Bodo Moeller*
13202 * Finish off removing the remaining LHASH function pointer casts.
13203 There should no longer be any prototype-casting required when using
13204 the LHASH abstraction, and any casts that remain are "bugs". See
13205 the callback types and macros at the head of lhash.h for details
13206 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13210 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13211 If /dev/[u]random devices are not available or do not return enough
13212 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13214 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13215 /etc/entropy will be queried once each in this sequence, querying stops
13216 when enough entropy was collected without querying more sockets.
13220 * Change the Unix RAND_poll() variant to be able to poll several
13221 random devices, as specified by DEVRANDOM, until a sufficient amount
13222 of data has been collected. We spend at most 10 ms on each file
13223 (select timeout) and read in non-blocking mode. DEVRANDOM now
13224 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13225 (previously it was just the string "/dev/urandom"), so on typical
13226 platforms the 10 ms delay will never occur.
13227 Also separate out the Unix variant to its own file, rand_unix.c.
13228 For VMS, there's a currently-empty rand_vms.c.
13232 * Move OCSP client related routines to ocsp_cl.c. These
13233 provide utility functions which an application needing
13234 to issue a request to an OCSP responder and analyse the
13235 response will typically need: as opposed to those which an
13236 OCSP responder itself would need which will be added later.
13238 OCSP_request_sign() signs an OCSP request with an API similar
13239 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13240 response. OCSP_response_get1_basic() extracts basic response
13241 from response. OCSP_resp_find_status(): finds and extracts status
13242 information from an OCSP_CERTID structure (which will be created
13243 when the request structure is built). These are built from lower
13244 level functions which work on OCSP_SINGLERESP structures but
13245 won't normally be used unless the application wishes to examine
13246 extensions in the OCSP response for example.
13248 Replace nonce routines with a pair of functions.
13249 OCSP_request_add1_nonce() adds a nonce value and optionally
13250 generates a random value. OCSP_check_nonce() checks the
13251 validity of the nonce in an OCSP response.
13255 * Change function OCSP_request_add() to OCSP_request_add0_id().
13256 This doesn't copy the supplied OCSP_CERTID and avoids the
13257 need to free up the newly created id. Change return type
13258 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13259 This can then be used to add extensions to the request.
13260 Deleted OCSP_request_new(), since most of its functionality
13261 is now in OCSP_REQUEST_new() (and the case insensitive name
13262 clash) apart from the ability to set the request name which
13263 will be added elsewhere.
13267 * Update OCSP API. Remove obsolete extensions argument from
13268 various functions. Extensions are now handled using the new
13269 OCSP extension code. New simple OCSP HTTP function which
13270 can be used to send requests and parse the response.
13274 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13275 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13276 uses the special reorder version of SET OF to sort the attributes
13277 and reorder them to match the encoded order. This resolves a long
13278 standing problem: a verify on a PKCS7 structure just after signing
13279 it used to fail because the attribute order did not match the
13280 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13281 it uses the received order. This is necessary to tolerate some broken
13282 software that does not order SET OF. This is handled by encoding
13283 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13284 to produce the required SET OF.
13288 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13289 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13290 files to get correct declarations of the ASN.1 item variables.
13294 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13295 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13296 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13297 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13298 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13299 ASN1_ITEM and no wrapper functions.
13303 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13304 replace the old function pointer based I/O routines. Change most of
13305 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13309 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13310 lines, recognize more "algorithms" that can be deselected, and make
13311 it complain about algorithm deselection that isn't recognised.
13315 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13316 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13317 to use new functions. Add NO_ASN1_OLD which can be set to remove
13318 some old style ASN1 functions: this can be used to determine if old
13319 code will still work when these eventually go away.
13323 * New extension functions for OCSP structures, these follow the
13324 same conventions as certificates and CRLs.
13328 * New function X509V3_add1_i2d(). This automatically encodes and
13329 adds an extension. Its behaviour can be customised with various
13330 flags to append, replace or delete. Various wrappers added for
13331 certificates and CRLs.
13335 * Fix to avoid calling the underlying ASN1 print routine when
13336 an extension cannot be parsed. Correct a typo in the
13337 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13341 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13342 entries for variables.
13346 * Add functionality to `apps/openssl.c` for detecting locking
13347 problems: As the program is single-threaded, all we have
13348 to do is register a locking callback using an array for
13349 storing which locks are currently held by the program.
13353 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13354 SSL_get_ex_data_X509_STORE_idx(), which is used in
13355 ssl_verify_cert_chain() and thus can be called at any time
13356 during TLS/SSL handshakes so that thread-safety is essential.
13357 Unfortunately, the ex_data design is not at all suited
13358 for multi-threaded use, so it probably should be abolished.
13362 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13364 *Broadcom, tweaked and integrated by Geoff Thorpe*
13366 * Move common extension printing code to new function
13367 X509V3_print_extensions(). Reorganise OCSP print routines and
13368 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13372 * New function X509_signature_print() to remove duplication in some
13377 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13378 set (this was treated exactly the same as SET OF previously). This
13379 is used to reorder the STACK representing the structure to match the
13380 encoding. This will be used to get round a problem where a PKCS7
13381 structure which was signed could not be verified because the STACK
13382 order did not reflect the encoded order.
13386 * Reimplement the OCSP ASN1 module using the new code.
13390 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13391 for its ASN1 operations. The old style function pointers still exist
13392 for now but they will eventually go away.
13396 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13397 completely replaces the old ASN1 functionality with a table driven
13398 encoder and decoder which interprets an ASN1_ITEM structure describing
13399 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13400 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13401 has also been converted to the new form.
13405 * Change BN_mod_exp_recp so that negative moduli are tolerated
13406 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13407 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13408 for negative moduli.
13412 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13413 of not touching the result's sign bit.
13417 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13422 * Changed the LHASH code to use prototypes for callbacks, and created
13423 macros to declare and implement thin (optionally static) functions
13424 that provide type-safety and avoid function pointer casting for the
13425 type-specific callbacks.
13429 * Added Kerberos Cipher Suites to be used with TLS, as written in
13431 *Veers Staats <staatsvr@asc.hpc.mil>,
13432 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13434 * Reformat the FAQ so the different questions and answers can be divided
13435 in sections depending on the subject.
13439 * Have the zlib compression code load ZLIB.DLL dynamically under
13444 * New function BN_mod_sqrt for computing square roots modulo a prime
13445 (using the probabilistic Tonelli-Shanks algorithm unless
13446 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13447 be handled deterministically).
13449 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13451 * Make BN_mod_inverse faster by explicitly handling small quotients
13452 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13453 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13457 * New function BN_kronecker.
13461 * Fix BN_gcd so that it works on negative inputs; the result is
13462 positive unless both parameters are zero.
13463 Previously something reasonably close to an infinite loop was
13464 possible because numbers could be growing instead of shrinking
13465 in the implementation of Euclid's algorithm.
13469 * Fix BN_is_word() and BN_is_one() macros to take into account the
13470 sign of the number in question.
13472 Fix BN_is_word(a,w) to work correctly for w == 0.
13474 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13475 because its test if the absolute value of 'a' equals 'w'.
13476 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13477 it exists mostly for use in the implementations of BN_is_zero(),
13478 BN_is_one(), and BN_is_word().
13482 * New function BN_swap.
13486 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13487 the exponentiation functions are more likely to produce reasonable
13488 results on negative inputs.
13492 * Change BN_mod_mul so that the result is always non-negative.
13493 Previously, it could be negative if one of the factors was negative;
13494 I don't think anyone really wanted that behaviour.
13498 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13499 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13500 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13501 and add new functions:
13510 BN_mod_lshift1_quick
13512 BN_mod_lshift_quick
13514 These functions always generate non-negative results.
13516 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13517 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13519 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13520 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13521 be reduced modulo `m`.
13523 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13526 The following entry accidentally appeared in the CHANGES file
13527 distributed with OpenSSL 0.9.7. The modifications described in
13528 it do *not* apply to OpenSSL 0.9.7.
13530 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13531 was actually never needed) and in BN_mul(). The removal in BN_mul()
13532 required a small change in bn_mul_part_recursive() and the addition
13533 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13534 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13535 bn_sub_words() and bn_add_words() except they take arrays with
13541 * In 'openssl passwd', verify passwords read from the terminal
13542 unless the '-salt' option is used (which usually means that
13543 verification would just waste user's time since the resulting
13544 hash is going to be compared with some given password hash)
13545 or the new '-noverify' option is used.
13547 This is an incompatible change, but it does not affect
13548 non-interactive use of 'openssl passwd' (passwords on the command
13549 line, '-stdin' option, '-in ...' option) and thus should not
13550 cause any problems.
13554 * Remove all references to RSAref, since there's no more need for it.
13558 * Make DSO load along a path given through an environment variable
13559 (SHLIB_PATH) with shl_load().
13563 * Constify the ENGINE code as a result of BIGNUM constification.
13564 Also constify the RSA code and most things related to it. In a
13565 few places, most notable in the depth of the ASN.1 code, ugly
13566 casts back to non-const were required (to be solved at a later
13571 * Make it so the openssl application has all engines loaded by default.
13575 * Constify the BIGNUM routines a little more.
13579 * Add the following functions:
13581 ENGINE_load_cswift()
13583 ENGINE_load_atalla()
13584 ENGINE_load_nuron()
13585 ENGINE_load_builtin_engines()
13587 That way, an application can itself choose if external engines that
13588 are built-in in OpenSSL shall ever be used or not. The benefit is
13589 that applications won't have to be linked with libdl or other dso
13590 libraries unless it's really needed.
13592 Changed 'openssl engine' to load all engines on demand.
13593 Changed the engine header files to avoid the duplication of some
13594 declarations (they differed!).
13598 * 'openssl engine' can now list capabilities.
13602 * Better error reporting in 'openssl engine'.
13606 * Never call load_dh_param(NULL) in s_server.
13610 * Add engine application. It can currently list engines by name and
13611 identity, and test if they are actually available.
13615 * Improve RPM specification file by forcing symbolic linking and making
13616 sure the installed documentation is also owned by root.root.
13618 *Damien Miller <djm@mindrot.org>*
13620 * Give the OpenSSL applications more possibilities to make use of
13621 keys (public as well as private) handled by engines.
13625 * Add OCSP code that comes from CertCo.
13629 * Add VMS support for the Rijndael code.
13633 * Added untested support for Nuron crypto accelerator.
13637 * Add support for external cryptographic devices. This code was
13638 previously distributed separately as the "engine" branch.
13640 *Geoff Thorpe, Richard Levitte*
13642 * Rework the filename-translation in the DSO code. It is now possible to
13643 have far greater control over how a "name" is turned into a filename
13644 depending on the operating environment and any oddities about the
13645 different shared library filenames on each system.
13649 * Support threads on FreeBSD-elf in Configure.
13653 * Fix for SHA1 assembly problem with MASM: it produces
13654 warnings about corrupt line number information when assembling
13655 with debugging information. This is caused by the overlapping
13658 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13661 NCONF_get_number() has no error checking at all. As a replacement,
13662 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13663 promoted strongly. The old NCONF_get_number is kept around for
13664 binary backward compatibility.
13665 Make it possible for methods to load from something other than a BIO,
13666 by providing a function pointer that is given a name instead of a BIO.
13667 For example, this could be used to load configuration data from an
13672 * Fix for non blocking accept BIOs. Added new I/O special reason
13673 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13674 with non blocking I/O was not possible because no retry code was
13675 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13680 * Added the beginnings of Rijndael support.
13684 * Fix for bug in DirectoryString mask setting. Add support for
13685 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13686 to allow certificate printing to more controllable, additional
13687 'certopt' option to 'x509' to allow new printing options to be
13692 * Clean old EAY MD5 hack from e_os.h.
13696 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13698 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13699 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13701 *Joe Orton, Steve Henson*
13703 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13705 * Fix additional bug revealed by the NISCC test suite:
13707 Stop bug triggering large recursion when presented with
13708 certain ASN.1 tags ([CVE-2003-0851])
13712 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13714 * Fix various bugs revealed by running the NISCC test suite:
13716 Stop out of bounds reads in the ASN1 code when presented with
13717 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13719 If verify callback ignores invalid public key errors don't try to check
13720 certificate signature with the NULL public key.
13724 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13725 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13730 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13731 extra data after the compression methods not only for TLS 1.0
13732 but also for SSL 3.0 (as required by the specification).
13734 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13736 * Change X509_certificate_type() to mark the key as exported/exportable
13737 when it's 512 *bits* long, not 512 bytes.
13741 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13743 * Countermeasure against the Klima-Pokorny-Rosa extension of
13744 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13745 a protocol version number mismatch like a decryption error
13746 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13750 * Turn on RSA blinding by default in the default implementation
13751 to avoid a timing attack. Applications that don't want it can call
13752 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13753 They would be ill-advised to do so in most cases.
13755 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13757 * Change RSA blinding code so that it works when the PRNG is not
13758 seeded (in this case, the secret RSA exponent is abused as
13759 an unpredictable seed -- if it is not unpredictable, there
13760 is no point in blinding anyway). Make RSA blinding thread-safe
13761 by remembering the creator's thread ID in rsa->blinding and
13762 having all other threads use local one-time blinding factors
13763 (this requires more computation than sharing rsa->blinding, but
13764 avoids excessive locking; and if an RSA object is not shared
13765 between threads, blinding will still be very fast).
13769 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13771 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13772 via timing by performing a MAC computation even if incorrect
13773 block cipher padding has been found. This is a countermeasure
13774 against active attacks where the attacker has to distinguish
13775 between bad padding and a MAC verification error. ([CVE-2003-0078])
13777 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13778 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13779 Martin Vuagnoux (EPFL, Ilion)*
13781 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13783 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13784 memory from its contents. This is done with a counter that will
13785 place alternating values in each byte. This can be used to solve
13786 two issues: 1) the removal of calls to memset() by highly optimizing
13787 compilers, and 2) cleansing with other values than 0, since those can
13788 be read through on certain media, for example a swap space on disk.
13792 * Bugfix: client side session caching did not work with external caching,
13793 because the session->cipher setting was not restored when reloading
13794 from the external cache. This problem was masked, when
13795 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13796 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13800 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13801 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13803 *Zeev Lieber <zeev-l@yahoo.com>*
13805 * Undo an undocumented change introduced in 0.9.6e which caused
13806 repeated calls to OpenSSL_add_all_ciphers() and
13807 OpenSSL_add_all_digests() to be ignored, even after calling
13812 * Change the default configuration reader to deal with last line not
13813 being properly terminated.
13817 * Change X509_NAME_cmp() so it applies the special rules on handling
13818 DN values that are of type PrintableString, as well as RDNs of type
13819 emailAddress where the value has the type ia5String.
13821 *stefank@valicert.com via Richard Levitte*
13823 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13824 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13825 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13826 the bitwise-OR of the two for use by the majority of applications
13827 wanting this behaviour, and update the docs. The documented
13828 behaviour and actual behaviour were inconsistent and had been
13829 changing anyway, so this is more a bug-fix than a behavioural
13832 *Geoff Thorpe, diagnosed by Nadav Har'El*
13834 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13835 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13839 * Fix initialization code race conditions in
13840 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13841 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13842 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13843 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13844 ssl2_get_cipher_by_char(),
13845 ssl3_get_cipher_by_char().
13847 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13849 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13850 the cached sessions are flushed, as the remove_cb() might use ex_data
13851 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13852 (see [openssl.org #212]).
13854 *Geoff Thorpe, Lutz Jaenicke*
13856 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13857 length, instead of the encoding length to d2i_ASN1_OBJECT.
13861 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13863 * [In 0.9.6g-engine release:]
13864 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13866 *Lynn Gazis <lgazis@rainbow.com>*
13868 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13870 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13871 and get fix the header length calculation.
13872 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13873 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13875 * Use proper error handling instead of 'assertions' in buffer
13876 overflow checks added in 0.9.6e. This prevents DoS (the
13877 assertions could call abort()).
13879 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13881 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13883 * Add various sanity checks to asn1_get_length() to reject
13884 the ASN1 length bytes if they exceed sizeof(long), will appear
13885 negative or the content length exceeds the length of the
13888 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13890 * Fix cipher selection routines: ciphers without encryption had no flags
13891 for the cipher strength set and where therefore not handled correctly
13892 by the selection routines (PR #130).
13896 * Fix EVP_dsa_sha macro.
13901 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13902 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13903 that was added in OpenSSL 0.9.6d.
13905 As the countermeasure turned out to be incompatible with some
13906 broken SSL implementations, the new option is part of SSL_OP_ALL.
13907 SSL_OP_ALL is usually employed when compatibility with weird SSL
13908 implementations is desired (e.g. '-bugs' option to 's_client' and
13909 's_server'), so the new option is automatically set in many
13914 * Changes in security patch:
13916 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13917 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13918 Air Force Materiel Command, USAF, under agreement number
13921 * Add various sanity checks to asn1_get_length() to reject
13922 the ASN1 length bytes if they exceed sizeof(long), will appear
13923 negative or the content length exceeds the length of the
13924 supplied buffer. ([CVE-2002-0659])
13926 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13928 * Assertions for various potential buffer overflows, not known to
13929 happen in practice.
13931 *Ben Laurie (CHATS)*
13933 * Various temporary buffers to hold ASCII versions of integers were
13934 too small for 64 bit platforms. ([CVE-2002-0655])
13935 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13937 * Remote buffer overflow in SSL3 protocol - an attacker could
13938 supply an oversized session ID to a client. ([CVE-2002-0656])
13940 *Ben Laurie (CHATS)*
13942 * Remote buffer overflow in SSL2 protocol - an attacker could
13943 supply an oversized client master key. ([CVE-2002-0656])
13945 *Ben Laurie (CHATS)*
13947 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13949 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13950 encoded as NULL) with id-dsa-with-sha1.
