4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
25 listed here are only a brief description.
26 The migration guide contains more detailed information related to new features,
27 breaking changes, and mappings for the large list of deprecated functions.
29 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
31 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
33 * The signatures of the functions to get and set options on SSL and
34 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
35 Some source code changes may be required.
39 * Client-initiated renegotiation is disabled by default. To allow it, use
40 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
41 flag, or the "ClientRenegotiation" config parameter as appropriate.
45 * Add "abspath" and "includedir" pragma's to config files, to prevent,
46 or modify relative pathname inclusion.
50 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
51 validated. Please consult the README-FIPS and
52 README-PROVIDERS files, as well as the migration guide.
54 *OpenSSL team members and many third party contributors*
56 * For the key types DH and DHX the allowed settable parameters are now different.
60 * The openssl commands that read keys, certificates, and CRLs now
61 automatically detect the PEM or DER format of the input files.
63 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
65 * Added enhanced PKCS#12 APIs which accept a library context.
69 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
73 * Added support for Kernel TLS (KTLS).
75 *Boris Pismenny, John Baldwin and Andrew Gallatin*
77 * Support for RFC 5746 secure renegotiation is now required by default for
78 SSL or TLS connections to succeed.
82 * The signature of the `copy` functional parameter of the
83 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
84 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
85 the signature of the `pub_decode` functional parameter of the
86 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
87 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
91 * The error return values from some control calls (ctrl) have changed.
95 * A public key check is now performed during EVP_PKEY_derive_set_peer().
99 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
100 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
101 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
106 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
109 * The output from the command line applications may have minor
114 * The output from numerous "printing" may have minor changes.
118 * Windows thread synchronization uses read/write primitives (SRWLock) when
119 supported by the OS, otherwise CriticalSection continues to be used.
123 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
124 work on read only BIO source/sinks that do not support these functions.
125 This allows piping or redirection of a file BIO using stdin to be buffered
126 into memory. This is used internally in OSSL_DECODER_from_bio().
130 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
131 this function would return one of the values OSSL_STORE_INFO_NAME,
132 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
133 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
134 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
135 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
136 using this function should be amended to handle the changed return value.
140 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
141 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
142 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
143 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
144 contains more than one certificate identifier: This means that all
145 certificates referenced there MUST be part of the validation chain.
149 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
150 RC5, DESX and DES have been moved to the legacy provider.
154 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
155 RIPEMD-160 have been moved to the legacy provider.
159 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
164 * The deprecated functions EVP_PKEY_get0_RSA(),
165 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
166 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
167 well as the similarly named "get1" functions behave differently in
172 * A number of functions handling low-level keys or engines were deprecated
173 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
174 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
175 EVP_PKEY_get0_siphash().
179 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
180 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
181 will need to load the legacy crypto provider. This includes these PBE
182 algorithms which use this KDF:
183 - NID_pbeWithMD2AndDES_CBC
184 - NID_pbeWithMD5AndDES_CBC
185 - NID_pbeWithSHA1AndRC2_CBC
186 - NID_pbeWithMD2AndRC2_CBC
187 - NID_pbeWithMD5AndRC2_CBC
188 - NID_pbeWithSHA1AndDES_CBC
192 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
193 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
197 * The RAND_METHOD APIs have been deprecated.
201 * The SRP APIs have been deprecated.
205 * Add a compile time option to prevent the caching of provider fetched
206 algorithms. This is enabled by including the no-cached-fetch option
207 at configuration time.
211 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
212 count of PKCS12_DEFAULT_ITER.
214 *Tomáš Mráz and Sahana Prasad*
216 * The openssl speed command does not use low-level API calls anymore.
220 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
223 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
225 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
229 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
230 providers may supply their own group implementations (using either the "key
231 exchange" or the "key encapsulation" methods) which will automatically be
232 detected and used by libssl.
234 *Matt Caswell, Nicola Tuveri*
236 * The undocumented function X509_certificate_type() has been deprecated;
240 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
244 * Removed RSA padding mode for SSLv23 (which was only used for
245 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
246 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
251 * Deprecated the obsolete X9.31 RSA key generation related functions.
255 * The default key generation method for the regular 2-prime RSA keys was
256 changed to the FIPS 186-4 B.3.6 method.
260 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
264 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
268 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
269 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
271 *Rich Salz, Richard Levitte, and David von Oheimb*
273 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
277 * Deprecated `OCSP_parse_url()`.
281 * Validation of SM2 keys has been separated from the validation of regular EC
286 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
287 switches: a validation failure triggers an early exit, returning a failure
288 exit status to the parent process.
292 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
293 to ignore unknown ciphers.
297 * The `-cipher-commands` and `-digest-commands` options
298 of the command line utility `list` have been deprecated.
299 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
303 * Added convenience functions for generating asymmetric key pairs:
304 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
305 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
309 * All of the low level EC_KEY functions have been deprecated.
311 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
313 * Deprecated all the libcrypto and libssl error string loading
318 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
319 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
324 * The `-crypt` option to the `passwd` command line tool has been removed.
328 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
333 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
337 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
338 EVP_PKEY_get1_tls_encodedpoint().
342 * The security callback, which can be customised by application code, supports
343 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
344 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
348 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
349 interface. Their functionality remains unchanged.
353 * Added new option for 'openssl list', '-providers', which will display the
354 list of loaded providers, their names, version and status. It optionally
355 displays their gettable parameters.
359 * Removed EVP_PKEY_set_alias_type().
363 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
364 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
368 * Changed all "STACK" functions to be macros instead of inline functions. Macro
369 parameters are still checked for type safety at compile time via helper
374 * Remove the RAND_DRBG API
376 *Paul Dale and Matthias St. Pierre*
378 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
379 as well as actual hostnames.
383 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
384 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
385 conversely, silently ignore DTLS protocol version bounds when configuring
386 TLS-based contexts. The commands can be repeated to set bounds of both
387 types. The same applies with the corresponding "min_protocol" and
388 "max_protocol" command-line switches, in case some application uses both TLS
391 SSL_CTX instances that are created for a fixed protocol version (e.g.
392 `TLSv1_server_method()`) also silently ignore version bounds. Previously
393 attempts to apply bounds to these protocol versions would result in an
394 error. Now only the "version-flexible" SSL_CTX instances are subject to
395 limits in configuration files in command-line options.
399 * Deprecated the `ENGINE` API. Engines should be replaced with providers
404 * Reworked the recorded ERR codes to make better space for system errors.
405 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
406 given code is a system error (true) or an OpenSSL error (false).
410 * Reworked the test perl framework to better allow parallel testing.
412 *Nicola Tuveri and David von Oheimb*
414 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
415 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
419 * 'Configure' has been changed to figure out the configuration target if
420 none is given on the command line. Consequently, the 'config' script is
421 now only a mere wrapper. All documentation is changed to only mention
424 *Rich Salz and Richard Levitte*
426 * Added a library context `OSSL_LIB_CTX` that applications as well as
427 other libraries can use to form a separate context within which
428 libcrypto operations are performed.
432 * Handshake now fails if Extended Master Secret extension is dropped
437 * Dropped interactive mode from the `openssl` program.
441 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
443 *David von Oheimb and Shane Lontis*
445 * Deprecated `EC_METHOD_get_field_type()`.
449 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
450 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
451 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
455 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
459 * Add CAdES-BES signature verification support, mostly derived
460 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
462 *Filipe Raimundo da Silva*
464 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
468 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
469 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
473 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
477 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
478 EC_KEY_precompute_mult().
482 * Deprecated EC_POINTs_mul().
486 * Removed FIPS_mode() and FIPS_mode_set().
490 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
494 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
495 EC_POINT_get_Jprojective_coordinates_GFp().
499 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
500 arrays to be more easily constructed via a series of utility functions.
501 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
502 the various push functions and finally convert to a passable OSSL_PARAM
503 array using OSSL_PARAM_BLD_to_param().
507 * The security strength of SHA1 and MD5 based signatures in TLS has been
512 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
513 contain a provider side internal key.
517 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
521 * Project text documents not yet having a proper file name extension
522 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
523 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
527 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
528 have been converted to Markdown with the goal to produce documents
529 which not only look pretty when viewed online in the browser, but
530 remain well readable inside a plain text editor.
532 To achieve this goal, a 'minimalistic' Markdown style has been applied
533 which avoids formatting elements that interfere too much with the
534 reading flow in the text file. For example, it
536 * avoids [ATX headings][] and uses [setext headings][] instead
537 (which works for `<h1>` and `<h2>` headings only).
538 * avoids [inline links][] and uses [reference links][] instead.
539 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
541 [ATX headings]: https://github.github.com/gfm/#atx-headings
542 [setext headings]: https://github.github.com/gfm/#setext-headings
543 [inline links]: https://github.github.com/gfm/#inline-link
544 [reference links]: https://github.github.com/gfm/#reference-link
545 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
546 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
548 *Matthias St. Pierre*
550 * The test suite is changed to preserve results of each test recipe.
551 A new directory test-runs/ with subdirectories named like the
552 test recipes are created in the build tree for this purpose.
556 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
557 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
558 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
560 *David von Oheimb, Martin Peylo*
562 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
563 It supports arbitrary request and response content types, GET redirection,
564 TLS, connections via HTTP(S) proxies, connections and exchange via
565 user-defined BIOs (allowing implicit connections), persistent connections,
566 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
567 The legacy OCSP-focused (and only partly documented) API
568 is retained for backward compatibility, while most of it is deprecated.
572 * Added `util/check-format.pl`, a tool for checking adherence to the
573 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
574 The checks performed are incomplete and yield some false positives.
575 Still the tool should be useful for detecting most typical glitches.
579 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
580 If domain name resolution yields multiple IP addresses all of them are tried
581 after `connect()` failures.
585 * All of the low level RSA functions have been deprecated.
589 * X509 certificates signed using SHA1 are no longer allowed at security
594 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
595 modified to use PKEY APIs. These commands are now in maintenance mode
596 and no new features will be added to them.
600 * The command line utility rsautl has been deprecated.
604 * The command line utilities genrsa and rsa have been modified to use PKEY
605 APIs. They now write PKCS#8 keys by default. These commands are now in
606 maintenance mode and no new features will be added to them.
610 * All of the low level DH functions have been deprecated.
612 *Paul Dale and Matt Caswell*
614 * All of the low level DSA functions have been deprecated.
618 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
619 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
623 * Deprecated low level ECDH and ECDSA functions.
627 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
631 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
632 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
633 a new formulation to include all the things it can be used for,
634 as well as words of caution.
638 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
642 * All of the low level HMAC functions have been deprecated.
644 *Paul Dale and David von Oheimb*
646 * Over two thousand fixes were made to the documentation, including:
647 - Common options (such as -rand/-writerand, TLS version control, etc)
648 were refactored and point to newly-enhanced descriptions in openssl.pod.
649 - Added style conformance for all options (with help from Richard Levitte),
650 documented all reported missing options, added a CI build to check
651 that all options are documented and that no unimplemented options
653 - Documented some internals, such as all use of environment variables.
654 - Addressed all internal broken L<> references.
658 * All of the low level CMAC functions have been deprecated.
662 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
663 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
665 *Paul Dale and David von Oheimb*
667 * Corrected the documentation of the return values from the `EVP_DigestSign*`
668 set of functions. The documentation mentioned negative values for some
669 errors, but this was never the case, so the mention of negative values
672 Code that followed the documentation and thereby check with something
673 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
677 * All of the low level cipher functions have been deprecated.
679 *Matt Caswell and Paul Dale*
681 * Removed include/openssl/opensslconf.h.in and replaced it with
682 include/openssl/configuration.h.in, which differs in not including
683 <openssl/macros.h>. A short header include/openssl/opensslconf.h
684 was added to include both.
686 This allows internal hacks where one might need to modify the set
687 of configured macros, for example this if deprecated symbols are
688 still supposed to be available internally:
690 #include <openssl/configuration.h>
692 #undef OPENSSL_NO_DEPRECATED
693 #define OPENSSL_SUPPRESS_DEPRECATED
695 #include <openssl/macros.h>
697 This should not be used by applications that use the exported
698 symbols, as that will lead to linking errors.
702 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
703 used in exponentiation with 512-bit moduli. No EC algorithms are
704 affected. Analysis suggests that attacks against 2-prime RSA1024,
705 3-prime RSA1536, and DSA1024 as a result of this defect would be very
706 difficult to perform and are not believed likely. Attacks against DH512
707 are considered just feasible. However, for an attack the target would
708 have to re-use the DH512 private key, which is not recommended anyway.
709 Also applications directly using the low-level API BN_mod_exp may be
710 affected if they use BN_FLG_CONSTTIME.
715 * Most memory-debug features have been deprecated, and the functionality
716 replaced with no-ops.
720 * Added documentation for the STACK API.
724 * Introduced a new method type and API, OSSL_ENCODER, to
725 represent generic encoders.
729 * Added a .pragma directive to the syntax of configuration files, to
730 allow varying behavior in a supported and predictable manner.
731 Currently added pragma:
735 This allows dollar signs to be a keyword character unless it's
736 followed by a opening brace or parenthesis. This is useful for
737 platforms where dollar signs are commonly used in names, such as
738 volume names and system directory names on VMS.
742 * Added functionality to create an EVP_PKEY from user data.
746 * Change the interpretation of the '--api' configuration option to
747 mean that this is a desired API compatibility level with no
748 further meaning. The previous interpretation, that this would
749 also mean to remove all deprecated symbols up to and including
750 the given version, no requires that 'no-deprecated' is also used
751 in the configuration.
753 When building applications, the desired API compatibility level
754 can be set with the OPENSSL_API_COMPAT macro like before. For
755 API compatibility version below 3.0, the old style numerical
756 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
757 For version 3.0 and on, the value is expected to be the decimal
758 value calculated from the major and minor version like this:
760 MAJOR * 10000 + MINOR * 100
764 -DOPENSSL_API_COMPAT=30000 For 3.0
765 -DOPENSSL_API_COMPAT=30200 For 3.2
767 To hide declarations that are deprecated up to and including the
768 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
769 given when building the application as well.
773 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
774 access to certificate and CRL stores via URIs and OSSL_STORE
777 This adds the following functions:
779 - X509_LOOKUP_store()
780 - X509_STORE_load_file()
781 - X509_STORE_load_path()
782 - X509_STORE_load_store()
783 - SSL_add_store_cert_subjects_to_stack()
784 - SSL_CTX_set_default_verify_store()
785 - SSL_CTX_load_verify_file()
786 - SSL_CTX_load_verify_dir()
787 - SSL_CTX_load_verify_store()
791 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
792 The presence of this system service is determined at run-time.
796 * Added functionality to create an EVP_PKEY context based on data
797 for methods from providers. This takes an algorithm name and a
798 property query string and simply stores them, with the intent
799 that any operation that uses this context will use those strings
800 to fetch the needed methods implicitly, thereby making the port
801 of application written for pre-3.0 OpenSSL easier.
805 * The undocumented function NCONF_WIN32() has been deprecated; for
806 conversion details see the HISTORY section of doc/man5/config.pod
810 * Introduced the new functions EVP_DigestSignInit_ex() and
811 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
812 EVP_DigestVerifyUpdate() have been converted to functions. See the man
813 pages for further details.
817 * Over two thousand fixes were made to the documentation, including:
818 adding missing command flags, better style conformance, documentation
821 *Rich Salz, Richard Levitte*
823 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
824 X25519, X448, Ed25519 and Ed448.
828 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
833 * Deprecated the public definition of `ERR_STATE` as well as the function
834 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
839 * Added ERR functionality to give callers access to the stored function
840 names that have replaced the older function code based functions.
842 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
843 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
844 ERR_peek_error_all() and ERR_peek_last_error_all().
846 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
847 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
848 ERR_func_error_string().
852 * Extended testing to be verbose for failing tests only. The make variables
853 VERBOSE_FAILURE or VF can be used to enable this:
855 $ make VF=1 test # Unix
856 $ mms /macro=(VF=1) test ! OpenVMS
857 $ nmake VF=1 test # Windows
861 * Added the `-copy_extensions` option to the `x509` command for use with
862 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
863 all extensions in the request are copied to the certificate or vice versa.
865 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
867 * Added the `-copy_extensions` option to the `req` command for use with
868 `-x509`. When given with the `copy` or `copyall` argument,
869 all extensions in the certification request are copied to the certificate.
873 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
874 they generate are by default RFC 5280 compliant in the following sense:
875 There is a subjectKeyIdentifier extension with a hash value of the public key
876 and for not self-signed certs there is an authorityKeyIdentifier extension
877 with a keyIdentifier field or issuer information identifying the signing key.
878 This is done unless some configuration overrides the new default behavior,
879 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
883 * Added several checks to `X509_verify_cert()` according to requirements in
884 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
885 (which may be done by using the CLI option `-x509_strict`):
886 * The basicConstraints of CA certificates must be marked critical.
887 * CA certificates must explicitly include the keyUsage extension.
888 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
889 * The issuer name of any certificate must not be empty.
890 * The subject name of CA certs, certs with keyUsage crlSign,
891 and certs without subjectAlternativeName must not be empty.
892 * If a subjectAlternativeName extension is given it must not be empty.
893 * The signatureAlgorithm field and the cert signature must be consistent.
894 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
895 must not be marked critical.
896 * The authorityKeyIdentifier must be given for X.509v3 certs
897 unless they are self-signed.
898 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
902 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
903 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
907 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
908 used even when parsing explicit parameters, when loading a encoded key
909 or calling `EC_GROUP_new_from_ecpkparameters()`/
910 `EC_GROUP_new_from_ecparameters()`.
911 This prevents bypass of security hardening and performance gains,
912 especially for curves with specialized EC_METHODs.
913 By default, if a key encoded with explicit parameters is loaded and later
914 encoded, the output is still encoded with explicit parameters, even if
915 internally a "named" EC_GROUP is used for computation.
919 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
920 this change, EC_GROUP_set_generator would accept order and/or cofactor as
921 NULL. After this change, only the cofactor parameter can be NULL. It also
922 does some minimal sanity checks on the passed order.
927 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
928 An attack is simple, if the first CMS_recipientInfo is valid but the
929 second CMS_recipientInfo is chosen ciphertext. If the second
930 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
931 encryption key will be replaced by garbage, and the message cannot be
932 decoded, but if the RSA decryption fails, the correct encryption key is
933 used and the recipient will not notice the attack.
934 As a work around for this potential attack the length of the decrypted
935 key must be equal to the cipher default key length, in case the
936 certifiate is not given and all recipientInfo are tried out.
937 The old behaviour can be re-enabled in the CMS code by setting the
938 CMS_DEBUG_DECRYPT flag.
942 * Early start up entropy quality from the DEVRANDOM seed source has been
943 improved for older Linux systems. The RAND subsystem will wait for
944 /dev/random to be producing output before seeding from /dev/urandom.
945 The seeded state is stored for future library initialisations using
946 a system global shared memory segment. The shared memory identifier
947 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
948 the desired value. The default identifier is 114.
952 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
953 when primes for RSA keys are computed.
954 Since we previously always generated primes == 2 (mod 3) for RSA keys,
955 the 2-prime and 3-prime RSA modules were easy to distinguish, since
956 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
957 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
958 This avoids possible fingerprinting of newly generated RSA modules.
962 * Correct the extended master secret constant on EBCDIC systems. Without this
963 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
964 negotiate EMS will fail. Unfortunately this also means that TLS connections
965 between EBCDIC systems with this fix, and EBCDIC systems without this
966 fix will fail if they negotiate EMS.
970 * Changed the library initialisation so that the config file is now loaded
971 by default. This was already the case for libssl. It now occurs for both
972 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
973 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
977 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
978 where the former acts as a replacement for `ERR_put_error()`, and the
979 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
980 `ERR_raise_data()` adds more flexibility by taking a format string and
981 an arbitrary number of arguments following it, to be processed with
986 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
987 to check if a named provider is loaded and available. When called, it
988 will also activate all fallback providers if such are still present.
992 * Enforce a minimum DH modulus size of 512 bits.
996 * Changed DH parameters to generate the order q subgroup instead of 2q.
997 Previously generated DH parameters are still accepted by DH_check
998 but DH_generate_key works around that by clearing bit 0 of the
999 private key for those. This avoids leaking bit 0 of the private key.
1003 * Significantly reduce secure memory usage by the randomness pools.
1007 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1012 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1013 algorithms. An implementation of a key exchange algorithm can be obtained
1014 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1015 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1016 the older EVP_PKEY_derive_init() function. See the man pages for the new
1017 functions for further details.
1021 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1025 * Removed the function names from error messages and deprecated the
1028 * Removed NextStep support and the macro OPENSSL_UNISTD
1032 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1033 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1034 Also removed "export var as function" capability; we do not export
1035 variables, only functions.
1039 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1040 an error and 1 indicating success. In previous versions of OpenSSL this
1041 was a void type. If a key was set longer than the maximum possible this
1046 * Support SM2 signing and verification schemes with X509 certificate.
1050 * Use SHA256 as the default digest for TS query in the `ts` app.
1054 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1058 * Default cipher lists/suites are now available via a function, the
1059 #defines are deprecated.
1063 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1064 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1065 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1069 * Join the directories crypto/x509 and crypto/x509v3
1073 * Added command 'openssl kdf' that uses the EVP_KDF API.
1077 * Added command 'openssl mac' that uses the EVP_MAC API.
1081 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1082 as default directories. Also added the command 'openssl info'
1083 for scripting purposes.
1087 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1092 * Add prediction resistance to the DRBG reseeding process.
1096 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1097 mandated by IEEE Std 1619-2018.
1101 * Added newline escaping functionality to a filename when using openssl dgst.
1102 This output format is to replicate the output format found in the `*sum`
1103 checksum programs. This aims to preserve backward compatibility.
1105 *Matt Eaton, Richard Levitte, and Paul Dale*
1107 * Removed the heartbeat message in DTLS feature, as it has very
1108 little usage and doesn't seem to fulfill a valuable purpose.
1109 The configuration option is now deprecated.
1113 * Changed the output of 'openssl {digestname} < file' to display the
1114 digest name in its output.
1118 * Added a new generic trace API which provides support for enabling
1119 instrumentation through trace output.
1121 *Richard Levitte & Matthias St. Pierre*
1123 * Added build tests for C++. These are generated files that only do one
1124 thing, to include one public OpenSSL head file each. This tests that
1125 the public header files can be usefully included in a C++ application.
1127 This test isn't enabled by default. It can be enabled with the option
1128 'enable-buildtest-c++'.
1132 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1136 * Add KMAC to EVP_MAC.
1140 * Added property based algorithm implementation selection framework to
1145 * Added SCA hardening for modular field inversion in EC_GROUP through
1146 a new dedicated field_inv() pointer in EC_METHOD.
1147 This also addresses a leakage affecting conversions from projective
1148 to affine coordinates.
1150 *Billy Bob Brumley, Nicola Tuveri*
1152 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1153 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1154 those algorithms that were already supported through the EVP_PKEY API
1155 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1156 and scrypt are now wrappers that call EVP_KDF.
1160 * Build devcrypto engine as a dynamic engine.
1162 *Eneas U de Queiroz*
1164 * Add keyed BLAKE2 to EVP_MAC.
1168 * Fix a bug in the computation of the endpoint-pair shared secret used
1169 by DTLS over SCTP. This breaks interoperability with older versions
1170 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1171 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1172 interoperability with such broken implementations. However, enabling
1173 this switch breaks interoperability with correct implementations.
1175 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1176 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1180 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1184 * Changed the license to the Apache License v2.0.
1188 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1190 - Major releases (indicated by incrementing the MAJOR release number)
1191 may introduce incompatible API/ABI changes.
1192 - Minor releases (indicated by incrementing the MINOR release number)
1193 may introduce new features but retain API/ABI compatibility.
1194 - Patch releases (indicated by incrementing the PATCH number)
1195 are intended for bug fixes and other improvements of existing
1196 features only (like improving performance or adding documentation)
1197 and retain API/ABI compatibility.
1201 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1205 * Remove the 'dist' target and add a tarball building script. The
1206 'dist' target has fallen out of use, and it shouldn't be
1207 necessary to configure just to create a source distribution.
1211 * Recreate the OS390-Unix config target. It no longer relies on a
1212 special script like it did for OpenSSL pre-1.1.0.
1216 * Instead of having the source directories listed in Configure, add
1217 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1222 * Add GMAC to EVP_MAC.
1226 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1230 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1231 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1232 to facilitate the continued use of MACs through raw private keys in
1233 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1237 * Deprecate ECDH_KDF_X9_62().
1241 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1242 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1243 are retained for backwards compatibility.
1247 * AES-XTS mode now enforces that its two keys are different to mitigate
1248 the attacked described in "Efficient Instantiations of Tweakable
1249 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1250 Details of this attack can be obtained from:
1251 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1255 * Rename the object files, i.e. give them other names than in previous
1256 versions. Their names now include the name of the final product, as
1257 well as its type mnemonic (bin, lib, shlib).
1261 * Added new option for 'openssl list', '-objects', which will display the
1262 list of built in objects, i.e. OIDs with names.
1266 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1267 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1272 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1273 improves application performance by removing data copies and providing
1274 applications with zero-copy system calls such as sendfile and splice.
1278 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1282 * `PKCS12_parse` now maintains the order of the parsed certificates
1283 when outputting them via `*ca` (rather than reversing it).
1287 * Deprecated pthread fork support methods.
1291 * Added support for FFDHE key exchange in TLS 1.3.
1298 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1300 * Fixed a problem with verifying a certificate chain when using the
1301 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1302 the certificates present in a certificate chain. It is not set by default.
1304 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1305 the chain that have explicitly encoded elliptic curve parameters was added
1306 as an additional strict check.
1308 An error in the implementation of this check meant that the result of a
1309 previous check to confirm that certificates in the chain are valid CA
1310 certificates was overwritten. This effectively bypasses the check
1311 that non-CA certificates must not be able to issue other certificates.
1313 If a "purpose" has been configured then there is a subsequent opportunity
1314 for checks that the certificate is a valid CA. All of the named "purpose"
1315 values implemented in libcrypto perform this check. Therefore, where
1316 a purpose is set the certificate chain will still be rejected even when the
1317 strict flag has been used. A purpose is set by default in libssl client and
1318 server certificate verification routines, but it can be overridden or
1319 removed by an application.
1321 In order to be affected, an application must explicitly set the
1322 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1323 for the certificate verification or, in the case of TLS client or server
1324 applications, override the default purpose.
1329 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1330 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1331 renegotiation ClientHello omits the signature_algorithms extension (where it
1332 was present in the initial ClientHello), but includes a
1333 signature_algorithms_cert extension then a NULL pointer dereference will
1334 result, leading to a crash and a denial of service attack.
1336 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1337 (which is the default configuration). OpenSSL TLS clients are not impacted by
1341 *Peter Kästle and Samuel Sapalski*
1343 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1345 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1346 create a unique hash value based on the issuer and serial number data
1347 contained within an X509 certificate. However it was failing to correctly
1348 handle any errors that may occur while parsing the issuer field (which might
1349 occur if the issuer field is maliciously constructed). This may subsequently
1350 result in a NULL pointer deref and a crash leading to a potential denial of
1356 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1357 padding mode to correctly check for rollback attacks. This is considered a
1358 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1363 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1364 functions. Previously they could overflow the output length argument in some
1365 cases where the input length is close to the maximum permissable length for
1366 an integer on the platform. In such cases the return value from the function
1367 call would be 1 (indicating success), but the output length value would be
1368 negative. This could cause applications to behave incorrectly or crash.
1373 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1374 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1375 could be exploited in a side channel attack to recover the password. Since
1376 the attack is local host only this is outside of the current OpenSSL
1377 threat model and therefore no CVE is assigned.
1379 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1384 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1386 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1387 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1388 If an attacker can control both items being compared then this could lead
1389 to a possible denial of service attack. OpenSSL itself uses the
1390 GENERAL_NAME_cmp function for two purposes:
1391 1) Comparing CRL distribution point names between an available CRL and a
1392 CRL distribution point embedded in an X509 certificate
1393 2) When verifying that a timestamp response token signer matches the
1394 timestamp authority name (exposed via the API functions
1395 TS_RESP_verify_response and TS_RESP_verify_token)
1400 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1402 * Certificates with explicit curve parameters are now disallowed in
1403 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1407 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1408 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1409 conversely, silently ignore DTLS protocol version bounds when configuring
1410 TLS-based contexts. The commands can be repeated to set bounds of both
1411 types. The same applies with the corresponding "min_protocol" and
1412 "max_protocol" command-line switches, in case some application uses both TLS
1415 SSL_CTX instances that are created for a fixed protocol version (e.g.
1416 TLSv1_server_method()) also silently ignore version bounds. Previously
1417 attempts to apply bounds to these protocol versions would result in an
1418 error. Now only the "version-flexible" SSL_CTX instances are subject to
1419 limits in configuration files in command-line options.
1423 * Handshake now fails if Extended Master Secret extension is dropped
1428 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1430 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1432 * Fixed segmentation fault in SSL_check_chain()
1433 Server or client applications that call the SSL_check_chain() function
1434 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1435 dereference as a result of incorrect handling of the
1436 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1437 or unrecognised signature algorithm is received from the peer. This could
1438 be exploited by a malicious peer in a Denial of Service attack.
1443 * Added AES consttime code for no-asm configurations
1444 an optional constant time support for AES was added
1445 when building openssl for no-asm.
1446 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1447 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1448 At this time this feature is by default disabled.
1449 It will be enabled by default in 3.0.
1453 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1455 * Revert the change of EOF detection while reading in libssl to avoid
1456 regressions in applications depending on the current way of reporting
1457 the EOF. As the existing method is not fully accurate the change to
1458 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1459 branch and will be present in the 3.0 release.
1463 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1464 when primes for RSA keys are computed.
1465 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1466 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1467 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1468 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1469 This avoids possible fingerprinting of newly generated RSA modules.
1473 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1475 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1476 while reading in libssl then we would report an error back to the
1477 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1478 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1479 therefore give a hint as to what went wrong.
1483 * Check that ed25519 and ed448 are allowed by the security level. Previously
1484 signature algorithms not using an MD were not being checked that they were
1485 allowed by the security level.
1489 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1490 was not quite right. The behaviour was not consistent between resumption
1491 and normal handshakes, and also not quite consistent with historical
1492 behaviour. The behaviour in various scenarios has been clarified and
1493 it has been updated to make it match historical behaviour as closely as
1498 * *[VMS only]* The header files that the VMS compilers include automatically,
1499 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1500 that the C++ compiler doesn't understand. This is a shortcoming in the
1501 compiler, but can be worked around with `__cplusplus` guards.
1503 C++ applications that use OpenSSL libraries must be compiled using the
1504 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1505 functions. Otherwise, only functions with symbols of less than 31
1506 characters can be used, as the linker will not be able to successfully
1507 resolve symbols with longer names.
1511 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1512 The presence of this system service is determined at run-time.
1516 * Added newline escaping functionality to a filename when using openssl dgst.
1517 This output format is to replicate the output format found in the `*sum`
1518 checksum programs. This aims to preserve backward compatibility.
1520 *Matt Eaton, Richard Levitte, and Paul Dale*
1522 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1527 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1529 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1530 number generator (RNG). This was intended to include protection in the
1531 event of a fork() system call in order to ensure that the parent and child
1532 processes did not share the same RNG state. However this protection was not
1533 being used in the default case.
1535 A partial mitigation for this issue is that the output from a high
1536 precision timer is mixed into the RNG state so the likelihood of a parent
1537 and child process sharing state is significantly reduced.
1539 If an application already calls OPENSSL_init_crypto() explicitly using
1540 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1543 *Matthias St. Pierre*
1545 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1546 used even when parsing explicit parameters, when loading a encoded key
1547 or calling `EC_GROUP_new_from_ecpkparameters()`/
1548 `EC_GROUP_new_from_ecparameters()`.
1549 This prevents bypass of security hardening and performance gains,
1550 especially for curves with specialized EC_METHODs.
1551 By default, if a key encoded with explicit parameters is loaded and later
1552 encoded, the output is still encoded with explicit parameters, even if
1553 internally a "named" EC_GROUP is used for computation.
1557 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1558 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1559 NULL. After this change, only the cofactor parameter can be NULL. It also
1560 does some minimal sanity checks on the passed order.
1565 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1566 An attack is simple, if the first CMS_recipientInfo is valid but the
1567 second CMS_recipientInfo is chosen ciphertext. If the second
1568 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1569 encryption key will be replaced by garbage, and the message cannot be
1570 decoded, but if the RSA decryption fails, the correct encryption key is
1571 used and the recipient will not notice the attack.
1572 As a work around for this potential attack the length of the decrypted
1573 key must be equal to the cipher default key length, in case the
1574 certifiate is not given and all recipientInfo are tried out.
1575 The old behaviour can be re-enabled in the CMS code by setting the
1576 CMS_DEBUG_DECRYPT flag.
1581 * Early start up entropy quality from the DEVRANDOM seed source has been
1582 improved for older Linux systems. The RAND subsystem will wait for
1583 /dev/random to be producing output before seeding from /dev/urandom.
1584 The seeded state is stored for future library initialisations using
1585 a system global shared memory segment. The shared memory identifier
1586 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1587 the desired value. The default identifier is 114.
1591 * Correct the extended master secret constant on EBCDIC systems. Without this
1592 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1593 negotiate EMS will fail. Unfortunately this also means that TLS connections
1594 between EBCDIC systems with this fix, and EBCDIC systems without this
1595 fix will fail if they negotiate EMS.
1599 * Use Windows installation paths in the mingw builds
1601 Mingw isn't a POSIX environment per se, which means that Windows
1602 paths should be used for installation.
1607 * Changed DH_check to accept parameters with order q and 2q subgroups.
1608 With order 2q subgroups the bit 0 of the private key is not secret
1609 but DH_generate_key works around that by clearing bit 0 of the
1610 private key for those. This avoids leaking bit 0 of the private key.
1614 * Significantly reduce secure memory usage by the randomness pools.
1618 * Revert the DEVRANDOM_WAIT feature for Linux systems
1620 The DEVRANDOM_WAIT feature added a select() call to wait for the
1621 /dev/random device to become readable before reading from the
1622 /dev/urandom device.
1624 It turned out that this change had negative side effects on
1625 performance which were not acceptable. After some discussion it
1626 was decided to revert this feature and leave it up to the OS
1627 resp. the platform maintainer to ensure a proper initialization
1628 during early boot time.
1630 *Matthias St. Pierre*
1632 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1634 * Add build tests for C++. These are generated files that only do one
1635 thing, to include one public OpenSSL head file each. This tests that
1636 the public header files can be usefully included in a C++ application.
1638 This test isn't enabled by default. It can be enabled with the option
1639 'enable-buildtest-c++'.
1643 * Enable SHA3 pre-hashing for ECDSA and DSA.
1647 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1648 This changes the size when using the `genpkey` command when no size is given.
1649 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1650 generation commands to use 2048 bits by default.
1654 * Reorganize the manual pages to consistently have RETURN VALUES,
1655 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1656 util/fix-doc-nits accordingly.
1658 *Paul Yang, Joshua Lock*
1660 * Add the missing accessor EVP_PKEY_get0_engine()
1664 * Have commands like `s_client` and `s_server` output the signature scheme
1665 along with other cipher suite parameters when debugging.
1669 * Make OPENSSL_config() error agnostic again.
1673 * Do the error handling in RSA decryption constant time.
1677 * Prevent over long nonces in ChaCha20-Poly1305.
1679 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1680 for every encryption operation. RFC 7539 specifies that the nonce value
1681 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1682 and front pads the nonce with 0 bytes if it is less than 12
1683 bytes. However it also incorrectly allows a nonce to be set of up to 16
1684 bytes. In this case only the last 12 bytes are significant and any
1685 additional leading bytes are ignored.
1687 It is a requirement of using this cipher that nonce values are
1688 unique. Messages encrypted using a reused nonce value are susceptible to
1689 serious confidentiality and integrity attacks. If an application changes
1690 the default nonce length to be longer than 12 bytes and then makes a
1691 change to the leading bytes of the nonce expecting the new value to be a
1692 new unique nonce then such an application could inadvertently encrypt
1693 messages with a reused nonce.
1695 Additionally the ignored bytes in a long nonce are not covered by the
1696 integrity guarantee of this cipher. Any application that relies on the
1697 integrity of these ignored leading bytes of a long nonce may be further
1698 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1699 is safe because no such use sets such a long nonce value. However user
1700 applications that use this cipher directly and set a non-default nonce
1701 length to be longer than 12 bytes may be vulnerable.
1703 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1709 * Add DEVRANDOM_WAIT feature for Linux systems
1711 On older Linux systems where the getrandom() system call is not available,
1712 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1713 Contrary to getrandom(), the /dev/urandom device will not block during
1714 early boot when the kernel CSPRNG has not been seeded yet.
1716 To mitigate this known weakness, use select() to wait for /dev/random to
1717 become readable before reading from /dev/urandom.
1719 * Ensure that SM2 only uses SM3 as digest algorithm
1723 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1725 * Change the info callback signals for the start and end of a post-handshake
1726 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1727 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1728 confused by this and assume that a TLSv1.2 renegotiation has started. This
1729 can break KeyUpdate handling. Instead we no longer signal the start and end
1730 of a post handshake message exchange (although the messages themselves are
1731 still signalled). This could break some applications that were expecting
1732 the old signals. However without this KeyUpdate is not usable for many
1737 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1739 * Timing vulnerability in DSA signature generation
1741 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1742 timing side channel attack. An attacker could use variations in the signing
1743 algorithm to recover the private key.
1745 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1750 * Timing vulnerability in ECDSA signature generation
1752 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1753 timing side channel attack. An attacker could use variations in the signing
1754 algorithm to recover the private key.
1756 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1761 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1762 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1763 of two gigabytes and the error handling improved.
1765 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1766 categorized as a normal bug, not a security issue, because the DRBG reseeds
1767 automatically and is fully functional even without additional randomness
1768 provided by the application.
1770 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1772 * Add a new ClientHello callback. Provides a callback interface that gives
1773 the application the ability to adjust the nascent SSL object at the
1774 earliest stage of ClientHello processing, immediately after extensions have
1775 been collected but before they have been processed. In particular, this
1776 callback can adjust the supported TLS versions in response to the contents
1781 * Add SM2 base algorithm support.
1785 * s390x assembly pack: add (improved) hardware-support for the following
1786 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1787 aes-cfb/cfb8, aes-ecb.
1791 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1792 parameter is no longer accepted, as it leads to a corrupt table. NULL
1793 pem_str is reserved for alias entries only.
1797 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1798 step for prime curves. The new implementation is based on formulae from
1799 differential addition-and-doubling in homogeneous projective coordinates
1800 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1801 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1802 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1803 to work in projective coordinates.
1805 *Billy Bob Brumley, Nicola Tuveri*
1807 * Change generating and checking of primes so that the error rate of not
1808 being prime depends on the intended use based on the size of the input.
1809 For larger primes this will result in more rounds of Miller-Rabin.
1810 The maximal error rate for primes with more than 1080 bits is lowered
1813 *Kurt Roeckx, Annie Yousar*
1815 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1819 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1820 moving between systems, and to avoid confusion when a Windows build is
1821 done with mingw vs with MSVC. For POSIX installs, there's still a
1822 symlink or copy named 'tsget' to avoid that confusion as well.
1826 * Revert blinding in ECDSA sign and instead make problematic addition
1827 length-invariant. Switch even to fixed-length Montgomery multiplication.
1831 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1832 step for binary curves. The new implementation is based on formulae from
1833 differential addition-and-doubling in mixed Lopez-Dahab projective
1834 coordinates, modified to independently blind the operands.
1836 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1838 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1839 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1840 EC_METHODs to implement their own specialized "ladder step", to take
1841 advantage of more favorable coordinate systems or more efficient
1842 differential addition-and-doubling algorithms.
1844 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1846 * Modified the random device based seed sources to keep the relevant
1847 file descriptors open rather than reopening them on each access.
1848 This allows such sources to operate in a chroot() jail without
1849 the associated device nodes being available. This behaviour can be
1850 controlled using RAND_keep_random_devices_open().
1854 * Numerous side-channel attack mitigations have been applied. This may have
1855 performance impacts for some algorithms for the benefit of improved
1856 security. Specific changes are noted in this change log by their respective
1861 * AIX shared library support overhaul. Switch to AIX "natural" way of
1862 handling shared libraries, which means collecting shared objects of
1863 different versions and bitnesses in one common archive. This allows to
1864 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1865 doesn't affect the way 3rd party applications are linked, only how
1866 multi-version installation is managed.
1870 * Make ec_group_do_inverse_ord() more robust and available to other
1871 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1872 mitigations are applied to the fallback BN_mod_inverse().
1873 When using this function rather than BN_mod_inverse() directly, new
1874 EC cryptosystem implementations are then safer-by-default.
1878 * Add coordinate blinding for EC_POINT and implement projective
1879 coordinate blinding for generic prime curves as a countermeasure to
1880 chosen point SCA attacks.
1882 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1884 * Add blinding to ECDSA and DSA signatures to protect against side channel
1885 attacks discovered by Keegan Ryan (NCC Group).
1889 * Enforce checking in the `pkeyutl` command to ensure that the input
1890 length does not exceed the maximum supported digest length when performing
1891 a sign, verify or verifyrecover operation.
1895 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1896 I/O in combination with something like select() or poll() will hang. This
1897 can be turned off again using SSL_CTX_clear_mode().
1898 Many applications do not properly handle non-application data records, and
1899 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1900 around the problems in those applications, but can also break some.
1901 It's recommended to read the manpages about SSL_read(), SSL_write(),
1902 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1903 SSL_CTX_set_read_ahead() again.
1907 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1908 now allow empty (zero character) pass phrases.
1912 * Apply blinding to binary field modular inversion and remove patent
1913 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1917 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1918 binary and prime elliptic curves.
1922 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1923 constant time fixed point multiplication.
1927 * Revise elliptic curve scalar multiplication with timing attack
1928 defenses: ec_wNAF_mul redirects to a constant time implementation
1929 when computing fixed point and variable point multiplication (which
1930 in OpenSSL are mostly used with secret scalars in keygen, sign,
1931 ECDH derive operations).
1932 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
1935 * Updated CONTRIBUTING
1939 * Updated DRBG / RAND to request nonce and additional low entropy
1940 randomness from the system.
1942 *Matthias St. Pierre*
1944 * Updated 'openssl rehash' to use OpenSSL consistent default.
1948 * Moved the load of the ssl_conf module to libcrypto, which helps
1949 loading engines that libssl uses before libssl is initialised.
1953 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1957 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1959 *Ingo Schwarze, Rich Salz*
1961 * Added output of accepting IP address and port for 'openssl s_server'
1965 * Added a new API for TLSv1.3 ciphersuites:
1966 SSL_CTX_set_ciphersuites()
1967 SSL_set_ciphersuites()
1971 * Memory allocation failures consistently add an error to the error
1976 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1977 in libcrypto when run as setuid/setgid.
1981 * Load any config file by default when libssl is used.
1985 * Added new public header file <openssl/rand_drbg.h> and documentation
1986 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1988 *Matthias St. Pierre*
1990 * QNX support removed (cannot find contributors to get their approval
1991 for the license change).
1995 * TLSv1.3 replay protection for early data has been implemented. See the
1996 SSL_read_early_data() man page for further details.
2000 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2001 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2002 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2003 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2004 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2005 configuration has been separated out. See the ciphers man page or the
2006 SSL_CTX_set_ciphersuites() man page for more information.
2010 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2011 in responder mode now supports the new "-multi" option, which
2012 spawns the specified number of child processes to handle OCSP
2013 requests. The "-timeout" option now also limits the OCSP
2014 responder's patience to wait to receive the full client request
2015 on a newly accepted connection. Child processes are respawned
2016 as needed, and the CA index file is automatically reloaded
2017 when changed. This makes it possible to run the "ocsp" responder
2018 as a long-running service, making the OpenSSL CA somewhat more
2019 feature-complete. In this mode, most diagnostic messages logged
2020 after entering the event loop are logged via syslog(3) rather than
2025 * Added support for X448 and Ed448. Heavily based on original work by
2030 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2031 objects loaded. This adds the functions OSSL_STORE_expect() and
2032 OSSL_STORE_find() as well as needed tools to construct searches and
2033 get the search data out of them.
2037 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2038 version of OpenSSL should review their configuration settings to ensure
2039 that they are still appropriate for TLSv1.3. For further information see:
2040 <https://wiki.openssl.org/index.php/TLS1.3>
2044 * Grand redesign of the OpenSSL random generator
2046 The default RAND method now utilizes an AES-CTR DRBG according to
2047 NIST standard SP 800-90Ar1. The new random generator is essentially
2048 a port of the default random generator from the OpenSSL FIPS 2.0
2049 object module. It is a hybrid deterministic random bit generator
2050 using an AES-CTR bit stream and which seeds and reseeds itself
2051 automatically using trusted system entropy sources.
2053 Some of its new features are:
2054 - Support for multiple DRBG instances with seed chaining.
2055 - The default RAND method makes use of a DRBG.
2056 - There is a public and private DRBG instance.
2057 - The DRBG instances are fork-safe.
2058 - Keep all global DRBG instances on the secure heap if it is enabled.
2059 - The public and private DRBG instance are per thread for lock free
2062 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2064 * Changed Configure so it only says what it does and doesn't dump
2065 so much data. Instead, ./configdata.pm should be used as a script
2066 to display all sorts of configuration data.
2070 * Added processing of "make variables" to Configure.
2074 * Added SHA512/224 and SHA512/256 algorithm support.
2078 * The last traces of Netware support, first removed in 1.1.0, have
2083 * Get rid of Makefile.shared, and in the process, make the processing
2084 of certain files (rc.obj, or the .def/.map/.opt files produced from
2085 the ordinal files) more visible and hopefully easier to trace and
2086 debug (or make silent).
2090 * Make it possible to have environment variable assignments as
2091 arguments to config / Configure.
2095 * Add multi-prime RSA (RFC 8017) support.
2099 * Add SM3 implemented according to GB/T 32905-2016
2100 *Jack Lloyd <jack.lloyd@ribose.com>,*
2101 *Ronald Tse <ronald.tse@ribose.com>,*
2102 *Erick Borsboom <erick.borsboom@ribose.com>*
2104 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2105 as documented in RFC6066.
2106 Based on a patch from Tomasz Moń
2108 *Filipe Raimundo da Silva*
2110 * Add SM4 implemented according to GB/T 32907-2016.
2111 *Jack Lloyd <jack.lloyd@ribose.com>,*
2112 *Ronald Tse <ronald.tse@ribose.com>,*
2113 *Erick Borsboom <erick.borsboom@ribose.com>*
2115 * Reimplement -newreq-nodes and ERR_error_string_n; the
2116 original author does not agree with the license change.
2120 * Add ARIA AEAD TLS support.
2124 * Some macro definitions to support VS6 have been removed. Visual
2125 Studio 6 has not worked since 1.1.0
2129 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2130 without clearing the errors.
2134 * Add "atfork" functions. If building on a system that without
2135 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2136 requirements. The RAND facility now uses/requires this.
2144 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2145 not possible to disable entirely. However, it's still possible to
2146 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2149 To disable, configure with 'no-ui-console'. 'no-ui' is still
2150 possible to use as an alias. Check at compile time with the
2151 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2152 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2156 * Add a STORE module, which implements a uniform and URI based reader of
2157 stores that can contain keys, certificates, CRLs and numerous other
2158 objects. The main API is loosely based on a few stdio functions,
2159 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2160 OSSL_STORE_error and OSSL_STORE_close.
2161 The implementation uses backends called "loaders" to implement arbitrary
2162 URI schemes. There is one built in "loader" for the 'file' scheme.
2166 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2167 then adjusted to work on FreeBSD 8.4 as well.
2168 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2169 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2173 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2174 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2175 error code calls like this:
2177 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2179 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2180 that can be encoded in C. For the foreseeable future, this will only
2183 *Richard Levitte and Tim Hudson*
2185 * Removed BSD cryptodev engine.
2189 * Add a build target 'build_all_generated', to build all generated files
2190 and only that. This can be used to prepare everything that requires
2191 things like perl for a system that lacks perl and then move everything
2192 to that system and do the rest of the build there.
2196 * In the UI interface, make it possible to duplicate the user data. This
2197 can be used by engines that need to retain the data for a longer time
2198 than just the call where this user data is passed.
2202 * Ignore the '-named_curve auto' value for compatibility of applications
2205 *Tomáš Mráz <tmraz@fedoraproject.org>*
2207 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2208 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2209 alerts across multiple records (some of which could be empty). In practice
2210 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2211 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2212 support this at all. Supporting it adds significant complexity to the
2213 record layer, and its removal is unlikely to cause interoperability
2218 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2219 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2220 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2225 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2226 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2228 *Richard Levitte, Andy Polyakov*
2230 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2235 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2236 platform rather than 'mingw'.
2240 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2241 success if they are asked to add an object which already exists
2242 in the store. This change cascades to other functions which load
2243 certificates and CRLs.
2247 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2248 facilitate stack unwinding even from assembly subroutines.
2252 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2253 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2257 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2258 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2259 which is the minimum version we support.
2263 * Certificate time validation (X509_cmp_time) enforces stricter
2264 compliance with RFC 5280. Fractional seconds and timezone offsets
2265 are no longer allowed.
2269 * Add support for ARIA
2273 * s_client will now send the Server Name Indication (SNI) extension by
2274 default unless the new "-noservername" option is used. The server name is
2275 based on the host provided to the "-connect" option unless overridden by
2276 using "-servername".
2280 * Add support for SipHash
2284 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2285 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2286 prevent issues where no progress is being made and the peer continually
2287 sends unrecognised record types, using up resources processing them.
2291 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2292 using the algorithm defined in
2293 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2297 * Heartbeat support has been removed; the ABI is changed for now.
2299 *Richard Levitte, Rich Salz*
2301 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2305 * The RSA "null" method, which was partially supported to avoid patent
2306 issues, has been replaced to always returns NULL.
2313 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2315 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2316 used even when parsing explicit parameters, when loading a encoded key
2317 or calling `EC_GROUP_new_from_ecpkparameters()`/
2318 `EC_GROUP_new_from_ecparameters()`.
2319 This prevents bypass of security hardening and performance gains,
2320 especially for curves with specialized EC_METHODs.
2321 By default, if a key encoded with explicit parameters is loaded and later
2322 encoded, the output is still encoded with explicit parameters, even if
2323 internally a "named" EC_GROUP is used for computation.
2327 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2328 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2329 NULL. After this change, only the cofactor parameter can be NULL. It also
2330 does some minimal sanity checks on the passed order.
2335 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2336 An attack is simple, if the first CMS_recipientInfo is valid but the
2337 second CMS_recipientInfo is chosen ciphertext. If the second
2338 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2339 encryption key will be replaced by garbage, and the message cannot be
2340 decoded, but if the RSA decryption fails, the correct encryption key is
2341 used and the recipient will not notice the attack.
2342 As a work around for this potential attack the length of the decrypted
2343 key must be equal to the cipher default key length, in case the
2344 certifiate is not given and all recipientInfo are tried out.
2345 The old behaviour can be re-enabled in the CMS code by setting the
2346 CMS_DEBUG_DECRYPT flag.
2351 * Use Windows installation paths in the mingw builds
2353 Mingw isn't a POSIX environment per se, which means that Windows
2354 paths should be used for installation.
2359 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2361 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2362 This changes the size when using the `genpkey` command when no size is given.
2363 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2364 generation commands to use 2048 bits by default.
2368 * Prevent over long nonces in ChaCha20-Poly1305.
2370 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2371 for every encryption operation. RFC 7539 specifies that the nonce value
2372 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2373 and front pads the nonce with 0 bytes if it is less than 12
2374 bytes. However it also incorrectly allows a nonce to be set of up to 16
2375 bytes. In this case only the last 12 bytes are significant and any
2376 additional leading bytes are ignored.
2378 It is a requirement of using this cipher that nonce values are
2379 unique. Messages encrypted using a reused nonce value are susceptible to
2380 serious confidentiality and integrity attacks. If an application changes
2381 the default nonce length to be longer than 12 bytes and then makes a
2382 change to the leading bytes of the nonce expecting the new value to be a
2383 new unique nonce then such an application could inadvertently encrypt
2384 messages with a reused nonce.
2386 Additionally the ignored bytes in a long nonce are not covered by the
2387 integrity guarantee of this cipher. Any application that relies on the
2388 integrity of these ignored leading bytes of a long nonce may be further
2389 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2390 is safe because no such use sets such a long nonce value. However user
2391 applications that use this cipher directly and set a non-default nonce
2392 length to be longer than 12 bytes may be vulnerable.
2394 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2400 * Added SCA hardening for modular field inversion in EC_GROUP through
2401 a new dedicated field_inv() pointer in EC_METHOD.
2402 This also addresses a leakage affecting conversions from projective
2403 to affine coordinates.
2405 *Billy Bob Brumley, Nicola Tuveri*
2407 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2408 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2412 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2416 * Remove the 'dist' target and add a tarball building script. The
2417 'dist' target has fallen out of use, and it shouldn't be
2418 necessary to configure just to create a source distribution.
2422 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2424 * Timing vulnerability in DSA signature generation
2426 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2427 timing side channel attack. An attacker could use variations in the signing
2428 algorithm to recover the private key.
2430 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2435 * Timing vulnerability in ECDSA signature generation
2437 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2438 timing side channel attack. An attacker could use variations in the signing
2439 algorithm to recover the private key.
2441 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2446 * Add coordinate blinding for EC_POINT and implement projective
2447 coordinate blinding for generic prime curves as a countermeasure to
2448 chosen point SCA attacks.
2450 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2452 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2454 * Client DoS due to large DH parameter
2456 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2457 malicious server can send a very large prime value to the client. This will
2458 cause the client to spend an unreasonably long period of time generating a
2459 key for this prime resulting in a hang until the client has finished. This
2460 could be exploited in a Denial Of Service attack.
2462 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2467 * Cache timing vulnerability in RSA Key Generation
2469 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2470 a cache timing side channel attack. An attacker with sufficient access to
2471 mount cache timing attacks during the RSA key generation process could
2472 recover the private key.
2474 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2475 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2480 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2481 parameter is no longer accepted, as it leads to a corrupt table. NULL
2482 pem_str is reserved for alias entries only.
2486 * Revert blinding in ECDSA sign and instead make problematic addition
2487 length-invariant. Switch even to fixed-length Montgomery multiplication.
2491 * Change generating and checking of primes so that the error rate of not
2492 being prime depends on the intended use based on the size of the input.
2493 For larger primes this will result in more rounds of Miller-Rabin.
2494 The maximal error rate for primes with more than 1080 bits is lowered
2497 *Kurt Roeckx, Annie Yousar*
2499 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2503 * Add blinding to ECDSA and DSA signatures to protect against side channel
2504 attacks discovered by Keegan Ryan (NCC Group).
2508 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2509 now allow empty (zero character) pass phrases.
2513 * Certificate time validation (X509_cmp_time) enforces stricter
2514 compliance with RFC 5280. Fractional seconds and timezone offsets
2515 are no longer allowed.