13952 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13954 * Check various `X509_...()` return values in `apps/req.c`.
13956 *Nils Larsch <nla@trustcenter.de>*
13958 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13959 an end-of-file condition would erroneously be flagged, when the CRLF
13960 was just at the end of a processed block. The bug was discovered when
13961 processing data through a buffering memory BIO handing the data to a
13962 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13963 <ptsekov@syntrex.com> and Nedelcho Stanev.
13967 * Implement a countermeasure against a vulnerability recently found
13968 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13969 before application data chunks to avoid the use of known IVs
13970 with data potentially chosen by the attacker.
13974 * Fix length checks in ssl3_get_client_hello().
13978 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13979 to prevent ssl3_read_internal() from incorrectly assuming that
13980 ssl3_read_bytes() found application data while handshake
13981 processing was enabled when in fact s->s3->in_read_app_data was
13982 merely automatically cleared during the initial handshake.
13984 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13986 * Fix object definitions for Private and Enterprise: they were not
13987 recognized in their shortname (=lowercase) representation. Extend
13988 obj_dat.pl to issue an error when using undefined keywords instead
13989 of silently ignoring the problem (Svenning Sorensen
13990 <sss@sss.dnsalias.net>).
13994 * Fix DH_generate_parameters() so that it works for 'non-standard'
13995 generators, i.e. generators other than 2 and 5. (Previously, the
13996 code did not properly initialise the 'add' and 'rem' values to
13997 BN_generate_prime().)
13999 In the new general case, we do not insist that 'generator' is
14000 actually a primitive root: This requirement is rather pointless;
14001 a generator of the order-q subgroup is just as good, if not
14006 * Map new X509 verification errors to alerts. Discovered and submitted by
14007 Tom Wu <tom@arcot.com>.
14011 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14012 returning non-zero before the data has been completely received
14013 when using non-blocking I/O.
14015 *Bodo Moeller; problem pointed out by John Hughes*
14017 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14019 *Ben Laurie, Lutz Jaenicke*
14021 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14022 Yoram Zahavi <YoramZ@gilian.com>).
14026 * Add information about CygWin 1.3 and on, and preserve proper
14027 configuration for the versions before that.
14029 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14031 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14032 check whether we deal with a copy of a session and do not delete from
14033 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14034 <izhar@checkpoint.com>.
14038 * Do not store session data into the internal session cache, if it
14039 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14040 flag is set). Proposed by Aslam <aslam@funk.com>.
14044 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14049 * [In 0.9.6d-engine release:]
14050 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14052 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14054 * Add the configuration target linux-s390x.
14056 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14058 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14059 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14060 variable as an indication that a ClientHello message has been
14061 received. As the flag value will be lost between multiple
14062 invocations of ssl3_accept when using non-blocking I/O, the
14063 function may not be aware that a handshake has actually taken
14064 place, thus preventing a new session from being added to the
14067 To avoid this problem, we now set s->new_session to 2 instead of
14068 using a local variable.
14070 *Lutz Jaenicke, Bodo Moeller*
14072 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14073 if the SSL_R_LENGTH_MISMATCH error is detected.
14075 *Geoff Thorpe, Bodo Moeller*
14077 * New 'shared_ldflag' column in Configure platform table.
14081 * Fix EVP_CIPHER_mode macro.
14083 *"Dan S. Camper" <dan@bti.net>*
14085 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14086 type, we must throw them away by setting rr->length to 0.
14088 *D P Chang <dpc@qualys.com>*
14090 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14092 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14093 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14094 worked incorrectly for those cases where range = `10..._2` and
14095 `3*range` is two bits longer than range.)
14099 * Only add signing time to PKCS7 structures if it is not already
14104 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14105 OBJ_ld_ce should be OBJ_id_ce.
14106 Also some ip-pda OIDs in crypto/objects/objects.txt were
14107 incorrect (cf. RFC 3039).
14109 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14111 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14112 returns early because it has nothing to do.
14114 *Andy Schneider <andy.schneider@bjss.co.uk>*
14116 * [In 0.9.6c-engine release:]
14117 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14119 *Andy Schneider <andy.schneider@bjss.co.uk>*
14121 * [In 0.9.6c-engine release:]
14122 Add support for Cryptographic Appliance's keyserver technology.
14123 (Use engine 'keyclient')
14125 *Cryptographic Appliances and Geoff Thorpe*
14127 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14128 is called via tools/c89.sh because arguments have to be
14129 rearranged (all '-L' options must appear before the first object
14132 *Richard Shapiro <rshapiro@abinitio.com>*
14134 * [In 0.9.6c-engine release:]
14135 Add support for Broadcom crypto accelerator cards, backported
14138 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14140 * [In 0.9.6c-engine release:]
14141 Add support for SureWare crypto accelerator cards from
14142 Baltimore Technologies. (Use engine 'sureware')
14144 *Baltimore Technologies and Mark Cox*
14146 * [In 0.9.6c-engine release:]
14147 Add support for crypto accelerator cards from Accelerated
14148 Encryption Processing, www.aep.ie. (Use engine 'aep')
14150 *AEP Inc. and Mark Cox*
14152 * Add a configuration entry for gcc on UnixWare.
14154 *Gary Benson <gbenson@redhat.com>*
14156 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14157 messages are stored in a single piece (fixed-length part and
14158 variable-length part combined) and fix various bugs found on the way.
14162 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14163 instead. BIO_gethostbyname() does not know what timeouts are
14164 appropriate, so entries would stay in cache even when they have
14166 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14168 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14169 faced with a pathologically small ClientHello fragment that does
14170 not contain client_version: Instead of aborting with an error,
14171 simply choose the highest available protocol version (i.e.,
14172 TLS 1.0 unless it is disabled). In practice, ClientHello
14173 messages are never sent like this, but this change gives us
14174 strictly correct behaviour at least for TLS.
14178 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14179 never resets s->method to s->ctx->method when called from within
14180 one of the SSL handshake functions.
14182 *Bodo Moeller; problem pointed out by Niko Baric*
14184 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14185 (sent using the client's version number) if client_version is
14186 smaller than the protocol version in use. Also change
14187 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14188 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14189 the client will at least see that alert.
14193 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14198 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14199 client receives HelloRequest while in a handshake.
14201 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14203 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14204 should end in 'break', not 'goto end' which circumvents various
14205 cleanups done in state SSL_ST_OK. But session related stuff
14206 must be disabled for SSL_ST_OK in the case that we just sent a
14209 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14210 before just sending a HelloRequest.
14212 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14214 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14215 reveal whether illegal block cipher padding was found or a MAC
14216 verification error occurred. (Neither SSLerr() codes nor alerts
14217 are directly visible to potential attackers, but the information
14218 may leak via logfiles.)
14220 Similar changes are not required for the SSL 2.0 implementation
14221 because the number of padding bytes is sent in clear for SSL 2.0,
14222 and the extra bytes are just ignored. However ssl/s2_pkt.c
14223 failed to verify that the purported number of padding bytes is in
14228 * Add OpenUNIX-8 support including shared libraries
14229 (Boyd Lynn Gerber <gerberb@zenez.com>).
14233 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14234 'wristwatch attack' using huge encoding parameters (cf.
14235 James H. Manger's CRYPTO 2001 paper). Note that the
14236 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14237 encoding parameters and hence was not vulnerable.
14241 * BN_sqr() bug fix.
14243 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14245 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14246 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14247 followed by modular reduction.
14249 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14251 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14252 equivalent based on BN_pseudo_rand() instead of BN_rand().
14256 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14257 This function was broken, as the check for a new client hello message
14258 to handle SGC did not allow these large messages.
14259 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14263 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14267 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14268 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14272 * Rework the configuration and shared library support for Tru64 Unix.
14273 The configuration part makes use of modern compiler features and
14274 still retains old compiler behavior for those that run older versions
14275 of the OS. The shared library support part includes a variant that
14276 uses the RPATH feature, and is available through the special
14277 configuration target "alpha-cc-rpath", which will never be selected
14280 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14282 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14283 with the same message size as in ssl3_get_certificate_request().
14284 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14285 messages might inadvertently be reject as too long.
14287 *Petr Lampa <lampa@fee.vutbr.cz>*
14289 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14293 * Modified SSL library such that the verify_callback that has been set
14294 specifically for an SSL object with SSL_set_verify() is actually being
14295 used. Before the change, a verify_callback set with this function was
14296 ignored and the verify_callback() set in the SSL_CTX at the time of
14297 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14298 to allow the necessary settings.
14302 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14303 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14304 done automatically (in contradiction to the requirements of the C
14305 standard). This made problems when used from OpenSSH.
14309 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14310 dh->length and always used
14312 BN_rand_range(priv_key, dh->p).
14314 BN_rand_range() is not necessary for Diffie-Hellman, and this
14315 specific range makes Diffie-Hellman unnecessarily inefficient if
14316 dh->length (recommended exponent length) is much smaller than the
14317 length of dh->p. We could use BN_rand_range() if the order of
14318 the subgroup was stored in the DH structure, but we only have
14323 BN_rand(priv_key, l, ...)
14325 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14332 RSA_eay_public_encrypt
14333 RSA_eay_private_decrypt
14334 RSA_eay_private_encrypt (signing)
14335 RSA_eay_public_decrypt (signature verification)
14337 (default implementations for RSA_public_encrypt,
14338 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14339 always reject numbers >= n.
14343 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14344 to synchronize access to 'locking_thread'. This is necessary on
14345 systems where access to 'locking_thread' (an 'unsigned long'
14346 variable) is not atomic.
14350 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14351 *before* setting the 'crypto_lock_rand' flag. The previous code had
14352 a race condition if 0 is a valid thread ID.
14354 *Travis Vitek <vitek@roguewave.com>*
14356 * Add support for shared libraries under Irix.
14358 *Albert Chin-A-Young <china@thewrittenword.com>*
14360 * Add configuration option to build on Linux on both big-endian and
14361 little-endian MIPS.
14363 *Ralf Baechle <ralf@uni-koblenz.de>*
14365 * Add the possibility to create shared libraries on HP-UX.
14369 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14371 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14372 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14373 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14374 PRNG state recovery was possible based on the output of
14375 one PRNG request appropriately sized to gain knowledge on
14376 'md' followed by enough consecutive 1-byte PRNG requests
14377 to traverse all of 'state'.
14379 1. When updating 'md_local' (the current thread's copy of 'md')
14380 during PRNG output generation, hash all of the previous
14381 'md_local' value, not just the half used for PRNG output.
14383 2. Make the number of bytes from 'state' included into the hash
14384 independent from the number of PRNG bytes requested.
14386 The first measure alone would be sufficient to avoid
14387 Markku-Juhani's attack. (Actually it had never occurred
14388 to me that the half of 'md_local' used for chaining was the
14389 half from which PRNG output bytes were taken -- I had always
14390 assumed that the secret half would be used.) The second
14391 measure makes sure that additional data from 'state' is never
14392 mixed into 'md_local' in small portions; this heuristically
14393 further strengthens the PRNG.
14397 * Fix crypto/bn/asm/mips3.s.
14401 * When only the key is given to "enc", the IV is undefined. Print out
14402 an error message in this case.
14406 * Handle special case when X509_NAME is empty in X509 printing routines.
14410 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14411 positive and less than q.
14415 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14416 used: it isn't thread safe and the add_lock_callback should handle
14419 *Paul Rose <Paul.Rose@bridge.com>*
14421 * Verify that incoming data obeys the block size in
14422 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14428 *Ulf Möller, Bodo Möller*
14430 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14431 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14432 when fixing the server behaviour for backwards-compatible 'client
14433 hello' messages. (Note that the attack is impractical against
14434 SSL 3.0 and TLS 1.0 anyway because length and version checking
14435 means that the probability of guessing a valid ciphertext is
14436 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14439 Before 0.9.5, the countermeasure (hide the error by generating a
14440 random 'decryption result') did not work properly because
14441 ERR_clear_error() was missing, meaning that SSL_get_error() would
14442 detect the supposedly ignored error.
14444 Both problems are now fixed.
14448 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14449 (previously it was 1024).
14453 * Fix for compatibility mode trust settings: ignore trust settings
14454 unless some valid trust or reject settings are present.
14458 * Fix for blowfish EVP: its a variable length cipher.
14462 * Fix various bugs related to DSA S/MIME verification. Handle missing
14463 parameters in DSA public key structures and return an error in the
14464 DSA routines if parameters are absent.
14468 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14469 in the current directory if neither $RANDFILE nor $HOME was set.
14470 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14471 caused some confusion to Windows users who haven't defined $HOME.
14472 Thus RAND_file_name() is changed again: e_os.h can define a
14473 DEFAULT_HOME, which will be used if $HOME is not set.
14474 For Windows, we use "C:"; on other platforms, we still require
14475 environment variables.
14477 * Move 'if (!initialized) RAND_poll()' into regions protected by
14478 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14479 having multiple threads call RAND_poll() concurrently.
14483 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14484 combination of a flag and a thread ID variable.
14485 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14486 flag), *other* threads can enter ssleay_add_bytes without obeying
14487 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14488 that they do not hold after the first thread unsets add_do_not_lock).
14492 * Change bctest again: '-x' expressions are not available in all
14493 versions of 'test'.
14497 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14499 * Fix a couple of memory leaks in PKCS7_dataDecode()
14501 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14503 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14504 the default extension for executables, if any. Also, make the perl
14505 scripts that use symlink() to test if it really exists and use "cp"
14506 if it doesn't. All this made OpenSSL compilable and installable in
14511 * Fix for asn1_GetSequence() for indefinite length constructed data.
14512 If SEQUENCE is length is indefinite just set c->slen to the total
14513 amount of data available.
14515 *Steve Henson, reported by shige@FreeBSD.org*
14517 *This change does not apply to 0.9.7.*
14519 * Change bctest to avoid here-documents inside command substitution
14520 (workaround for FreeBSD /bin/sh bug).
14521 For compatibility with Ultrix, avoid shell functions (introduced
14522 in the bctest version that searches along $PATH).
14526 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14527 with des_encrypt() defined on some operating systems, like Solaris
14532 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14533 On the Importance of Eliminating Errors in Cryptographic
14534 Computations, J. Cryptology 14 (2001) 2, 101-119,
14535 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14539 * MIPS assembler BIGNUM division bug fix.
14543 * Disabled incorrect Alpha assembler code.
14547 * Fix PKCS#7 decode routines so they correctly update the length
14548 after reading an EOC for the EXPLICIT tag.
14552 *This change does not apply to 0.9.7.*
14554 * Fix bug in PKCS#12 key generation routines. This was triggered
14555 if a 3DES key was generated with a 0 initial byte. Include
14556 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14557 (but broken) behaviour.
14561 * Enhance bctest to search for a working bc along $PATH and print
14564 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14566 * Fix memory leaks in err.c: free err_data string if necessary;
14567 don't write to the wrong index in ERR_set_error_data.
14571 * Implement ssl23_peek (analogous to ssl23_read), which previously
14576 * Replace rdtsc with `_emit` statements for VC++ version 5.
14578 *Jeremy Cooper <jeremy@baymoo.org>*
14580 * Make it possible to reuse SSLv2 sessions.
14584 * In copy_email() check for >= 0 as a return value for
14585 X509_NAME_get_index_by_NID() since 0 is a valid index.
14587 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14589 * Avoid coredump with unsupported or invalid public keys by checking if
14590 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14591 PKCS7_verify() fails with non detached data.
14595 * Don't use getenv in library functions when run as setuid/setgid.
14596 New function OPENSSL_issetugid().
14600 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14601 due to incorrect handling of multi-threading:
14603 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14605 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14607 3. Count how many times MemCheck_off() has been called so that
14608 nested use can be treated correctly. This also avoids
14609 inband-signalling in the previous code (which relied on the
14610 assumption that thread ID 0 is impossible).
14614 * Add "-rand" option also to s_client and s_server.
14618 * Fix CPU detection on Irix 6.x.
14619 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14620 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14622 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14627 *This change does not apply to 0.9.7.*
14629 * Use the cached encoding of an X509_NAME structure rather than
14630 copying it. This is apparently the reason for the libsafe "errors"
14631 but the code is actually correct.
14635 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14636 Bleichenbacher's DSA attack.
14637 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14638 to be set and top=0 forces the highest bit to be set; top=-1 is new
14639 and leaves the highest bit random.
14641 *Ulf Moeller, Bodo Moeller*
14643 * In the `NCONF_...`-based implementations for `CONF_...` queries
14644 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14645 a temporary CONF structure with the data component set to NULL
14646 (which gives segmentation faults in lh_retrieve).
14647 Instead, use NULL for the CONF pointer in CONF_get_string and
14648 CONF_get_number (which may use environment variables) and directly
14649 return NULL from CONF_get_section.
14653 * Fix potential buffer overrun for EBCDIC.
14657 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14658 keyUsage if basicConstraints absent for a CA.
14662 * Make SMIME_write_PKCS7() write mail header values with a format that
14663 is more generally accepted (no spaces before the semicolon), since
14664 some programs can't parse those values properly otherwise. Also make
14665 sure BIO's that break lines after each write do not create invalid
14670 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14671 macros previously used would not encode an empty SEQUENCE OF
14672 and break the signature.