2519 * Fixed a text canonicalisation bug in CMS
2521 Where a CMS detached signature is used with text content the text goes
2522 through a canonicalisation process first prior to signing or verifying a
2523 signature. This process strips trailing space at the end of lines, converts
2524 line terminators to CRLF and removes additional trailing line terminators
2525 at the end of a file. A bug in the canonicalisation process meant that
2526 some characters, such as form-feed, were incorrectly treated as whitespace
2527 and removed. This is contrary to the specification (RFC5485). This fix
2528 could mean that detached text data signed with an earlier version of
2529 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2530 signed with a fixed OpenSSL may fail to verify with an earlier version of
2531 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2532 and use the "-binary" flag (for the "cms" command line application) or set
2533 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2537 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2539 * Constructed ASN.1 types with a recursive definition could exceed the stack
2541 Constructed ASN.1 types with a recursive definition (such as can be found
2542 in PKCS7) could eventually exceed the stack given malicious input with
2543 excessive recursion. This could result in a Denial Of Service attack. There
2544 are no such structures used within SSL/TLS that come from untrusted sources
2545 so this is considered safe.
2547 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2553 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2555 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2556 effectively reduced to only comparing the least significant bit of each
2557 byte. This allows an attacker to forge messages that would be considered as
2558 authenticated in an amount of tries lower than that guaranteed by the
2559 security claims of the scheme. The module can only be compiled by the
2560 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2562 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2568 * Add a build target 'build_all_generated', to build all generated files
2569 and only that. This can be used to prepare everything that requires
2570 things like perl for a system that lacks perl and then move everything
2571 to that system and do the rest of the build there.
2575 * Backport SSL_OP_NO_RENGOTIATION
2577 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2578 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2579 changes this is no longer possible in 1.1.0. Therefore the new
2580 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2581 1.1.0 to provide equivalent functionality.
2583 Note that if an application built against 1.1.0h headers (or above) is run
2584 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2585 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2589 * Removed the OS390-Unix config target. It relied on a script that doesn't
2594 * rsaz_1024_mul_avx2 overflow bug on x86_64
2596 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2597 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2598 Analysis suggests that attacks against RSA and DSA as a result of this
2599 defect would be very difficult to perform and are not believed likely.
2600 Attacks against DH1024 are considered just feasible, because most of the
2601 work necessary to deduce information about a private key may be performed
2602 offline. The amount of resources required for such an attack would be
2603 significant. However, for an attack on TLS to be meaningful, the server
2604 would have to share the DH1024 private key among multiple clients, which is
2605 no longer an option since CVE-2016-0701.
2607 This only affects processors that support the AVX2 but not ADX extensions
2608 like Intel Haswell (4th generation).
2610 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2611 was originally found via the OSS-Fuzz project.
2616 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2618 * bn_sqrx8x_internal carry bug on x86_64
2620 There is a carry propagating bug in the x86_64 Montgomery squaring
2621 procedure. No EC algorithms are affected. Analysis suggests that attacks
2622 against RSA and DSA as a result of this defect would be very difficult to
2623 perform and are not believed likely. Attacks against DH are considered just
2624 feasible (although very difficult) because most of the work necessary to
2625 deduce information about a private key may be performed offline. The amount
2626 of resources required for such an attack would be very significant and
2627 likely only accessible to a limited number of attackers. An attacker would
2628 additionally need online access to an unpatched system using the target
2629 private key in a scenario with persistent DH parameters and a private
2630 key that is shared between multiple clients.
2632 This only affects processors that support the BMI1, BMI2 and ADX extensions
2633 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2635 This issue was reported to OpenSSL by the OSS-Fuzz project.
2640 * Malformed X.509 IPAddressFamily could cause OOB read
2642 If an X.509 certificate has a malformed IPAddressFamily extension,
2643 OpenSSL could do a one-byte buffer overread. The most likely result
2644 would be an erroneous display of the certificate in text format.
2646 This issue was reported to OpenSSL by the OSS-Fuzz project.
2651 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2653 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2654 platform rather than 'mingw'.
2658 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2659 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2660 which is the minimum version we support.
2664 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2666 * Encrypt-Then-Mac renegotiation crash
2668 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2669 negotiated where it was not in the original handshake (or vice-versa) then
2670 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2671 and servers are affected.
2673 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2678 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2680 * Truncated packet could crash via OOB read
2682 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2683 cipher is being used, then a truncated packet can cause that host to
2684 perform an out-of-bounds read, usually resulting in a crash.
2686 This issue was reported to OpenSSL by Robert Święcki of Google.
2691 * Bad (EC)DHE parameters cause a client crash
2693 If a malicious server supplies bad parameters for a DHE or ECDHE key
2694 exchange then this can result in the client attempting to dereference a
2695 NULL pointer leading to a client crash. This could be exploited in a Denial
2698 This issue was reported to OpenSSL by Guido Vranken.
2703 * BN_mod_exp may produce incorrect results on x86_64
2705 There is a carry propagating bug in the x86_64 Montgomery squaring
2706 procedure. No EC algorithms are affected. Analysis suggests that attacks
2707 against RSA and DSA as a result of this defect would be very difficult to
2708 perform and are not believed likely. Attacks against DH are considered just
2709 feasible (although very difficult) because most of the work necessary to
2710 deduce information about a private key may be performed offline. The amount
2711 of resources required for such an attack would be very significant and
2712 likely only accessible to a limited number of attackers. An attacker would
2713 additionally need online access to an unpatched system using the target
2714 private key in a scenario with persistent DH parameters and a private
2715 key that is shared between multiple clients. For example this can occur by
2716 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2717 similar to CVE-2015-3193 but must be treated as a separate problem.
2719 This issue was reported to OpenSSL by the OSS-Fuzz project.
2724 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2726 * ChaCha20/Poly1305 heap-buffer-overflow
2728 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2729 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2730 crash. This issue is not considered to be exploitable beyond a DoS.
2732 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2737 * CMS Null dereference
2739 Applications parsing invalid CMS structures can crash with a NULL pointer
2740 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2741 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2742 structure callback if an attempt is made to free certain invalid encodings.
2743 Only CHOICE structures using a callback which do not handle NULL value are
2746 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2751 * Montgomery multiplication may produce incorrect results
2753 There is a carry propagating bug in the Broadwell-specific Montgomery
2754 multiplication procedure that handles input lengths divisible by, but
2755 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2756 and DH private keys are impossible. This is because the subroutine in
2757 question is not used in operations with the private key itself and an input
2758 of the attacker's direct choice. Otherwise the bug can manifest itself as
2759 transient authentication and key negotiation failures or reproducible
2760 erroneous outcome of public-key operations with specially crafted input.
2761 Among EC algorithms only Brainpool P-512 curves are affected and one
2762 presumably can attack ECDH key negotiation. Impact was not analyzed in
2763 detail, because pre-requisites for attack are considered unlikely. Namely
2764 multiple clients have to choose the curve in question and the server has to
2765 share the private key among them, neither of which is default behaviour.
2766 Even then only clients that chose the curve will be affected.
2768 This issue was publicly reported as transient failures and was not
2769 initially recognized as a security issue. Thanks to Richard Morgan for
2770 providing reproducible case.
2775 * Removed automatic addition of RPATH in shared libraries and executables,
2776 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2780 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2782 * Fix Use After Free for large message sizes
2784 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2785 message larger than approx 16k is received then the underlying buffer to
2786 store the incoming message is reallocated and moved. Unfortunately a
2787 dangling pointer to the old location is left which results in an attempt to
2788 write to the previously freed location. This is likely to result in a
2789 crash, however it could potentially lead to execution of arbitrary code.
2791 This issue only affects OpenSSL 1.1.0a.
2793 This issue was reported to OpenSSL by Robert Święcki.
2798 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2800 * OCSP Status Request extension unbounded memory growth
2802 A malicious client can send an excessively large OCSP Status Request
2803 extension. If that client continually requests renegotiation, sending a
2804 large OCSP Status Request extension each time, then there will be unbounded
2805 memory growth on the server. This will eventually lead to a Denial Of
2806 Service attack through memory exhaustion. Servers with a default
2807 configuration are vulnerable even if they do not support OCSP. Builds using
2808 the "no-ocsp" build time option are not affected.
2810 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2815 * SSL_peek() hang on empty record
2817 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2818 sends an empty record. This could be exploited by a malicious peer in a
2819 Denial Of Service attack.
2821 This issue was reported to OpenSSL by Alex Gaynor.
2826 * Excessive allocation of memory in tls_get_message_header() and
2827 dtls1_preprocess_fragment()
2829 A (D)TLS message includes 3 bytes for its length in the header for the
2830 message. This would allow for messages up to 16Mb in length. Messages of
2831 this length are excessive and OpenSSL includes a check to ensure that a
2832 peer is sending reasonably sized messages in order to avoid too much memory
2833 being consumed to service a connection. A flaw in the logic of version
2834 1.1.0 means that memory for the message is allocated too early, prior to
2835 the excessive message length check. Due to way memory is allocated in
2836 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2837 to service a connection. This could lead to a Denial of Service through
2838 memory exhaustion. However, the excessive message length check still takes
2839 place, and this would cause the connection to immediately fail. Assuming
2840 that the application calls SSL_free() on the failed connection in a timely
2841 manner then the 21Mb of allocated memory will then be immediately freed
2842 again. Therefore the excessive memory allocation will be transitory in
2843 nature. This then means that there is only a security impact if:
2845 1) The application does not call SSL_free() in a timely manner in the event
2846 that the connection fails
2848 2) The application is working in a constrained environment where there is
2849 very little free memory
2851 3) The attacker initiates multiple connection attempts such that there are
2852 multiple connections in a state where memory has been allocated for the
2853 connection; SSL_free() has not yet been called; and there is insufficient
2854 memory to service the multiple requests.
2856 Except in the instance of (1) above any Denial Of Service is likely to be
2857 transitory because as soon as the connection fails the memory is
2858 subsequently freed again in the SSL_free() call. However there is an
2859 increased risk during this period of application crashes due to the lack of
2860 memory - which would then mean a more serious Denial of Service.
2862 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2863 (CVE-2016-6307 and CVE-2016-6308)
2867 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2868 had to be removed. Primary reason is that vendor assembler can't
2869 assemble our modules with -KPIC flag. As result it, assembly
2870 support, was not even available as option. But its lack means
2871 lack of side-channel resistant code, which is incompatible with
2872 security by todays standards. Fortunately gcc is readily available
2873 prepackaged option, which we firmly point at...
2877 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2879 * Windows command-line tool supports UTF-8 opt-in option for arguments
2880 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2881 (to any value) allows Windows user to access PKCS#12 file generated
2882 with Windows CryptoAPI and protected with non-ASCII password, as well
2883 as files generated under UTF-8 locale on Linux also protected with
2888 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2889 have been disabled by default and removed from DEFAULT, just like RC4.
2890 See the RC4 item below to re-enable both.
2894 * The method for finding the storage location for the Windows RAND seed file
2895 has changed. First we check %RANDFILE%. If that is not set then we check
2896 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2897 all else fails we fall back to C:\.
2901 * The EVP_EncryptUpdate() function has had its return type changed from void
2902 to int. A return of 0 indicates and error while a return of 1 indicates
2907 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2908 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2909 off the constant time implementation for RSA, DSA and DH have been made
2910 no-ops and deprecated.
2914 * Windows RAND implementation was simplified to only get entropy by
2915 calling CryptGenRandom(). Various other RAND-related tickets
2918 *Joseph Wylie Yandle, Rich Salz*
2920 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2921 and `OPENSSL_LH_`, respectively. The old names are available
2922 with API compatibility. They new names are now completely documented.
2926 * Unify TYPE_up_ref(obj) methods signature.
2927 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2928 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2929 int (instead of void) like all others TYPE_up_ref() methods.
2930 So now these methods also check the return value of CRYPTO_atomic_add(),
2931 and the validity of object reference counter.
2933 *fdasilvayy@gmail.com*
2935 * With Windows Visual Studio builds, the .pdb files are installed
2936 alongside the installed libraries and executables. For a static
2937 library installation, ossl_static.pdb is the associate compiler
2938 generated .pdb file to be used when linking programs.
2942 * Remove openssl.spec. Packaging files belong with the packagers.
2946 * Automatic Darwin/OSX configuration has had a refresh, it will now
2947 recognise x86_64 architectures automatically. You can still decide
2948 to build for a different bitness with the environment variable
2949 KERNEL_BITS (can be 32 or 64), for example:
2951 KERNEL_BITS=32 ./config
2955 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2956 256 bit AES and HMAC with SHA256.
2960 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2964 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2968 * To enable users to have their own config files and build file templates,
2969 Configure looks in the directory indicated by the environment variable
2970 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2971 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2972 name and is used as is.
2976 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2977 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2978 X509_CERT_FILE_CTX was removed.
2982 * "shared" builds are now the default. To create only static libraries use
2983 the "no-shared" Configure option.
2987 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2988 All of these option have not worked for some while and are fundamental
2993 * Make various cleanup routines no-ops and mark them as deprecated. Most
2994 global cleanup functions are no longer required because they are handled
2995 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2996 Explicitly de-initing can cause problems (e.g. where a library that uses
2997 OpenSSL de-inits, but an application is still using it). The affected
2998 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2999 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3000 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3001 COMP_zlib_cleanup().
3005 * --strict-warnings no longer enables runtime debugging options
3006 such as REF_DEBUG. Instead, debug options are automatically
3007 enabled with '--debug' builds.
3009 *Andy Polyakov, Emilia Käsper*
3011 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3012 have been moved out of the public header files. New functions for managing
3013 these have been added.
3017 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3018 objects have been moved out of the public header files. New
3019 functions for managing these have been added.
3023 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3024 have been moved out of the public header files. New functions for managing
3025 these have been added.
3029 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3030 moved out of the public header files. New functions for managing these
3035 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3039 * Removed the mk1mf build scripts.
3043 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3044 it is always safe to #include a header now.
3048 * Removed the aged BC-32 config and all its supporting scripts
3052 * Removed support for Ultrix, Netware, and OS/2.
3056 * Add support for HKDF.
3058 *Alessandro Ghedini*
3060 * Add support for blake2b and blake2s
3064 * Added support for "pipelining". Ciphers that have the
3065 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3066 encryptions/decryptions simultaneously. There are currently no built-in
3067 ciphers with this property but the expectation is that engines will be able
3068 to offer it to significantly improve throughput. Support has been extended
3069 into libssl so that multiple records for a single connection can be
3070 processed in one go (for >=TLS 1.1).
3074 * Added the AFALG engine. This is an async capable engine which is able to
3075 offload work to the Linux kernel. In this initial version it only supports
3076 AES128-CBC. The kernel must be version 4.1.0 or greater.
3080 * OpenSSL now uses a new threading API. It is no longer necessary to
3081 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3082 are two supported threading models: pthreads and windows threads. It is
3083 also possible to configure OpenSSL at compile time for "no-threads". The
3084 old threading API should no longer be used. The functions have been
3085 replaced with "no-op" compatibility macros.
3087 *Alessandro Ghedini, Matt Caswell*
3089 * Modify behavior of ALPN to invoke callback after SNI/servername
3090 callback, such that updates to the SSL_CTX affect ALPN.
3094 * Add SSL_CIPHER queries for authentication and key-exchange.
3098 * Changes to the DEFAULT cipherlist:
3099 - Prefer (EC)DHE handshakes over plain RSA.
3100 - Prefer AEAD ciphers over legacy ciphers.
3101 - Prefer ECDSA over RSA when both certificates are available.
3102 - Prefer TLSv1.2 ciphers/PRF.
3103 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3108 * Change the ECC default curve list to be this, in order: x25519,
3109 secp256r1, secp521r1, secp384r1.
3113 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3114 disabled by default. They can be re-enabled using the
3115 enable-weak-ssl-ciphers option to Configure.
3119 * If the server has ALPN configured, but supports no protocols that the
3120 client advertises, send a fatal "no_application_protocol" alert.
3121 This behaviour is SHALL in RFC 7301, though it isn't universally
3122 implemented by other servers.
3126 * Add X25519 support.
3127 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3128 for public and private key encoding using the format documented in
3129 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3130 key generation and key derivation.
3132 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3137 * Deprecate SRP_VBASE_get_by_user.
3138 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3139 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3140 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3141 seed, even if the seed is configured.
3143 Users should use SRP_VBASE_get1_by_user instead. Note that in
3144 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3145 also that even though configuring the SRP seed attempts to hide
3146 invalid usernames by continuing the handshake with fake
3147 credentials, this behaviour is not constant time and no strong
3148 guarantees are made that the handshake is indistinguishable from
3149 that of a valid user.
3153 * Configuration change; it's now possible to build dynamic engines
3154 without having to build shared libraries and vice versa. This
3155 only applies to the engines in `engines/`, those in `crypto/engine/`
3156 will always be built into libcrypto (i.e. "static").
3158 Building dynamic engines is enabled by default; to disable, use
3159 the configuration option "disable-dynamic-engine".
3161 The only requirements for building dynamic engines are the
3162 presence of the DSO module and building with position independent
3163 code, so they will also automatically be disabled if configuring
3164 with "disable-dso" or "disable-pic".
3166 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3167 are also taken away from openssl/opensslconf.h, as they are
3172 * Configuration change; if there is a known flag to compile
3173 position independent code, it will always be applied on the
3174 libcrypto and libssl object files, and never on the application
3175 object files. This means other libraries that use routines from
3176 libcrypto / libssl can be made into shared libraries regardless
3177 of how OpenSSL was configured.
3179 If this isn't desirable, the configuration options "disable-pic"
3180 or "no-pic" can be used to disable the use of PIC. This will
3181 also disable building shared libraries and dynamic engines.
3185 * Removed JPAKE code. It was experimental and has no wide use.
3189 * The INSTALL_PREFIX Makefile variable has been renamed to
3190 DESTDIR. That makes for less confusion on what this variable
3191 is for. Also, the configuration option --install_prefix is
3196 * Heartbeat for TLS has been removed and is disabled by default
3197 for DTLS; configure with enable-heartbeats. Code that uses the
3198 old #define's might need to be updated.
3200 *Emilia Käsper, Rich Salz*
3202 * Rename REF_CHECK to REF_DEBUG.
3206 * New "unified" build system
3208 The "unified" build system is aimed to be a common system for all
3209 platforms we support. With it comes new support for VMS.
3211 This system builds supports building in a different directory tree
3212 than the source tree. It produces one Makefile (for unix family
3213 or lookalikes), or one descrip.mms (for VMS).
3215 The source of information to make the Makefile / descrip.mms is
3216 small files called 'build.info', holding the necessary
3217 information for each directory with source to compile, and a
3218 template in Configurations, like unix-Makefile.tmpl or
3221 With this change, the library names were also renamed on Windows
3222 and on VMS. They now have names that are closer to the standard
3223 on Unix, and include the major version number, and in certain
3224 cases, the architecture they are built for. See "Notes on shared
3225 libraries" in INSTALL.
3227 We rely heavily on the perl module Text::Template.
3231 * Added support for auto-initialisation and de-initialisation of the library.
3232 OpenSSL no longer requires explicit init or deinit routines to be called,
3233 except in certain circumstances. See the OPENSSL_init_crypto() and
3234 OPENSSL_init_ssl() man pages for further information.
3238 * The arguments to the DTLSv1_listen function have changed. Specifically the
3239 "peer" argument is now expected to be a BIO_ADDR object.
3241 * Rewrite of BIO networking library. The BIO library lacked consistent
3242 support of IPv6, and adding it required some more extensive
3243 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3244 which hold all types of addresses and chains of address information.
3245 It also introduces a new API, with functions like BIO_socket,
3246 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3247 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3248 have been adapted accordingly.
3252 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3257 * CRIME protection: disable compression by default, even if OpenSSL is
3258 compiled with zlib enabled. Applications can still enable compression
3259 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3260 using the SSL_CONF library to configure compression.
3264 * The signature of the session callback configured with
3265 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3266 was explicitly marked as `const unsigned char*` instead of
3271 * Always DPURIFY. Remove the use of uninitialized memory in the
3272 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3276 * Removed many obsolete configuration items, including
3277 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3278 MD2_CHAR, MD2_INT, MD2_LONG
3280 IDEA_SHORT, IDEA_LONG
3281 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3283 *Rich Salz, with advice from Andy Polyakov*
3285 * Many BN internals have been moved to an internal header file.
3287 *Rich Salz with help from Andy Polyakov*
3289 * Configuration and writing out the results from it has changed.
3290 Files such as Makefile include/openssl/opensslconf.h and are now
3291 produced through general templates, such as Makefile.in and
3292 crypto/opensslconf.h.in and some help from the perl module
3295 Also, the center of configuration information is no longer
3296 Makefile. Instead, Configure produces a perl module in
3297 configdata.pm which holds most of the config data (in the hash
3298 table %config), the target data that comes from the target
3299 configuration in one of the `Configurations/*.conf` files (in
3304 * To clarify their intended purposes, the Configure options
3305 --prefix and --openssldir change their semantics, and become more
3306 straightforward and less interdependent.
3308 --prefix shall be used exclusively to give the location INSTALLTOP
3309 where programs, scripts, libraries, include files and manuals are
3310 going to be installed. The default is now /usr/local.
3312 --openssldir shall be used exclusively to give the default
3313 location OPENSSLDIR where certificates, private keys, CRLs are
3314 managed. This is also where the default openssl.cnf gets
3316 If the directory given with this option is a relative path, the
3317 values of both the --prefix value and the --openssldir value will
3318 be combined to become OPENSSLDIR.
3319 The default for --openssldir is INSTALLTOP/ssl.
3321 Anyone who uses --openssldir to specify where OpenSSL is to be
3322 installed MUST change to use --prefix instead.
3326 * The GOST engine was out of date and therefore it has been removed. An up
3327 to date GOST engine is now being maintained in an external repository.
3328 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3329 support for GOST ciphersuites (these are only activated if a GOST engine
3334 * EGD is no longer supported by default; use enable-egd when
3337 *Ben Kaduk and Rich Salz*
3339 * The distribution now has Makefile.in files, which are used to
3340 create Makefile's when Configure is run. *Configure must be run
3341 before trying to build now.*
3345 * The return value for SSL_CIPHER_description() for error conditions
3350 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3352 Obtaining and performing DNSSEC validation of TLSA records is
3353 the application's responsibility. The application provides
3354 the TLSA records of its choice to OpenSSL, and these are then
3355 used to authenticate the peer.
3357 The TLSA records need not even come from DNS. They can, for
3358 example, be used to implement local end-entity certificate or
3359 trust-anchor "pinning", where the "pin" data takes the form
3360 of TLSA records, which can augment or replace verification
3361 based on the usual WebPKI public certification authorities.
3365 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3366 continues to support deprecated interfaces in default builds.
3367 However, applications are strongly advised to compile their
3368 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3369 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3370 or the 1.1.0 releases.
3372 In environments in which all applications have been ported to
3373 not use any deprecated interfaces OpenSSL's Configure script
3374 should be used with the --api=1.1.0 option to entirely remove
3375 support for the deprecated features from the library and
3376 unconditionally disable them in the installed headers.
3377 Essentially the same effect can be achieved with the "no-deprecated"
3378 argument to Configure, except that this will always restrict
3379 the build to just the latest API, rather than a fixed API
3382 As applications are ported to future revisions of the API,
3383 they should update their compile-time OPENSSL_API_COMPAT define
3384 accordingly, but in most cases should be able to continue to
3385 compile with later releases.
3387 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3388 0x10000000L and 0x00908000L, respectively. However those
3389 versions did not support the OPENSSL_API_COMPAT feature, and
3390 so applications are not typically tested for explicit support
3391 of just the undeprecated features of either release.
3395 * Add support for setting the minimum and maximum supported protocol.
3396 It can bet set via the SSL_set_min_proto_version() and
3397 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3398 MaxProtocol. It's recommended to use the new APIs to disable
3399 protocols instead of disabling individual protocols using
3400 SSL_set_options() or SSL_CONF's Protocol. This change also
3401 removes support for disabling TLS 1.2 in the OpenSSL TLS
3402 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3406 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3410 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3411 and integrates ECDSA and ECDH functionality into EC. Implementations can
3412 now redirect key generation and no longer need to convert to or from
3415 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3416 include the ec.h header file instead.
3420 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3421 ciphers who are no longer supported and drops support the ephemeral RSA key
3422 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3426 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3427 opaque. For HMAC_CTX, the following constructors and destructors
3430 HMAC_CTX *HMAC_CTX_new(void);
3431 void HMAC_CTX_free(HMAC_CTX *ctx);
3433 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3434 destroy such methods has been added. See EVP_MD_meth_new(3) and
3435 EVP_CIPHER_meth_new(3) for documentation.
3438 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3439 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3440 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3441 an already created structure.
3442 2) For consistency with the majority of our object creators and
3443 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3444 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3445 for deprecated builds.
3449 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3450 cryptographic operations to be performed asynchronously as long as an
3451 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3452 further details. Libssl has also had this capability integrated with the
3453 introduction of the new mode SSL_MODE_ASYNC and associated error
3454 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3455 pages. This work was developed in partnership with Intel Corp.
3459 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3460 always enabled now. If you want to disable the support you should
3461 exclude it using the list of supported ciphers. This also means that the
3462 "-no_ecdhe" option has been removed from s_server.
3466 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3467 SSL_{CTX_}set1_curves() which can set a list.
3471 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3472 curve you want to support using SSL_{CTX_}set1_curves().
3476 * State machine rewrite. The state machine code has been significantly
3477 refactored in order to remove much duplication of code and solve issues
3478 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3479 further details). This change does have some associated API changes.
3480 Notably the SSL_state() function has been removed and replaced by
3481 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3482 SSL_set_state() has been removed altogether. The previous handshake states
3483 defined in ssl.h and ssl3.h have also been removed.
3487 * All instances of the string "ssleay" in the public API were replaced
3488 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3489 Some error codes related to internal RSA_eay API's were renamed.
3493 * The demo files in crypto/threads were moved to demo/threads.
3497 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3500 *Matt Caswell, Rich Salz*
3502 * New ASN.1 embed macro.
3504 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3505 structure is not allocated: it is part of the parent. That is instead of
3513 This reduces memory fragmentation and make it impossible to accidentally
3514 set a mandatory field to NULL.
3516 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3517 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3518 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3523 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3527 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3528 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3529 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3530 DES and RC4 ciphersuites.
3534 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3535 This changes the decoding behaviour for some invalid messages,
3536 though the change is mostly in the more lenient direction, and
3537 legacy behaviour is preserved as much as possible.
3541 * Fix no-stdio build.
3542 *David Woodhouse <David.Woodhouse@intel.com> and also*
3543 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3545 * New testing framework
3546 The testing framework has been largely rewritten and is now using
3547 perl and the perl modules Test::Harness and an extended variant of
3548 Test::More called OpenSSL::Test to do its work. All test scripts in
3549 test/ have been rewritten into test recipes, and all direct calls to
3550 executables in test/Makefile have become individual recipes using the
3551 simplified testing OpenSSL::Test::Simple.
3553 For documentation on our testing modules, do:
3555 perldoc test/testlib/OpenSSL/Test/Simple.pm
3556 perldoc test/testlib/OpenSSL/Test.pm
3560 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3561 are used; the latter aborts on memory leaks (usually checked on exit).
3562 Some undocumented "set malloc, etc., hooks" functions were removed
3563 and others were changed. All are now documented.
3567 * In DSA_generate_parameters_ex, if the provided seed is too short,
3570 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3572 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3573 from RFC4279, RFC4785, RFC5487, RFC5489.
3575 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3576 original RSA_PSK patch.
3580 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3581 era flag was never set throughout the codebase (only read). Also removed
3582 SSL3_FLAGS_POP_BUFFER which was only used if
3583 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3587 * Changed the default name options in the "ca", "crl", "req" and "x509"
3588 to be "oneline" instead of "compat".
3592 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3593 not aware of clients that still exhibit this bug, and the workaround
3594 hasn't been working properly for a while.