14676 *This change does not apply to 0.9.7.*
14678 * Zero the premaster secret after deriving the master secret in
14683 * Add some EVP_add_digest_alias registrations (as found in
14684 OpenSSL_add_all_digests()) to SSL_library_init()
14685 aka OpenSSL_add_ssl_algorithms(). This provides improved
14686 compatibility with peers using X.509 certificates
14687 with unconventional AlgorithmIdentifier OIDs.
14691 * Fix for Irix with NO_ASM.
14693 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14695 * ./config script fixes.
14697 *Ulf Moeller, Richard Levitte*
14699 * Fix 'openssl passwd -1'.
14703 * Change PKCS12_key_gen_asc() so it can cope with non null
14704 terminated strings whose length is passed in the passlen
14705 parameter, for example from PEM callbacks. This was done
14706 by adding an extra length parameter to asc2uni().
14708 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14710 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14711 call failed, free the DSA structure.
14715 * Fix to uni2asc() to cope with zero length Unicode strings.
14716 These are present in some PKCS#12 files.
14720 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14721 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14722 when writing a 32767 byte record.
14724 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14726 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14727 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14729 (RSA objects have a reference count access to which is protected
14730 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14731 so they are meant to be shared between threads.)
14732 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14733 "Reddie, Steven" <Steven.Reddie@ca.com>*
14735 * Fix a deadlock in CRYPTO_mem_leaks().
14739 * Use better test patterns in bntest.
14743 * rand_win.c fix for Borland C.
14747 * BN_rshift bugfix for n == 0.
14751 * Add a 'bctest' script that checks for some known 'bc' bugs
14752 so that 'make test' does not abort just because 'bc' is broken.
14756 * Store verify_result within SSL_SESSION also for client side to
14757 avoid potential security hole. (Re-used sessions on the client side
14758 always resulted in verify_result==X509_V_OK, not using the original
14759 result of the server certificate verification.)
14763 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14764 SSL3_RT_APPLICATION_DATA, return 0.
14765 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14770 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14771 releases, have been re-implemented by renaming the previous
14772 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14773 and ssl3_read_internal, respectively, and adding 'peek' parameters
14774 to them. The new ssl[23]_{read,peek} functions are calls to
14775 ssl[23]_read_internal with the 'peek' flag set appropriately.
14776 A 'peek' parameter has also been added to ssl3_read_bytes, which
14777 does the actual work for ssl3_read_internal.
14781 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14782 the method-specific "init()" handler. Also clean up ex_data after
14783 calling the method-specific "finish()" handler. Previously, this was
14784 happening the other way round.
14788 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14789 The previous value, 12, was not always sufficient for BN_mod_exp().
14793 * Make sure that shared libraries get the internal name engine with
14794 the full version number and not just 0. This should mark the
14795 shared libraries as not backward compatible. Of course, this should
14796 be changed again when we can guarantee backward binary compatibility.
14800 * Fix typo in get_cert_by_subject() in by_dir.c
14802 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14804 * Rework the system to generate shared libraries:
14806 - Make note of the expected extension for the shared libraries and
14807 if there is a need for symbolic links from for example libcrypto.so.0
14808 to libcrypto.so.0.9.7. There is extended info in Configure for
14811 - Make as few rebuilds of the shared libraries as possible.
14813 - Still avoid linking the OpenSSL programs with the shared libraries.
14815 - When installing, install the shared libraries separately from the
14820 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14822 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14823 and not in SSL_clear because the latter is also used by the
14824 accept/connect functions; previously, the settings made by
14825 SSL_set_read_ahead would be lost during the handshake.
14827 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14829 * Correct util/mkdef.pl to be selective about disabled algorithms.
14830 Previously, it would create entries for disabled algorithms no
14835 * Added several new manual pages for SSL_* function.
14839 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14841 * In ssl23_get_client_hello, generate an error message when faced
14842 with an initial SSL 3.0/TLS record that is too small to contain the
14843 first two bytes of the ClientHello message, i.e. client_version.
14844 (Note that this is a pathologic case that probably has never happened
14845 in real life.) The previous approach was to use the version number
14846 from the record header as a substitute; but our protocol choice
14847 should not depend on that one because it is not authenticated
14848 by the Finished messages.
14852 * More robust randomness gathering functions for Windows.
14854 *Jeffrey Altman <jaltman@columbia.edu>*
14856 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14857 not set then we don't setup the error code for issuer check errors
14858 to avoid possibly overwriting other errors which the callback does
14859 handle. If an application does set the flag then we assume it knows
14860 what it is doing and can handle the new informational codes
14865 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14866 a general "ANY" type, as such it should be able to decode anything
14867 including tagged types. However it didn't check the class so it would
14868 wrongly interpret tagged types in the same way as their universal
14869 counterpart and unknown types were just rejected. Changed so that the
14870 tagged and unknown types are handled in the same way as a SEQUENCE:
14871 that is the encoding is stored intact. There is also a new type
14872 "V_ASN1_OTHER" which is used when the class is not universal, in this
14873 case we have no idea what the actual type is so we just lump them all
14878 * On VMS, stdout may very well lead to a file that is written to
14879 in a record-oriented fashion. That means that every write() will
14880 write a separate record, which will be read separately by the
14881 programs trying to read from it. This can be very confusing.
14883 The solution is to put a BIO filter in the way that will buffer
14884 text until a linefeed is reached, and then write everything a
14885 line at a time, so every record written will be an actual line,
14886 not chunks of lines and not (usually doesn't happen, but I've
14887 seen it once) several lines in one record. BIO_f_linebuffer() is
14890 Currently, it's a VMS-only method, because that's where it has
14891 been tested well enough.
14895 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14896 it can return incorrect results.
14897 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14898 but it was in 0.9.6-beta[12].)
14902 * Disable the check for content being present when verifying detached
14903 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14904 include zero length content when signing messages.
14908 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14909 BIO_ctrl (for BIO pairs).
14913 * Add DSO method for VMS.
14917 * Bug fix: Montgomery multiplication could produce results with the
14922 * Add RPM specification openssl.spec and modify it to build three
14923 packages. The default package contains applications, application
14924 documentation and run-time libraries. The devel package contains
14925 include files, static libraries and function documentation. The
14926 doc package contains the contents of the doc directory. The original
14927 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14931 * Add a large number of documentation files for many SSL routines.
14933 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14935 * Add a configuration entry for Sony News 4.
14937 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14939 * Don't set the two most significant bits to one when generating a
14940 random number < q in the DSA library.
14944 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14945 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14946 the underlying transport is blocking) if a handshake took place.
14947 (The default behaviour is needed by applications such as s_client
14948 and s_server that use select() to determine when to use SSL_read;
14949 but for applications that know in advance when to expect data, it
14950 just makes things more complicated.)
14954 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14959 * Add a few more EBCDIC conditionals that make `req` and `x509`
14960 work better on such systems.
14962 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14964 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14965 Update PKCS12_parse() so it copies the friendlyName and the
14966 keyid to the certificates aux info.
14970 * Fix bug in PKCS7_verify() which caused an infinite loop
14971 if there was more than one signature.
14973 *Sven Uszpelkat <su@celocom.de>*
14975 * Major change in util/mkdef.pl to include extra information
14976 about each symbol, as well as presenting variables as well
14977 as functions. This change means that there's n more need
14978 to rebuild the .num files when some algorithms are excluded.
14982 * Allow the verify time to be set by an application,
14983 rather than always using the current time.
14987 * Phase 2 verify code reorganisation. The certificate
14988 verify code now looks up an issuer certificate by a
14989 number of criteria: subject name, authority key id
14990 and key usage. It also verifies self signed certificates
14991 by the same criteria. The main comparison function is
14992 X509_check_issued() which performs these checks.
14994 Lot of changes were necessary in order to support this
14995 without completely rewriting the lookup code.
14997 Authority and subject key identifier are now cached.
14999 The LHASH 'certs' is X509_STORE has now been replaced
15000 by a STACK_OF(X509_OBJECT). This is mainly because an
15001 LHASH can't store or retrieve multiple objects with
15002 the same hash value.
15004 As a result various functions (which were all internal
15005 use only) have changed to handle the new X509_STORE
15006 structure. This will break anything that messed round
15007 with X509_STORE internally.
15009 The functions X509_STORE_add_cert() now checks for an
15010 exact match, rather than just subject name.
15012 The X509_STORE API doesn't directly support the retrieval
15013 of multiple certificates matching a given criteria, however
15014 this can be worked round by performing a lookup first
15015 (which will fill the cache with candidate certificates)
15016 and then examining the cache for matches. This is probably
15017 the best we can do without throwing out X509_LOOKUP
15018 entirely (maybe later...).
15020 The X509_VERIFY_CTX structure has been enhanced considerably.
15022 All certificate lookup operations now go via a get_issuer()
15023 callback. Although this currently uses an X509_STORE it
15024 can be replaced by custom lookups. This is a simple way
15025 to bypass the X509_STORE hackery necessary to make this
15026 work and makes it possible to use more efficient techniques
15027 in future. A very simple version which uses a simple
15028 STACK for its trusted certificate store is also provided
15029 using X509_STORE_CTX_trusted_stack().
15031 The verify_cb() and verify() callbacks now have equivalents
15032 in the X509_STORE_CTX structure.
15034 X509_STORE_CTX also has a 'flags' field which can be used
15035 to customise the verify behaviour.
15039 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15040 excludes S/MIME capabilities.
15044 * When a certificate request is read in keep a copy of the
15045 original encoding of the signed data and use it when outputting
15046 again. Signatures then use the original encoding rather than
15047 a decoded, encoded version which may cause problems if the
15048 request is improperly encoded.
15052 * For consistency with other BIO_puts implementations, call
15053 buffer_write(b, ...) directly in buffer_puts instead of calling
15056 In BIO_puts, increment b->num_write as in BIO_write.
15058 *Peter.Sylvester@EdelWeb.fr*
15060 * Fix BN_mul_word for the case where the word is 0. (We have to use
15061 BN_zero, we may not return a BIGNUM with an array consisting of
15062 words set to zero.)
15066 * Avoid calling abort() from within the library when problems are
15067 detected, except if preprocessor symbols have been defined
15068 (such as REF_CHECK, BN_DEBUG etc.).
15072 * New openssl application 'rsautl'. This utility can be
15073 used for low-level RSA operations. DER public key
15074 BIO/fp routines also added.
15078 * New Configure entry and patches for compiling on QNX 4.
15080 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15082 * A demo state-machine implementation was sponsored by
15083 Nuron (<http://www.nuron.com/>) and is now available in
15084 demos/state_machine.
15088 * New options added to the 'dgst' utility for signature
15089 generation and verification.
15093 * Unrecognized PKCS#7 content types are now handled via a
15094 catch all ASN1_TYPE structure. This allows unsupported
15095 types to be stored as a "blob" and an application can
15096 encode and decode it manually.
15100 * Fix various signed/unsigned issues to make a_strex.c
15101 compile under VC++.
15103 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15105 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15106 length if passed a buffer. ASN1_INTEGER_to_BN failed
15107 if passed a NULL BN and its argument was negative.
15109 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15111 * Modification to PKCS#7 encoding routines to output definite
15112 length encoding. Since currently the whole structures are in
15113 memory there's not real point in using indefinite length
15114 constructed encoding. However if OpenSSL is compiled with
15115 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15119 * Added BIO_vprintf() and BIO_vsnprintf().
15123 * Added more prefixes to parse for in the strings written
15124 through a logging bio, to cover all the levels that are available
15125 through syslog. The prefixes are now:
15127 PANIC, EMERG, EMR => LOG_EMERG
15128 ALERT, ALR => LOG_ALERT
15129 CRIT, CRI => LOG_CRIT
15130 ERROR, ERR => LOG_ERR
15131 WARNING, WARN, WAR => LOG_WARNING
15132 NOTICE, NOTE, NOT => LOG_NOTICE
15133 INFO, INF => LOG_INFO
15134 DEBUG, DBG => LOG_DEBUG
15136 and as before, if none of those prefixes are present at the
15137 beginning of the string, LOG_ERR is chosen.
15139 On Win32, the `LOG_*` levels are mapped according to this:
15141 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15142 LOG_WARNING => EVENTLOG_WARNING_TYPE
15143 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15147 * Made it possible to reconfigure with just the configuration
15148 argument "reconf" or "reconfigure". The command line arguments
15149 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15150 and are retrieved from there when reconfiguring.
15156 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15158 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15162 * The obj_dat.pl script was messing up the sorting of object
15163 names. The reason was that it compared the quoted version
15164 of strings as a result "OCSP" > "OCSP Signing" because
15165 " > SPACE. Changed script to store unquoted versions of
15166 names and add quotes on output. It was also omitting some
15167 names from the lookup table if they were given a default
15168 value (that is if SN is missing it is given the same
15169 value as LN and vice versa), these are now added on the
15170 grounds that if an object has a name we should be able to
15171 look it up. Finally added warning output when duplicate
15172 short or long names are found.
15176 * Changes needed for Tandem NSK.
15178 *Scott Uroff <scott@xypro.com>*
15180 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15181 RSA_padding_check_SSLv23(), special padding was never detected
15182 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15183 version rollback attacks was not effective.
15185 In s23_clnt.c, don't use special rollback-attack detection padding
15186 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15187 client; similarly, in s23_srvr.c, don't do the rollback check if
15188 SSL 2.0 is the only protocol enabled in the server.
15192 * Make it possible to get hexdumps of unprintable data with 'openssl
15193 asn1parse'. By implication, the functions ASN1_parse_dump() and
15194 BIO_dump_indent() are added.
15198 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15199 these print out strings and name structures based on various
15200 flags including RFC2253 support and proper handling of
15201 multibyte characters. Added options to the 'x509' utility
15202 to allow the various flags to be set.
15206 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15207 Also change the functions X509_cmp_current_time() and
15208 X509_gmtime_adj() work with an ASN1_TIME structure,
15209 this will enable certificates using GeneralizedTime in validity
15210 dates to be checked.
15214 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15215 negative public key encodings) on by default,
15216 NO_NEG_PUBKEY_BUG can be set to disable it.
15220 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15221 content octets. An i2c_ASN1_OBJECT is unnecessary because
15222 the encoding can be trivially obtained from the structure.
15226 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15227 not read locks (`CRYPTO_r_[un]lock`).
15231 * A first attempt at creating official support for shared
15232 libraries through configuration. I've kept it so the
15233 default is static libraries only, and the OpenSSL programs
15234 are always statically linked for now, but there are
15235 preparations for dynamic linking in place.
15236 This has been tested on Linux and Tru64.
15240 * Randomness polling function for Win9x, as described in:
15241 Peter Gutmann, Software Generation of Practically Strong
15246 * Fix so PRNG is seeded in req if using an already existing
15251 * New options to smime application. -inform and -outform
15252 allow alternative formats for the S/MIME message including
15253 PEM and DER. The -content option allows the content to be
15254 specified separately. This should allow things like Netscape
15255 form signing output easier to verify.
15259 * Fix the ASN1 encoding of tags using the 'long form'.
15263 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15264 STRING types. These convert content octets to and from the
15265 underlying type. The actual tag and length octets are
15266 already assumed to have been read in and checked. These
15267 are needed because all other string types have virtually
15268 identical handling apart from the tag. By having versions
15269 of the ASN1 functions that just operate on content octets
15270 IMPLICIT tagging can be handled properly. It also allows
15271 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15272 and ASN1_INTEGER are identical apart from the tag.
15276 * Change the handling of OID objects as follows:
15278 - New object identifiers are inserted in objects.txt, following
15279 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15280 - objects.pl is used to process obj_mac.num and create a new
15282 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15285 This is currently kind of a hack, and the perl code in objects.pl
15286 isn't very elegant, but it works as I intended. The simplest way
15287 to check that it worked correctly is to look in obj_dat.h and
15288 check the array nid_objs and make sure the objects haven't moved
15289 around (this is important!). Additions are OK, as well as
15290 consistent name changes.
15294 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15298 * Addition of the command line parameter '-rand file' to 'openssl req'.
15299 The given file adds to whatever has already been seeded into the
15300 random pool through the RANDFILE configuration file option or
15301 environment variable, or the default random state file.
15305 * mkstack.pl now sorts each macro group into lexical order.
15306 Previously the output order depended on the order the files
15307 appeared in the directory, resulting in needless rewriting
15312 * Patches to make OpenSSL compile under Win32 again. Mostly
15313 work arounds for the VC++ problem that it treats func() as
15314 func(void). Also stripped out the parts of mkdef.pl that
15315 added extra typesafe functions: these no longer exist.
15319 * Reorganisation of the stack code. The macros are now all
15320 collected in safestack.h . Each macro is defined in terms of
15321 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15322 DEBUG_SAFESTACK is now handled in terms of function casts,
15323 this has the advantage of retaining type safety without the
15324 use of additional functions. If DEBUG_SAFESTACK is not defined
15325 then the non typesafe macros are used instead. Also modified the
15326 mkstack.pl script to handle the new form. Needs testing to see
15327 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15328 the default if no major problems. Similar behaviour for ASN1_SET_OF
15329 and PKCS12_STACK_OF.
15333 * When some versions of IIS use the 'NET' form of private key the
15334 key derivation algorithm is different. Normally MD5(password) is
15335 used as a 128 bit RC4 key. In the modified case
15336 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15337 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15338 as the old Netscape_RSA functions except they have an additional
15339 'sgckey' parameter which uses the modified algorithm. Also added
15340 an -sgckey command line option to the rsa utility. Thanks to
15341 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15342 algorithm to openssl-dev.
15346 * The evp_local.h macros were using 'c.##kname' which resulted in
15347 invalid expansion on some systems (SCO 5.0.5 for example).
15348 Corrected to 'c.kname'.
15350 *Phillip Porch <root@theporch.com>*
15352 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15353 a STACK of email addresses from a certificate or request, these look
15354 in the subject name and the subject alternative name extensions and
15355 omit any duplicate addresses.
15359 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15360 This makes DSA verification about 2 % faster.
15364 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15365 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15366 plus overhead for 1024 bit moduli).
15367 This makes exponentiations about 0.5 % faster for 1024 bit
15368 exponents (as measured by "openssl speed rsa2048").
15372 * Rename memory handling macros to avoid conflicts with other
15374 Malloc => OPENSSL_malloc
15375 Malloc_locked => OPENSSL_malloc_locked
15376 Realloc => OPENSSL_realloc
15377 Free => OPENSSL_free
15381 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15382 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15386 * CygWin32 support.
15388 *John Jarvie <jjarvie@newsguy.com>*
15390 * The type-safe stack code has been rejigged. It is now only compiled
15391 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15392 by default all type-specific stack functions are "#define"d back to
15393 standard stack functions. This results in more streamlined output
15394 but retains the type-safety checking possibilities of the original
15399 * The STACK code has been cleaned up, and certain type declarations
15400 that didn't make a lot of sense have been brought in line. This has
15401 also involved a cleanup of sorts in safestack.h to more correctly
15402 map type-safe stack functions onto their plain stack counterparts.
15403 This work has also resulted in a variety of "const"ifications of
15404 lots of the code, especially `_cmp` operations which should normally
15405 be prototyped with "const" parameters anyway.
15409 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15410 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15411 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15412 where all of 'md' is used each time the PRNG is used, but 'state'
15413 is used only indexed by a cyclic counter. As entropy may not be
15414 well distributed from the beginning, 'md' is important as a
15415 chaining variable. However, the output function chains only half
15416 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15417 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15418 in all of 'state' being rewritten, with the new values depending
15419 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15423 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15424 the handshake is continued after ssl_verify_cert_chain();
15425 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15426 can lead to 'unexplainable' connection aborts later.
15428 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15430 * Major EVP API cipher revision.
15431 Add hooks for extra EVP features. This allows various cipher
15432 parameters to be set in the EVP interface. Support added for variable
15433 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15434 setting of RC2 and RC5 parameters.
15436 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15439 Remove lots of duplicated code from the EVP library. For example *every*
15440 cipher init() function handles the 'iv' in the same way according to the
15441 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15442 for CFB and OFB modes they zero ctx->num.
15444 New functionality allows removal of S/MIME code RC2 hack.
15446 Most of the routines have the same form and so can be declared in terms
15449 By shifting this to the top level EVP_CipherInit() it can be removed from
15450 all individual ciphers. If the cipher wants to handle IVs or keys
15451 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15454 Change lots of functions like EVP_EncryptUpdate() to now return a
15455 value: although software versions of the algorithms cannot fail
15456 any installed hardware versions can.
15460 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15461 this option is set, tolerate broken clients that send the negotiated
15462 protocol version number instead of the requested protocol version
15467 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15468 i.e. non-zero for export ciphersuites, zero otherwise.
15469 Previous versions had this flag inverted, inconsistent with
15470 rsa_tmp_cb (..._TMP_RSA_CB).
15472 *Bodo Moeller; problem reported by Amit Chopra*
15474 * Add missing DSA library text string. Work around for some IIS
15475 key files with invalid SEQUENCE encoding.
15479 * Add a document (doc/standards.txt) that list all kinds of standards
15480 and so on that are implemented in OpenSSL.
15484 * Enhance c_rehash script. Old version would mishandle certificates
15485 with the same subject name hash and wouldn't handle CRLs at all.
15486 Added -fingerprint option to crl utility, to support new c_rehash
15491 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15495 * Fix for SSL server purpose checking. Server checking was
15496 rejecting certificates which had extended key usage present
15497 but no ssl client purpose.
15499 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15501 * Make PKCS#12 code work with no password. The PKCS#12 spec
15502 is a little unclear about how a blank password is handled.
15503 Since the password in encoded as a BMPString with terminating
15504 double NULL a zero length password would end up as just the
15505 double NULL. However no password at all is different and is
15506 handled differently in the PKCS#12 key generation code. NS
15507 treats a blank password as zero length. MSIE treats it as no
15508 password on export: but it will try both on import. We now do
15509 the same: PKCS12_parse() tries zero length and no password if
15510 the password is set to "" or NULL (NULL is now a valid password:
15511 it wasn't before) as does the pkcs12 application.
15515 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15516 perror when PEM_read_bio_X509_REQ fails, the error message must
15517 be obtained from the error queue.
15521 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15522 it in ERR_remove_state if appropriate, and change ERR_get_state
15523 accordingly to avoid race conditions (this is necessary because
15524 thread_hash is no longer constant once set).
15528 * Bugfix for linux-elf makefile.one.
15532 * RSA_get_default_method() will now cause a default
15533 RSA_METHOD to be chosen if one doesn't exist already.
15534 Previously this was only set during a call to RSA_new()
15535 or RSA_new_method(NULL) meaning it was possible for
15536 RSA_get_default_method() to return NULL.
15540 * Added native name translation to the existing DSO code
15541 that will convert (if the flag to do so is set) filenames
15542 that are sufficiently small and have no path information
15543 into a canonical native form. Eg. "blah" converted to
15544 "libblah.so" or "blah.dll" etc.
15548 * New function ERR_error_string_n(e, buf, len) which is like
15549 ERR_error_string(e, buf), but writes at most 'len' bytes
15550 including the 0 terminator. For ERR_error_string_n, 'buf'
15553 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15555 * CONF library reworked to become more general. A new CONF
15556 configuration file reader "class" is implemented as well as a
15557 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15558 old `CONF_*` functions are still there, but are reimplemented to
15559 work in terms of the new functions. Also, a set of functions
15560 to handle the internal storage of the configuration data is
15561 provided to make it easier to write new configuration file
15562 reader "classes" (I can definitely see something reading a
15563 configuration file in XML format, for example), called `_CONF_*`,
15564 or "the configuration storage API"...
15566 The new configuration file reading functions are:
15568 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15569 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15571 NCONF_default, NCONF_WIN32
15573 NCONF_dump_fp, NCONF_dump_bio
15575 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15576 NCONF_new creates a new CONF object. This works in the same way
15577 as other interfaces in OpenSSL, like the BIO interface.
15578 `NCONF_dump_*` dump the internal storage of the configuration file,
15579 which is useful for debugging. All other functions take the same
15580 arguments as the old `CONF_*` functions with the exception of the
15581 first that must be a `CONF *` instead of a `LHASH *`.
15583 To make it easier to use the new classes with the old `CONF_*` functions,
15584 the function CONF_set_default_method is provided.
15588 * Add '-tls1' option to 'openssl ciphers', which was already
15589 mentioned in the documentation but had not been implemented.
15590 (This option is not yet really useful because even the additional
15591 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15595 * Initial DSO code added into libcrypto for letting OpenSSL (and
15596 OpenSSL-based applications) load shared libraries and bind to
15597 them in a portable way.
15599 *Geoff Thorpe, with contributions from Richard Levitte*
15601 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15603 * Make sure _lrotl and _lrotr are only used with MSVC.
15605 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15606 (the default implementation of RAND_status).
15608 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15609 to '-clrext' (= clear extensions), as intended and documented.
15610 *Bodo Moeller; inconsistency pointed out by Michael Attili
15611 <attili@amaxo.com>*
15613 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15614 was larger than the MD block size.
15616 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15618 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15619 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15620 using the passed key: if the passed key was a private key the result
15621 of X509_print(), for example, would be to print out all the private key
15626 * des_quad_cksum() byte order bug fix.
15627 *Ulf Möller, using the problem description in krb4-0.9.7, where
15628 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15630 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15633 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15635 * For easily testing in shell scripts whether some command
15636 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15637 returns with exit code 0 iff no command of the given name is available.
15638 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15639 the output goes to stdout and nothing is printed to stderr.
15640 Additional arguments are always ignored.
15642 Since for each cipher there is a command of the same name,
15643 the 'no-cipher' compilation switches can be tested this way.
15645 ('openssl no-XXX' is not able to detect pseudo-commands such
15646 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15650 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15654 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15655 is set; it will be thrown away anyway because each handshake creates
15657 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15658 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15659 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15660 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15664 * New s_client option -ign_eof: EOF at stdin is ignored, and
15665 'Q' and 'R' lose their special meanings (quit/renegotiate).
15666 This is part of what -quiet does; unlike -quiet, -ign_eof
15667 does not suppress any output.
15671 * Add compatibility options to the purpose and trust code. The
15672 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15673 accepts a certificate or CA, this was the previous behaviour,
15674 with all the associated security issues.
15676 X509_TRUST_COMPAT is the old trust behaviour: only and
15677 automatically trust self signed roots in certificate store. A
15678 new trust setting X509_TRUST_DEFAULT is used to specify that
15679 a purpose has no associated trust setting and it should instead
15680 use the value in the default purpose.
15684 * Fix the PKCS#8 DSA private key code so it decodes keys again
15685 and fix a memory leak.
15689 * In util/mkerr.pl (which implements 'make errors'), preserve
15690 reason strings from the previous version of the .c file, as
15691 the default to have only downcase letters (and digits) in
15692 automatically generated reasons codes is not always appropriate.
15696 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15697 using strerror. Previously, ERR_reason_error_string() returned
15698 library names as reason strings for SYSerr; but SYSerr is a special
15699 case where small numbers are errno values, not library numbers.
15703 * Add '-dsaparam' option to 'openssl dhparam' application. This
15704 converts DSA parameters into DH parameters. (When creating parameters,
15705 DSA_generate_parameters is used.)
15709 * Include 'length' (recommended exponent length) in C code generated
15710 by 'openssl dhparam -C'.
15714 * The second argument to set_label in perlasm was already being used
15715 so couldn't be used as a "file scope" flag. Moved to third argument
15720 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15721 instead of RAND_bytes for encryption IVs and salts.
15725 * Include RAND_status() into RAND_METHOD instead of implementing
15726 it only for md_rand.c Otherwise replacing the PRNG by calling
15727 RAND_set_rand_method would be impossible.
15731 * Don't let DSA_generate_key() enter an infinite loop if the random
15732 number generation fails.
15736 * New 'rand' application for creating pseudo-random output.
15740 * Added configuration support for Linux/IA64
15742 *Rolf Haberrecker <rolf@suse.de>*
15744 * Assembler module support for Mingw32.
15748 * Shared library support for HPUX (in shlib/).
15750 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15752 * Shared library support for Solaris gcc.
15754 *Lutz Behnke <behnke@trustcenter.de>*
15756 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15758 * PKCS7_encrypt() was adding text MIME headers twice because they
15759 were added manually and by SMIME_crlf_copy().
15763 * In bntest.c don't call BN_rand with zero bits argument.
15765 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15767 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15768 case was implemented. This caused BN_div_recp() to fail occasionally.
15772 * Add an optional second argument to the set_label() in the perl
15773 assembly language builder. If this argument exists and is set
15774 to 1 it signals that the assembler should use a symbol whose
15775 scope is the entire file, not just the current function. This
15776 is needed with MASM which uses the format label:: for this scope.
15778 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15780 * Change the ASN1 types so they are typedefs by default. Before
15781 almost all types were #define'd to ASN1_STRING which was causing
15782 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15787 * Change names of new functions to the new get1/get0 naming
15788 convention: After 'get1', the caller owns a reference count
15789 and has to call `..._free`; 'get0' returns a pointer to some
15790 data structure without incrementing reference counters.
15791 (Some of the existing 'get' functions increment a reference
15792 counter, some don't.)
15793 Similarly, 'set1' and 'add1' functions increase reference
15794 counters or duplicate objects.
15798 * Allow for the possibility of temp RSA key generation failure:
15799 the code used to assume it always worked and crashed on failure.
15803 * Fix potential buffer overrun problem in BIO_printf().
15804 *Ulf Möller, using public domain code by Patrick Powell; problem
15805 pointed out by David Sacerdote <das33@cornell.edu>*
15807 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15808 RAND_egd() and RAND_status(). In the command line application,
15809 the EGD socket can be specified like a seed file using RANDFILE
15814 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15815 Some CAs (e.g. Verisign) distribute certificates in this form.
15819 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15820 list to exclude them. This means that no special compilation option
15821 is needed to use anonymous DH: it just needs to be included in the
15826 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15827 EVP_MD_type. The old functionality is available in a new macro called
15828 EVP_MD_md(). Change code that uses it and update docs.
15832 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15833 where the `void *` argument is replaced by a function pointer argument.
15834 Previously `void *` was abused to point to functions, which works on
15835 many platforms, but is not correct. As these functions are usually
15836 called by macros defined in OpenSSL header files, most source code
15837 should work without changes.
15841 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15842 sections with information on -D... compiler switches used for
15843 compiling the library so that applications can see them. To enable
15844 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15845 must be defined. E.g.,
15846 #define OPENSSL_ALGORITHM_DEFINES
15847 #include <openssl/opensslconf.h>
15848 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15850 *Richard Levitte, Ulf and Bodo Möller*
15852 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15857 * Change the 'other' type in certificate aux info to a STACK_OF
15858 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15859 the required ASN1 format: arbitrary types determined by an OID.
15863 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15864 argument to 'req'. This is not because the function is newer or
15865 better than others it just uses the work 'NEW' in the certificate
15866 request header lines. Some software needs this.
15870 * Reorganise password command line arguments: now passwords can be
15871 obtained from various sources. Delete the PEM_cb function and make
15872 it the default behaviour: i.e. if the callback is NULL and the
15873 usrdata argument is not NULL interpret it as a null terminated pass
15874 phrase. If usrdata and the callback are NULL then the pass phrase
15875 is prompted for as usual.
15879 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15880 the support is automatically enabled. The resulting binaries will
15881 autodetect the card and use it if present.
15883 *Ben Laurie and Compaq Inc.*
15885 * Work around for Netscape hang bug. This sends certificate request
15886 and server done in one record. Since this is perfectly legal in the
15887 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15888 the bugs/SSLv3 entry for more info.
15892 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15896 * Add -rand argument to smime and pkcs12 applications and read/write
15901 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15905 * Add command line password options to the remaining applications.
15909 * Bug fix for BN_div_recp() for numerators with an even number of
15914 * More tests in bntest.c, and changed test_bn output.
15918 * ./config recognizes MacOS X now.
15922 * Bug fix for BN_div() when the first words of num and divisor are
15923 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15927 * Add support for various broken PKCS#8 formats, and command line
15928 options to produce them.
15932 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15933 get temporary BIGNUMs from a BN_CTX.
15937 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15942 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15943 include a #define from the old name to the new. The original intent
15944 was that statically linked binaries could for example just call
15945 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15946 link with digests. This never worked because SSLeay_add_all_digests()
15947 and SSLeay_add_all_ciphers() were in the same source file so calling
15948 one would link with the other. They are now in separate source files.
15952 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15956 * Use a less unusual form of the Miller-Rabin primality test (it used
15957 a binary algorithm for exponentiation integrated into the Miller-Rabin
15958 loop, our standard modexp algorithms are faster).
15962 * Support for the EBCDIC character set completed.
15964 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15966 * Source code cleanups: use const where appropriate, eliminate casts,
15967 use `void *` instead of `char *` in lhash.
15971 * Bugfix: ssl3_send_server_key_exchange was not restartable
15972 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15973 this the server could overwrite ephemeral keys that the client
15978 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15979 using 50 iterations of the Rabin-Miller test.
15981 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15982 iterations of the Rabin-Miller test as required by the appendix
15983 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15984 As BN_is_prime_fasttest includes trial division, DSA parameter
15985 generation becomes much faster.
15987 This implies a change for the callback functions in DSA_is_prime
15988 and DSA_generate_parameters: The callback function is called once
15989 for each positive witness in the Rabin-Miller test, not just
15990 occasionally in the inner loop; and the parameters to the
15991 callback function now provide an iteration count for the outer
15992 loop rather than for the current invocation of the inner loop.
15993 DSA_generate_parameters additionally can call the callback
15994 function with an 'iteration count' of -1, meaning that a
15995 candidate has passed the trial division test (when q is generated
15996 from an application-provided seed, trial division is skipped).
16000 * New function BN_is_prime_fasttest that optionally does trial
16001 division before starting the Rabin-Miller test and has
16002 an additional BN_CTX * argument (whereas BN_is_prime always
16003 has to allocate at least one BN_CTX).
16004 'callback(1, -1, cb_arg)' is called when a number has passed the
16005 trial division stage.
16009 * Fix for bug in CRL encoding. The validity dates weren't being handled
16014 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16018 * New function BN_pseudo_rand().
16022 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16023 bignum version of BN_from_montgomery() with the working code from
16024 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16029 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16030 made it impossible to use the same SSL_SESSION data structure in
16031 SSL2 clients in multiple threads.
16035 * The return value of RAND_load_file() no longer counts bytes obtained
16036 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16037 to seed the PRNG (previously an explicit byte count was required).
16039 *Ulf Möller, Bodo Möller*
16041 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16042 used `char *` instead of `void *` and had casts all over the place.
16046 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16050 * Retain source code compatibility for BN_prime_checks macro:
16051 BN_is_prime(..., BN_prime_checks, ...) now uses
16052 BN_prime_checks_for_size to determine the appropriate number of
16053 Rabin-Miller iterations.
16057 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16058 DH_CHECK_P_NOT_SAFE_PRIME.
16059 (Check if this is true? OpenPGP calls them "strong".)
16063 * Merge the functionality of "dh" and "gendh" programs into a new program
16064 "dhparam". The old programs are retained for now but will handle DH keys
16065 (instead of parameters) in future.