3598 * The return type of BIO_number_read() and BIO_number_written() as well as
3599 the corresponding num_read and num_write members in the BIO structure has
3600 changed from unsigned long to uint64_t. On platforms where an unsigned
3601 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3606 * Given the pervasive nature of TLS extensions it is inadvisable to run
3607 OpenSSL without support for them. It also means that maintaining
3608 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3609 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3613 * Removed support for the two export grade static DH ciphersuites
3614 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3615 were newly added (along with a number of other static DH ciphersuites) to
3616 1.0.2. However the two export ones have *never* worked since they were
3617 introduced. It seems strange in any case to be adding new export
3618 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3622 * Version negotiation has been rewritten. In particular SSLv23_method(),
3623 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3624 and turned into macros which simply call the new preferred function names
3625 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3626 should use the new names instead. Also as part of this change the ssl23.h
3627 header file has been removed.
3631 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3632 code and the associated standard is no longer considered fit-for-purpose.
3636 * RT2547 was closed. When generating a private key, try to make the
3637 output file readable only by the owner. This behavior change might
3638 be noticeable when interacting with other software.
3640 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3645 * Added HTTP GET support to the ocsp command.
3649 * Changed default digest for the dgst and enc commands from MD5 to
3654 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3658 * Added support for TLS extended master secret from
3659 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3660 initial patch which was a great help during development.
3664 * All libssl internal structures have been removed from the public header
3665 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3666 now redundant). Users should not attempt to access internal structures
3667 directly. Instead they should use the provided API functions.
3671 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3672 Access to deprecated functions can be re-enabled by running config with
3673 "enable-deprecated". In addition applications wishing to use deprecated
3674 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3675 will, by default, disable some transitive includes that previously existed
3676 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3680 * Added support for OCB mode. OpenSSL has been granted a patent license
3681 compatible with the OpenSSL license for use of OCB. Details are available
3682 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3683 for OCB can be removed by calling config with no-ocb.
3687 * SSLv2 support has been removed. It still supports receiving a SSLv2
3688 compatible client hello.
3692 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3693 done while fixing the error code for the key-too-small case.
3695 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3697 * CA.sh has been removed; use CA.pl instead.
3701 * Removed old DES API.
3705 * Remove various unsupported platforms:
3711 Sinix/ReliantUNIX RM400
3716 16-bit platforms such as WIN16
3720 * Clean up OPENSSL_NO_xxx #define's
3721 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3722 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3723 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3724 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3725 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3726 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3727 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3728 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3729 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3730 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3734 * Cleaned up dead code
3735 Remove all but one '#ifdef undef' which is to be looked at.
3739 * Clean up calling of xxx_free routines.
3740 Just like free(), fix most of the xxx_free routines to accept
3741 NULL. Remove the non-null checks from callers. Save much code.
3745 * Add secure heap for storage of private keys (when possible).
3746 Add BIO_s_secmem(), CBIGNUM, etc.
3747 Contributed by Akamai Technologies under our Corporate CLA.
3751 * Experimental support for a new, fast, unbiased prime candidate generator,
3752 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3754 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3756 * New output format NSS in the sess_id command line tool. This allows
3757 exporting the session id and the master key in NSS keylog format.
3759 *Martin Kaiser <martin@kaiser.cx>*
3761 * Harmonize version and its documentation. -f flag is used to display
3764 *mancha <mancha1@zoho.com>*
3766 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3767 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3769 *mancha <mancha1@zoho.com>*
3771 * Fix some double frees. These are not thought to be exploitable.
3773 *mancha <mancha1@zoho.com>*
3775 * A missing bounds check in the handling of the TLS heartbeat extension
3776 can be used to reveal up to 64k of memory to a connected client or
3779 Thanks for Neel Mehta of Google Security for discovering this bug and to
3780 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3781 preparing the fix ([CVE-2014-0160])
3783 *Adam Langley, Bodo Moeller*
3785 * Fix for the attack described in the paper "Recovering OpenSSL
3786 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3787 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3788 <http://eprint.iacr.org/2014/140>
3790 Thanks to Yuval Yarom and Naomi Benger for discovering this
3791 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3793 *Yuval Yarom and Naomi Benger*
3795 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3796 this fixes a limitation in previous versions of OpenSSL.
3800 * Experimental encrypt-then-mac support.
3802 Experimental support for encrypt then mac from
3803 draft-gutmann-tls-encrypt-then-mac-02.txt
3805 To enable it set the appropriate extension number (0x42 for the test
3806 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3808 For non-compliant peers (i.e. just about everything) this should have no
3811 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3815 * Add EVP support for key wrapping algorithms, to avoid problems with
3816 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3817 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3818 algorithms and include tests cases.
3822 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3827 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3828 MGF1 digest and OAEP label.
3832 * Make openssl verify return errors.
3834 *Chris Palmer <palmer@google.com> and Ben Laurie*
3836 * New function ASN1_TIME_diff to calculate the difference between two
3837 ASN1_TIME structures or one structure and the current time.
3841 * Update fips_test_suite to support multiple command line options. New
3842 test to induce all self test errors in sequence and check expected
3847 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3848 sign or verify all in one operation.
3852 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3853 test programs and fips_test_suite. Includes functionality to parse
3854 the minimal script output of fipsalgest.pl directly.
3858 * Add authorisation parameter to FIPS_module_mode_set().
3862 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3866 * Use separate DRBG fields for internal and external flags. New function
3867 FIPS_drbg_health_check() to perform on demand health checking. Add
3868 generation tests to fips_test_suite with reduced health check interval to
3869 demonstrate periodic health checking. Add "nodh" option to
3870 fips_test_suite to skip very slow DH test.
3874 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3879 * More extensive health check for DRBG checking many more failure modes.
3880 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3881 combination: call this in fips_test_suite.
3885 * Add support for canonical generation of DSA parameter 'g'. See
3888 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3889 POST to handle HMAC cases.
3893 * Add functions FIPS_module_version() and FIPS_module_version_text()
3894 to return numerical and string versions of the FIPS module number.
3898 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3899 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3900 outside the validated module in the FIPS capable OpenSSL.
3904 * Minor change to DRBG entropy callback semantics. In some cases
3905 there is no multiple of the block length between min_len and
3906 max_len. Allow the callback to return more than max_len bytes
3907 of entropy but discard any extra: it is the callback's responsibility
3908 to ensure that the extra data discarded does not impact the
3909 requested amount of entropy.
3913 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3914 information in FIPS186-3, SP800-57 and SP800-131A.
3918 * CCM support via EVP. Interface is very similar to GCM case except we
3919 must supply all data in one chunk (i.e. no update, final) and the
3920 message length must be supplied if AAD is used. Add algorithm test
3925 * Initial version of POST overhaul. Add POST callback to allow the status
3926 of POST to be monitored and/or failures induced. Modify fips_test_suite
3927 to use callback. Always run all selftests even if one fails.
3931 * XTS support including algorithm test driver in the fips_gcmtest program.
3932 Note: this does increase the maximum key length from 32 to 64 bytes but
3933 there should be no binary compatibility issues as existing applications
3934 will never use XTS mode.
3938 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3939 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3940 performs algorithm blocking for unapproved PRNG types. Also do not
3941 set PRNG type in FIPS_mode_set(): leave this to the application.
3942 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3943 the standard OpenSSL PRNG: set additional data to a date time vector.
3947 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3948 This shouldn't present any incompatibility problems because applications
3949 shouldn't be using these directly and any that are will need to rethink
3950 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3954 * Extensive self tests and health checking required by SP800-90 DRBG.
3955 Remove strength parameter from FIPS_drbg_instantiate and always
3956 instantiate at maximum supported strength.
3960 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3964 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3968 * New function DH_compute_key_padded() to compute a DH key and pad with
3969 leading zeroes if needed: this complies with SP800-56A et al.
3973 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3974 anything, incomplete, subject to change and largely untested at present.
3978 * Modify fipscanisteronly build option to only build the necessary object
3979 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3983 * Add experimental option FIPSSYMS to give all symbols in
3984 fipscanister.o and FIPS or fips prefix. This will avoid
3985 conflicts with future versions of OpenSSL. Add perl script
3986 util/fipsas.pl to preprocess assembly language source files
3987 and rename any affected symbols.
3991 * Add selftest checks and algorithm block of non-fips algorithms in
3992 FIPS mode. Remove DES2 from selftests.
3996 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3997 return internal method without any ENGINE dependencies. Add new
3998 tiny fips sign and verify functions.
4002 * New build option no-ec2m to disable characteristic 2 code.
4006 * New build option "fipscanisteronly". This only builds fipscanister.o
4007 and (currently) associated fips utilities. Uses the file Makefile.fips
4008 instead of Makefile.org as the prototype.
4012 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4013 Update fips_gcmtest to use IV generator.
4017 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4018 setting output buffer to NULL. The `*Final` function must be
4019 called although it will not retrieve any additional data. The tag
4020 can be set or retrieved with a ctrl. The IV length is by default 12
4021 bytes (96 bits) but can be set to an alternative value. If the IV
4022 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4027 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4028 underlying do_cipher function handles all cipher semantics itself
4029 including padding and finalisation. This is useful if (for example)
4030 an ENGINE cipher handles block padding itself. The behaviour of
4031 do_cipher is subtly changed if this flag is set: the return value
4032 is the number of characters written to the output buffer (zero is
4033 no longer an error code) or a negative error code. Also if the
4034 input buffer is NULL and length 0 finalisation should be performed.
4038 * If a candidate issuer certificate is already part of the constructed
4039 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4043 * Improve forward-security support: add functions
4045 void SSL_CTX_set_not_resumable_session_callback(
4046 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4047 void SSL_set_not_resumable_session_callback(
4048 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4050 for use by SSL/TLS servers; the callback function will be called whenever a
4051 new session is created, and gets to decide whether the session may be
4052 cached to make it resumable (return 0) or not (return 1). (As by the
4053 SSL/TLS protocol specifications, the session_id sent by the server will be
4054 empty to indicate that the session is not resumable; also, the server will
4055 not generate RFC 4507 (RFC 5077) session tickets.)
4057 A simple reasonable callback implementation is to return is_forward_secure.
4058 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4059 by the SSL/TLS server library, indicating whether it can provide forward
4062 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4064 * New -verify_name option in command line utilities to set verification
4069 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4070 Add CMAC pkey methods.
4074 * Experimental renegotiation in s_server -www mode. If the client
4075 browses /reneg connection is renegotiated. If /renegcert it is
4076 renegotiated requesting a certificate.
4080 * Add an "external" session cache for debugging purposes to s_server. This
4081 should help trace issues which normally are only apparent in deployed
4082 multi-process servers.
4086 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4087 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4088 BIO_set_cipher() and some obscure PEM functions were changed so they
4089 can now return an error. The RAND changes required a change to the
4090 RAND_METHOD structure.
4094 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4095 a gcc attribute to warn if the result of a function is ignored. This
4096 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4097 whose return value is often ignored.
4101 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4102 These allow SCTs (signed certificate timestamps) to be requested and
4103 validated when establishing a connection.
4105 *Rob Percival <robpercival@google.com>*
4110 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4112 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4113 used even when parsing explicit parameters, when loading a encoded key
4114 or calling `EC_GROUP_new_from_ecpkparameters()`/
4115 `EC_GROUP_new_from_ecparameters()`.
4116 This prevents bypass of security hardening and performance gains,
4117 especially for curves with specialized EC_METHODs.
4118 By default, if a key encoded with explicit parameters is loaded and later
4119 encoded, the output is still encoded with explicit parameters, even if
4120 internally a "named" EC_GROUP is used for computation.
4124 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4125 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4126 NULL. After this change, only the cofactor parameter can be NULL. It also
4127 does some minimal sanity checks on the passed order.
4132 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4133 An attack is simple, if the first CMS_recipientInfo is valid but the
4134 second CMS_recipientInfo is chosen ciphertext. If the second
4135 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4136 encryption key will be replaced by garbage, and the message cannot be
4137 decoded, but if the RSA decryption fails, the correct encryption key is
4138 used and the recipient will not notice the attack.
4139 As a work around for this potential attack the length of the decrypted
4140 key must be equal to the cipher default key length, in case the
4141 certifiate is not given and all recipientInfo are tried out.
4142 The old behaviour can be re-enabled in the CMS code by setting the
4143 CMS_DEBUG_DECRYPT flag.
4148 * Document issue with installation paths in diverse Windows builds
4150 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4151 binaries and run-time config file.
4156 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4158 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4159 This changes the size when using the `genpkey` command when no size is given.
4160 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4161 generation commands to use 2048 bits by default.
4165 * Add FIPS support for Android Arm 64-bit
4167 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4168 Module in Version 2.0.10. For some reason, the corresponding target
4169 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4170 built with FIPS support on Android Arm 64-bit. This omission has been
4173 *Matthias St. Pierre*
4175 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4177 * 0-byte record padding oracle
4179 If an application encounters a fatal protocol error and then calls
4180 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4181 then OpenSSL can respond differently to the calling application if a 0 byte
4182 record is received with invalid padding compared to if a 0 byte record is
4183 received with an invalid MAC. If the application then behaves differently
4184 based on that in a way that is detectable to the remote peer, then this
4185 amounts to a padding oracle that could be used to decrypt data.
4187 In order for this to be exploitable "non-stitched" ciphersuites must be in
4188 use. Stitched ciphersuites are optimised implementations of certain
4189 commonly used ciphersuites. Also the application must call SSL_shutdown()
4190 twice even if a protocol error has occurred (applications should not do
4191 this but some do anyway).
4193 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4194 Aviram, with additional investigation by Steven Collison and Andrew
4195 Hourselt. It was reported to OpenSSL on 10th December 2018.
4200 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4204 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4206 * Microarchitecture timing vulnerability in ECC scalar multiplication
4208 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4209 shown to be vulnerable to a microarchitecture timing side channel attack.
4210 An attacker with sufficient access to mount local timing attacks during
4211 ECDSA signature generation could recover the private key.
4213 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4214 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4220 * Timing vulnerability in DSA signature generation
4222 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4223 timing side channel attack. An attacker could use variations in the signing
4224 algorithm to recover the private key.
4226 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4231 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4232 Module, accidentally introduced while backporting security fixes from the
4233 development branch and hindering the use of ECC in FIPS mode.
4237 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4239 * Client DoS due to large DH parameter
4241 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4242 malicious server can send a very large prime value to the client. This will
4243 cause the client to spend an unreasonably long period of time generating a
4244 key for this prime resulting in a hang until the client has finished. This
4245 could be exploited in a Denial Of Service attack.
4247 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4252 * Cache timing vulnerability in RSA Key Generation
4254 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4255 a cache timing side channel attack. An attacker with sufficient access to
4256 mount cache timing attacks during the RSA key generation process could
4257 recover the private key.
4259 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4260 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4265 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4266 parameter is no longer accepted, as it leads to a corrupt table. NULL
4267 pem_str is reserved for alias entries only.
4271 * Revert blinding in ECDSA sign and instead make problematic addition
4272 length-invariant. Switch even to fixed-length Montgomery multiplication.
4276 * Change generating and checking of primes so that the error rate of not
4277 being prime depends on the intended use based on the size of the input.
4278 For larger primes this will result in more rounds of Miller-Rabin.
4279 The maximal error rate for primes with more than 1080 bits is lowered
4282 *Kurt Roeckx, Annie Yousar*
4284 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4288 * Add blinding to ECDSA and DSA signatures to protect against side channel
4289 attacks discovered by Keegan Ryan (NCC Group).
4293 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4294 now allow empty (zero character) pass phrases.
4298 * Certificate time validation (X509_cmp_time) enforces stricter
4299 compliance with RFC 5280. Fractional seconds and timezone offsets
4300 are no longer allowed.
4304 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4306 * Constructed ASN.1 types with a recursive definition could exceed the stack
4308 Constructed ASN.1 types with a recursive definition (such as can be found
4309 in PKCS7) could eventually exceed the stack given malicious input with
4310 excessive recursion. This could result in a Denial Of Service attack. There
4311 are no such structures used within SSL/TLS that come from untrusted sources
4312 so this is considered safe.
4314 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4320 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4322 * Read/write after SSL object in error state
4324 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4325 mechanism. The intent was that if a fatal error occurred during a handshake
4326 then OpenSSL would move into the error state and would immediately fail if
4327 you attempted to continue the handshake. This works as designed for the
4328 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4329 SSL_connect()), however due to a bug it does not work correctly if
4330 SSL_read() or SSL_write() is called directly. In that scenario, if the
4331 handshake fails then a fatal error will be returned in the initial function
4332 call. If SSL_read()/SSL_write() is subsequently called by the application
4333 for the same SSL object then it will succeed and the data is passed without
4334 being decrypted/encrypted directly from the SSL/TLS record layer.
4336 In order to exploit this issue an application bug would have to be present
4337 that resulted in a call to SSL_read()/SSL_write() being issued after having
4338 already received a fatal error.
4340 This issue was reported to OpenSSL by David Benjamin (Google).
4345 * rsaz_1024_mul_avx2 overflow bug on x86_64
4347 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4348 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4349 Analysis suggests that attacks against RSA and DSA as a result of this
4350 defect would be very difficult to perform and are not believed likely.
4351 Attacks against DH1024 are considered just feasible, because most of the
4352 work necessary to deduce information about a private key may be performed
4353 offline. The amount of resources required for such an attack would be
4354 significant. However, for an attack on TLS to be meaningful, the server
4355 would have to share the DH1024 private key among multiple clients, which is
4356 no longer an option since CVE-2016-0701.
4358 This only affects processors that support the AVX2 but not ADX extensions
4359 like Intel Haswell (4th generation).
4361 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4362 was originally found via the OSS-Fuzz project.
4367 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4369 * bn_sqrx8x_internal carry bug on x86_64
4371 There is a carry propagating bug in the x86_64 Montgomery squaring
4372 procedure. No EC algorithms are affected. Analysis suggests that attacks
4373 against RSA and DSA as a result of this defect would be very difficult to
4374 perform and are not believed likely. Attacks against DH are considered just
4375 feasible (although very difficult) because most of the work necessary to
4376 deduce information about a private key may be performed offline. The amount
4377 of resources required for such an attack would be very significant and
4378 likely only accessible to a limited number of attackers. An attacker would
4379 additionally need online access to an unpatched system using the target
4380 private key in a scenario with persistent DH parameters and a private
4381 key that is shared between multiple clients.
4383 This only affects processors that support the BMI1, BMI2 and ADX extensions
4384 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4386 This issue was reported to OpenSSL by the OSS-Fuzz project.
4391 * Malformed X.509 IPAddressFamily could cause OOB read
4393 If an X.509 certificate has a malformed IPAddressFamily extension,
4394 OpenSSL could do a one-byte buffer overread. The most likely result
4395 would be an erroneous display of the certificate in text format.
4397 This issue was reported to OpenSSL by the OSS-Fuzz project.
4401 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4403 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4404 platform rather than 'mingw'.
4408 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4410 * Truncated packet could crash via OOB read
4412 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4413 cipher is being used, then a truncated packet can cause that host to
4414 perform an out-of-bounds read, usually resulting in a crash.
4416 This issue was reported to OpenSSL by Robert Święcki of Google.
4421 * BN_mod_exp may produce incorrect results on x86_64
4423 There is a carry propagating bug in the x86_64 Montgomery squaring
4424 procedure. No EC algorithms are affected. Analysis suggests that attacks
4425 against RSA and DSA as a result of this defect would be very difficult to
4426 perform and are not believed likely. Attacks against DH are considered just
4427 feasible (although very difficult) because most of the work necessary to
4428 deduce information about a private key may be performed offline. The amount
4429 of resources required for such an attack would be very significant and
4430 likely only accessible to a limited number of attackers. An attacker would
4431 additionally need online access to an unpatched system using the target
4432 private key in a scenario with persistent DH parameters and a private
4433 key that is shared between multiple clients. For example this can occur by
4434 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4435 similar to CVE-2015-3193 but must be treated as a separate problem.
4437 This issue was reported to OpenSSL by the OSS-Fuzz project.
4442 * Montgomery multiplication may produce incorrect results
4444 There is a carry propagating bug in the Broadwell-specific Montgomery
4445 multiplication procedure that handles input lengths divisible by, but
4446 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4447 and DH private keys are impossible. This is because the subroutine in
4448 question is not used in operations with the private key itself and an input
4449 of the attacker's direct choice. Otherwise the bug can manifest itself as
4450 transient authentication and key negotiation failures or reproducible
4451 erroneous outcome of public-key operations with specially crafted input.
4452 Among EC algorithms only Brainpool P-512 curves are affected and one
4453 presumably can attack ECDH key negotiation. Impact was not analyzed in
4454 detail, because pre-requisites for attack are considered unlikely. Namely
4455 multiple clients have to choose the curve in question and the server has to
4456 share the private key among them, neither of which is default behaviour.
4457 Even then only clients that chose the curve will be affected.
4459 This issue was publicly reported as transient failures and was not
4460 initially recognized as a security issue. Thanks to Richard Morgan for
4461 providing reproducible case.
4466 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4467 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4468 prevent issues where no progress is being made and the peer continually
4469 sends unrecognised record types, using up resources processing them.
4473 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4475 * Missing CRL sanity check
4477 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4478 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4479 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4481 This issue only affects the OpenSSL 1.0.2i
4486 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4488 * OCSP Status Request extension unbounded memory growth
4490 A malicious client can send an excessively large OCSP Status Request
4491 extension. If that client continually requests renegotiation, sending a
4492 large OCSP Status Request extension each time, then there will be unbounded
4493 memory growth on the server. This will eventually lead to a Denial Of
4494 Service attack through memory exhaustion. Servers with a default
4495 configuration are vulnerable even if they do not support OCSP. Builds using
4496 the "no-ocsp" build time option are not affected.
4498 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4503 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4506 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4512 * OOB write in MDC2_Update()
4514 An overflow can occur in MDC2_Update() either if called directly or
4515 through the EVP_DigestUpdate() function using MDC2. If an attacker
4516 is able to supply very large amounts of input data after a previous
4517 call to EVP_EncryptUpdate() with a partial block then a length check
4518 can overflow resulting in a heap corruption.
4520 The amount of data needed is comparable to SIZE_MAX which is impractical
4523 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4528 * Malformed SHA512 ticket DoS
4530 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4531 DoS attack where a malformed ticket will result in an OOB read which will
4534 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4535 a custom server callback and ticket lookup mechanism.
4537 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4542 * OOB write in BN_bn2dec()
4544 The function BN_bn2dec() does not check the return value of BN_div_word().
4545 This can cause an OOB write if an application uses this function with an
4546 overly large BIGNUM. This could be a problem if an overly large certificate
4547 or CRL is printed out from an untrusted source. TLS is not affected because
4548 record limits will reject an oversized certificate before it is parsed.
4550 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4555 * OOB read in TS_OBJ_print_bio()
4557 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4558 the total length the OID text representation would use and not the amount
4559 of data written. This will result in OOB reads when large OIDs are
4562 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4567 * Pointer arithmetic undefined behaviour
4569 Avoid some undefined pointer arithmetic
4571 A common idiom in the codebase is to check limits in the following manner:
4574 Where "p" points to some malloc'd data of SIZE bytes and
4577 "len" here could be from some externally supplied data (e.g. from a TLS
4580 The rules of C pointer arithmetic are such that "p + len" is only well
4581 defined where len <= SIZE. Therefore the above idiom is actually
4582 undefined behaviour.
4584 For example this could cause problems if some malloc implementation
4585 provides an address for "p" such that "p + len" actually overflows for
4586 values of len that are too big and therefore p + len < limit.
4588 This issue was reported to OpenSSL by Guido Vranken
4593 * Constant time flag not preserved in DSA signing
4595 Operations in the DSA signing algorithm should run in constant time in
4596 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4597 implementation means that a non-constant time codepath is followed for
4598 certain operations. This has been demonstrated through a cache-timing
4599 attack to be sufficient for an attacker to recover the private DSA key.
4601 This issue was reported by César Pereida (Aalto University), Billy Brumley
4602 (Tampere University of Technology), and Yuval Yarom (The University of
4603 Adelaide and NICTA).
4608 * DTLS buffered message DoS
4610 In a DTLS connection where handshake messages are delivered out-of-order
4611 those messages that OpenSSL is not yet ready to process will be buffered
4612 for later use. Under certain circumstances, a flaw in the logic means that
4613 those messages do not get removed from the buffer even though the handshake
4614 has been completed. An attacker could force up to approx. 15 messages to
4615 remain in the buffer when they are no longer required. These messages will
4616 be cleared when the DTLS connection is closed. The default maximum size for
4617 a message is 100k. Therefore the attacker could force an additional 1500k
4618 to be consumed per connection. By opening many simulataneous connections an
4619 attacker could cause a DoS attack through memory exhaustion.
4621 This issue was reported to OpenSSL by Quan Luo.
4626 * DTLS replay protection DoS
4628 A flaw in the DTLS replay attack protection mechanism means that records
4629 that arrive for future epochs update the replay protection "window" before
4630 the MAC for the record has been validated. This could be exploited by an
4631 attacker by sending a record for the next epoch (which does not have to
4632 decrypt or have a valid MAC), with a very large sequence number. This means
4633 that all subsequent legitimate packets are dropped causing a denial of
4634 service for a specific DTLS connection.
4636 This issue was reported to OpenSSL by the OCAP audit team.
4641 * Certificate message OOB reads
4643 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4644 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4645 theoretical DoS risk but this has not been observed in practice on common
4648 The messages affected are client certificate, client certificate request
4649 and server certificate. As a result the attack can only be performed
4650 against a client or a server which enables client authentication.
4652 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4657 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4659 * Prevent padding oracle in AES-NI CBC MAC check
4661 A MITM attacker can use a padding oracle attack to decrypt traffic
4662 when the connection uses an AES CBC cipher and the server support
4665 This issue was introduced as part of the fix for Lucky 13 padding
4666 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4667 constant time by making sure that always the same bytes are read and
4668 compared against either the MAC or padding bytes. But it no longer
4669 checked that there was enough data to have both the MAC and padding
4672 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4676 * Fix EVP_EncodeUpdate overflow
4678 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4679 Base64 encoding of binary data. If an attacker is able to supply very large
4680 amounts of input data then a length check can overflow resulting in a heap
4683 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4684 the `PEM_write_bio*` family of functions. These are mainly used within the
4685 OpenSSL command line applications, so any application which processes data
4686 from an untrusted source and outputs it as a PEM file should be considered
4687 vulnerable to this issue. User applications that call these APIs directly
4688 with large amounts of untrusted data may also be vulnerable.
4690 This issue was reported by Guido Vranken.
4695 * Fix EVP_EncryptUpdate overflow
4697 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4698 is able to supply very large amounts of input data after a previous call to
4699 EVP_EncryptUpdate() with a partial block then a length check can overflow
4700 resulting in a heap corruption. Following an analysis of all OpenSSL
4701 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4702 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4703 the first called function after an EVP_EncryptInit(), and therefore that
4704 specific call must be safe. The second form is where the length passed to
4705 EVP_EncryptUpdate() can be seen from the code to be some small value and
4706 therefore there is no possibility of an overflow. Since all instances are
4707 one of these two forms, it is believed that there can be no overflows in
4708 internal code due to this problem. It should be noted that
4709 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4710 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4711 of these calls have also been analysed too and it is believed there are no
4712 instances in internal usage where an overflow could occur.
4714 This issue was reported by Guido Vranken.
4719 * Prevent ASN.1 BIO excessive memory allocation
4721 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4722 a short invalid encoding can cause allocation of large amounts of memory
4723 potentially consuming excessive resources or exhausting memory.
4725 Any application parsing untrusted data through d2i BIO functions is
4726 affected. The memory based functions such as d2i_X509() are *not* affected.
4727 Since the memory based functions are used by the TLS library, TLS
4728 applications are not affected.