16069 * Make the ciphers, s_server and s_client programs check the return values
16070 when a new cipher list is set.
16074 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16075 ciphers. Before when the 56bit ciphers were enabled the sorting was
16078 The syntax for the cipher sorting has been extended to support sorting by
16079 cipher-strength (using the strength_bits hard coded in the tables).
16080 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16082 Fix a bug in the cipher-command parser: when supplying a cipher command
16083 string with an "undefined" symbol (neither command nor alphanumeric
16084 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16085 an error is flagged.
16087 Due to the strength-sorting extension, the code of the
16088 ssl_create_cipher_list() function was completely rearranged. I hope that
16089 the readability was also increased :-)
16091 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16093 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16094 for the first serial number and places 2 in the serial number file. This
16095 avoids problems when the root CA is created with serial number zero and
16096 the first user certificate has the same issuer name and serial number
16101 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16102 the new code. Add documentation for this stuff.
16106 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16107 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16108 structures and behave in an analogous way to the X509v3 functions:
16109 they shouldn't be called directly but wrapper functions should be used
16112 So we also now have some wrapper functions that call the X509at functions
16113 when passed certificate requests. (TO DO: similar things can be done with
16114 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16115 things. Some of these need some d2i or i2d and print functionality
16116 because they handle more complex structures.)
16120 * Add missing #ifndefs that caused missing symbols when building libssl
16121 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16122 NO_RSA in `ssl/s2*.c`.
16124 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16126 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16127 has a return value which indicates the quality of the random data
16128 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16129 error queue. New function RAND_pseudo_bytes() generates output that is
16130 guaranteed to be unique but not unpredictable. RAND_add is like
16131 RAND_seed, but takes an extra argument for an entropy estimate
16132 (RAND_seed always assumes full entropy).
16136 * Do more iterations of Rabin-Miller probable prime test (specifically,
16137 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16138 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16139 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16140 false-positive rate of at most 2^-80 for random input.
16144 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16148 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16149 in the 0.9.5 release), this returns the chain
16150 from an X509_CTX structure with a dup of the stack and all
16151 the X509 reference counts upped: so the stack will exist
16152 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16155 Also make SSL_SESSION_print() print out the verify return
16160 * Add manpage for the pkcs12 command. Also change the default
16161 behaviour so MAC iteration counts are used unless the new
16162 -nomaciter option is used. This improves file security and
16163 only older versions of MSIE (4.0 for example) need it.
16167 * Honor the no-xxx Configure options when creating .DEF files.
16171 * Add PKCS#10 attributes to field table: challengePassword,
16172 unstructuredName and unstructuredAddress. These are taken from
16173 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16174 international characters are used.
16176 More changes to X509_ATTRIBUTE code: allow the setting of types
16177 based on strings. Remove the 'loc' parameter when adding
16178 attributes because these will be a SET OF encoding which is sorted
16183 * Initial changes to the 'req' utility to allow request generation
16184 automation. This will allow an application to just generate a template
16185 file containing all the field values and have req construct the
16188 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16189 used all over the place including certificate requests and PKCS#7
16190 structures. They are currently handled manually where necessary with
16191 some primitive wrappers for PKCS#7. The new functions behave in a
16192 manner analogous to the X509 extension functions: they allow
16193 attributes to be looked up by NID and added.
16195 Later something similar to the X509V3 code would be desirable to
16196 automatically handle the encoding, decoding and printing of the
16197 more complex types. The string types like challengePassword can
16198 be handled by the string table functions.
16200 Also modified the multi byte string table handling. Now there is
16201 a 'global mask' which masks out certain types. The table itself
16202 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16203 is useful when for example there is only one permissible type
16204 (as in countryName) and using the mask might result in no valid
16209 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16210 SSL_get_peer_finished to allow applications to obtain the latest
16211 Finished messages sent to the peer or expected from the peer,
16212 respectively. (SSL_get_peer_finished is usually the Finished message
16213 actually received from the peer, otherwise the protocol will be aborted.)
16215 As the Finished message are message digests of the complete handshake
16216 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16217 be used for external authentication procedures when the authentication
16218 provided by SSL/TLS is not desired or is not enough.
16222 * Enhanced support for Alpha Linux is added. Now ./config checks if
16223 the host supports BWX extension and if Compaq C is present on the
16224 $PATH. Just exploiting of the BWX extension results in 20-30%
16225 performance kick for some algorithms, e.g. DES and RC4 to mention
16226 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16231 * Add support for MS "fast SGC". This is arguably a violation of the
16232 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16233 weak crypto and after checking the certificate is SGC a second one
16234 with strong crypto. MS SGC stops the first handshake after receiving
16235 the server certificate message and sends a second client hello. Since
16236 a server will typically do all the time consuming operations before
16237 expecting any further messages from the client (server key exchange
16238 is the most expensive) there is little difference between the two.
16240 To get OpenSSL to support MS SGC we have to permit a second client
16241 hello message after we have sent server done. In addition we have to
16242 reset the MAC if we do get this second client hello.
16246 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16247 if a DER encoded private key is RSA or DSA traditional format. Changed
16248 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16249 format DER encoded private key. Newer code should use PKCS#8 format which
16250 has the key type encoded in the ASN1 structure. Added DER private key
16251 support to pkcs8 application.
16255 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16256 ciphersuites has been selected (as required by the SSL 3/TLS 1
16257 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16258 is set, we interpret this as a request to violate the specification
16259 (the worst that can happen is a handshake failure, and 'correct'
16260 behaviour would result in a handshake failure anyway).
16264 * In SSL_CTX_add_session, take into account that there might be multiple
16265 SSL_SESSION structures with the same session ID (e.g. when two threads
16266 concurrently obtain them from an external cache).
16267 The internal cache can handle only one SSL_SESSION with a given ID,
16268 so if there's a conflict, we now throw out the old one to achieve
16273 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16274 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16275 some routines that use cipher OIDs: some ciphers do not have OIDs
16276 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16281 * Simplify the trust setting structure and code. Now we just have
16282 two sequences of OIDs for trusted and rejected settings. These will
16283 typically have values the same as the extended key usage extension
16284 and any application specific purposes.
16286 The trust checking code now has a default behaviour: it will just
16287 check for an object with the same NID as the passed id. Functions can
16288 be provided to override either the default behaviour or the behaviour
16289 for a given id. SSL client, server and email already have functions
16290 in place for compatibility: they check the NID and also return "trusted"
16291 if the certificate is self signed.
16295 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16296 traditional format into an EVP_PKEY structure.
16300 * Add a password callback function PEM_cb() which either prompts for
16301 a password if usr_data is NULL or otherwise assumes it is a null
16302 terminated password. Allow passwords to be passed on command line
16303 environment or config files in a few more utilities.
16307 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16308 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16309 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16310 Update documentation.
16314 * Support for ASN1 "NULL" type. This could be handled before by using
16315 ASN1_TYPE but there wasn't any function that would try to read a NULL
16316 and produce an error if it couldn't. For compatibility we also have
16317 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16318 don't allocate anything because they don't need to.
16322 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16325 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16327 * Rebuild of the memory allocation routines used by OpenSSL code and
16328 possibly others as well. The purpose is to make an interface that
16329 provide hooks so anyone can build a separate set of allocation and
16330 deallocation routines to be used by OpenSSL, for example memory
16331 pool implementations, or something else, which was previously hard
16332 since Malloc(), Realloc() and Free() were defined as macros having
16333 the values malloc, realloc and free, respectively (except for Win32
16334 compilations). The same is provided for memory debugging code.
16335 OpenSSL already comes with functionality to find memory leaks, but
16336 this gives people a chance to debug other memory problems.
16338 With these changes, a new set of functions and macros have appeared:
16340 CRYPTO_set_mem_debug_functions() [F]
16341 CRYPTO_get_mem_debug_functions() [F]
16342 CRYPTO_dbg_set_options() [F]
16343 CRYPTO_dbg_get_options() [F]
16344 CRYPTO_malloc_debug_init() [M]
16346 The memory debug functions are NULL by default, unless the library
16347 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16348 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16349 gives the standard debugging functions that come with OpenSSL) or
16350 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16351 provided by the library user) must be used. When the standard
16352 debugging functions are used, CRYPTO_dbg_set_options can be used to
16353 request additional information:
16354 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16355 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16357 Also, things like CRYPTO_set_mem_functions will always give the
16358 expected result (the new set of functions is used for allocation
16359 and deallocation) at all times, regardless of platform and compiler
16362 To finish it up, some functions that were never use in any other
16363 way than through macros have a new API and new semantic:
16365 CRYPTO_dbg_malloc()
16366 CRYPTO_dbg_realloc()
16369 All macros of value have retained their old syntax.
16371 *Richard Levitte and Bodo Moeller*
16373 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16374 ordering of SMIMECapabilities wasn't in "strength order" and there
16375 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16380 * Some ASN1 types with illegal zero length encoding (INTEGER,
16381 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16383 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16385 * Merge in my S/MIME library for OpenSSL. This provides a simple
16386 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16387 functionality to handle multipart/signed properly) and a utility
16388 called 'smime' to call all this stuff. This is based on code I
16389 originally wrote for Celo who have kindly allowed it to be
16390 included in OpenSSL.
16394 * Add variants des_set_key_checked and des_set_key_unchecked of
16395 des_set_key (aka des_key_sched). Global variable des_check_key
16396 decides which of these is called by des_set_key; this way
16397 des_check_key behaves as it always did, but applications and
16398 the library itself, which was buggy for des_check_key == 1,
16399 have a cleaner way to pick the version they need.
16403 * New function PKCS12_newpass() which changes the password of a
16408 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16409 dynamic mix. In both cases the ids can be used as an index into the
16410 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16411 functions so they accept a list of the field values and the
16412 application doesn't need to directly manipulate the X509_TRUST
16417 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16422 * Modify the way the V3 extension code looks up extensions. This now
16423 works in a similar way to the object code: we have some "standard"
16424 extensions in a static table which is searched with OBJ_bsearch()
16425 and the application can add dynamic ones if needed. The file
16426 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16427 updated whenever a new extension is added to the core code and kept
16428 in ext_nid order. There is a simple program 'tabtest.c' which checks
16429 this. New extensions are not added too often so this file can readily
16430 be maintained manually.
16432 There are two big advantages in doing things this way. The extensions
16433 can be looked up immediately and no longer need to be "added" using
16434 X509V3_add_standard_extensions(): this function now does nothing.
16435 Side note: I get *lots* of email saying the extension code doesn't
16436 work because people forget to call this function.
16437 Also no dynamic allocation is done unless new extensions are added:
16438 so if we don't add custom extensions there is no need to call
16439 X509V3_EXT_cleanup().
16443 * Modify enc utility's salting as follows: make salting the default. Add a
16444 magic header, so unsalted files fail gracefully instead of just decrypting
16445 to garbage. This is because not salting is a big security hole, so people
16446 should be discouraged from doing it.
16450 * Fixes and enhancements to the 'x509' utility. It allowed a message
16451 digest to be passed on the command line but it only used this
16452 parameter when signing a certificate. Modified so all relevant
16453 operations are affected by the digest parameter including the
16454 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16455 DSA key was used because it didn't fix the digest.
16459 * Initial certificate chain verify code. Currently tests the untrusted
16460 certificates for consistency with the verify purpose (which is set
16461 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16463 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16464 this is because it will reject chains with invalid extensions whereas
16465 every previous version of OpenSSL and SSLeay made no checks at all.
16467 Trust code: checks the root CA for the relevant trust settings. Trust
16468 settings have an initial value consistent with the verify purpose: e.g.
16469 if the verify purpose is for SSL client use it expects the CA to be
16470 trusted for SSL client use. However the default value can be changed to
16471 permit custom trust settings: one example of this would be to only trust
16472 certificates from a specific "secure" set of CAs.
16474 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16475 which should be used for version portability: especially since the
16476 verify structure is likely to change more often now.
16478 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16479 to set them. If not set then assume SSL clients will verify SSL servers
16482 Two new options to the verify program: -untrusted allows a set of
16483 untrusted certificates to be passed in and -purpose which sets the
16484 intended purpose of the certificate. If a purpose is set then the
16485 new chain verify code is used to check extension consistency.
16489 * Support for the authority information access extension.
16493 * Modify RSA and DSA PEM read routines to transparently handle
16494 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16495 public keys in a format compatible with certificate
16496 SubjectPublicKeyInfo structures. Unfortunately there were already
16497 functions called *_PublicKey_* which used various odd formats so
16498 these are retained for compatibility: however the DSA variants were
16499 never in a public release so they have been deleted. Changed dsa/rsa
16500 utilities to handle the new format: note no releases ever handled public
16501 keys so we should be OK.
16503 The primary motivation for this change is to avoid the same fiasco
16504 that dogs private keys: there are several incompatible private key
16505 formats some of which are standard and some OpenSSL specific and
16506 require various evil hacks to allow partial transparent handling and
16507 even then it doesn't work with DER formats. Given the option anything
16508 other than PKCS#8 should be dumped: but the other formats have to
16509 stay in the name of compatibility.
16511 With public keys and the benefit of hindsight one standard format
16512 is used which works with EVP_PKEY, RSA or DSA structures: though
16513 it clearly returns an error if you try to read the wrong kind of key.
16515 Added a -pubkey option to the 'x509' utility to output the public key.
16516 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16517 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16518 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16519 that do the same as the `EVP_PKEY_assign_*()` except they up the
16520 reference count of the added key (they don't "swallow" the
16525 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16526 CRLs would fail if the file contained no certificates or no CRLs:
16527 added a new function to read in both types and return the number
16528 read: this means that if none are read it will be an error. The
16529 DER versions of the certificate and CRL reader would always fail
16530 because it isn't possible to mix certificates and CRLs in DER format
16531 without choking one or the other routine. Changed this to just read
16532 a certificate: this is the best we can do. Also modified the code
16533 in `apps/verify.c` to take notice of return codes: it was previously
16534 attempting to read in certificates from NULL pointers and ignoring
16535 any errors: this is one reason why the cert and CRL reader seemed
16536 to work. It doesn't check return codes from the default certificate
16537 routines: these may well fail if the certificates aren't installed.
16541 * Code to support otherName option in GeneralName.
16545 * First update to verify code. Change the verify utility
16546 so it warns if it is passed a self signed certificate:
16547 for consistency with the normal behaviour. X509_verify
16548 has been modified to it will now verify a self signed
16549 certificate if *exactly* the same certificate appears
16550 in the store: it was previously impossible to trust a
16551 single self signed certificate. This means that:
16552 openssl verify ss.pem
16553 now gives a warning about a self signed certificate but
16554 openssl verify -CAfile ss.pem ss.pem
16559 * For servers, store verify_result in SSL_SESSION data structure
16560 (and add it to external session representation).
16561 This is needed when client certificate verifications fails,
16562 but an application-provided verification callback (set by
16563 SSL_CTX_set_cert_verify_callback) allows accepting the session
16564 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16565 but returns 1): When the session is reused, we have to set
16566 ssl->verify_result to the appropriate error code to avoid
16569 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16571 * Fix a bug in the new PKCS#7 code: it didn't consider the
16572 case in PKCS7_dataInit() where the signed PKCS7 structure
16573 didn't contain any existing data because it was being created.
16575 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16577 * Add a salt to the key derivation routines in enc.c. This
16578 forms the first 8 bytes of the encrypted file. Also add a
16579 -S option to allow a salt to be input on the command line.
16583 * New function X509_cmp(). Oddly enough there wasn't a function
16584 to compare two certificates. We do this by working out the SHA1
16585 hash and comparing that. X509_cmp() will be needed by the trust
16590 * SSL_get1_session() is like SSL_get_session(), but increments
16591 the reference count in the SSL_SESSION returned.
16593 *Geoff Thorpe <geoff@eu.c2.net>*
16595 * Fix for 'req': it was adding a null to request attributes.
16596 Also change the X509_LOOKUP and X509_INFO code to handle
16597 certificate auxiliary information.
16601 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16606 * Add the possibility to add extra information to the memory leak
16607 detecting output, to form tracebacks, showing from where each
16608 allocation was originated: CRYPTO_push_info("constant string") adds
16609 the string plus current file name and line number to a per-thread
16610 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16611 is like calling CYRPTO_pop_info() until the stack is empty.
16612 Also updated memory leak detection code to be multi-thread-safe.
16616 * Add options -text and -noout to pkcs7 utility and delete the
16617 encryption options which never did anything. Update docs.
16621 * Add options to some of the utilities to allow the pass phrase
16622 to be included on either the command line (not recommended on
16623 OSes like Unix) or read from the environment. Update the
16624 manpages and fix a few bugs.
16628 * Add a few manpages for some of the openssl commands.
16632 * Fix the -revoke option in ca. It was freeing up memory twice,
16633 leaking and not finding already revoked certificates.
16637 * Extensive changes to support certificate auxiliary information.
16638 This involves the use of X509_CERT_AUX structure and X509_AUX
16639 functions. An X509_AUX function such as PEM_read_X509_AUX()
16640 can still read in a certificate file in the usual way but it
16641 will also read in any additional "auxiliary information". By
16642 doing things this way a fair degree of compatibility can be
16643 retained: existing certificates can have this information added
16644 using the new 'x509' options.
16646 Current auxiliary information includes an "alias" and some trust
16647 settings. The trust settings will ultimately be used in enhanced
16648 certificate chain verification routines: currently a certificate
16649 can only be trusted if it is self signed and then it is trusted
16654 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16655 The problem was that one of the replacement routines had not been working
16656 since SSLeay releases. For now the offending routine has been replaced
16657 with non-optimised assembler. Even so, this now gives around 95%
16658 performance improvement for 1024 bit RSA signs.
16662 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16663 handling. Most clients have the effective key size in bits equal to
16664 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16665 A few however don't do this and instead use the size of the decrypted key
16666 to determine the RC2 key length and the AlgorithmIdentifier to determine
16667 the effective key length. In this case the effective key length can still
16668 be 40 bits but the key length can be 168 bits for example. This is fixed
16669 by manually forcing an RC2 key into the EVP_PKEY structure because the
16670 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16671 the key length and effective key length are equal.
16675 * Add a bunch of functions that should simplify the creation of
16676 X509_NAME structures. Now you should be able to do:
16677 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16678 and have it automatically work out the correct field type and fill in
16679 the structures. The more adventurous can try:
16680 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16681 and it will (hopefully) work out the correct multibyte encoding.
16685 * Change the 'req' utility to use the new field handling and multibyte
16686 copy routines. Before the DN field creation was handled in an ad hoc
16687 way in req, ca, and x509 which was rather broken and didn't support
16688 BMPStrings or UTF8Strings. Since some software doesn't implement
16689 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16690 using the dirstring_type option. See the new comment in the default
16691 openssl.cnf for more info.
16695 * Make crypto/rand/md_rand.c more robust:
16696 - Assure unique random numbers after fork().
16697 - Make sure that concurrent threads access the global counter and
16698 md serializably so that we never lose entropy in them
16699 or use exactly the same state in multiple threads.
16700 Access to the large state is not always serializable because
16701 the additional locking could be a performance killer, and
16702 md should be large enough anyway.
16706 * New file `apps/app_rand.c` with commonly needed functionality
16707 for handling the random seed file.
16709 Use the random seed file in some applications that previously did not:
16711 dsaparam -genkey (which also ignored its '-rand' option),
16714 x509 (when signing).
16715 Except on systems with /dev/urandom, it is crucial to have a random
16716 seed file at least for key creation, DSA signing, and for DH exchanges;
16717 for RSA signatures we could do without one.
16719 gendh and gendsa (unlike genrsa) used to read only the first byte
16720 of each file listed in the '-rand' option. The function as previously
16721 found in genrsa is now in app_rand.c and is used by all programs
16722 that support '-rand'.
16726 * In RAND_write_file, use mode 0600 for creating files;
16727 don't just chmod when it may be too late.
16731 * Report an error from X509_STORE_load_locations
16732 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16736 * New function ASN1_mbstring_copy() this copies a string in either
16737 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16738 into an ASN1_STRING type. A mask of permissible types is passed
16739 and it chooses the "minimal" type to use or an error if not type
16744 * Add function equivalents to the various macros in asn1.h. The old
16745 macros are retained with an `M_` prefix. Code inside the library can
16746 use the `M_` macros. External code (including the openssl utility)
16747 should *NOT* in order to be "shared library friendly".
16751 * Add various functions that can check a certificate's extensions
16752 to see if it usable for various purposes such as SSL client,
16753 server or S/MIME and CAs of these types. This is currently
16754 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16755 verification. Also added a -purpose flag to x509 utility to
16756 print out all the purposes.
16760 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16765 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16766 for, obtain and decode and extension and obtain its critical flag.
16767 This allows all the necessary extension code to be handled in a
16768 single function call.
16772 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16773 platforms. See crypto/rc4/rc4_enc.c for further details.
16777 * New -noout option to asn1parse. This causes no output to be produced
16778 its main use is when combined with -strparse and -out to extract data
16779 from a file (which may not be in ASN.1 format).
16783 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16784 when producing the local key id.
16786 *Richard Levitte <levitte@stacken.kth.se>*
16788 * New option -dhparam in s_server. This allows a DH parameter file to be
16789 stated explicitly. If it is not stated then it tries the first server
16790 certificate file. The previous behaviour hard coded the filename
16795 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16796 a public key to be input or output. For example:
16797 openssl rsa -in key.pem -pubout -out pubkey.pem
16798 Also added necessary DSA public key functions to handle this.
16802 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16803 in the message. This was handled by allowing
16804 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16806 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16808 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16809 to the end of the strings whereas this didn't. This would cause problems
16810 if strings read with d2i_ASN1_bytes() were later modified.
16812 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16814 * Fix for base64 decode bug. When a base64 bio reads only one line of
16815 data and it contains EOF it will end up returning an error. This is
16816 caused by input 46 bytes long. The cause is due to the way base64
16817 BIOs find the start of base64 encoded data. They do this by trying a
16818 trial decode on each line until they find one that works. When they
16819 do a flag is set and it starts again knowing it can pass all the
16820 data directly through the decoder. Unfortunately it doesn't reset
16821 the context it uses. This means that if EOF is reached an attempt
16822 is made to pass two EOFs through the context and this causes the
16823 resulting error. This can also cause other problems as well. As is
16824 usual with these problems it takes *ages* to find and the fix is
16825 trivial: move one line.
16827 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16829 * Ugly workaround to get s_client and s_server working under Windows. The
16830 old code wouldn't work because it needed to select() on sockets and the
16831 tty (for keypresses and to see if data could be written). Win32 only
16832 supports select() on sockets so we select() with a 1s timeout on the
16833 sockets and then see if any characters are waiting to be read, if none
16834 are present then we retry, we also assume we can always write data to
16835 the tty. This isn't nice because the code then blocks until we've
16836 received a complete line of data and it is effectively polling the
16837 keyboard at 1s intervals: however it's quite a bit better than not
16838 working at all :-) A dedicated Windows application might handle this
16839 with an event loop for example.
16843 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16844 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16845 will be called when RSA_sign() and RSA_verify() are used. This is useful
16846 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16847 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16848 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16849 This necessitated the support of an extra signature type NID_md5_sha1
16850 for SSL signatures and modifications to the SSL library to use it instead
16851 of calling RSA_public_decrypt() and RSA_private_encrypt().
16855 * Add new -verify -CAfile and -CApath options to the crl program, these
16856 will lookup a CRL issuers certificate and verify the signature in a
16857 similar way to the verify program. Tidy up the crl program so it
16858 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16859 less strict. It will now permit CRL extensions even if it is not
16860 a V2 CRL: this will allow it to tolerate some broken CRLs.
16864 * Initialize all non-automatic variables each time one of the openssl
16865 sub-programs is started (this is necessary as they may be started
16866 multiple times from the "OpenSSL>" prompt).
16868 *Lennart Bang, Bodo Moeller*
16870 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16871 removing all other RSA functionality (this is what NO_RSA does). This
16872 is so (for example) those in the US can disable those operations covered
16873 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16878 * Non-copying interface to BIO pairs.
16879 (still largely untested)
16883 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16884 ASCII string. This was handled independently in various places before.
16888 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16889 UTF8 strings a character at a time.
16893 * Use client_version from client hello to select the protocol
16894 (s23_srvr.c) and for RSA client key exchange verification
16895 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16899 * Add various utility functions to handle SPKACs, these were previously
16900 handled by poking round in the structure internals. Added new function
16901 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16902 print, verify and generate SPKACs. Based on an original idea from
16903 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16907 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16911 * Allow the config file extension section to be overwritten on the
16912 command line. Based on an original idea from Massimiliano Pala
16913 <madwolf@comune.modena.it>. The new option is called -extensions
16914 and can be applied to ca, req and x509. Also -reqexts to override
16915 the request extensions in req and -crlexts to override the crl extensions
16920 * Add new feature to the SPKAC handling in ca. Now you can include
16921 the same field multiple times by preceding it by "XXXX." for example:
16924 this is the same syntax as used in the req config file.
16928 * Allow certificate extensions to be added to certificate requests. These
16929 are specified in a 'req_extensions' option of the req section of the
16930 config file. They can be printed out with the -text option to req but
16931 are otherwise ignored at present.
16935 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16936 data read consists of only the final block it would not decrypted because
16937 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16938 A misplaced 'break' also meant the decrypted final block might not be
16939 copied until the next read.
16943 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16944 a few extra parameters to the DH structure: these will be useful if
16945 for example we want the value of 'q' or implement X9.42 DH.
16949 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16950 provides hooks that allow the default DSA functions or functions on a
16951 "per key" basis to be replaced. This allows hardware acceleration and
16952 hardware key storage to be handled without major modification to the
16953 library. Also added low-level modexp hooks and CRYPTO_EX structure and
16954 associated functions.
16958 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16959 as "read only": it can't be written to and the buffer it points to will
16960 not be freed. Reading from a read only BIO is much more efficient than
16961 a normal memory BIO. This was added because there are several times when
16962 an area of memory needs to be read from a BIO. The previous method was
16963 to create a memory BIO and write the data to it, this results in two
16964 copies of the data and an O(n^2) reading algorithm. There is a new
16965 function BIO_new_mem_buf() which creates a read only memory BIO from
16966 an area of memory. Also modified the PKCS#7 routines to use read only
16971 * Bugfix: ssl23_get_client_hello did not work properly when called in
16972 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16973 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16974 but a retry condition occurred while trying to read the rest.
16978 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16979 NID_pkcs7_encrypted by default: this was wrong since this should almost
16980 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16981 the encrypted data type: this is a more sensible place to put it and it
16982 allows the PKCS#12 code to be tidied up that duplicated this
16987 * Changed obj_dat.pl script so it takes its input and output files on
16988 the command line. This should avoid shell escape redirection problems
16993 * Initial support for certificate extension requests, these are included
16994 in things like Xenroll certificate requests. Included functions to allow
16995 extensions to be obtained and added.
16999 * -crlf option to s_client and s_server for sending newlines as
17000 CRLF (as required by many protocols).
17004 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17006 * Install libRSAglue.a when OpenSSL is built with RSAref.
17008 *Ralf S. Engelschall*
17010 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17012 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17014 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17019 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17020 DH parameters/keys (q is lost during that conversion, but the resulting
17021 DH parameters contain its length).
17023 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17024 much faster than DH_generate_parameters (which creates parameters
17025 where `p = 2*q + 1`), and also the smaller q makes DH computations
17026 much more efficient (160-bit exponentiation instead of 1024-bit
17027 exponentiation); so this provides a convenient way to support DHE
17028 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17029 utter importance to use
17030 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17032 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17033 when such DH parameters are used, because otherwise small subgroup
17034 attacks may become possible!
17038 * Avoid memory leak in i2d_DHparams.
17042 * Allow the -k option to be used more than once in the enc program:
17043 this allows the same encrypted message to be read by multiple recipients.
17047 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17048 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17049 it will always use the numerical form of the OID, even if it has a short
17054 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17055 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17056 otherwise bn_mod_exp was called. In the case of hardware keys for example
17057 no private key components need be present and it might store extra data
17058 in the RSA structure, which cannot be accessed from bn_mod_exp.
17059 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17060 private key operations.
17064 * Added support for SPARC Linux.
17068 * pem_password_cb function type incompatibly changed from
17069 typedef int pem_password_cb(char *buf, int size, int rwflag);
17071 ....(char *buf, int size, int rwflag, void *userdata);
17072 so that applications can pass data to their callbacks:
17073 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17074 additional void * argument, which is just handed through whenever
17075 the password callback is called.
17077 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17079 New function SSL_CTX_set_default_passwd_cb_userdata.
17081 Compatibility note: As many C implementations push function arguments
17082 onto the stack in reverse order, the new library version is likely to
17083 interoperate with programs that have been compiled with the old
17084 pem_password_cb definition (PEM_whatever takes some data that
17085 happens to be on the stack as its last argument, and the callback
17086 just ignores this garbage); but there is no guarantee whatsoever that
17089 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17090 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17091 problems not only on Windows, but also on some Unix platforms.
17092 To avoid problematic command lines, these definitions are now in an
17093 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17094 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17098 * MIPS III/IV assembler module is reimplemented.
17102 * More DES library cleanups: remove references to srand/rand and
17103 delete an unused file.
17107 * Add support for the free Netwide assembler (NASM) under Win32,
17108 since not many people have MASM (ml) and it can be hard to obtain.
17109 This is currently experimental but it seems to work OK and pass all
17110 the tests. Check out INSTALL.W32 for info.
17114 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17115 without temporary keys kept an extra copy of the server key,
17116 and connections with temporary keys did not free everything in case
17121 * New function RSA_check_key and new openssl rsa option -check
17122 for verifying the consistency of RSA keys.
17124 *Ulf Moeller, Bodo Moeller*
17126 * Various changes to make Win32 compile work:
17127 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17128 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17129 comparison" warnings.
17130 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17134 * Add a debugging option to PKCS#5 v2 key generation function: when
17135 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17136 derived keys are printed to stderr.
17140 * Copy the flags in ASN1_STRING_dup().
17142 *Roman E. Pavlov <pre@mo.msk.ru>*
17144 * The x509 application mishandled signing requests containing DSA
17145 keys when the signing key was also DSA and the parameters didn't match.
17147 It was supposed to omit the parameters when they matched the signing key:
17148 the verifying software was then supposed to automatically use the CA's
17149 parameters if they were absent from the end user certificate.
17151 Omitting parameters is no longer recommended. The test was also
17152 the wrong way round! This was probably due to unusual behaviour in
17153 EVP_cmp_parameters() which returns 1 if the parameters match.
17154 This meant that parameters were omitted when they *didn't* match and
17155 the certificate was useless. Certificates signed with 'ca' didn't have
17158 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17160 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17161 The interface is as follows:
17162 Applications can use
17163 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17164 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17165 "off" is now the default.
17166 The library internally uses
17167 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17168 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17169 to disable memory-checking temporarily.
17171 Some inconsistent states that previously were possible (and were
17172 even the default) are now avoided.
17174 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17175 with each memory chunk allocated; this is occasionally more helpful
17176 than just having a counter.
17178 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17180 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17185 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17186 which largely parallels "options", but is for changing API behaviour,
17187 whereas "options" are about protocol behaviour.
17188 Initial "mode" flags are:
17190 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17191 a single record has been written.
17192 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17193 retries use the same buffer location.
17194 (But all of the contents must be
17199 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17202 * Fix problems with no-hmac etc.
17204 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17206 * New functions RSA_get_default_method(), RSA_set_method() and
17207 RSA_get_method(). These allows replacement of RSA_METHODs without having
17208 to mess around with the internals of an RSA structure.
17212 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17213 Also really enable memory leak checks in openssl.c and in some
17216 *Chad C. Mulligan, Bodo Moeller*
17218 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17219 up the length of negative integers. This has now been simplified to just
17220 store the length when it is first determined and use it later, rather
17221 than trying to keep track of where data is copied and updating it to
17223 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17225 * Add a new function PKCS7_signatureVerify. This allows the verification
17226 of a PKCS#7 signature but with the signing certificate passed to the
17227 function itself. This contrasts with PKCS7_dataVerify which assumes the
17228 certificate is present in the PKCS#7 structure. This isn't always the
17229 case: certificates can be omitted from a PKCS#7 structure and be
17230 distributed by "out of band" means (such as a certificate database).
17234 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17235 function prototypes in pem.h, also change util/mkdef.pl to add the
17236 necessary function names.
17240 * mk1mf.pl (used by Windows builds) did not properly read the
17241 options set by Configure in the top level Makefile, and Configure
17242 was not even able to write more than one option correctly.
17243 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17247 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17248 file to be loaded from a BIO or FILE pointer. The BIO version will
17249 for example allow memory BIOs to contain config info.
17253 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17254 Whoever hopes to achieve shared-library compatibility across versions
17255 must use this, not the compile-time macro.
17256 (Exercise 0.9.4: Which is the minimum library version required by
17258 Note: All this applies only to multi-threaded programs, others don't
17263 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17264 through a BIO pair triggered the default case, i.e.
17265 SSLerr(...,SSL_R_UNKNOWN_STATE).
17269 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17270 can use the SSL library even if none of the specific BIOs is
17275 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17276 for the encoded length.
17278 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17280 * Add initial documentation of the X509V3 functions.
17284 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17285 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17286 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17287 secure PKCS#8 private key format with a high iteration count.
17291 * Fix determination of Perl interpreter: A perl or perl5
17292 *directory* in $PATH was also accepted as the interpreter.
17294 *Ralf S. Engelschall*
17296 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17297 wrong with it but it was very old and did things like calling
17298 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17299 unusual formatting.
17303 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17304 to use the new extension code.
17308 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17309 with macros. This should make it easier to change their form, add extra
17310 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17315 * Add to configuration table a new entry that can specify an alternative
17316 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17317 according to Mark Crispin <MRC@Panda.COM>.
17321 * DES CBC did not update the IV. Weird.
17325 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17326 Changing the behaviour of the former might break existing programs --
17327 where IV updating is needed, des_ncbc_encrypt can be used.
17330 * When bntest is run from "make test" it drives bc to check its
17331 calculations, as well as internally checking them. If an internal check
17332 fails, it needs to cause bc to give a non-zero result or make test carries
17333 on without noticing the failure. Fixed.
17337 * DES library cleanups.
17341 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17342 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17343 ciphers. NOTE: although the key derivation function has been verified
17344 against some published test vectors it has not been extensively tested
17345 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17350 * Instead of "mkdir -p", which is not fully portable, use new
17351 Perl script "util/mkdir-p.pl".
17355 * Rewrite the way password based encryption (PBE) is handled. It used to
17356 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17357 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17358 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17359 the 'parameter' field of the AlgorithmIdentifier is passed to the
17360 underlying key generation function so it must do its own ASN1 parsing.