4730 This issue was reported by Brian Carpenter.
4737 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4738 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4739 in arbitrary stack data being returned in the buffer.
4741 This issue was reported by Guido Vranken.
4746 * Modify behavior of ALPN to invoke callback after SNI/servername
4747 callback, such that updates to the SSL_CTX affect ALPN.
4751 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4756 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4757 methods are enabled and ssl2 is disabled the methods return NULL.
4761 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4763 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4764 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4765 provide any "EXPORT" or "LOW" strength ciphers.
4769 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4770 is by default disabled at build-time. Builds that are not configured with
4771 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4772 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4773 will need to explicitly call either of:
4775 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4777 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4779 as appropriate. Even if either of those is used, or the application
4780 explicitly uses the version-specific SSLv2_method() or its client and
4781 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4782 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4783 ciphers, and SSLv2 56-bit DES are no longer available.
4788 * Fix a double-free in DSA code
4790 A double free bug was discovered when OpenSSL parses malformed DSA private
4791 keys and could lead to a DoS attack or memory corruption for applications
4792 that receive DSA private keys from untrusted sources. This scenario is
4795 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4801 * Disable SRP fake user seed to address a server memory leak.
4803 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4805 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4806 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4807 was changed to ignore the "fake user" SRP seed, even if the seed
4810 Users should use SRP_VBASE_get1_by_user instead. Note that in
4811 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4812 also that even though configuring the SRP seed attempts to hide
4813 invalid usernames by continuing the handshake with fake
4814 credentials, this behaviour is not constant time and no strong
4815 guarantees are made that the handshake is indistinguishable from
4816 that of a valid user.
4821 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4823 In the BN_hex2bn function the number of hex digits is calculated using an
4824 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4825 large values of `i` this can result in `bn_expand` not allocating any
4826 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4827 field as NULL leading to a subsequent NULL ptr deref. For very large values
4828 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4829 In this case memory is allocated to the internal BIGNUM data field, but it
4830 is insufficiently sized leading to heap corruption. A similar issue exists
4831 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4832 is ever called by user applications with very large untrusted hex/dec data.
4833 This is anticipated to be a rare occurrence.
4835 All OpenSSL internal usage of these functions use data that is not expected
4836 to be untrusted, e.g. config file data or application command line
4837 arguments. If user developed applications generate config file data based
4838 on untrusted data then it is possible that this could also lead to security
4839 consequences. This is also anticipated to be rare.
4841 This issue was reported to OpenSSL by Guido Vranken.
4846 * Fix memory issues in `BIO_*printf` functions
4848 The internal `fmtstr` function used in processing a "%s" format string in
4849 the `BIO_*printf` functions could overflow while calculating the length of a
4850 string and cause an OOB read when printing very long strings.
4852 Additionally the internal `doapr_outch` function can attempt to write to an
4853 OOB memory location (at an offset from the NULL pointer) in the event of a
4854 memory allocation failure. In 1.0.2 and below this could be caused where
4855 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4856 could be in processing a very long "%s" format string. Memory leaks can
4859 The first issue may mask the second issue dependent on compiler behaviour.
4860 These problems could enable attacks where large amounts of untrusted data
4861 is passed to the `BIO_*printf` functions. If applications use these functions
4862 in this way then they could be vulnerable. OpenSSL itself uses these
4863 functions when printing out human-readable dumps of ASN.1 data. Therefore
4864 applications that print this data could be vulnerable if the data is from
4865 untrusted sources. OpenSSL command line applications could also be
4866 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4867 as command line arguments.
4869 Libssl is not considered directly vulnerable. Additionally certificates etc
4870 received via remote connections via libssl are also unlikely to be able to
4871 trigger these issues because of message size limits enforced within libssl.
4873 This issue was reported to OpenSSL Guido Vranken.
4878 * Side channel attack on modular exponentiation
4880 A side-channel attack was found which makes use of cache-bank conflicts on
4881 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4882 of RSA keys. The ability to exploit this issue is limited as it relies on
4883 an attacker who has control of code in a thread running on the same
4884 hyper-threaded core as the victim thread which is performing decryptions.
4886 This issue was reported to OpenSSL by Yuval Yarom, The University of
4887 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4888 Nadia Heninger, University of Pennsylvania with more information at
4889 <http://cachebleed.info>.
4894 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
4895 if no keysize is specified with default_bits. This fixes an
4896 omission in an earlier change that changed all RSA/DSA key generation
4897 commands to use 2048 bits by default.
4901 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4903 * DH small subgroups
4905 Historically OpenSSL only ever generated DH parameters based on "safe"
4906 primes. More recently (in version 1.0.2) support was provided for
4907 generating X9.42 style parameter files such as those required for RFC 5114
4908 support. The primes used in such files may not be "safe". Where an
4909 application is using DH configured with parameters based on primes that are
4910 not "safe" then an attacker could use this fact to find a peer's private
4911 DH exponent. This attack requires that the attacker complete multiple
4912 handshakes in which the peer uses the same private DH exponent. For example
4913 this could be used to discover a TLS server's private DH exponent if it's
4914 reusing the private DH exponent or it's using a static DH ciphersuite.
4916 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4917 TLS. It is not on by default. If the option is not set then the server
4918 reuses the same private DH exponent for the life of the server process and
4919 would be vulnerable to this attack. It is believed that many popular
4920 applications do set this option and would therefore not be at risk.
4922 The fix for this issue adds an additional check where a "q" parameter is
4923 available (as is the case in X9.42 based parameters). This detects the
4924 only known attack, and is the only possible defense for static DH
4925 ciphersuites. This could have some performance impact.
4927 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4928 default and cannot be disabled. This could have some performance impact.
4930 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4935 * SSLv2 doesn't block disabled ciphers
4937 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4938 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4939 been disabled, provided that the SSLv2 protocol was not also disabled via
4942 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4943 and Sebastian Schinzel.
4948 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4950 * BN_mod_exp may produce incorrect results on x86_64
4952 There is a carry propagating bug in the x86_64 Montgomery squaring
4953 procedure. No EC algorithms are affected. Analysis suggests that attacks
4954 against RSA and DSA as a result of this defect would be very difficult to
4955 perform and are not believed likely. Attacks against DH are considered just
4956 feasible (although very difficult) because most of the work necessary to
4957 deduce information about a private key may be performed offline. The amount
4958 of resources required for such an attack would be very significant and
4959 likely only accessible to a limited number of attackers. An attacker would
4960 additionally need online access to an unpatched system using the target
4961 private key in a scenario with persistent DH parameters and a private
4962 key that is shared between multiple clients. For example this can occur by
4963 default in OpenSSL DHE based SSL/TLS ciphersuites.
4965 This issue was reported to OpenSSL by Hanno Böck.
4970 * Certificate verify crash with missing PSS parameter
4972 The signature verification routines will crash with a NULL pointer
4973 dereference if presented with an ASN.1 signature using the RSA PSS
4974 algorithm and absent mask generation function parameter. Since these
4975 routines are used to verify certificate signature algorithms this can be
4976 used to crash any certificate verification operation and exploited in a
4977 DoS attack. Any application which performs certificate verification is
4978 vulnerable including OpenSSL clients and servers which enable client
4981 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4986 * X509_ATTRIBUTE memory leak
4988 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4989 memory. This structure is used by the PKCS#7 and CMS routines so any
4990 application which reads PKCS#7 or CMS data from untrusted sources is
4991 affected. SSL/TLS is not affected.
4993 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4999 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5000 This changes the decoding behaviour for some invalid messages,
5001 though the change is mostly in the more lenient direction, and
5002 legacy behaviour is preserved as much as possible.
5006 * In DSA_generate_parameters_ex, if the provided seed is too short,
5009 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5011 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5013 * Alternate chains certificate forgery
5015 During certificate verification, OpenSSL will attempt to find an
5016 alternative certificate chain if the first attempt to build such a chain
5017 fails. An error in the implementation of this logic can mean that an
5018 attacker could cause certain checks on untrusted certificates to be
5019 bypassed, such as the CA flag, enabling them to use a valid leaf
5020 certificate to act as a CA and "issue" an invalid certificate.
5022 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5027 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5029 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5030 incompatibility in the handling of HMAC. The previous ABI has now been
5035 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5037 * Malformed ECParameters causes infinite loop
5039 When processing an ECParameters structure OpenSSL enters an infinite loop
5040 if the curve specified is over a specially malformed binary polynomial
5043 This can be used to perform denial of service against any
5044 system which processes public keys, certificate requests or
5045 certificates. This includes TLS clients and TLS servers with
5046 client authentication enabled.
5048 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5053 * Exploitable out-of-bounds read in X509_cmp_time
5055 X509_cmp_time does not properly check the length of the ASN1_TIME
5056 string and can read a few bytes out of bounds. In addition,
5057 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5060 An attacker can use this to craft malformed certificates and CRLs of
5061 various sizes and potentially cause a segmentation fault, resulting in
5062 a DoS on applications that verify certificates or CRLs. TLS clients
5063 that verify CRLs are affected. TLS clients and servers with client
5064 authentication enabled may be affected if they use custom verification
5067 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5068 independently by Hanno Böck.
5073 * PKCS7 crash with missing EnvelopedContent
5075 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5076 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5077 with missing content and trigger a NULL pointer dereference on parsing.
5079 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5080 structures from untrusted sources are affected. OpenSSL clients and
5081 servers are not affected.
5083 This issue was reported to OpenSSL by Michal Zalewski (Google).
5088 * CMS verify infinite loop with unknown hash function
5090 When verifying a signedData message the CMS code can enter an infinite loop
5091 if presented with an unknown hash function OID. This can be used to perform
5092 denial of service against any system which verifies signedData messages using
5094 This issue was reported to OpenSSL by Johannes Bauer.
5099 * Race condition handling NewSessionTicket
5101 If a NewSessionTicket is received by a multi-threaded client when attempting to
5102 reuse a previous ticket then a race condition can occur potentially leading to
5103 a double free of the ticket data.
5108 * Only support 256-bit or stronger elliptic curves with the
5109 'ecdh_auto' setting (server) or by default (client). Of supported
5110 curves, prefer P-256 (both).
5114 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5116 * ClientHello sigalgs DoS fix
5118 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5119 invalid signature algorithms extension a NULL pointer dereference will
5120 occur. This can be exploited in a DoS attack against the server.
5122 This issue was was reported to OpenSSL by David Ramos of Stanford
5126 *Stephen Henson and Matt Caswell*
5128 * Multiblock corrupted pointer fix
5130 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5131 feature only applies on 64 bit x86 architecture platforms that support AES
5132 NI instructions. A defect in the implementation of "multiblock" can cause
5133 OpenSSL's internal write buffer to become incorrectly set to NULL when
5134 using non-blocking IO. Typically, when the user application is using a
5135 socket BIO for writing, this will only result in a failed connection.
5136 However if some other BIO is used then it is likely that a segmentation
5137 fault will be triggered, thus enabling a potential DoS attack.
5139 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5144 * Segmentation fault in DTLSv1_listen fix
5146 The DTLSv1_listen function is intended to be stateless and processes the
5147 initial ClientHello from many peers. It is common for user code to loop
5148 over the call to DTLSv1_listen until a valid ClientHello is received with
5149 an associated cookie. A defect in the implementation of DTLSv1_listen means
5150 that state is preserved in the SSL object from one invocation to the next
5151 that can lead to a segmentation fault. Errors processing the initial
5152 ClientHello can trigger this scenario. An example of such an error could be
5153 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5156 This issue was reported to OpenSSL by Per Allansson.
5161 * Segmentation fault in ASN1_TYPE_cmp fix
5163 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5164 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5165 certificate signature algorithm consistency this can be used to crash any
5166 certificate verification operation and exploited in a DoS attack. Any
5167 application which performs certificate verification is vulnerable including
5168 OpenSSL clients and servers which enable client authentication.
5173 * Segmentation fault for invalid PSS parameters fix
5175 The signature verification routines will crash with a NULL pointer
5176 dereference if presented with an ASN.1 signature using the RSA PSS
5177 algorithm and invalid parameters. Since these routines are used to verify
5178 certificate signature algorithms this can be used to crash any
5179 certificate verification operation and exploited in a DoS attack. Any
5180 application which performs certificate verification is vulnerable including
5181 OpenSSL clients and servers which enable client authentication.
5183 This issue was was reported to OpenSSL by Brian Carpenter.
5188 * ASN.1 structure reuse memory corruption fix
5190 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5191 memory corruption via an invalid write. Such reuse is and has been
5192 strongly discouraged and is believed to be rare.
5194 Applications that parse structures containing CHOICE or ANY DEFINED BY
5195 components may be affected. Certificate parsing (d2i_X509 and related
5196 functions) are however not affected. OpenSSL clients and servers are
5202 * PKCS7 NULL pointer dereferences fix
5204 The PKCS#7 parsing code does not handle missing outer ContentInfo
5205 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5206 missing content and trigger a NULL pointer dereference on parsing.
5208 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5209 otherwise parse PKCS#7 structures from untrusted sources are
5210 affected. OpenSSL clients and servers are not affected.
5212 This issue was reported to OpenSSL by Michal Zalewski (Google).
5217 * DoS via reachable assert in SSLv2 servers fix
5219 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5220 servers that both support SSLv2 and enable export cipher suites by sending
5221 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5223 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5224 (OpenSSL development team).
5229 * Empty CKE with client auth and DHE fix
5231 If client auth is used then a server can seg fault in the event of a DHE
5232 ciphersuite being selected and a zero length ClientKeyExchange message
5233 being sent by the client. This could be exploited in a DoS attack.
5238 * Handshake with unseeded PRNG fix
5240 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5241 with an unseeded PRNG. The conditions are:
5242 - The client is on a platform where the PRNG has not been seeded
5243 automatically, and the user has not seeded manually
5244 - A protocol specific client method version has been used (i.e. not
5245 SSL_client_methodv23)
5246 - A ciphersuite is used that does not require additional random data from
5247 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5249 If the handshake succeeds then the client random that has been used will
5250 have been generated from a PRNG with insufficient entropy and therefore the
5251 output may be predictable.
5253 For example using the following command with an unseeded openssl will
5254 succeed on an unpatched platform:
5256 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5261 * Use After Free following d2i_ECPrivatekey error fix
5263 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5264 could cause a use after free condition. This, in turn, could cause a double
5265 free in several private key parsing functions (such as d2i_PrivateKey
5266 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5267 for applications that receive EC private keys from untrusted
5268 sources. This scenario is considered rare.
5270 This issue was discovered by the BoringSSL project and fixed in their
5276 * X509_to_X509_REQ NULL pointer deref fix
5278 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5279 the certificate key is invalid. This function is rarely used in practice.
5281 This issue was discovered by Brian Carpenter.
5286 * Removed the export ciphers from the DEFAULT ciphers
5290 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5292 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5293 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5294 So far those who have to target multiple platforms would compromise
5295 and argue that binary targeting say ARMv5 would still execute on
5296 ARMv8. "Universal" build resolves this compromise by providing
5297 near-optimal performance even on newer platforms.
5301 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5302 (other platforms pending).
5304 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5306 * Add support for the SignedCertificateTimestampList certificate and
5307 OCSP response extensions from RFC6962.
5311 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5312 for corner cases. (Certain input points at infinity could lead to
5313 bogus results, with non-infinity inputs mapped to infinity too.)
5317 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5318 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5319 common cases are optimized and there still is room for further
5320 improvements. Vector Permutation AES for Altivec is also added.
5324 * Add support for little-endian ppc64 Linux target.
5326 *Marcelo Cerri (IBM)*
5328 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5329 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5330 are optimized and there still is room for further improvements.
5331 Both 32- and 64-bit modes are supported.
5333 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5335 * Improved ARMv7 NEON support.
5339 * Support for SPARC Architecture 2011 crypto extensions, first
5340 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5341 SHA256/512, MD5, GHASH and modular exponentiation.
5343 *Andy Polyakov, David Miller*
5345 * Accelerated modular exponentiation for Intel processors, a.k.a.
5348 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5350 * Support for new and upcoming Intel processors, including AVX2,
5351 BMI and SHA ISA extensions. This includes additional "stitched"
5352 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5355 This work was sponsored by Intel Corp.
5359 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5360 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5361 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5365 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5366 this fixes a limitation in previous versions of OpenSSL.
5370 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5371 MGF1 digest and OAEP label.
5375 * Add EVP support for key wrapping algorithms, to avoid problems with
5376 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5377 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5378 algorithms and include tests cases.
5382 * Add functions to allocate and set the fields of an ECDSA_METHOD
5385 *Douglas E. Engert, Steve Henson*
5387 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5388 difference in days and seconds between two tm or ASN1_TIME structures.
5392 * Add -rev test option to s_server to just reverse order of characters
5393 received by client and send back to server. Also prints an abbreviated
5394 summary of the connection parameters.
5398 * New option -brief for s_client and s_server to print out a brief summary
5399 of connection parameters.
5403 * Add callbacks for arbitrary TLS extensions.
5405 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5407 * New option -crl_download in several openssl utilities to download CRLs
5408 from CRLDP extension in certificates.
5412 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5416 * New function X509_CRL_diff to generate a delta CRL from the difference
5417 of two full CRLs. Add support to "crl" utility.
5421 * New functions to set lookup_crls function and to retrieve
5422 X509_STORE from X509_STORE_CTX.
5426 * Print out deprecated issuer and subject unique ID fields in
5431 * Extend OCSP I/O functions so they can be used for simple general purpose
5432 HTTP as well as OCSP. New wrapper function which can be used to download
5433 CRLs using the OCSP API.
5437 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5441 * `SSL_CONF*` functions. These provide a common framework for application
5442 configuration using configuration files or command lines.
5446 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5447 message callback and prints the results. Needs compile time option
5448 "enable-ssl-trace". New options to s_client and s_server to enable
5453 * New ctrl and macro to retrieve supported points extensions.
5454 Print out extension in s_server and s_client.
5458 * New functions to retrieve certificate signature and signature
5463 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5468 * New Suite B modes for TLS code. These use and enforce the requirements
5469 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5470 only use Suite B curves. The Suite B modes can be set by using the
5471 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5475 * New chain verification flags for Suite B levels of security. Check
5476 algorithms are acceptable when flags are set in X509_verify_cert.
5480 * Make tls1_check_chain return a set of flags indicating checks passed
5481 by a certificate chain. Add additional tests to handle client
5482 certificates: checks for matching certificate type and issuer name
5487 * If an attempt is made to use a signature algorithm not in the peer
5488 preference list abort the handshake. If client has no suitable
5489 signature algorithms in response to a certificate request do not
5490 use the certificate.
5494 * If server EC tmp key is not in client preference list abort handshake.
5498 * Add support for certificate stores in CERT structure. This makes it
5499 possible to have different stores per SSL structure or one store in
5500 the parent SSL_CTX. Include distinct stores for certificate chain
5501 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5502 to build and store a certificate chain in CERT structure: returning
5503 an error if the chain cannot be built: this will allow applications
5504 to test if a chain is correctly configured.
5506 Note: if the CERT based stores are not set then the parent SSL_CTX
5507 store is used to retain compatibility with existing behaviour.
5511 * New function ssl_set_client_disabled to set a ciphersuite disabled
5512 mask based on the current session, check mask when sending client
5513 hello and checking the requested ciphersuite.
5517 * New ctrls to retrieve and set certificate types in a certificate
5518 request message. Print out received values in s_client. If certificate
5519 types is not set with custom values set sensible values based on
5520 supported signature algorithms.
5524 * Support for distinct client and server supported signature algorithms.
5528 * Add certificate callback. If set this is called whenever a certificate
5529 is required by client or server. An application can decide which
5530 certificate chain to present based on arbitrary criteria: for example
5531 supported signature algorithms. Add very simple example to s_server.
5532 This fixes many of the problems and restrictions of the existing client
5533 certificate callback: for example you can now clear an existing
5534 certificate and specify the whole chain.
5538 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5539 the certificate can be used for (if anything). Set valid_flags field
5540 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5541 to have similar checks in it.
5543 Add new "cert_flags" field to CERT structure and include a "strict mode".
5544 This enforces some TLS certificate requirements (such as only permitting
5545 certificate signature algorithms contained in the supported algorithms
5546 extension) which some implementations ignore: this option should be used
5547 with caution as it could cause interoperability issues.
5551 * Update and tidy signature algorithm extension processing. Work out
5552 shared signature algorithms based on preferences and peer algorithms
5553 and print them out in s_client and s_server. Abort handshake if no
5554 shared signature algorithms.
5558 * Add new functions to allow customised supported signature algorithms
5559 for SSL and SSL_CTX structures. Add options to s_client and s_server
5564 * New function SSL_certs_clear() to delete all references to certificates
5565 from an SSL structure. Before this once a certificate had been added
5566 it couldn't be removed.
5570 * Integrate hostname, email address and IP address checking with certificate
5571 verification. New verify options supporting checking in openssl utility.
5575 * Fixes and wildcard matching support to hostname and email checking
5576 functions. Add manual page.
5578 *Florian Weimer (Red Hat Product Security Team)*
5580 * New functions to check a hostname email or IP address against a
5581 certificate. Add options x509 utility to print results of checks against
5586 * Fix OCSP checking.
5588 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5590 * Initial experimental support for explicitly trusted non-root CAs.
5591 OpenSSL still tries to build a complete chain to a root but if an
5592 intermediate CA has a trust setting included that is used. The first
5593 setting is used: whether to trust (e.g., -addtrust option to the x509
5598 * Add -trusted_first option which attempts to find certificates in the
5599 trusted store even if an untrusted chain is also supplied.
5603 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5604 platform support for Linux and Android.
5608 * Support for linux-x32, ILP32 environment in x86_64 framework.
5612 * Experimental multi-implementation support for FIPS capable OpenSSL.
5613 When in FIPS mode the approved implementations are used as normal,
5614 when not in FIPS mode the internal unapproved versions are used instead.
5615 This means that the FIPS capable OpenSSL isn't forced to use the
5616 (often lower performance) FIPS implementations outside FIPS mode.
5620 * Transparently support X9.42 DH parameters when calling
5621 PEM_read_bio_DHparameters. This means existing applications can handle
5622 the new parameter format automatically.
5626 * Initial experimental support for X9.42 DH parameter format: mainly
5627 to support use of 'q' parameter for RFC5114 parameters.
5631 * Add DH parameters from RFC5114 including test data to dhtest.
5635 * Support for automatic EC temporary key parameter selection. If enabled
5636 the most preferred EC parameters are automatically used instead of
5637 hardcoded fixed parameters. Now a server just has to call:
5638 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5639 support ECDH and use the most appropriate parameters.
5643 * Enhance and tidy EC curve and point format TLS extension code. Use
5644 static structures instead of allocation if default values are used.
5645 New ctrls to set curves we wish to support and to retrieve shared curves.
5646 Print out shared curves in s_server. New options to s_server and s_client
5647 to set list of supported curves.
5651 * New ctrls to retrieve supported signature algorithms and
5652 supported curve values as an array of NIDs. Extend openssl utility
5653 to print out received values.
5657 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5658 between NIDs and the more common NIST names such as "P-256". Enhance
5659 ecparam utility and ECC method to recognise the NIST names for curves.
5663 * Enhance SSL/TLS certificate chain handling to support different
5664 chains for each certificate instead of one chain in the parent SSL_CTX.
5668 * Support for fixed DH ciphersuite client authentication: where both
5669 server and client use DH certificates with common parameters.
5673 * Support for fixed DH ciphersuites: those requiring DH server
5678 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5680 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5681 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5682 X509_CINF_get_signature were reverted post internal team review.
5687 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5689 * OCSP Status Request extension unbounded memory growth
5691 A malicious client can send an excessively large OCSP Status Request
5692 extension. If that client continually requests renegotiation, sending a
5693 large OCSP Status Request extension each time, then there will be unbounded
5694 memory growth on the server. This will eventually lead to a Denial Of
5695 Service attack through memory exhaustion. Servers with a default
5696 configuration are vulnerable even if they do not support OCSP. Builds using
5697 the "no-ocsp" build time option are not affected.
5699 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5704 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5707 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5713 * OOB write in MDC2_Update()
5715 An overflow can occur in MDC2_Update() either if called directly or
5716 through the EVP_DigestUpdate() function using MDC2. If an attacker
5717 is able to supply very large amounts of input data after a previous
5718 call to EVP_EncryptUpdate() with a partial block then a length check
5719 can overflow resulting in a heap corruption.
5721 The amount of data needed is comparable to SIZE_MAX which is impractical
5724 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5729 * Malformed SHA512 ticket DoS
5731 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5732 DoS attack where a malformed ticket will result in an OOB read which will
5735 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5736 a custom server callback and ticket lookup mechanism.
5738 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5743 * OOB write in BN_bn2dec()
5745 The function BN_bn2dec() does not check the return value of BN_div_word().
5746 This can cause an OOB write if an application uses this function with an
5747 overly large BIGNUM. This could be a problem if an overly large certificate
5748 or CRL is printed out from an untrusted source. TLS is not affected because
5749 record limits will reject an oversized certificate before it is parsed.
5751 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5756 * OOB read in TS_OBJ_print_bio()
5758 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5759 the total length the OID text representation would use and not the amount
5760 of data written. This will result in OOB reads when large OIDs are
5763 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5768 * Pointer arithmetic undefined behaviour
5770 Avoid some undefined pointer arithmetic
5772 A common idiom in the codebase is to check limits in the following manner:
5775 Where "p" points to some malloc'd data of SIZE bytes and
5778 "len" here could be from some externally supplied data (e.g. from a TLS
5781 The rules of C pointer arithmetic are such that "p + len" is only well
5782 defined where len <= SIZE. Therefore the above idiom is actually
5783 undefined behaviour.
5785 For example this could cause problems if some malloc implementation
5786 provides an address for "p" such that "p + len" actually overflows for
5787 values of len that are too big and therefore p + len < limit.
5789 This issue was reported to OpenSSL by Guido Vranken
5794 * Constant time flag not preserved in DSA signing
5796 Operations in the DSA signing algorithm should run in constant time in
5797 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5798 implementation means that a non-constant time codepath is followed for
5799 certain operations. This has been demonstrated through a cache-timing
5800 attack to be sufficient for an attacker to recover the private DSA key.
5802 This issue was reported by César Pereida (Aalto University), Billy Brumley
5803 (Tampere University of Technology), and Yuval Yarom (The University of
5804 Adelaide and NICTA).
5809 * DTLS buffered message DoS
5811 In a DTLS connection where handshake messages are delivered out-of-order
5812 those messages that OpenSSL is not yet ready to process will be buffered
5813 for later use. Under certain circumstances, a flaw in the logic means that
5814 those messages do not get removed from the buffer even though the handshake
5815 has been completed. An attacker could force up to approx. 15 messages to
5816 remain in the buffer when they are no longer required. These messages will
5817 be cleared when the DTLS connection is closed. The default maximum size for
5818 a message is 100k. Therefore the attacker could force an additional 1500k
5819 to be consumed per connection. By opening many simulataneous connections an
5820 attacker could cause a DoS attack through memory exhaustion.
5822 This issue was reported to OpenSSL by Quan Luo.
5827 * DTLS replay protection DoS
5829 A flaw in the DTLS replay attack protection mechanism means that records
5830 that arrive for future epochs update the replay protection "window" before
5831 the MAC for the record has been validated. This could be exploited by an
5832 attacker by sending a record for the next epoch (which does not have to
5833 decrypt or have a valid MAC), with a very large sequence number. This means
5834 that all subsequent legitimate packets are dropped causing a denial of
5835 service for a specific DTLS connection.
5837 This issue was reported to OpenSSL by the OCAP audit team.
5842 * Certificate message OOB reads
5844 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5845 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5846 theoretical DoS risk but this has not been observed in practice on common
5849 The messages affected are client certificate, client certificate request
5850 and server certificate. As a result the attack can only be performed
5851 against a client or a server which enables client authentication.