17361 This has also changed the EVP_PBE_CipherInit() function which now has a
17362 'parameter' argument instead of literal salt and iteration count values
17363 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17367 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17368 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17369 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17370 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17371 value was just used as a "magic string" and not used directly its
17372 value doesn't matter.
17376 * Introduce some semblance of const correctness to BN. Shame C doesn't
17381 * "linux-sparc64" configuration (ultrapenguin).
17383 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17384 "linux-sparc" configuration.
17386 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17388 * config now generates no-xxx options for missing ciphers.
17392 * Support the EBCDIC character set (work in progress).
17393 File ebcdic.c not yet included because it has a different license.
17395 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17397 * Support BS2000/OSD-POSIX.
17399 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17401 * Make callbacks for key generation use `void *` instead of `char *`.
17405 * Make S/MIME samples compile (not yet tested).
17409 * Additional typesafe stacks.
17413 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17417 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17419 * New configuration variant "sco5-gcc".
17421 * Updated some demos.
17423 *Sean O Riordain, Wade Scholine*
17425 * Add missing BIO_free at exit of pkcs12 application.
17429 * Fix memory leak in conf.c.
17433 * Updates for Win32 to assembler version of MD5.
17437 * Set #! path to perl in `apps/der_chop` to where we found it
17438 instead of using a fixed path.
17442 * SHA library changes for irix64-mips4-cc.
17446 * Improvements for VMS support.
17450 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17452 * Bignum library bug fix. IRIX 6 passes "make test" now!
17453 This also avoids the problems with SC4.2 and unpatched SC5.
17455 *Andy Polyakov <appro@fy.chalmers.se>*
17457 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17458 These are required because of the typesafe stack would otherwise break
17459 existing code. If old code used a structure member which used to be STACK
17460 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17461 sk_num or sk_value it would produce an error because the num, data members
17462 are not present in STACK_OF. Now it just produces a warning. sk_set
17463 replaces the old method of assigning a value to sk_value
17464 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17465 that does this will no longer work (and should use sk_set instead) but
17466 this could be regarded as a "questionable" behaviour anyway.
17470 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17471 correctly handle encrypted S/MIME data.
17475 * Change type of various DES function arguments from des_cblock
17476 (which means, in function argument declarations, pointer to char)
17477 to des_cblock * (meaning pointer to array with 8 char elements),
17478 which allows the compiler to do more typechecking; it was like
17479 that back in SSLeay, but with lots of ugly casts.
17481 Introduce new type const_des_cblock.
17485 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17486 problems: find RecipientInfo structure that matches recipient certificate
17487 and initialise the ASN1 structures properly based on passed cipher.
17491 * Belatedly make the BN tests actually check the results.
17495 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17496 to and from BNs: it was completely broken. New compilation option
17497 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17498 key elements as negative integers.
17502 * Reorganize and speed up MD5.
17504 *Andy Polyakov <appro@fy.chalmers.se>*
17508 *Richard Levitte <richard@levitte.org>*
17510 * New option -out to asn1parse to allow the parsed structure to be
17511 output to a file. This is most useful when combined with the -strparse
17512 option to examine the output of things like OCTET STRINGS.
17516 * Make SSL library a little more fool-proof by not requiring any longer
17517 that `SSL_set_{accept,connect}_state` be called before
17518 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17519 in many applications because usually everything *appeared* to work as
17520 intended anyway -- now it really works as intended).
17524 * Move openssl.cnf out of lib/.
17528 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17529 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17530 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17532 *Ralf S. Engelschall*
17534 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17535 handle PKCS#7 enveloped data properly.
17537 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17539 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17540 copying pointers. The cert_st handling is changed by this in
17541 various ways (and thus what used to be known as ctx->default_cert
17542 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17543 any longer when s->cert does not give us what we need).
17544 ssl_cert_instantiate becomes obsolete by this change.
17545 As soon as we've got the new code right (possibly it already is?),
17546 we have solved a couple of bugs of the earlier code where s->cert
17547 was used as if it could not have been shared with other SSL structures.
17549 Note that using the SSL API in certain dirty ways now will result
17550 in different behaviour than observed with earlier library versions:
17551 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17552 does not influence s as it used to.
17554 In order to clean up things more thoroughly, inside SSL_SESSION
17555 we don't use CERT any longer, but a new structure SESS_CERT
17556 that holds per-session data (if available); currently, this is
17557 the peer's certificate chain and, for clients, the server's certificate
17558 and temporary key. CERT holds only those values that can have
17559 meaningful defaults in an SSL_CTX.
17563 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17564 from the internal representation. Various PKCS#7 fixes: remove some
17565 evil casts and set the enc_dig_alg field properly based on the signing
17570 * Allow PKCS#12 password to be set from the command line or the
17571 environment. Let 'ca' get its config file name from the environment
17572 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17577 * Allow certificate policies extension to use an IA5STRING for the
17578 organization field. This is contrary to the PKIX definition but
17579 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17584 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17585 without disallowing inline assembler and the like for non-pedantic builds.
17589 * Support Borland C++ builder.
17591 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17597 * SHA-1 cleanups and performance enhancements.
17599 *Andy Polyakov <appro@fy.chalmers.se>*
17601 * Sparc v8plus assembler for the bignum library.
17603 *Andy Polyakov <appro@fy.chalmers.se>*
17605 * Accept any -xxx and +xxx compiler options in Configure.
17609 * Update HPUX configuration.
17613 * Add missing `sk_<type>_unshift()` function to safestack.h
17615 *Ralf S. Engelschall*
17617 * New function SSL_CTX_use_certificate_chain_file that sets the
17618 "extra_cert"s in addition to the certificate. (This makes sense
17619 only for "PEM" format files, as chains as a whole are not
17624 * Support verify_depth from the SSL API.
17625 x509_vfy.c had what can be considered an off-by-one-error:
17626 Its depth (which was not part of the external interface)
17627 was actually counting the number of certificates in a chain;
17628 now it really counts the depth.
17632 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17633 instead of X509err, which often resulted in confusing error
17634 messages since the error codes are not globally unique
17635 (e.g. an alleged error in ssl3_accept when a certificate
17636 didn't match the private key).
17638 * New function SSL_CTX_set_session_id_context that allows to set a default
17639 value (so that you don't need SSL_set_session_id_context for each
17640 connection using the SSL_CTX).
17644 * OAEP decoding bug fix.
17648 * Support INSTALL_PREFIX for package builders, as proposed by
17653 * New Configure options "threads" and "no-threads". For systems
17654 where the proper compiler options are known (currently Solaris
17655 and Linux), "threads" is the default.
17659 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17663 * Install various scripts to $(OPENSSLDIR)/misc, not to
17664 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17665 such as /usr/local/bin.
17669 * "make linux-shared" to build shared libraries.
17671 *Niels Poppe <niels@netbox.org>*
17673 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17677 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17678 extension adding in x509 utility.
17682 * Remove NOPROTO sections and error code comments.
17686 * Partial rewrite of the DEF file generator to now parse the ANSI
17691 * New Configure options --prefix=DIR and --openssldir=DIR.
17695 * Complete rewrite of the error code script(s). It is all now handled
17696 by one script at the top level which handles error code gathering,
17697 header rewriting and C source file generation. It should be much better
17698 than the old method: it now uses a modified version of Ulf's parser to
17699 read the ANSI prototypes in all header files (thus the old K&R definitions
17700 aren't needed for error creation any more) and do a better job of
17701 translating function codes into names. The old 'ASN1 error code embedded
17702 in a comment' is no longer necessary and it doesn't use .err files which
17703 have now been deleted. Also the error code call doesn't have to appear all
17704 on one line (which resulted in some large lines...).
17708 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17712 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17713 0 (which usually indicates a closed connection), but continue reading.
17717 * Fix some race conditions.
17721 * Add support for CRL distribution points extension. Add Certificate
17722 Policies and CRL distribution points documentation.
17726 * Move the autogenerated header file parts to crypto/opensslconf.h.
17730 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17731 8 of keying material. Merlin has also confirmed interop with this fix
17732 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17734 *Merlin Hughes <merlin@baltimore.ie>*
17736 * Fix lots of warnings.
17738 *Richard Levitte <levitte@stacken.kth.se>*
17740 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17741 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17743 *Richard Levitte <levitte@stacken.kth.se>*
17745 * Fix problems with sizeof(long) == 8.
17747 *Andy Polyakov <appro@fy.chalmers.se>*
17749 * Change functions to ANSI C.
17753 * Fix typos in error codes.
17755 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17757 * Remove defunct assembler files from Configure.
17761 * SPARC v8 assembler BIGNUM implementation.
17763 *Andy Polyakov <appro@fy.chalmers.se>*
17765 * Support for Certificate Policies extension: both print and set.
17766 Various additions to support the r2i method this uses.
17770 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17771 return a const string when you are expecting an allocated buffer.
17775 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17776 types DirectoryString and DisplayText.
17780 * Add code to allow r2i extensions to access the configuration database,
17781 add an LHASH database driver and add several ctx helper functions.
17785 * Fix an evil bug in bn_expand2() which caused various BN functions to
17786 fail when they extended the size of a BIGNUM.
17790 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17791 support typesafe stack.
17795 * Fix typo in SSL_[gs]et_options().
17797 *Nils Frostberg <nils@medcom.se>*
17799 * Delete various functions and files that belonged to the (now obsolete)
17800 old X509V3 handling code.
17804 * New Configure option "rsaref".
17808 * Don't auto-generate pem.h.
17812 * Introduce type-safe ASN.1 SETs.
17816 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17818 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17820 * Introduce type-safe STACKs. This will almost certainly break lots of code
17821 that links with OpenSSL (well at least cause lots of warnings), but fear
17822 not: the conversion is trivial, and it eliminates loads of evil casts. A
17823 few STACKed things have been converted already. Feel free to convert more.
17824 In the fullness of time, I'll do away with the STACK type altogether.
17828 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17829 specified in `<certfile>` by updating the entry in the index.txt file.
17830 This way one no longer has to edit the index.txt file manually for
17831 revoking a certificate. The -revoke option does the gory details now.
17833 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17835 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17836 `-text` option at all and this way the `-noout -text` combination was
17837 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17839 *Ralf S. Engelschall*
17841 * Make sure a corresponding plain text error message exists for the
17842 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17843 verify callback function determined that a certificate was revoked.
17845 *Ralf S. Engelschall*
17847 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17848 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17849 all available ciphers including rc5, which was forgotten until now.
17850 In order to let the testing shell script know which algorithms
17851 are available, a new (up to now undocumented) command
17852 `openssl list-cipher-commands` is used.
17856 * Bugfix: s_client occasionally would sleep in select() when
17857 it should have checked SSL_pending() first.
17861 * New functions DSA_do_sign and DSA_do_verify to provide access to
17862 the raw DSA values prior to ASN.1 encoding.
17866 * Tweaks to Configure
17868 *Niels Poppe <niels@netbox.org>*
17870 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17875 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17879 * New config option to avoid instructions that are illegal on the 80386.
17880 The default code is faster, but requires at least a 486.
17884 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17885 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17886 same as SSL2_VERSION anyway.
17890 * New "-showcerts" option for s_client.
17894 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17895 application. Various cleanups and fixes.
17899 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17900 modify error routines to work internally. Add error codes and PBE init
17901 to library startup routines.
17905 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17906 packing functions to asn1 and evp. Changed function names and error
17907 codes along the way.
17911 * PKCS12 integration: and so it begins... First of several patches to
17912 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17913 objects to objects.h
17917 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17918 and display support for Thawte strong extranet extension.
17922 * Add LinuxPPC support.
17924 *Jeff Dubrule <igor@pobox.org>*
17926 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17927 bn_div_words in alpha.s.
17929 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17931 * Make sure the RSA OAEP test is skipped under -DRSAref because
17932 OAEP isn't supported when OpenSSL is built with RSAref.
17934 *Ulf Moeller <ulf@fitug.de>*
17936 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17937 so they no longer are missing under -DNOPROTO.
17939 *Soren S. Jorvang <soren@t.dk>*
17941 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17943 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17944 doesn't work when the session is reused. Coming soon!
17948 * Fix a security hole, that allows sessions to be reused in the wrong
17949 context thus bypassing client cert protection! All software that uses
17950 client certs and session caches in multiple contexts NEEDS PATCHING to
17951 allow session reuse! A fuller solution is in the works.
17953 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17955 * Some more source tree cleanups (removed obsolete files
17956 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17957 permission on "config" script to be executable) and a fix for the INSTALL
17960 *Ulf Moeller <ulf@fitug.de>*
17962 * Remove some legacy and erroneous uses of malloc, free instead of
17965 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17967 * Make rsa_oaep_test return non-zero on error.
17969 *Ulf Moeller <ulf@fitug.de>*
17971 * Add support for native Solaris shared libraries. Configure
17972 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17973 if someone would make that last step automatic.
17975 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17977 * ctx_size was not built with the right compiler during "make links". Fixed.
17981 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17982 except NULL ciphers". This means the default cipher list will no longer
17983 enable NULL ciphers. They need to be specifically enabled e.g. with
17984 the string "DEFAULT:eNULL".
17988 * Fix to RSA private encryption routines: if p < q then it would
17989 occasionally produce an invalid result. This will only happen with
17990 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17994 * Be less restrictive and allow also `perl util/perlpath.pl
17995 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17996 because this way one can also use an interpreter named `perl5` (which is
17997 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17998 installed as `perl`).
18000 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18002 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18004 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18006 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18007 advapi32.lib to Win32 build and change the pem test comparison
18008 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18009 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18010 and crypto/des/ede_cbcm_enc.c.
18014 * DES quad checksum was broken on big-endian architectures. Fixed.
18018 * Comment out two functions in bio.h that aren't implemented. Fix up the
18019 Win32 test batch file so it (might) work again. The Win32 test batch file
18020 is horrible: I feel ill....
18024 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18025 in e_os.h. Audit of header files to check ANSI and non ANSI
18026 sections: 10 functions were absent from non ANSI section and not exported
18027 from Windows DLLs. Fixed up libeay.num for new functions.
18031 * Make `openssl version` output lines consistent.
18033 *Ralf S. Engelschall*
18035 * Fix Win32 symbol export lists for BIO functions: Added
18036 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18037 to ms/libeay{16,32}.def.
18039 *Ralf S. Engelschall*
18041 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18042 fine under Unix and passes some trivial tests I've now added. But the
18043 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18044 added to make sure no one expects that this stuff really works in the
18045 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18046 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18049 *Ralf S. Engelschall*
18051 * Fix the generation of two part addresses in perl.
18053 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18055 * Add config entry for Linux on MIPS.
18057 *John Tobey <jtobey@channel1.com>*
18059 * Make links whenever Configure is run, unless we are on Windoze.
18063 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18064 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18069 * Add a useful kludge to allow package maintainers to specify compiler and
18070 other platforms details on the command line without having to patch the
18071 Configure script every time: One now can use
18072 `perl Configure <id>:<details>`,
18073 i.e. platform ids are allowed to have details appended
18074 to them (separated by colons). This is treated as there would be a static
18075 pre-configured entry in Configure's %table under key `<id>` with value
18076 `<details>` and `perl Configure <id>` is called. So, when you want to
18077 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18078 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18079 now, which overrides the FreeBSD-elf entry on-the-fly.
18081 *Ralf S. Engelschall*
18083 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18087 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18088 on the `perl Configure ...` command line. This way one can compile
18089 OpenSSL libraries with Position Independent Code (PIC) which is needed
18090 for linking it into DSOs.
18092 *Ralf S. Engelschall*
18094 * Remarkably, export ciphers were totally broken and no-one had noticed!
18099 * Cleaned up the LICENSE document: The official contact for any license
18100 questions now is the OpenSSL core team under openssl-core@openssl.org.
18101 And add a paragraph about the dual-license situation to make sure people
18102 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18103 to the OpenSSL toolkit.
18105 *Ralf S. Engelschall*
18107 * General source tree makefile cleanups: Made `making xxx in yyy...`
18108 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18109 Additionally cleaned up the `make links` target: Remove unnecessary
18110 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18111 to speed processing and no longer clutter the display with confusing
18112 stuff. Instead only the actually done links are displayed.
18114 *Ralf S. Engelschall*
18116 * Permit null encryption ciphersuites, used for authentication only. It used
18117 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18118 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18123 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18124 signed attributes when verifying signatures (this would break them),
18125 the detached data encoding was wrong and public keys obtained using
18126 X509_get_pubkey() weren't freed.
18130 * Add text documentation for the BUFFER functions. Also added a work around
18131 to a Win95 console bug. This was triggered by the password read stuff: the
18132 last character typed gets carried over to the next fread(). If you were
18133 generating a new cert request using 'req' for example then the last
18134 character of the passphrase would be CR which would then enter the first
18139 * Added the new 'Includes OpenSSL Cryptography Software' button as
18140 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18141 button and can be used by applications based on OpenSSL to show the
18142 relationship to the OpenSSL project.
18144 *Ralf S. Engelschall*
18146 * Remove confusing variables in function signatures in files
18147 ssl/ssl_lib.c and ssl/ssl.h.
18149 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18151 * Don't install bss_file.c under PREFIX/include/
18153 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18155 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18156 functions that return function pointers and has support for NT specific
18157 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18158 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18159 unsigned to signed types: this was killing the Win32 compile.
18163 * Add new certificate file to stack functions,
18164 SSL_add_dir_cert_subjects_to_stack() and
18165 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18166 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18167 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18168 This means that Apache-SSL and similar packages don't have to mess around
18169 to add as many CAs as they want to the preferred list.
18173 * Experiment with doxygen documentation. Currently only partially applied to
18175 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18176 openssl.doxy as the configuration file.
18180 * Get rid of remaining C++-style comments which strict C compilers hate.
18182 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18184 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18185 compiled in by default: it has problems with large keys.
18189 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18190 DH private keys and/or callback functions which directly correspond to
18191 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18192 is needed for applications which have to configure certificates on a
18193 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18195 For the RSA certificate situation is makes no difference, but
18196 for the DSA certificate situation this fixes the "no shared cipher"
18197 problem where the OpenSSL cipher selection procedure failed because the
18198 temporary keys were not overtaken from the context and the API provided
18199 no way to reconfigure them.
18200 The new functions now let applications reconfigure the stuff and they
18201 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18202 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18203 non-public-API function ssl_cert_instantiate() is used as a helper
18204 function and also to reduce code redundancy inside ssl_rsa.c.
18206 *Ralf S. Engelschall*
18208 * Move s_server -dcert and -dkey options out of the undocumented feature
18209 area because they are useful for the DSA situation and should be
18210 recognized by the users.
18212 *Ralf S. Engelschall*
18214 * Fix the cipher decision scheme for export ciphers: the export bits are
18215 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18216 SSL_EXP_MASK. So, the original variable has to be used instead of the
18217 already masked variable.
18219 *Richard Levitte <levitte@stacken.kth.se>*
18221 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18223 *Richard Levitte <levitte@stacken.kth.se>*
18225 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18226 from `int` to `unsigned int` because it is a length and initialized by
18227 EVP_DigestFinal() which expects an `unsigned int *`.
18229 *Richard Levitte <levitte@stacken.kth.se>*
18231 * Don't hard-code path to Perl interpreter on shebang line of Configure
18232 script. Instead use the usual Shell->Perl transition trick.
18234 *Ralf S. Engelschall*
18236 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18237 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18238 -noout -modulus` as it's already the case for `openssl rsa -noout
18239 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18240 currently the public key is printed (a decision which was already done by
18241 `openssl dsa -modulus` in the past) which serves a similar purpose.
18242 Additionally the NO_RSA no longer completely removes the whole -modulus
18243 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18246 *Ralf S. Engelschall*
18248 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18249 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18251 *Arne Ansper <arne@ats.cyber.ee>*
18253 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18254 to be added. Now both 'req' and 'ca' can use new objects defined in the
18259 * Add cool BIO that does syslog (or event log on NT).
18261 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18263 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18264 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18265 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18266 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18270 * Add preliminary config info for new extension code.
18274 * Make RSA_NO_PADDING really use no padding.
18276 *Ulf Moeller <ulf@fitug.de>*
18278 * Generate errors when private/public key check is done.
18282 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18283 for some CRL extensions and new objects added.
18287 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18288 key usage extension and fuller support for authority key id.
18292 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18293 padding method for RSA, which is recommended for new applications in PKCS
18294 #1 v2.0 (RFC 2437, October 1998).
18295 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18296 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18297 against Bleichbacher's attack on RSA.
18298 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18301 * Updates to the new SSL compression code
18303 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18305 * Fix so that the version number in the master secret, when passed
18306 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18307 (because the server will not accept higher), that the version number
18308 is 0x03,0x01, not 0x03,0x00
18310 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18312 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18313 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18314 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18318 * Support for RAW extensions where an arbitrary extension can be
18319 created by including its DER encoding. See `apps/openssl.cnf` for
18324 * Make sure latest Perl versions don't interpret some generated C array
18325 code as Perl array code in the crypto/err/err_genc.pl script.
18327 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18329 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18330 not many people have the assembler. Various Win32 compilation fixes and
18331 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18332 build instructions.
18336 * Modify configure script 'Configure' to automatically create crypto/date.h
18337 file under Win32 and also build pem.h from pem.org. New script
18338 util/mkfiles.pl to create the MINFO file on environments that can't do a
18339 'make files': perl util/mkfiles.pl >MINFO should work.
18343 * Major rework of DES function declarations, in the pursuit of correctness
18344 and purity. As a result, many evil casts evaporated, and some weirdness,
18345 too. You may find this causes warnings in your code. Zapping your evil
18346 casts will probably fix them. Mostly.
18350 * Fix for a typo in asn1.h. Bug fix to object creation script
18351 obj_dat.pl. It considered a zero in an object definition to mean
18352 "end of object": none of the objects in objects.h have any zeros
18353 so it wasn't spotted.
18355 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18357 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18358 Masking (CBCM). In the absence of test vectors, the best I have been able
18359 to do is check that the decrypt undoes the encrypt, so far. Send me test
18360 vectors if you have them.
18364 * Correct calculation of key length for export ciphers (too much space was
18365 allocated for null ciphers). This has not been tested!
18369 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18370 message is now correct (it understands "crypto" and "ssl" on its
18371 command line). There is also now an "update" option. This will update
18372 the util/ssleay.num and util/libeay.num files with any new functions.
18374 perl util/mkdef.pl crypto ssl update
18375 it will update them.
18379 * Overhauled the Perl interface:
18380 - ported BN stuff to OpenSSL's different BN library
18381 - made the perl/ source tree CVS-aware
18382 - renamed the package from SSLeay to OpenSSL (the files still contain
18383 their history because I've copied them in the repository)
18384 - removed obsolete files (the test scripts will be replaced
18385 by better Test::Harness variants in the future)
18387 *Ralf S. Engelschall*
18389 * First cut for a very conservative source tree cleanup:
18390 1. merge various obsolete readme texts into doc/ssleay.txt
18391 where we collect the old documents and readme texts.
18392 2. remove the first part of files where I'm already sure that we no
18393 longer need them because of three reasons: either they are just temporary
18394 files which were left by Eric or they are preserved original files where
18395 I've verified that the diff is also available in the CVS via "cvs diff
18396 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18397 the crypto/md/ stuff).
18399 *Ralf S. Engelschall*
18401 * More extension code. Incomplete support for subject and issuer alt
18402 name, issuer and authority key id. Change the i2v function parameters
18403 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18404 what that's for :-) Fix to ASN1 macro which messed up
18405 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18409 * Preliminary support for ENUMERATED type. This is largely copied from the
18414 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18416 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18418 * Make sure `make rehash` target really finds the `openssl` program.
18420 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18422 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18423 like to hear about it if this slows down other processors.
18427 * Add CygWin32 platform information to Configure script.
18429 *Alan Batie <batie@aahz.jf.intel.com>*
18431 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18433 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18435 * New program nseq to manipulate netscape certificate sequences
18439 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18444 * Fixes to BN code. Previously the default was to define BN_RECURSION
18445 but the BN code had some problems that would cause failures when
18446 doing certificate verification and some other functions.
18448 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18450 * Add ASN1 and PEM code to support netscape certificate sequences.
18454 * Add ASN1 and PEM code to support netscape certificate sequences.
18458 * Add several PKIX and private extended key usage OIDs.
18462 * Modify the 'ca' program to handle the new extension code. Modify
18463 openssl.cnf for new extension format, add comments.
18467 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18468 and add a sample to openssl.cnf so req -x509 now adds appropriate
18473 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18474 error code, add initial support to X509_print() and x509 application.
18478 * Takes a deep breath and start adding X509 V3 extension support code. Add
18479 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18480 stuff is currently isolated and isn't even compiled yet.
18484 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18485 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18486 Removed the versions check from X509 routines when loading extensions:
18487 this allows certain broken certificates that don't set the version
18488 properly to be processed.
18492 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18493 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18494 can still be regenerated with "make depend".
18498 * Spelling mistake in C version of CAST-128.
18500 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18502 * Changes to the error generation code. The perl script err-code.pl
18503 now reads in the old error codes and retains the old numbers, only
18504 adding new ones if necessary. It also only changes the .err files if new
18505 codes are added. The makefiles have been modified to only insert errors
18506 when needed (to avoid needlessly modifying header files). This is done
18507 by only inserting errors if the .err file is newer than the auto generated
18508 C file. To rebuild all the error codes from scratch (the old behaviour)
18509 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18510 or delete all the .err files.
18514 * CAST-128 was incorrectly implemented for short keys. The C version has
18515 been fixed, but is untested. The assembler versions are also fixed, but
18516 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18517 to regenerate it if needed.
18518 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18519 Hagino <itojun@kame.net>*
18521 * File was opened incorrectly in randfile.c.
18523 *Ulf Möller <ulf@fitug.de>*
18525 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18526 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18527 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18528 al: it's just almost always a UTCTime. Note this patch adds new error
18529 codes so do a "make errors" if there are problems.
18533 * Correct Linux 1 recognition in config.
18535 *Ulf Möller <ulf@fitug.de>*
18537 * Remove pointless MD5 hash when using DSA keys in ca.
18539 *Anonymous <nobody@replay.com>*
18541 * Generate an error if given an empty string as a cert directory. Also
18542 generate an error if handed NULL (previously returned 0 to indicate an
18543 error, but didn't set one).
18545 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18547 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18551 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18552 parameters. This was causing a warning which killed off the Win32 compile.
18556 * Remove C++ style comments from crypto/bn/bn_local.h.
18558 *Neil Costigan <neil.costigan@celocom.com>*
18560 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18561 based on a text string, looking up short and long names and finally
18562 "dot" format. The "dot" format stuff didn't work. Added new function
18563 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18564 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18565 OID is not part of the table.
18569 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18570 X509_LOOKUP_by_alias().
18574 * Sort openssl functions by name.
18578 * Get the `gendsa` command working and add it to the `list` command. Remove
18579 encryption from sample DSA keys (in case anyone is interested the password
18584 * Make *all* `*_free` functions accept a NULL pointer.
18586 *Frans Heymans <fheymans@isaserver.be>*
18588 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18591 *Anonymous <nobody@replay.com>*
18593 * s_server should send the CAfile as acceptable CAs, not its own cert.
18595 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18597 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18599 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18601 * Temp key "for export" tests were wrong in s3_srvr.c.
18603 *Anonymous <nobody@replay.com>*
18605 * Add prototype for temp key callback functions
18606 SSL_CTX_set_tmp_{rsa,dh}_callback().
18610 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18611 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18615 * X509_name_add_entry() freed the wrong thing after an error.
18617 *Arne Ansper <arne@ats.cyber.ee>*
18619 * rsa_eay.c would attempt to free a NULL context.
18621 *Arne Ansper <arne@ats.cyber.ee>*
18623 * BIO_s_socket() had a broken should_retry() on Windoze.
18625 *Arne Ansper <arne@ats.cyber.ee>*
18627 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18629 *Arne Ansper <arne@ats.cyber.ee>*
18631 * Make sure the already existing X509_STORE->depth variable is initialized
18632 in X509_STORE_new(), but document the fact that this variable is still
18633 unused in the certificate verification process.
18635 *Ralf S. Engelschall*
18637 * Fix the various library and `apps/` files to free up pkeys obtained from
18638 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18642 * Fix reference counting in X509_PUBKEY_get(). This makes
18643 demos/maurice/example2.c work, amongst others, probably.
18645 *Steve Henson and Ben Laurie*
18647 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18648 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18649 are no longer created. This way we have a single and consistent command
18650 line interface `openssl <command>`, similar to `cvs <command>`.
18652 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18654 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18655 BIT STRING wrapper always have zero unused bits.
18659 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18663 * Make the top-level INSTALL documentation easier to understand.
18667 * Makefiles updated to exit if an error occurs in a sub-directory
18668 make (including if user presses ^C) [Paul Sutton]
18670 * Make Montgomery context stuff explicit in RSA data structure.
18674 * Fix build order of pem and err to allow for generated pem.h.
18678 * Fix renumbering bug in X509_NAME_delete_entry().
18682 * Enhanced the err-ins.pl script so it makes the error library number
18683 global and can add a library name. This is needed for external ASN1 and
18684 other error libraries.
18688 * Fixed sk_insert which never worked properly.
18692 * Fix ASN1 macros so they can handle indefinite length constructed
18693 EXPLICIT tags. Some non standard certificates use these: they can now
18698 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18699 into a single doc/ssleay.txt bundle. This way the information is still
18700 preserved but no longer messes up this directory. Now it's new room for
18701 the new set of documentation files.
18703 *Ralf S. Engelschall*
18705 * SETs were incorrectly DER encoded. This was a major pain, because they
18706 shared code with SEQUENCEs, which aren't coded the same. This means that
18707 almost everything to do with SETs or SEQUENCEs has either changed name or
18708 number of arguments.
18710 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18712 * Fix test data to work with the above.
18716 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18717 was already fixed by Eric for 0.9.1 it seems.
18719 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18721 * Autodetect FreeBSD3.
18725 * Fix various bugs in Configure. This affects the following platforms:
18729 unixware-2.0-pentium
18734 * Eliminate generated files from CVS. Reorder tests to regenerate files
18735 before they are needed.
18739 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18743 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18745 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18746 changed SSLeay to OpenSSL in version strings.
18748 *Ralf S. Engelschall*
18750 * Some fixups to the top-level documents.
18754 * Fixed the nasty bug where rsaref.h was not found under compile-time
18755 because the symlink to include/ was missing.
18757 *Ralf S. Engelschall*
18759 * Incorporated the popular no-RSA/DSA-only patches
18760 which allow to compile a RSA-free SSLeay.
18762 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18764 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18765 when "ssleay" is still not found.
18767 *Ralf S. Engelschall*
18769 * Added more platforms to Configure: Cray T3E, HPUX 11,
18771 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18773 * Updated the README file.
18775 *Ralf S. Engelschall*
18777 * Added various .cvsignore files in the CVS repository subdirs
18778 to make a "cvs update" really silent.
18780 *Ralf S. Engelschall*
18782 * Recompiled the error-definition header files and added
18783 missing symbols to the Win32 linker tables.
18785 *Ralf S. Engelschall*
18787 * Cleaned up the top-level documents;
18788 o new files: CHANGES and LICENSE
18789 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18790 o merged COPYRIGHT into LICENSE
18791 o removed obsolete TODO file
18792 o renamed MICROSOFT to INSTALL.W32
18794 *Ralf S. Engelschall*
18796 * Removed dummy files from the 0.9.1b source tree:
18797 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18798 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18799 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18800 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18801 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18803 *Ralf S. Engelschall*
18805 * Added various platform portability fixes.
18809 * The Genesis of the OpenSSL rpject:
18810 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18811 Young and Tim J. Hudson created while they were working for C2Net until
18814 *The OpenSSL Project*
18816 ### Changes between 0.9.0b and 0.9.1b [not released]
18818 * Updated a few CA certificates under certs/
18822 * Changed some BIGNUM api stuff.
18826 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18827 DGUX x86, Linux Alpha, etc.
18831 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18832 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18837 * Add -strparse option to asn1pars program which parses nested
18840 *Dr Stephen Henson <shenson@bigfoot.com>*
18842 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18846 * DSA fix for "ca" program.
18850 * Added "-genkey" option to "dsaparam" program.
18854 * Added RIPE MD160 (rmd160) message digest.
18858 * Added -a (all) option to "ssleay version" command.
18862 * Added PLATFORM define which is the id given to Configure.
18866 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18870 * Extended the ASN.1 parser routines.
18874 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18878 * Added a BN_CTX to the BN library.
18882 * Fixed the weak key values in DES library
18886 * Changed API in EVP library for cipher aliases.
18890 * Added support for RC2/64bit cipher.
18894 * Converted the lhash library to the crypto/mem.c functions.
18898 * Added more recognized ASN.1 object ids.
18902 * Added more RSA padding checks for SSL/TLS.
18906 * Added BIO proxy/filter functionality.
18910 * Added extra_certs to SSL_CTX which can be used
18911 send extra CA certificates to the client in the CA cert chain sending
18912 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18916 * Now Fortezza is denied in the authentication phase because
18917 this is key exchange mechanism is not supported by SSLeay at all.
18921 * Additional PKCS1 checks.
18925 * Support the string "TLSv1" for all TLS v1 ciphers.
18929 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18930 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18934 * Fixed a few memory leaks.
18938 * Fixed various code and comment typos.
18942 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18943 bytes sent in the client random.
18945 *Edward Bishop <ebishop@spyglass.com>*
18949 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18950 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18951 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18952 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18953 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18954 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18955 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18956 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18957 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18958 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18959 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18960 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18961 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18962 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18963 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18964 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18965 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18966 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18967 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18968 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18969 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18970 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18971 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18972 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18973 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18974 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18975 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18976 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18977 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18978 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18979 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18980 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18981 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18982 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18983 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18984 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18985 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18986 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18987 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18988 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18989 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18990 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18991 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18992 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18993 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18994 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18995 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18996 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18997 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18998 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18999 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19000 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19001 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19002 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19003 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19004 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19005 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19006 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19007 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19008 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19009 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19010 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19011 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19012 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19013 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19014 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19015 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19016 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19017 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19018 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19019 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19020 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19021 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19022 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19023 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19024 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19025 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19026 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19027 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19028 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19029 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19030 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19031 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19032 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19033 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19034 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19035 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19036 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19037 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19038 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19039 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19040 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19041 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19042 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19043 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19044 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19045 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19046 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19047 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19048 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19049 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19050 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19051 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19052 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19053 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19054 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19055 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19056 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19057 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19058 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19059 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19060 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19061 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19062 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19063 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19064 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19065 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19066 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19067 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19068 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19069 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19070 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19071 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19072 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19073 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19074 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19075 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19076 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19077 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19078 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19079 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19080 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19081 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19082 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19083 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19084 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19085 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19086 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19087 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19088 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19089 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19090 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19091 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19092 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19093 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19094 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19095 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19096 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19097 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19098 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19099 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19100 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19101 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19102 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19103 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19104 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19105 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19106 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19107 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19108 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19109 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19110 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655