5853 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5858 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5860 * Prevent padding oracle in AES-NI CBC MAC check
5862 A MITM attacker can use a padding oracle attack to decrypt traffic
5863 when the connection uses an AES CBC cipher and the server support
5866 This issue was introduced as part of the fix for Lucky 13 padding
5867 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5868 constant time by making sure that always the same bytes are read and
5869 compared against either the MAC or padding bytes. But it no longer
5870 checked that there was enough data to have both the MAC and padding
5873 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5878 * Fix EVP_EncodeUpdate overflow
5880 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5881 Base64 encoding of binary data. If an attacker is able to supply very large
5882 amounts of input data then a length check can overflow resulting in a heap
5885 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5886 the `PEM_write_bio*` family of functions. These are mainly used within the
5887 OpenSSL command line applications, so any application which processes data
5888 from an untrusted source and outputs it as a PEM file should be considered
5889 vulnerable to this issue. User applications that call these APIs directly
5890 with large amounts of untrusted data may also be vulnerable.
5892 This issue was reported by Guido Vranken.
5897 * Fix EVP_EncryptUpdate overflow
5899 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5900 is able to supply very large amounts of input data after a previous call to
5901 EVP_EncryptUpdate() with a partial block then a length check can overflow
5902 resulting in a heap corruption. Following an analysis of all OpenSSL
5903 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5904 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5905 the first called function after an EVP_EncryptInit(), and therefore that
5906 specific call must be safe. The second form is where the length passed to
5907 EVP_EncryptUpdate() can be seen from the code to be some small value and
5908 therefore there is no possibility of an overflow. Since all instances are
5909 one of these two forms, it is believed that there can be no overflows in
5910 internal code due to this problem. It should be noted that
5911 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5912 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5913 of these calls have also been analysed too and it is believed there are no
5914 instances in internal usage where an overflow could occur.
5916 This issue was reported by Guido Vranken.
5921 * Prevent ASN.1 BIO excessive memory allocation
5923 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5924 a short invalid encoding can casuse allocation of large amounts of memory
5925 potentially consuming excessive resources or exhausting memory.
5927 Any application parsing untrusted data through d2i BIO functions is
5928 affected. The memory based functions such as d2i_X509() are *not* affected.
5929 Since the memory based functions are used by the TLS library, TLS
5930 applications are not affected.
5932 This issue was reported by Brian Carpenter.
5939 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5940 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5941 in arbitrary stack data being returned in the buffer.
5943 This issue was reported by Guido Vranken.
5948 * Modify behavior of ALPN to invoke callback after SNI/servername
5949 callback, such that updates to the SSL_CTX affect ALPN.
5953 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5958 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5959 methods are enabled and ssl2 is disabled the methods return NULL.
5963 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5965 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5966 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5967 provide any "EXPORT" or "LOW" strength ciphers.
5971 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5972 is by default disabled at build-time. Builds that are not configured with
5973 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5974 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5975 will need to explicitly call either of:
5977 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5979 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5981 as appropriate. Even if either of those is used, or the application
5982 explicitly uses the version-specific SSLv2_method() or its client and
5983 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5984 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5985 ciphers, and SSLv2 56-bit DES are no longer available.
5990 * Fix a double-free in DSA code
5992 A double free bug was discovered when OpenSSL parses malformed DSA private
5993 keys and could lead to a DoS attack or memory corruption for applications
5994 that receive DSA private keys from untrusted sources. This scenario is
5997 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6003 * Disable SRP fake user seed to address a server memory leak.
6005 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6007 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6008 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6009 was changed to ignore the "fake user" SRP seed, even if the seed
6012 Users should use SRP_VBASE_get1_by_user instead. Note that in
6013 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6014 also that even though configuring the SRP seed attempts to hide
6015 invalid usernames by continuing the handshake with fake
6016 credentials, this behaviour is not constant time and no strong
6017 guarantees are made that the handshake is indistinguishable from
6018 that of a valid user.
6023 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6025 In the BN_hex2bn function the number of hex digits is calculated using an
6026 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6027 large values of `i` this can result in `bn_expand` not allocating any
6028 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6029 field as NULL leading to a subsequent NULL ptr deref. For very large values
6030 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6031 In this case memory is allocated to the internal BIGNUM data field, but it
6032 is insufficiently sized leading to heap corruption. A similar issue exists
6033 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6034 is ever called by user applications with very large untrusted hex/dec data.
6035 This is anticipated to be a rare occurrence.
6037 All OpenSSL internal usage of these functions use data that is not expected
6038 to be untrusted, e.g. config file data or application command line
6039 arguments. If user developed applications generate config file data based
6040 on untrusted data then it is possible that this could also lead to security
6041 consequences. This is also anticipated to be rare.
6043 This issue was reported to OpenSSL by Guido Vranken.
6048 * Fix memory issues in `BIO_*printf` functions
6050 The internal `fmtstr` function used in processing a "%s" format string in
6051 the `BIO_*printf` functions could overflow while calculating the length of a
6052 string and cause an OOB read when printing very long strings.
6054 Additionally the internal `doapr_outch` function can attempt to write to an
6055 OOB memory location (at an offset from the NULL pointer) in the event of a
6056 memory allocation failure. In 1.0.2 and below this could be caused where
6057 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6058 could be in processing a very long "%s" format string. Memory leaks can
6061 The first issue may mask the second issue dependent on compiler behaviour.
6062 These problems could enable attacks where large amounts of untrusted data
6063 is passed to the `BIO_*printf` functions. If applications use these functions
6064 in this way then they could be vulnerable. OpenSSL itself uses these
6065 functions when printing out human-readable dumps of ASN.1 data. Therefore
6066 applications that print this data could be vulnerable if the data is from
6067 untrusted sources. OpenSSL command line applications could also be
6068 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6069 as command line arguments.
6071 Libssl is not considered directly vulnerable. Additionally certificates etc
6072 received via remote connections via libssl are also unlikely to be able to
6073 trigger these issues because of message size limits enforced within libssl.
6075 This issue was reported to OpenSSL Guido Vranken.
6080 * Side channel attack on modular exponentiation
6082 A side-channel attack was found which makes use of cache-bank conflicts on
6083 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6084 of RSA keys. The ability to exploit this issue is limited as it relies on
6085 an attacker who has control of code in a thread running on the same
6086 hyper-threaded core as the victim thread which is performing decryptions.
6088 This issue was reported to OpenSSL by Yuval Yarom, The University of
6089 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6090 Nadia Heninger, University of Pennsylvania with more information at
6091 <http://cachebleed.info>.
6096 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6097 if no keysize is specified with default_bits. This fixes an
6098 omission in an earlier change that changed all RSA/DSA key generation
6099 commands to use 2048 bits by default.
6103 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6105 * Protection for DH small subgroup attacks
6107 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6108 switched on by default and cannot be disabled. This could have some
6113 * SSLv2 doesn't block disabled ciphers
6115 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6116 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6117 been disabled, provided that the SSLv2 protocol was not also disabled via
6120 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6121 and Sebastian Schinzel.
6126 * Reject DH handshakes with parameters shorter than 1024 bits.
6130 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6132 * Certificate verify crash with missing PSS parameter
6134 The signature verification routines will crash with a NULL pointer
6135 dereference if presented with an ASN.1 signature using the RSA PSS
6136 algorithm and absent mask generation function parameter. Since these
6137 routines are used to verify certificate signature algorithms this can be
6138 used to crash any certificate verification operation and exploited in a
6139 DoS attack. Any application which performs certificate verification is
6140 vulnerable including OpenSSL clients and servers which enable client
6143 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6148 * X509_ATTRIBUTE memory leak
6150 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6151 memory. This structure is used by the PKCS#7 and CMS routines so any
6152 application which reads PKCS#7 or CMS data from untrusted sources is
6153 affected. SSL/TLS is not affected.
6155 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6161 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6162 This changes the decoding behaviour for some invalid messages,
6163 though the change is mostly in the more lenient direction, and
6164 legacy behaviour is preserved as much as possible.
6168 * In DSA_generate_parameters_ex, if the provided seed is too short,
6169 use a random seed, as already documented.
6171 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6173 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6175 * Alternate chains certificate forgery
6177 During certificate verfification, OpenSSL will attempt to find an
6178 alternative certificate chain if the first attempt to build such a chain
6179 fails. An error in the implementation of this logic can mean that an
6180 attacker could cause certain checks on untrusted certificates to be
6181 bypassed, such as the CA flag, enabling them to use a valid leaf
6182 certificate to act as a CA and "issue" an invalid certificate.
6184 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6190 * Race condition handling PSK identify hint
6192 If PSK identity hints are received by a multi-threaded client then
6193 the values are wrongly updated in the parent SSL_CTX structure. This can
6194 result in a race condition potentially leading to a double free of the
6200 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6202 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6203 incompatibility in the handling of HMAC. The previous ABI has now been
6206 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6208 * Malformed ECParameters causes infinite loop
6210 When processing an ECParameters structure OpenSSL enters an infinite loop
6211 if the curve specified is over a specially malformed binary polynomial
6214 This can be used to perform denial of service against any
6215 system which processes public keys, certificate requests or
6216 certificates. This includes TLS clients and TLS servers with
6217 client authentication enabled.
6219 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6224 * Exploitable out-of-bounds read in X509_cmp_time
6226 X509_cmp_time does not properly check the length of the ASN1_TIME
6227 string and can read a few bytes out of bounds. In addition,
6228 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6231 An attacker can use this to craft malformed certificates and CRLs of
6232 various sizes and potentially cause a segmentation fault, resulting in
6233 a DoS on applications that verify certificates or CRLs. TLS clients
6234 that verify CRLs are affected. TLS clients and servers with client
6235 authentication enabled may be affected if they use custom verification
6238 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6239 independently by Hanno Böck.
6244 * PKCS7 crash with missing EnvelopedContent
6246 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6247 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6248 with missing content and trigger a NULL pointer dereference on parsing.
6250 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6251 structures from untrusted sources are affected. OpenSSL clients and
6252 servers are not affected.
6254 This issue was reported to OpenSSL by Michal Zalewski (Google).
6259 * CMS verify infinite loop with unknown hash function
6261 When verifying a signedData message the CMS code can enter an infinite loop
6262 if presented with an unknown hash function OID. This can be used to perform
6263 denial of service against any system which verifies signedData messages using
6265 This issue was reported to OpenSSL by Johannes Bauer.
6270 * Race condition handling NewSessionTicket
6272 If a NewSessionTicket is received by a multi-threaded client when attempting to
6273 reuse a previous ticket then a race condition can occur potentially leading to
6274 a double free of the ticket data.
6279 * Reject DH handshakes with parameters shorter than 768 bits.
6281 *Kurt Roeckx and Emilia Kasper*
6283 * dhparam: generate 2048-bit parameters by default.
6285 *Kurt Roeckx and Emilia Kasper*
6287 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6289 * Segmentation fault in ASN1_TYPE_cmp fix
6291 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6292 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6293 certificate signature algorithm consistency this can be used to crash any
6294 certificate verification operation and exploited in a DoS attack. Any
6295 application which performs certificate verification is vulnerable including
6296 OpenSSL clients and servers which enable client authentication.
6301 * ASN.1 structure reuse memory corruption fix
6303 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6304 memory corruption via an invalid write. Such reuse is and has been
6305 strongly discouraged and is believed to be rare.
6307 Applications that parse structures containing CHOICE or ANY DEFINED BY
6308 components may be affected. Certificate parsing (d2i_X509 and related
6309 functions) are however not affected. OpenSSL clients and servers are
6315 * PKCS7 NULL pointer dereferences fix
6317 The PKCS#7 parsing code does not handle missing outer ContentInfo
6318 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6319 missing content and trigger a NULL pointer dereference on parsing.
6321 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6322 otherwise parse PKCS#7 structures from untrusted sources are
6323 affected. OpenSSL clients and servers are not affected.
6325 This issue was reported to OpenSSL by Michal Zalewski (Google).
6330 * DoS via reachable assert in SSLv2 servers fix
6332 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6333 servers that both support SSLv2 and enable export cipher suites by sending
6334 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6336 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6337 (OpenSSL development team).
6342 * Use After Free following d2i_ECPrivatekey error fix
6344 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6345 could cause a use after free condition. This, in turn, could cause a double
6346 free in several private key parsing functions (such as d2i_PrivateKey
6347 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6348 for applications that receive EC private keys from untrusted
6349 sources. This scenario is considered rare.
6351 This issue was discovered by the BoringSSL project and fixed in their
6357 * X509_to_X509_REQ NULL pointer deref fix
6359 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6360 the certificate key is invalid. This function is rarely used in practice.
6362 This issue was discovered by Brian Carpenter.
6367 * Removed the export ciphers from the DEFAULT ciphers
6371 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6373 * Build fixes for the Windows and OpenVMS platforms
6375 *Matt Caswell and Richard Levitte*
6377 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6379 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6380 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6381 dereference. This could lead to a Denial Of Service attack. Thanks to
6382 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6387 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6388 dtls1_buffer_record function under certain conditions. In particular this
6389 could occur if an attacker sent repeated DTLS records with the same
6390 sequence number but for the next epoch. The memory leak could be exploited
6391 by an attacker in a Denial of Service attack through memory exhaustion.
6392 Thanks to Chris Mueller for reporting this issue.
6397 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6398 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6399 method would be set to NULL which could later result in a NULL pointer
6400 dereference. Thanks to Frank Schmirler for reporting this issue.
6405 * Abort handshake if server key exchange message is omitted for ephemeral
6408 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6409 reporting this issue.
6414 * Remove non-export ephemeral RSA code on client and server. This code
6415 violated the TLS standard by allowing the use of temporary RSA keys in
6416 non-export ciphersuites and could be used by a server to effectively
6417 downgrade the RSA key length used to a value smaller than the server
6418 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6419 INRIA or reporting this issue.
6424 * Fixed issue where DH client certificates are accepted without verification.
6425 An OpenSSL server will accept a DH certificate for client authentication
6426 without the certificate verify message. This effectively allows a client to
6427 authenticate without the use of a private key. This only affects servers
6428 which trust a client certificate authority which issues certificates
6429 containing DH keys: these are extremely rare and hardly ever encountered.
6430 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6436 * Ensure that the session ID context of an SSL is updated when its
6437 SSL_CTX is updated via SSL_set_SSL_CTX.
6439 The session ID context is typically set from the parent SSL_CTX,
6440 and can vary with the CTX.
6444 * Fix various certificate fingerprint issues.
6446 By using non-DER or invalid encodings outside the signed portion of a
6447 certificate the fingerprint can be changed without breaking the signature.
6448 Although no details of the signed portion of the certificate can be changed
6449 this can cause problems with some applications: e.g. those using the
6450 certificate fingerprint for blacklists.
6452 1. Reject signatures with non zero unused bits.
6454 If the BIT STRING containing the signature has non zero unused bits reject
6455 the signature. All current signature algorithms require zero unused bits.
6457 2. Check certificate algorithm consistency.
6459 Check the AlgorithmIdentifier inside TBS matches the one in the
6460 certificate signature. NB: this will result in signature failure
6461 errors for some broken certificates.
6463 Thanks to Konrad Kraszewski from Google for reporting this issue.
6465 3. Check DSA/ECDSA signatures use DER.
6467 Re-encode DSA/ECDSA signatures and compare with the original received
6468 signature. Return an error if there is a mismatch.
6470 This will reject various cases including garbage after signature
6471 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6472 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6473 (negative or with leading zeroes).
6475 Further analysis was conducted and fixes were developed by Stephen Henson
6476 of the OpenSSL core team.
6482 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6483 results on some platforms, including x86_64. This bug occurs at random
6484 with a very low probability, and is not known to be exploitable in any
6485 way, though its exact impact is difficult to determine. Thanks to Pieter
6486 Wuille (Blockstream) who reported this issue and also suggested an initial
6487 fix. Further analysis was conducted by the OpenSSL development team and
6488 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6489 the OpenSSL core team.
6494 * Do not resume sessions on the server if the negotiated protocol
6495 version does not match the session's version. Resuming with a different
6496 version, while not strictly forbidden by the RFC, is of questionable
6497 sanity and breaks all known clients.
6499 *David Benjamin, Emilia Käsper*
6501 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6502 early CCS messages during renegotiation. (Note that because
6503 renegotiation is encrypted, this early CCS was not exploitable.)
6507 * Tighten client-side session ticket handling during renegotiation:
6508 ensure that the client only accepts a session ticket if the server sends
6509 the extension anew in the ServerHello. Previously, a TLS client would
6510 reuse the old extension state and thus accept a session ticket if one was
6511 announced in the initial ServerHello.
6513 Similarly, ensure that the client requires a session ticket if one
6514 was advertised in the ServerHello. Previously, a TLS client would
6515 ignore a missing NewSessionTicket message.
6519 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6523 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6524 sends a carefully crafted handshake message, to cause OpenSSL to fail
6525 to free up to 64k of memory causing a memory leak. This could be
6526 exploited in a Denial Of Service attack. This issue affects OpenSSL
6527 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6528 whether SRTP is used or configured. Implementations of OpenSSL that
6529 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6531 The fix was developed by the OpenSSL team.
6536 * Session Ticket Memory Leak.
6538 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6539 integrity of that ticket is first verified. In the event of a session
6540 ticket integrity check failing, OpenSSL will fail to free memory
6541 causing a memory leak. By sending a large number of invalid session
6542 tickets an attacker could exploit this issue in a Denial Of Service
6548 * Build option no-ssl3 is incomplete.
6550 When OpenSSL is configured with "no-ssl3" as a build option, servers
6551 could accept and complete a SSL 3.0 handshake, and clients could be
6552 configured to send them.
6555 *Akamai and the OpenSSL team*
6557 * Add support for TLS_FALLBACK_SCSV.
6558 Client applications doing fallback retries should call
6559 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6562 *Adam Langley, Bodo Moeller*
6564 * Add additional DigestInfo checks.
6566 Re-encode DigestInto in DER and check against the original when
6567 verifying RSA signature: this will reject any improperly encoded
6568 DigestInfo structures.
6570 Note: this is a precautionary measure and no attacks are currently known.
6574 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6576 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6577 SRP code can be overrun an internal buffer. Add sanity check that
6578 g, A, B < N to SRP code.
6580 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6581 Group for discovering this issue.
6586 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6587 TLS 1.0 instead of higher protocol versions when the ClientHello message
6588 is badly fragmented. This allows a man-in-the-middle attacker to force a
6589 downgrade to TLS 1.0 even if both the server and the client support a
6590 higher protocol version, by modifying the client's TLS records.
6592 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6593 researching this issue.
6598 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6599 to a denial of service attack. A malicious server can crash the client
6600 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6601 ciphersuite and sending carefully crafted handshake messages.
6603 Thanks to Felix Gröbert (Google) for discovering and researching this
6609 * By sending carefully crafted DTLS packets an attacker could cause openssl
6610 to leak memory. This can be exploited through a Denial of Service attack.
6611 Thanks to Adam Langley for discovering and researching this issue.
6616 * An attacker can force openssl to consume large amounts of memory whilst
6617 processing DTLS handshake messages. This can be exploited through a
6618 Denial of Service attack.
6619 Thanks to Adam Langley for discovering and researching this issue.
6624 * An attacker can force an error condition which causes openssl to crash
6625 whilst processing DTLS packets due to memory being freed twice. This
6626 can be exploited through a Denial of Service attack.
6627 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6633 * If a multithreaded client connects to a malicious server using a resumed
6634 session and the server sends an ec point format extension it could write
6635 up to 255 bytes to freed memory.
6637 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6643 * A malicious server can crash an OpenSSL client with a null pointer
6644 dereference (read) by specifying an SRP ciphersuite even though it was not
6645 properly negotiated with the client. This can be exploited through a
6646 Denial of Service attack.
6648 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6649 discovering and researching this issue.
6654 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6655 X509_name_oneline, X509_name_print_ex et al. to leak some information
6656 from the stack. Applications may be affected if they echo pretty printing
6657 output to the attacker.
6659 Thanks to Ivan Fratric (Google) for discovering this issue.
6662 *Emilia Käsper, and Steve Henson*
6664 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6665 for corner cases. (Certain input points at infinity could lead to
6666 bogus results, with non-infinity inputs mapped to infinity too.)
6670 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6672 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6673 handshake can force the use of weak keying material in OpenSSL
6674 SSL/TLS clients and servers.
6676 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6677 researching this issue. ([CVE-2014-0224])
6679 *KIKUCHI Masashi, Steve Henson*
6681 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6682 OpenSSL DTLS client the code can be made to recurse eventually crashing
6685 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6688 *Imre Rad, Steve Henson*
6690 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6691 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6692 client or server. This is potentially exploitable to run arbitrary
6693 code on a vulnerable client or server.
6695 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6697 *Jüri Aedla, Steve Henson*
6699 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6700 are subject to a denial of service attack.
6702 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6703 this issue. ([CVE-2014-3470])
6705 *Felix Gröbert, Ivan Fratric, Steve Henson*
6707 * Harmonize version and its documentation. -f flag is used to display
6710 *mancha <mancha1@zoho.com>*
6712 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6713 in i2d_ECPrivateKey.
6715 *mancha <mancha1@zoho.com>*
6717 * Fix some double frees. These are not thought to be exploitable.
6719 *mancha <mancha1@zoho.com>*
6721 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6723 * A missing bounds check in the handling of the TLS heartbeat extension
6724 can be used to reveal up to 64k of memory to a connected client or
6727 Thanks for Neel Mehta of Google Security for discovering this bug and to
6728 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6729 preparing the fix ([CVE-2014-0160])
6731 *Adam Langley, Bodo Moeller*
6733 * Fix for the attack described in the paper "Recovering OpenSSL
6734 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6735 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6736 <http://eprint.iacr.org/2014/140>
6738 Thanks to Yuval Yarom and Naomi Benger for discovering this
6739 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6741 *Yuval Yarom and Naomi Benger*
6743 * TLS pad extension: draft-agl-tls-padding-03
6745 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6746 TLS client Hello record length value would otherwise be > 255 and
6747 less that 512 pad with a dummy extension containing zeroes so it
6748 is at least 512 bytes long.
6750 *Adam Langley, Steve Henson*
6752 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6754 * Fix for TLS record tampering bug. A carefully crafted invalid
6755 handshake could crash OpenSSL with a NULL pointer exception.
6756 Thanks to Anton Johansson for reporting this issues.
6759 * Keep original DTLS digest and encryption contexts in retransmission
6760 structures so we can use the previous session parameters if they need
6761 to be resent. ([CVE-2013-6450])
6765 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6766 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6767 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6768 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6769 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6770 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6772 *Rob Stradling, Adam Langley*
6774 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6776 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6777 supporting platforms or when small records were transferred.
6779 *Andy Polyakov, Steve Henson*
6781 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6783 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6785 This addresses the flaw in CBC record processing discovered by
6786 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6787 at: <http://www.isg.rhul.ac.uk/tls/>
6789 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6790 Security Group at Royal Holloway, University of London
6791 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6792 Emilia Käsper for the initial patch.
6795 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6797 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6798 ciphersuites which can be exploited in a denial of service attack.
6799 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6800 and detecting this bug and to Wolfgang Ettlinger
6801 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6806 * Return an error when checking OCSP signatures when key is NULL.
6807 This fixes a DoS attack. ([CVE-2013-0166])
6811 * Make openssl verify return errors.
6813 *Chris Palmer <palmer@google.com> and Ben Laurie*
6815 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6816 the right response is stapled. Also change SSL_get_certificate()
6817 so it returns the certificate actually sent.
6818 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6820 *Rob Stradling <rob.stradling@comodo.com>*
6822 * Fix possible deadlock when decoding public keys.
6826 * Don't use TLS 1.0 record version number in initial client hello
6831 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6833 * Sanity check record length before skipping explicit IV in TLS
6834 1.2, 1.1 and DTLS to fix DoS attack.
6836 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6837 fuzzing as a service testing platform.
6842 * Initialise tkeylen properly when encrypting CMS messages.
6843 Thanks to Solar Designer of Openwall for reporting this issue.
6847 * In FIPS mode don't try to use composite ciphers as they are not
6852 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6854 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6855 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6856 mean any application compiled against OpenSSL 1.0.0 headers setting
6857 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6858 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6859 0x10000000L Any application which was previously compiled against
6860 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6861 will need to be recompiled as a result. Letting be results in
6862 inability to disable specifically TLS 1.1 and in client context,
6863 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6867 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6868 disable just protocol X, but all protocols above X *if* there are
6869 protocols *below* X still enabled. In more practical terms it means
6870 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6871 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6872 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6877 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6879 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6880 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6881 in CRYPTO_realloc_clean.
6883 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6884 issue and to Adam Langley <agl@chromium.org> for fixing it.
6887 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6889 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6893 * Workarounds for some broken servers that "hang" if a client hello
6894 record length exceeds 255 bytes.
6896 1. Do not use record version number > TLS 1.0 in initial client
6897 hello: some (but not all) hanging servers will now work.
6898 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6899 the number of ciphers sent in the client hello. This should be
6900 set to an even number, such as 50, for example by passing:
6901 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6902 Most broken servers should now work.
6903 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6904 TLS 1.2 client support entirely.
6908 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6912 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6914 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6915 STRING form instead of a DigestInfo.
6919 * The format used for MDC2 RSA signatures is inconsistent between EVP
6920 and the RSA_sign/RSA_verify functions. This was made more apparent when
6921 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6922 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6923 the correct format in RSA_verify so both forms transparently work.
6927 * Some servers which support TLS 1.0 can choke if we initially indicate
6928 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6929 encrypted premaster secret. As a workaround use the maximum permitted
6930 client version in client hello, this should keep such servers happy
6931 and still work with previous versions of OpenSSL.
6935 * Add support for TLS/DTLS heartbeats.
6937 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6939 * Add support for SCTP.
6941 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6943 * Improved PRNG seeding for VOS.
6945 *Paul Green <Paul.Green@stratus.com>*
6947 * Extensive assembler packs updates, most notably:
6949 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6950 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6951 - x86_64: bit-sliced AES implementation;
6952 - ARM: NEON support, contemporary platforms optimizations;
6953 - s390x: z196 support;
6954 - `*`: GHASH and GF(2^m) multiplication implementations;
6958 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6959 (removal of unnecessary code)
6961 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6963 * Add TLS key material exporter from RFC 5705.
6967 * Add DTLS-SRTP negotiation from RFC 5764.
6971 * Add Next Protocol Negotiation,
6972 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6973 disabled with a no-npn flag to config or Configure. Code donated
6976 *Adam Langley <agl@google.com> and Ben Laurie*
6978 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6979 NIST-P256, NIST-P521, with constant-time single point multiplication on
6980 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6981 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6982 Code made available under Apache License version 2.0.
6984 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6985 line to include this in your build of OpenSSL, and run "make depend" (or
6986 "make update"). This enables the following EC_METHODs:
6988 EC_GFp_nistp224_method()
6989 EC_GFp_nistp256_method()
6990 EC_GFp_nistp521_method()
6992 EC_GROUP_new_by_curve_name() will automatically use these (while
6993 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6996 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6998 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6999 all platforms. Move ssize_t definition from e_os.h to the public
7000 header file e_os2.h as it now appears in public header file cms.h
7004 * New -sigopt option to the ca, req and x509 utilities. Additional
7005 signature parameters can be passed using this option and in
7010 * Add RSA PSS signing function. This will generate and set the
7011 appropriate AlgorithmIdentifiers for PSS based on those in the
7012 corresponding EVP_MD_CTX structure. No application support yet.
7016 * Support for companion algorithm specific ASN1 signing routines.
7017 New function ASN1_item_sign_ctx() signs a pre-initialised
7018 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7019 the appropriate parameters.
7023 * Add new algorithm specific ASN1 verification initialisation function
7024 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7025 handling will be the same no matter what EVP_PKEY_METHOD is used.
7026 Add a PSS handler to support verification of PSS signatures: checked
7027 against a number of sample certificates.
7031 * Add signature printing for PSS. Add PSS OIDs.
7033 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7035 * Add algorithm specific signature printing. An individual ASN1 method
7036 can now print out signatures instead of the standard hex dump.
7038 More complex signatures (e.g. PSS) can print out more meaningful
7039 information. Include DSA version that prints out the signature
7044 * Password based recipient info support for CMS library: implementing
7049 * Split password based encryption into PBES2 and PBKDF2 functions. This
7050 neatly separates the code into cipher and PBE sections and is required
7051 for some algorithms that split PBES2 into separate pieces (such as
7052 password based CMS).
7056 * Session-handling fixes:
7057 - Fix handling of connections that are resuming with a session ID,
7058 but also support Session Tickets.
7059 - Fix a bug that suppressed issuing of a new ticket if the client
7060 presented a ticket with an expired session.
7061 - Try to set the ticket lifetime hint to something reasonable.
7062 - Make tickets shorter by excluding irrelevant information.
7063 - On the client side, don't ignore renewed tickets.
7065 *Adam Langley, Bodo Moeller (Google)*
7067 * Fix PSK session representation.
7071 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7073 This work was sponsored by Intel.
7077 * Add GCM support to TLS library. Some custom code is needed to split
7078 the IV between the fixed (from PRF) and explicit (from TLS record)
7079 portions. This adds all GCM ciphersuites supported by RFC5288 and
7080 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7081 add a special AESGCM string for GCM only.
7085 * Expand range of ctrls for AES GCM. Permit setting invocation
7086 field on decrypt and retrieval of invocation field only on encrypt.
7090 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7091 As required by RFC5289 these ciphersuites cannot be used if for
7092 versions of TLS earlier than 1.2.
7096 * For FIPS capable OpenSSL interpret a NULL default public key method
7097 as unset and return the appropriate default but do *not* set the default.
7098 This means we can return the appropriate method in applications that
7099 switch between FIPS and non-FIPS modes.
7103 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7104 ENGINE is used then we cannot handle that in the FIPS module so we
7105 keep original code iff non-FIPS operations are allowed.
7109 * Add -attime option to openssl utilities.
7111 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7113 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7117 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7118 FIPS EC methods unconditionally for now.
7122 * New build option no-ec2m to disable characteristic 2 code.
7126 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7127 all cases can be covered as some introduce binary incompatibilities.
7131 * Redirect RSA operations to FIPS module including keygen,
7132 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7136 * Add similar low-level API blocking to ciphers.
7140 * low-level digest APIs are not approved in FIPS mode: any attempt
7141 to use these will cause a fatal error. Applications that *really* want
7142 to use them can use the `private_*` version instead.
7146 * Redirect cipher operations to FIPS module for FIPS builds.
7150 * Redirect digest operations to FIPS module for FIPS builds.
7154 * Update build system to add "fips" flag which will link in fipscanister.o
7155 for static and shared library builds embedding a signature if needed.
7159 * Output TLS supported curves in preference order instead of numerical
7160 order. This is currently hardcoded for the highest order curves first.
7161 This should be configurable so applications can judge speed vs strength.
7165 * Add TLS v1.2 server support for client authentication.
7169 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7174 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7175 FIPS modules versions.
7179 * Add TLS v1.2 client side support for client authentication. Keep cache
7180 of handshake records longer as we don't know the hash algorithm to use
7181 until after the certificate request message is received.
7185 * Initial TLS v1.2 client support. Add a default signature algorithms
7186 extension including all the algorithms we support. Parse new signature
7187 format in client key exchange. Relax some ECC signing restrictions for
7188 TLS v1.2 as indicated in RFC5246.
7192 * Add server support for TLS v1.2 signature algorithms extension. Switch
7193 to new signature format when needed using client digest preference.
7194 All server ciphersuites should now work correctly in TLS v1.2. No client
7195 support yet and no support for client certificates.
7199 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7200 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7201 ciphersuites. At present only RSA key exchange ciphersuites work with
7202 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7203 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7204 and version checking.
7208 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7209 with this defined it will not be affected by any changes to ssl internal
7210 structures. Add several utility functions to allow openssl application
7211 to work with OPENSSL_NO_SSL_INTERN defined.
7215 * A long standing patch to add support for SRP from EdelWeb (Peter
7216 Sylvester and Christophe Renou) was integrated.
7217 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7218 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7221 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7225 * Permit abbreviated handshakes when renegotiating using the function
7226 SSL_renegotiate_abbreviated().
7228 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7230 * Add call to ENGINE_register_all_complete() to
7231 ENGINE_load_builtin_engines(), so some implementations get used
7232 automatically instead of needing explicit application support.
7236 * Add support for TLS key exporter as described in RFC5705.
7238 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7240 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7241 a few changes are required:
7243 Add SSL_OP_NO_TLSv1_1 flag.
7244 Add TLSv1_1 methods.
7245 Update version checking logic to handle version 1.1.
7246 Add explicit IV handling (ported from DTLS code).
7247 Add command line options to s_client/s_server.
7254 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7256 * X509_ATTRIBUTE memory leak
7258 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7259 memory. This structure is used by the PKCS#7 and CMS routines so any
7260 application which reads PKCS#7 or CMS data from untrusted sources is
7261 affected. SSL/TLS is not affected.
7263 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7269 * Race condition handling PSK identify hint
7271 If PSK identity hints are received by a multi-threaded client then
7272 the values are wrongly updated in the parent SSL_CTX structure. This can
7273 result in a race condition potentially leading to a double free of the
7279 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7281 * Malformed ECParameters causes infinite loop
7283 When processing an ECParameters structure OpenSSL enters an infinite loop
7284 if the curve specified is over a specially malformed binary polynomial
7287 This can be used to perform denial of service against any
7288 system which processes public keys, certificate requests or
7289 certificates. This includes TLS clients and TLS servers with
7290 client authentication enabled.
7292 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7297 * Exploitable out-of-bounds read in X509_cmp_time
7299 X509_cmp_time does not properly check the length of the ASN1_TIME
7300 string and can read a few bytes out of bounds. In addition,
7301 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7304 An attacker can use this to craft malformed certificates and CRLs of
7305 various sizes and potentially cause a segmentation fault, resulting in
7306 a DoS on applications that verify certificates or CRLs. TLS clients
7307 that verify CRLs are affected. TLS clients and servers with client
7308 authentication enabled may be affected if they use custom verification
7311 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7312 independently by Hanno Böck.
7317 * PKCS7 crash with missing EnvelopedContent
7319 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7320 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7321 with missing content and trigger a NULL pointer dereference on parsing.
7323 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7324 structures from untrusted sources are affected. OpenSSL clients and
7325 servers are not affected.
7327 This issue was reported to OpenSSL by Michal Zalewski (Google).
7332 * CMS verify infinite loop with unknown hash function
7334 When verifying a signedData message the CMS code can enter an infinite loop
7335 if presented with an unknown hash function OID. This can be used to perform
7336 denial of service against any system which verifies signedData messages using
7338 This issue was reported to OpenSSL by Johannes Bauer.
7343 * Race condition handling NewSessionTicket
7345 If a NewSessionTicket is received by a multi-threaded client when attempting to
7346 reuse a previous ticket then a race condition can occur potentially leading to
7347 a double free of the ticket data.
7352 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7354 * Segmentation fault in ASN1_TYPE_cmp fix
7356 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7357 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7358 certificate signature algorithm consistency this can be used to crash any
7359 certificate verification operation and exploited in a DoS attack. Any
7360 application which performs certificate verification is vulnerable including
7361 OpenSSL clients and servers which enable client authentication.
7366 * ASN.1 structure reuse memory corruption fix
7368 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7369 memory corruption via an invalid write. Such reuse is and has been
7370 strongly discouraged and is believed to be rare.
7372 Applications that parse structures containing CHOICE or ANY DEFINED BY
7373 components may be affected. Certificate parsing (d2i_X509 and related
7374 functions) are however not affected. OpenSSL clients and servers are
7380 * PKCS7 NULL pointer dereferences fix
7382 The PKCS#7 parsing code does not handle missing outer ContentInfo
7383 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7384 missing content and trigger a NULL pointer dereference on parsing.
7386 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7387 otherwise parse PKCS#7 structures from untrusted sources are
7388 affected. OpenSSL clients and servers are not affected.
7390 This issue was reported to OpenSSL by Michal Zalewski (Google).
7395 * DoS via reachable assert in SSLv2 servers fix
7397 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7398 servers that both support SSLv2 and enable export cipher suites by sending
7399 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7401 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7402 (OpenSSL development team).
7407 * Use After Free following d2i_ECPrivatekey error fix
7409 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7410 could cause a use after free condition. This, in turn, could cause a double
7411 free in several private key parsing functions (such as d2i_PrivateKey
7412 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7413 for applications that receive EC private keys from untrusted
7414 sources. This scenario is considered rare.
7416 This issue was discovered by the BoringSSL project and fixed in their
7422 * X509_to_X509_REQ NULL pointer deref fix
7424 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7425 the certificate key is invalid. This function is rarely used in practice.
7427 This issue was discovered by Brian Carpenter.
7432 * Removed the export ciphers from the DEFAULT ciphers
7436 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7438 * Build fixes for the Windows and OpenVMS platforms
7440 *Matt Caswell and Richard Levitte*
7442 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7444 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7445 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7446 dereference. This could lead to a Denial Of Service attack. Thanks to
7447 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7452 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7453 dtls1_buffer_record function under certain conditions. In particular this
7454 could occur if an attacker sent repeated DTLS records with the same
7455 sequence number but for the next epoch. The memory leak could be exploited
7456 by an attacker in a Denial of Service attack through memory exhaustion.
7457 Thanks to Chris Mueller for reporting this issue.
7462 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7463 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7464 method would be set to NULL which could later result in a NULL pointer
7465 dereference. Thanks to Frank Schmirler for reporting this issue.
7470 * Abort handshake if server key exchange message is omitted for ephemeral
7473 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7474 reporting this issue.
7479 * Remove non-export ephemeral RSA code on client and server. This code
7480 violated the TLS standard by allowing the use of temporary RSA keys in
7481 non-export ciphersuites and could be used by a server to effectively
7482 downgrade the RSA key length used to a value smaller than the server
7483 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7484 INRIA or reporting this issue.
7489 * Fixed issue where DH client certificates are accepted without verification.
7490 An OpenSSL server will accept a DH certificate for client authentication
7491 without the certificate verify message. This effectively allows a client to
7492 authenticate without the use of a private key. This only affects servers
7493 which trust a client certificate authority which issues certificates
7494 containing DH keys: these are extremely rare and hardly ever encountered.
7495 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7501 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7502 results on some platforms, including x86_64. This bug occurs at random
7503 with a very low probability, and is not known to be exploitable in any
7504 way, though its exact impact is difficult to determine. Thanks to Pieter
7505 Wuille (Blockstream) who reported this issue and also suggested an initial
7506 fix. Further analysis was conducted by the OpenSSL development team and
7507 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7508 the OpenSSL core team.
7513 * Fix various certificate fingerprint issues.
7515 By using non-DER or invalid encodings outside the signed portion of a
7516 certificate the fingerprint can be changed without breaking the signature.
7517 Although no details of the signed portion of the certificate can be changed
7518 this can cause problems with some applications: e.g. those using the
7519 certificate fingerprint for blacklists.
7521 1. Reject signatures with non zero unused bits.
7523 If the BIT STRING containing the signature has non zero unused bits reject
7524 the signature. All current signature algorithms require zero unused bits.
7526 2. Check certificate algorithm consistency.
7528 Check the AlgorithmIdentifier inside TBS matches the one in the
7529 certificate signature. NB: this will result in signature failure
7530 errors for some broken certificates.
7532 Thanks to Konrad Kraszewski from Google for reporting this issue.
7534 3. Check DSA/ECDSA signatures use DER.
7536 Reencode DSA/ECDSA signatures and compare with the original received
7537 signature. Return an error if there is a mismatch.
7539 This will reject various cases including garbage after signature
7540 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7541 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7542 (negative or with leading zeroes).
7544 Further analysis was conducted and fixes were developed by Stephen Henson
7545 of the OpenSSL core team.
7551 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7553 * Session Ticket Memory Leak.
7555 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7556 integrity of that ticket is first verified. In the event of a session
7557 ticket integrity check failing, OpenSSL will fail to free memory
7558 causing a memory leak. By sending a large number of invalid session
7559 tickets an attacker could exploit this issue in a Denial Of Service
7565 * Build option no-ssl3 is incomplete.
7567 When OpenSSL is configured with "no-ssl3" as a build option, servers
7568 could accept and complete a SSL 3.0 handshake, and clients could be
7569 configured to send them.
7572 *Akamai and the OpenSSL team*
7574 * Add support for TLS_FALLBACK_SCSV.
7575 Client applications doing fallback retries should call
7576 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7579 *Adam Langley, Bodo Moeller*
7581 * Add additional DigestInfo checks.
7583 Reencode DigestInto in DER and check against the original when
7584 verifying RSA signature: this will reject any improperly encoded
7585 DigestInfo structures.
7587 Note: this is a precautionary measure and no attacks are currently known.
7591 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7593 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7594 to a denial of service attack. A malicious server can crash the client
7595 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7596 ciphersuite and sending carefully crafted handshake messages.
7598 Thanks to Felix Gröbert (Google) for discovering and researching this
7604 * By sending carefully crafted DTLS packets an attacker could cause openssl
7605 to leak memory. This can be exploited through a Denial of Service attack.
7606 Thanks to Adam Langley for discovering and researching this issue.
7611 * An attacker can force openssl to consume large amounts of memory whilst
7612 processing DTLS handshake messages. This can be exploited through a
7613 Denial of Service attack.
7614 Thanks to Adam Langley for discovering and researching this issue.
7619 * An attacker can force an error condition which causes openssl to crash
7620 whilst processing DTLS packets due to memory being freed twice. This
7621 can be exploited through a Denial of Service attack.
7622 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7628 * If a multithreaded client connects to a malicious server using a resumed
7629 session and the server sends an ec point format extension it could write
7630 up to 255 bytes to freed memory.
7632 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7638 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7639 X509_name_oneline, X509_name_print_ex et al. to leak some information
7640 from the stack. Applications may be affected if they echo pretty printing
7641 output to the attacker.
7643 Thanks to Ivan Fratric (Google) for discovering this issue.
7646 *Emilia Käsper, and Steve Henson*
7648 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7649 for corner cases. (Certain input points at infinity could lead to
7650 bogus results, with non-infinity inputs mapped to infinity too.)
7654 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7656 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7657 handshake can force the use of weak keying material in OpenSSL
7658 SSL/TLS clients and servers.
7660 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7661 researching this issue. ([CVE-2014-0224])
7663 *KIKUCHI Masashi, Steve Henson*
7665 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7666 OpenSSL DTLS client the code can be made to recurse eventually crashing
7669 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7672 *Imre Rad, Steve Henson*
7674 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7675 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7676 client or server. This is potentially exploitable to run arbitrary
7677 code on a vulnerable client or server.
7679 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7681 *Jüri Aedla, Steve Henson*
7683 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7684 are subject to a denial of service attack.
7686 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7687 this issue. ([CVE-2014-3470])
7689 *Felix Gröbert, Ivan Fratric, Steve Henson*
7691 * Harmonize version and its documentation. -f flag is used to display
7694 *mancha <mancha1@zoho.com>*
7696 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7697 in i2d_ECPrivateKey.
7699 *mancha <mancha1@zoho.com>*
7701 * Fix some double frees. These are not thought to be exploitable.
7703 *mancha <mancha1@zoho.com>*
7705 * Fix for the attack described in the paper "Recovering OpenSSL
7706 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7707 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7708 <http://eprint.iacr.org/2014/140>
7710 Thanks to Yuval Yarom and Naomi Benger for discovering this
7711 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7713 *Yuval Yarom and Naomi Benger*
7715 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7717 * Keep original DTLS digest and encryption contexts in retransmission
7718 structures so we can use the previous session parameters if they need
7719 to be resent. ([CVE-2013-6450])
7723 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7724 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7725 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7726 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7727 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7728 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7730 *Rob Stradling, Adam Langley*
7732 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7734 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7736 This addresses the flaw in CBC record processing discovered by
7737 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7738 at: <http://www.isg.rhul.ac.uk/tls/>
7740 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7741 Security Group at Royal Holloway, University of London
7742 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7743 Emilia Käsper for the initial patch.
7746 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7748 * Return an error when checking OCSP signatures when key is NULL.
7749 This fixes a DoS attack. ([CVE-2013-0166])
7753 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7754 the right response is stapled. Also change SSL_get_certificate()
7755 so it returns the certificate actually sent.
7756 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7757 (This is a backport)
7759 *Rob Stradling <rob.stradling@comodo.com>*
7761 * Fix possible deadlock when decoding public keys.
7765 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7767 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7770 * Sanity check record length before skipping explicit IV in DTLS
7773 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7774 fuzzing as a service testing platform.
7779 * Initialise tkeylen properly when encrypting CMS messages.
7780 Thanks to Solar Designer of Openwall for reporting this issue.
7784 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7786 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7787 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7788 in CRYPTO_realloc_clean.
7790 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7791 issue and to Adam Langley <agl@chromium.org> for fixing it.
7794 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7796 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7798 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7799 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7800 content decryption and always return the same error. Note: this attack
7801 needs on average 2^20 messages so it only affects automated senders. The
7802 old behaviour can be re-enabled in the CMS code by setting the
7803 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7804 an MMA defence is not necessary.
7805 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7806 this issue. ([CVE-2012-0884])
7810 * Fix CVE-2011-4619: make sure we really are receiving a
7811 client hello before rejecting multiple SGC restarts. Thanks to
7812 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7816 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7818 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7819 Thanks to Antonio Martin, Enterprise Secure Access Research and
7820 Development, Cisco Systems, Inc. for discovering this bug and
7821 preparing a fix. ([CVE-2012-0050])
7825 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7827 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7828 of the Vaudenay padding oracle attack on CBC mode encryption
7829 which enables an efficient plaintext recovery attack against
7830 the OpenSSL implementation of DTLS. Their attack exploits timing
7831 differences arising during decryption processing. A research
7832 paper describing this attack can be found at:
7833 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7834 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7835 Security Group at Royal Holloway, University of London
7836 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7837 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7838 for preparing the fix. ([CVE-2011-4108])
7840 *Robin Seggelmann, Michael Tuexen*
7842 * Clear bytes used for block padding of SSL 3.0 records.
7845 *Adam Langley (Google)*
7847 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7848 Kadianakis <desnacked@gmail.com> for discovering this issue and
7849 Adam Langley for preparing the fix. ([CVE-2011-4619])
7851 *Adam Langley (Google)*
7853 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7855 *Andrey Kulikov <amdeich@gmail.com>*
7857 * Prevent malformed RFC3779 data triggering an assertion failure.
7858 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7859 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7861 *Rob Austein <sra@hactrn.net>*
7863 * Improved PRNG seeding for VOS.
7865 *Paul Green <Paul.Green@stratus.com>*
7867 * Fix ssl_ciph.c set-up race.
7869 *Adam Langley (Google)*
7871 * Fix spurious failures in ecdsatest.c.
7873 *Emilia Käsper (Google)*
7875 * Fix the BIO_f_buffer() implementation (which was mixing different
7876 interpretations of the `..._len` fields).
7878 *Adam Langley (Google)*
7880 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7881 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7882 threads won't reuse the same blinding coefficients.
7884 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7885 lock to call BN_BLINDING_invert_ex, and avoids one use of
7886 BN_BLINDING_update for each BN_BLINDING structure (previously,
7887 the last update always remained unused).
7889 *Emilia Käsper (Google)*
7891 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7893 *Bob Buckholz (Google)*
7895 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7897 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7898 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7900 *Kaspar Brand <ossl@velox.ch>*
7902 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7903 for multi-threaded use of ECDH. ([CVE-2011-3210])
7905 *Adam Langley (Google)*
7907 * Fix x509_name_ex_d2i memory leak on bad inputs.
7911 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7912 signature public key algorithm by using OID xref utilities instead.
7913 Before this you could only use some ECC ciphersuites with SHA1 only.
7917 * Add protection against ECDSA timing attacks as mentioned in the paper
7918 by Billy Bob Brumley and Nicola Tuveri, see:
7919 <http://eprint.iacr.org/2011/232.pdf>
7921 *Billy Bob Brumley and Nicola Tuveri*
7923 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7925 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7927 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7929 * Fix bug in string printing code: if *any* escaping is enabled we must
7930 escape the escape character (backslash) or the resulting string is
7935 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7937 * Disable code workaround for ancient and obsolete Netscape browsers
7938 and servers: an attacker can use it in a ciphersuite downgrade attack.
7939 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7943 * Fixed J-PAKE implementation error, originally discovered by
7944 Sebastien Martini, further info and confirmation from Stefan
7945 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7949 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7951 * Fix extension code to avoid race conditions which can result in a buffer
7952 overrun vulnerability: resumed sessions must not be modified as they can
7953 be shared by multiple threads. CVE-2010-3864
7957 * Fix WIN32 build system to correctly link an ENGINE directory into
7962 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7964 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7967 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7969 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7971 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7972 context. The operation can be customised via the ctrl mechanism in
7973 case ENGINEs want to include additional functionality.
7977 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7981 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7982 output hashes compatible with older versions of OpenSSL.
7984 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7986 * Fix compression algorithm handling: if resuming a session use the
7987 compression algorithm of the resumed session instead of determining
7988 it from client hello again. Don't allow server to change algorithm.
7992 * Add load_crls() function to commands tidying load_certs() too. Add option
7993 to verify utility to allow additional CRLs to be included.
7997 * Update OCSP request code to permit adding custom headers to the request:
7998 some responders need this.
8002 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8005 *Julia Lawall <julia@diku.dk>*
8007 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8008 needlessly dereferenced structures, used obsolete functions and
8009 didn't handle all updated verify codes correctly.
8013 * Disable MD2 in the default configuration.
8017 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8018 indicate the initial BIO being pushed or popped. This makes it possible
8019 to determine whether the BIO is the one explicitly called or as a result
8020 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8021 it handles reference counts correctly and doesn't zero out the I/O bio
8022 when it is not being explicitly popped. WARNING: applications which
8023 included workarounds for the old buggy behaviour will need to be modified
8024 or they could free up already freed BIOs.
8028 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8029 renaming to all platforms (within the 0.9.8 branch, this was
8030 done conditionally on Netware platforms to avoid a name clash).
8032 *Guenter <lists@gknw.net>*
8034 * Add ECDHE and PSK support to DTLS.
8036 *Michael Tuexen <tuexen@fh-muenster.de>*
8038 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8043 * Add "missing" function EVP_MD_flags() (without this the only way to
8044 retrieve a digest flags is by accessing the structure directly. Update
8045 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8046 or cipher is registered as in the "from" argument. Print out all
8047 registered digests in the dgst usage message instead of manually
8048 attempting to work them out.
8052 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8053 this allows the use of compression and extensions. Change default cipher
8054 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8055 by default unless an application cipher string requests it.
8059 * Alter match criteria in PKCS12_parse(). It used to try to use local
8060 key ids to find matching certificates and keys but some PKCS#12 files
8061 don't follow the (somewhat unwritten) rules and this strategy fails.
8062 Now just gather all certificates together and the first private key
8063 then look for the first certificate that matches the key.
8067 * Support use of registered digest and cipher names for dgst and cipher
8068 commands instead of having to add each one as a special case. So now
8075 openssl dgst -sha256 foo
8077 and this works for ENGINE based algorithms too.
8081 * Update Gost ENGINE to support parameter files.
8083 *Victor B. Wagner <vitus@cryptocom.ru>*
8085 * Support GeneralizedTime in ca utility.
8087 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8089 * Enhance the hash format used for certificate directory links. The new
8090 form uses the canonical encoding (meaning equivalent names will work
8091 even if they aren't identical) and uses SHA1 instead of MD5. This form
8092 is incompatible with the older format and as a result c_rehash should
8093 be used to rebuild symbolic links.
8097 * Make PKCS#8 the default write format for private keys, replacing the
8098 traditional format. This form is standardised, more secure and doesn't
8099 include an implicit MD5 dependency.
8103 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8104 committed to OpenSSL should pass this lot as a minimum.
8108 * Add session ticket override functionality for use by EAP-FAST.
8110 *Jouni Malinen <j@w1.fi>*
8112 * Modify HMAC functions to return a value. Since these can be implemented
8113 in an ENGINE errors can occur.
8117 * Type-checked OBJ_bsearch_ex.
8121 * Type-checked OBJ_bsearch. Also some constification necessitated
8122 by type-checking. Still to come: TXT_DB, bsearch(?),
8123 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8128 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8129 seconds to a tm structure directly, instead of going through OS
8130 specific date routines. This avoids any issues with OS routines such
8131 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8132 and X509_time_adj_ex() to cover the extended range. The existing
8133 X509_time_adj() is still usable and will no longer have any date issues.
8137 * Delta CRL support. New use deltas option which will attempt to locate
8138 and search any appropriate delta CRLs available.
8140 This work was sponsored by Google.
8144 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8145 code and add additional score elements. Validate alternate CRL paths
8146 as part of the CRL checking and indicate a new error "CRL path validation
8147 error" in this case. Applications wanting additional details can use
8148 the verify callback and check the new "parent" field. If this is not
8149 NULL CRL path validation is taking place. Existing applications won't
8150 see this because it requires extended CRL support which is off by
8153 This work was sponsored by Google.
8157 * Support for freshest CRL extension.
8159 This work was sponsored by Google.
8163 * Initial indirect CRL support. Currently only supported in the CRLs
8164 passed directly and not via lookup. Process certificate issuer
8165 CRL entry extension and lookup CRL entries by bother issuer name
8166 and serial number. Check and process CRL issuer entry in IDP extension.
8168 This work was sponsored by Google.
8172 * Add support for distinct certificate and CRL paths. The CRL issuer
8173 certificate is validated separately in this case. Only enabled if
8174 an extended CRL support flag is set: this flag will enable additional
8175 CRL functionality in future.
8177 This work was sponsored by Google.
8181 * Add support for policy mappings extension.
8183 This work was sponsored by Google.
8187 * Fixes to pathlength constraint, self issued certificate handling,
8188 policy processing to align with RFC3280 and PKITS tests.
8190 This work was sponsored by Google.
8194 * Support for name constraints certificate extension. DN, email, DNS
8195 and URI types are currently supported.
8197 This work was sponsored by Google.
8201 * To cater for systems that provide a pointer-based thread ID rather
8202 than numeric, deprecate the current numeric thread ID mechanism and
8203 replace it with a structure and associated callback type. This
8204 mechanism allows a numeric "hash" to be extracted from a thread ID in
8205 either case, and on platforms where pointers are larger than 'long',
8206 mixing is done to help ensure the numeric 'hash' is usable even if it
8207 can't be guaranteed unique. The default mechanism is to use "&errno"
8208 as a pointer-based thread ID to distinguish between threads.
8210 Applications that want to provide their own thread IDs should now use
8211 CRYPTO_THREADID_set_callback() to register a callback that will call
8212 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8214 Note that ERR_remove_state() is now deprecated, because it is tied
8215 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8216 to free the current thread's error state should be replaced by
8217 ERR_remove_thread_state(NULL).
8219 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8220 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8221 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8222 application was previously providing a numeric thread callback that
8223 was inappropriate for distinguishing threads, then uniqueness might
8224 have been obtained with &errno that happened immediately in the
8225 intermediate development versions of OpenSSL; this is no longer the
8226 case, the numeric thread callback will now override the automatic use
8229 *Geoff Thorpe, with help from Bodo Moeller*
8231 * Initial support for different CRL issuing certificates. This covers a
8232 simple case where the self issued certificates in the chain exist and
8233 the real CRL issuer is higher in the existing chain.
8235 This work was sponsored by Google.
8239 * Removed effectively defunct crypto/store from the build.
8243 * Revamp of STACK to provide stronger type-checking. Still to come:
8244 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8245 ASN1_STRING, CONF_VALUE.
8249 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8250 RAM on SSL connections. This option can save about 34k per idle SSL.
8254 * Revamp of LHASH to provide stronger type-checking. Still to come:
8255 STACK, TXT_DB, bsearch, qsort.
8259 * Initial support for Cryptographic Message Syntax (aka CMS) based
8260 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8261 support for data, signedData, compressedData, digestedData and
8262 encryptedData, envelopedData types included. Scripts to check against
8263 RFC4134 examples draft and interop and consistency checks of many
8264 content types and variants.
8268 * Add options to enc utility to support use of zlib compression BIO.
8272 * Extend mk1mf to support importing of options and assembly language
8273 files from Configure script, currently only included in VC-WIN32.
8274 The assembly language rules can now optionally generate the source
8275 files from the associated perl scripts.
8279 * Implement remaining functionality needed to support GOST ciphersuites.
8280 Interop testing has been performed using CryptoPro implementations.
8282 *Victor B. Wagner <vitus@cryptocom.ru>*
8284 * s390x assembler pack.
8288 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8293 * Implement Opaque PRF Input TLS extension as specified in
8294 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8295 official specification yet and no extension type assignment by
8296 IANA exists, this extension (for now) will have to be explicitly
8297 enabled when building OpenSSL by providing the extension number
8298 to use. For example, specify an option
8300 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8302 to the "config" or "Configure" script to enable the extension,
8303 assuming extension number 0x9527 (which is a completely arbitrary
8304 and unofficial assignment based on the MD5 hash of the Internet
8305 Draft). Note that by doing so, you potentially lose
8306 interoperability with other TLS implementations since these might
8307 be using the same extension number for other purposes.
8309 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8310 opaque PRF input value to use in the handshake. This will create
8311 an internal copy of the length-'len' string at 'src', and will
8312 return non-zero for success.
8314 To get more control and flexibility, provide a callback function
8317 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8318 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8322 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8325 Callback function 'cb' will be called in handshakes, and is
8326 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8327 Argument 'arg' is for application purposes (the value as given to
8328 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8329 be provided to the callback function). The callback function
8330 has to return non-zero to report success: usually 1 to use opaque
8331 PRF input just if possible, or 2 to enforce use of the opaque PRF
8332 input. In the latter case, the library will abort the handshake
8333 if opaque PRF input is not successfully negotiated.
8335 Arguments 'peerinput' and 'len' given to the callback function
8336 will always be NULL and 0 in the case of a client. A server will
8337 see the client's opaque PRF input through these variables if
8338 available (NULL and 0 otherwise). Note that if the server
8339 provides an opaque PRF input, the length must be the same as the
8340 length of the client's opaque PRF input.
8342 Note that the callback function will only be called when creating
8343 a new session (session resumption can resume whatever was
8344 previously negotiated), and will not be called in SSL 2.0
8345 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8346 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8347 for applications that need to enforce opaque PRF input.
8351 * Update ssl code to support digests other than SHA1+MD5 for handshake
8354 *Victor B. Wagner <vitus@cryptocom.ru>*
8356 * Add RFC4507 support to OpenSSL. This includes the corrections in
8357 RFC4507bis. The encrypted ticket format is an encrypted encoded
8358 SSL_SESSION structure, that way new session features are automatically
8361 If a client application caches session in an SSL_SESSION structure
8362 support is transparent because tickets are now stored in the encoded
8365 The SSL_CTX structure automatically generates keys for ticket
8366 protection in servers so again support should be possible
8367 with no application modification.
8369 If a client or server wishes to disable RFC4507 support then the option
8370 SSL_OP_NO_TICKET can be set.
8372 Add a TLS extension debugging callback to allow the contents of any client
8373 or server extensions to be examined.
8375 This work was sponsored by Google.
8379 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8380 OpenSSL should now compile cleanly on gcc 4.2
8382 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8384 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8385 support including streaming MAC support: this is required for GOST
8386 ciphersuite support.
8388 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8390 * Add option -stream to use PKCS#7 streaming in smime utility. New
8391 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8392 to output in BER and PEM format.
8396 * Experimental support for use of HMAC via EVP_PKEY interface. This
8397 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8398 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8399 ENGINE support for HMAC keys which are unextractable. New -mac and
8400 -macopt options to dgst utility.
8404 * New option -sigopt to dgst utility. Update dgst to use
8405 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8406 alternative signing parameters such as X9.31 or PSS in the dgst
8411 * Change ssl_cipher_apply_rule(), the internal function that does
8412 the work each time a ciphersuite string requests enabling
8413 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8414 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8415 the order of disabled ciphersuites such that those ciphersuites
8416 that most recently went from enabled to disabled not only stay
8417 in order with respect to each other, but also have higher priority
8418 than other disabled ciphersuites the next time ciphersuites are
8421 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8422 the same ciphersuites as with "HIGH" alone, but in a specific
8423 order where the PSK ciphersuites come first (since they are the
8424 most recently disabled ciphersuites when "HIGH" is parsed).
8426 Also, change ssl_create_cipher_list() (using this new
8427 functionality) such that between otherwise identical
8428 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8433 * Change ssl_create_cipher_list() so that it automatically
8434 arranges the ciphersuites in reasonable order before starting
8435 to process the rule string. Thus, the definition for "DEFAULT"
8436 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8437 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8438 This makes it much easier to arrive at a reasonable default order
8439 in applications for which anonymous ciphers are OK (meaning
8440 that you can't actually use DEFAULT).
8442 *Bodo Moeller; suggested by Victor Duchovni*
8444 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8445 processing) into multiple integers instead of setting
8446 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8447 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8448 (These masks as well as the individual bit definitions are hidden
8449 away into the non-exported interface ssl/ssl_locl.h, so this
8450 change to the definition of the SSL_CIPHER structure shouldn't
8451 affect applications.) This give us more bits for each of these
8452 categories, so there is no longer a need to coagulate AES128 and
8453 AES256 into a single algorithm bit, and to coagulate Camellia128
8454 and Camellia256 into a single algorithm bit, which has led to all
8457 Thus, among other things, the kludge introduced in 0.9.7m and
8458 0.9.8e for masking out AES256 independently of AES128 or masking
8459 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8461 With the change, we also introduce new ciphersuite aliases that
8462 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8467 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8468 Use the leftmost N bytes of the signature input if the input is
8469 larger than the prime q (with N being the size in bytes of q).
8473 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8474 it yet and it is largely untested.
8478 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8482 * Initial incomplete changes to avoid need for function casts in OpenSSL
8483 some compilers (gcc 4.2 and later) reject their use. Safestack is
8484 reimplemented. Update ASN1 to avoid use of legacy functions.
8488 * Win32/64 targets are linked with Winsock2.
8492 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8493 to external functions. This can be used to increase CRL handling
8494 efficiency especially when CRLs are very large by (for example) storing
8495 the CRL revoked certificates in a database.
8499 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8500 new CRLs added to a directory can be used. New command line option
8501 -verify_return_error to s_client and s_server. This causes real errors
8502 to be returned by the verify callback instead of carrying on no matter
8503 what. This reflects the way a "real world" verify callback would behave.
8507 * GOST engine, supporting several GOST algorithms and public key formats.
8508 Kindly donated by Cryptocom.
8512 * Partial support for Issuing Distribution Point CRL extension. CRLs
8513 partitioned by DP are handled but no indirect CRL or reason partitioning
8514 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8515 selected via a scoring technique which handles IDP and AKID in CRLs.
8519 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8520 will ultimately be used for all verify operations: this will remove the
8521 X509_STORE dependency on certificate verification and allow alternative
8522 lookup methods. X509_STORE based implementations of these two callbacks.
8526 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8527 Modify get_crl() to find a valid (unexpired) CRL if possible.
8531 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8532 this would be called X509_CRL_cmp() but that name is already used by
8533 a function that just compares CRL issuer names. Cache several CRL
8534 extensions in X509_CRL structure and cache CRLDP in X509.
8538 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8539 this maps equivalent X509_NAME structures into a consistent structure.
8540 Name comparison can then be performed rapidly using memcmp().
8544 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8549 * Allow digests to supply their own micalg string for S/MIME type using
8550 the ctrl EVP_MD_CTRL_MICALG.
8554 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8555 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8556 ctrl. It can then customise the structure before and/or after signing
8561 * New function OBJ_add_sigid() to allow application defined signature OIDs
8562 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8563 to free up any added signature OIDs.
8567 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8568 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8569 digest and cipher tables. New options added to openssl utility:
8570 list-message-digest-algorithms and list-cipher-algorithms.
8574 * Change the array representation of binary polynomials: the list
8575 of degrees of non-zero coefficients is now terminated with -1.
8576 Previously it was terminated with 0, which was also part of the
8577 value; thus, the array representation was not applicable to
8578 polynomials where t^0 has coefficient zero. This change makes
8579 the array representation useful in a more general context.
8583 * Various modifications and fixes to SSL/TLS cipher string
8584 handling. For ECC, the code now distinguishes between fixed ECDH
8585 with RSA certificates on the one hand and with ECDSA certificates
8586 on the other hand, since these are separate ciphersuites. The
8587 unused code for Fortezza ciphersuites has been removed.
8589 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8590 (not "ECDHE"). For consistency with the code for DH
8591 certificates, use of ECDH certificates is now considered ECDH
8592 authentication, not RSA or ECDSA authentication (the latter is
8593 merely the CA's signing algorithm and not actively used in the
8596 The temporary ciphersuite alias "ECCdraft" is no longer
8597 available, and ECC ciphersuites are no longer excluded from "ALL"
8598 and "DEFAULT". The following aliases now exist for RFC 4492
8599 ciphersuites, most of these by analogy with the DH case:
8601 kECDHr - ECDH cert, signed with RSA
8602 kECDHe - ECDH cert, signed with ECDSA
8603 kECDH - ECDH cert (signed with either RSA or ECDSA)
8604 kEECDH - ephemeral ECDH
8605 ECDH - ECDH cert or ephemeral ECDH
8611 AECDH - anonymous ECDH
8612 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8616 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8617 Use correct micalg parameters depending on digest(s) in signed message.
8621 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8622 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8626 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8627 an engine to register a method. Add ENGINE lookups for methods and
8628 functional reference processing.
8632 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8633 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8638 * New -resign option to smime utility. This adds one or more signers
8639 to an existing PKCS#7 signedData structure. Also -md option to use an
8640 alternative message digest algorithm for signing.
8644 * Tidy up PKCS#7 routines and add new functions to make it easier to
8645 create PKCS7 structures containing multiple signers. Update smime
8646 application to support multiple signers.
8650 * New -macalg option to pkcs12 utility to allow setting of an alternative
8655 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8656 Reorganize PBE internals to lookup from a static table using NIDs,
8657 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8658 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8659 PRF which will be automatically used with PBES2.
8663 * Replace the algorithm specific calls to generate keys in "req" with the
8668 * Update PKCS#7 enveloped data routines to use new API. This is now
8669 supported by any public key method supporting the encrypt operation. A
8670 ctrl is added to allow the public key algorithm to examine or modify
8671 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8676 * Add a ctrl to asn1 method to allow a public key algorithm to express
8677 a default digest type to use. In most cases this will be SHA1 but some
8678 algorithms (such as GOST) need to specify an alternative digest. The
8679 return value indicates how strong the preference is 1 means optional and
8680 2 is mandatory (that is it is the only supported type). Modify
8681 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8682 use the default md. Update openssl utilities to use the default digest
8683 type for signing if it is not explicitly indicated.
8687 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8688 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8689 signing method from the key type. This effectively removes the link
8690 between digests and public key types.
8694 * Add an OID cross reference table and utility functions. Its purpose is to
8695 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8696 rsaEncryption. This will allow some of the algorithm specific hackery
8697 needed to use the correct OID to be removed.
8701 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8702 structures for PKCS7_sign(). They are now set up by the relevant public
8707 * Add provisional EC pkey method with support for ECDSA and ECDH.
8711 * Add support for key derivation (agreement) in the API, DH method and
8716 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8717 public and private key formats. As a side effect these add additional
8718 command line functionality not previously available: DSA signatures can be
8719 generated and verified using pkeyutl and DH key support and generation in
8726 *Oliver Tappe <zooey@hirschkaefer.de>*
8728 * New make target "install_html_docs" installs HTML renditions of the
8731 *Oliver Tappe <zooey@hirschkaefer.de>*
8733 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8734 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8735 support key and parameter generation and add initial key generation
8736 functionality for RSA.
8740 * Add functions for main EVP_PKEY_method operations. The undocumented
8741 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8742 `EVP_PKEY_{encrypt,decrypt}_old`.
8746 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8747 key API, doesn't do much yet.
8751 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8752 public key algorithms. New option to openssl utility:
8753 "list-public-key-algorithms" to print out info.
8757 * Implement the Supported Elliptic Curves Extension for
8758 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8762 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8763 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8767 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8768 utilities such as rsa, dsa, dsaparam etc except they process any key
8773 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8774 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8775 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8780 * Initial support for pluggable public key ASN1.
8781 De-spaghettify the public key ASN1 handling. Move public and private
8782 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8783 algorithm specific handling to a single module within the relevant
8784 algorithm directory. Add functions to allow (near) opaque processing
8785 of public and private key structures.
8789 * Implement the Supported Point Formats Extension for
8790 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8794 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8795 for the psk identity [hint] and the psk callback functions to the
8796 SSL_SESSION, SSL and SSL_CTX structure.
8799 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8803 SSL_CTX_use_psk_identity_hint
8804 SSL_get_psk_identity_hint
8805 SSL_get_psk_identity
8806 SSL_use_psk_identity_hint
8808 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8810 * Add RFC 3161 compliant time stamp request creation, response generation
8811 and response verification functionality.
8813 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8815 * Add initial support for TLS extensions, specifically for the server_name
8816 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8817 have new members for a host name. The SSL data structure has an
8818 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8819 stored in that context to allow for session resumption, even after the
8820 SSL has been switched to a new SSL_CTX in reaction to a client's
8821 server_name extension.
8823 New functions (subject to change):
8825 SSL_get_servername()
8826 SSL_get_servername_type()
8829 New CTRL codes and macros (subject to change):
8831 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8832 - SSL_CTX_set_tlsext_servername_callback()
8833 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8834 - SSL_CTX_set_tlsext_servername_arg()
8835 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8837 openssl s_client has a new '-servername ...' option.
8839 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8840 '-key2 ...', '-servername_fatal' (subject to change). This allows
8841 testing the HostName extension for a specific single host name ('-cert'
8842 and '-key' remain fallbacks for handshakes without HostName
8843 negotiation). If the unrecognized_name alert has to be sent, this by
8844 default is a warning; it becomes fatal with the '-servername_fatal'
8847 *Peter Sylvester, Remy Allais, Christophe Renou*
8849 * Whirlpool hash implementation is added.
8853 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8854 bn(64,32). Because of instruction set limitations it doesn't have
8855 any negative impact on performance. This was done mostly in order
8856 to make it possible to share assembler modules, such as bn_mul_mont
8857 implementations, between 32- and 64-bit builds without hassle.
8861 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8862 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8867 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8868 dedicated Montgomery multiplication procedure, is introduced.
8869 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8870 "64-bit" performance on certain 32-bit targets.
8874 * New option SSL_OP_NO_COMP to disable use of compression selectively
8875 in SSL structures. New SSL ctrl to set maximum send fragment size.
8876 Save memory by setting the I/O buffer sizes dynamically instead of
8877 using the maximum available value.
8881 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8882 in addition to the text details.
8886 * Very, very preliminary EXPERIMENTAL support for printing of general
8887 ASN1 structures. This currently produces rather ugly output and doesn't
8888 handle several customised structures at all.
8892 * Integrated support for PVK file format and some related formats such
8893 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8894 these in the 'rsa' and 'dsa' utilities.
8898 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8902 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8903 place for the (very old) "NETSCAPE" format certificates which are now
8904 handled using new ASN1 code equivalents.
8908 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8909 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8910 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8914 * Modify CRL distribution points extension code to print out previously
8915 unsupported fields. Enhance extension setting code to allow setting of
8920 * Add print and set support for Issuing Distribution Point CRL extension.
8924 * Change 'Configure' script to enable Camellia by default.
8931 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8933 * When rejecting SSL/TLS records due to an incorrect version number, never
8934 update s->server with a new major version number. As of
8935 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8936 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8937 the previous behavior could result in a read attempt at NULL when
8938 receiving specific incorrect SSL/TLS records once record payload
8939 protection is active. ([CVE-2010-0740])
8941 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8943 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8944 could be crashed if the relevant tables were not present (e.g. chrooted).
8946 *Tomas Hoger <thoger@redhat.com>*
8948 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8950 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8952 *Martin Olsson, Neel Mehta*
8954 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8955 accommodate for stack sorting, always a write lock!).
8959 * On some versions of WIN32 Heap32Next is very slow. This can cause
8960 excessive delays in the RAND_poll(): over a minute. As a workaround
8961 include a time check in the inner Heap32Next loop too.
8965 * The code that handled flushing of data in SSL/TLS originally used the
8966 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8967 the problem outlined in PR#1949. The fix suggested there however can
8968 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8969 of Apache). So instead simplify the code to flush unconditionally.
8970 This should be fine since flushing with no data to flush is a no op.
8974 * Handle TLS versions 2.0 and later properly and correctly use the
8975 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8976 off ancient servers have a habit of sticking around for a while...
8980 * Modify compression code so it frees up structures without using the
8981 ex_data callbacks. This works around a problem where some applications
8982 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8983 restarting) then use compression (e.g. SSL with compression) later.
8984 This results in significant per-connection memory leaks and
8985 has caused some security issues including CVE-2008-1678 and
8990 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8991 change when encrypting or decrypting.
8995 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8996 connect and renegotiate with servers which do not support RI.
8997 Until RI is more widely deployed this option is enabled by default.
9001 * Add "missing" ssl ctrls to clear options and mode.
9005 * If client attempts to renegotiate and doesn't support RI respond with
9006 a no_renegotiation alert as required by RFC5746. Some renegotiating
9007 TLS clients will continue a connection gracefully when they receive
9008 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9009 waiting for a server hello which it will never receive. Now we treat a
9010 received no_renegotiation alert as a fatal error. This is because
9011 applications requesting a renegotiation might well expect it to succeed
9012 and would have no code in place to handle the server denying it so the
9013 only safe thing to do is to terminate the connection.
9017 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9018 peer supports secure renegotiation and 0 otherwise. Print out peer
9019 renegotiation support in s_client/s_server.
9023 * Replace the highly broken and deprecated SPKAC certification method with
9024 the updated NID creation version. This should correctly handle UTF8.
9028 * Implement RFC5746. Re-enable renegotiation but require the extension
9029 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9030 turns out to be a bad idea. It has been replaced by
9031 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9032 SSL_CTX_set_options(). This is really not recommended unless you
9033 know what you are doing.
9035 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9037 * Fixes to stateless session resumption handling. Use initial_ctx when
9038 issuing and attempting to decrypt tickets in case it has changed during
9039 servername handling. Use a non-zero length session ID when attempting
9040 stateless session resumption: this makes it possible to determine if
9041 a resumption has occurred immediately after receiving server hello
9042 (several places in OpenSSL subtly assume this) instead of later in
9047 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9048 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9049 fixes for a few places where the return code is not checked
9052 *Julia Lawall <julia@diku.dk>*
9054 * Add --strict-warnings option to Configure script to include devteam
9055 warnings in other configurations.
9059 * Add support for --libdir option and LIBDIR variable in makefiles. This
9060 makes it possible to install openssl libraries in locations which
9061 have names other than "lib", for example "/usr/lib64" which some
9064 *Steve Henson, based on patch from Jeremy Utley*
9066 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9067 X690 8.9.12 and can produce some misleading textual output of OIDs.
9069 *Steve Henson, reported by Dan Kaminsky*
9071 * Delete MD2 from algorithm tables. This follows the recommendation in
9072 several standards that it is not used in new applications due to
9073 several cryptographic weaknesses. For binary compatibility reasons
9074 the MD2 API is still compiled in by default.
9078 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9083 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9084 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9087 *Guenter <lists@gknw.net>*
9089 * Fix the server certificate chain building code to use X509_verify_cert(),
9090 it used to have an ad-hoc builder which was unable to cope with anything
9091 other than a simple chain.
9093 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9095 * Don't check self signed certificate signatures in X509_verify_cert()
9096 by default (a flag can override this): it just wastes time without
9097 adding any security. As a useful side effect self signed root CAs
9098 with non-FIPS digests are now usable in FIPS mode.
9102 * In dtls1_process_out_of_seq_message() the check if the current message
9103 is already buffered was missing. For every new message was memory
9104 allocated, allowing an attacker to perform an denial of service attack
9105 with sending out of seq handshake messages until there is no memory
9106 left. Additionally every future message was buffered, even if the
9107 sequence number made no sense and would be part of another handshake.
9108 So only messages with sequence numbers less than 10 in advance will be
9109 buffered. ([CVE-2009-1378])
9111 *Robin Seggelmann, discovered by Daniel Mentz*
9113 * Records are buffered if they arrive with a future epoch to be
9114 processed after finishing the corresponding handshake. There is
9115 currently no limitation to this buffer allowing an attacker to perform
9116 a DOS attack with sending records with future epochs until there is no
9117 memory left. This patch adds the pqueue_size() function to determine
9118 the size of a buffer and limits the record buffer to 100 entries.
9121 *Robin Seggelmann, discovered by Daniel Mentz*
9123 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9124 parent structure is freed. ([CVE-2009-1379])
9128 * Handle non-blocking I/O properly in SSL_shutdown() call.
9130 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9132 * Add `2.5.4.*` OIDs
9134 *Ilya O. <vrghost@gmail.com>*
9136 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9138 * Disable renegotiation completely - this fixes a severe security
9139 problem ([CVE-2009-3555]) at the cost of breaking all
9140 renegotiation. Renegotiation can be re-enabled by setting
9141 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9142 run-time. This is really not recommended unless you know what
9147 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9149 * Don't set val to NULL when freeing up structures, it is freed up by
9150 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9151 zeroing past the valid field. ([CVE-2009-0789])
9153 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9155 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9156 checked correctly. This would allow some invalid signed attributes to
9157 appear to verify correctly. ([CVE-2009-0591])
9159 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9161 * Reject UniversalString and BMPString types with invalid lengths. This
9162 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9163 a legal length. ([CVE-2009-0590])
9167 * Set S/MIME signing as the default purpose rather than setting it
9168 unconditionally. This allows applications to override it at the store
9173 * Permit restricted recursion of ASN1 strings. This is needed in practice
9174 to handle some structures.
9178 * Improve efficiency of mem_gets: don't search whole buffer each time
9181 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9183 * New -hex option for openssl rand.
9187 * Print out UTF8String and NumericString when parsing ASN1.
9191 * Support NumericString type for name components.
9195 * Allow CC in the environment to override the automatically chosen
9196 compiler. Note that nothing is done to ensure flags work with the
9201 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9203 * Properly check EVP_VerifyFinal() and similar return values
9206 *Ben Laurie, Bodo Moeller, Google Security Team*
9208 * Enable TLS extensions by default.
9212 * Allow the CHIL engine to be loaded, whether the application is
9213 multithreaded or not. (This does not release the developer from the
9214 obligation to set up the dynamic locking callbacks.)
9216 *Sander Temme <sander@temme.net>*
9218 * Use correct exit code if there is an error in dgst command.
9220 *Steve Henson; problem pointed out by Roland Dirlewanger*
9222 * Tweak Configure so that you need to say "experimental-jpake" to enable
9223 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9227 * Add experimental JPAKE support, including demo authentication in
9228 s_client and s_server.
9232 * Set the comparison function in v3_addr_canonize().
9234 *Rob Austein <sra@hactrn.net>*
9236 * Add support for XMPP STARTTLS in s_client.
9238 *Philip Paeps <philip@freebsd.org>*
9240 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9241 to ensure that even with this option, only ciphersuites in the
9242 server's preference list will be accepted. (Note that the option
9243 applies only when resuming a session, so the earlier behavior was
9244 just about the algorithm choice for symmetric cryptography.)
9248 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9250 * Fix NULL pointer dereference if a DTLS server received
9251 ChangeCipherSpec as first record ([CVE-2009-1386]).
9255 * Fix a state transition in s3_srvr.c and d1_srvr.c
9256 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9260 * The fix in 0.9.8c that supposedly got rid of unsafe
9261 double-checked locking was incomplete for RSA blinding,
9262 addressing just one layer of what turns out to have been
9263 doubly unsafe triple-checked locking.
9265 So now fix this for real by retiring the MONT_HELPER macro
9266 in crypto/rsa/rsa_eay.c.
9268 *Bodo Moeller; problem pointed out by Marius Schilder*
9270 * Various precautionary measures:
9272 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9274 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9275 (NB: This would require knowledge of the secret session ticket key
9276 to exploit, in which case you'd be SOL either way.)
9278 - Change bn_nist.c so that it will properly handle input BIGNUMs
9279 outside the expected range.
9281 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9284 *Neel Mehta, Bodo Moeller*
9286 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9287 the load fails. Useful for distros.
9289 *Ben Laurie and the FreeBSD team*
9291 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9295 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9299 * Expand ENGINE to support engine supplied SSL client certificate functions.
9301 This work was sponsored by Logica.
9305 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9306 keystores. Support for SSL/TLS client authentication too.
9307 Not compiled unless enable-capieng specified to Configure.
9309 This work was sponsored by Logica.
9313 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9314 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9315 attribute creation routines such as certificate requests and PKCS#12
9320 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9322 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9323 handshake which could lead to a client crash as found using the
9324 Codenomicon TLS test suite ([CVE-2008-1672])
9326 *Steve Henson, Mark Cox*
9328 * Fix double free in TLS server name extensions which could lead to
9329 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9333 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9335 Clear the error queue to ensure that error entries left from
9336 older function calls do not interfere with the correct operation.
9338 *Lutz Jaenicke, Erik de Castro Lopo*
9340 * Remove root CA certificates of commercial CAs:
9342 The OpenSSL project does not recommend any specific CA and does not
9343 have any policy with respect to including or excluding any CA.
9344 Therefore it does not make any sense to ship an arbitrary selection
9345 of root CA certificates with the OpenSSL software.
9349 * RSA OAEP patches to fix two separate invalid memory reads.
9350 The first one involves inputs when 'lzero' is greater than
9351 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9352 before the beginning of from). The second one involves inputs where
9353 the 'db' section contains nothing but zeroes (there is a one-byte
9354 invalid read after the end of 'db').
9356 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9358 * Partial backport from 0.9.9-dev:
9360 Introduce bn_mul_mont (dedicated Montgomery multiplication
9361 procedure) as a candidate for BIGNUM assembler implementation.
9362 While 0.9.9-dev uses assembler for various architectures, only
9363 x86_64 is available by default here in the 0.9.8 branch, and
9364 32-bit x86 is available through a compile-time setting.
9366 To try the 32-bit x86 assembler implementation, use Configure
9367 option "enable-montasm" (which exists only for this backport).
9369 As "enable-montasm" for 32-bit x86 disclaims code stability
9370 anyway, in this constellation we activate additional code
9371 backported from 0.9.9-dev for further performance improvements,
9372 namely BN_from_montgomery_word. (To enable this otherwise,
9373 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9375 *Andy Polyakov (backport partially by Bodo Moeller)*
9377 * Add TLS session ticket callback. This allows an application to set
9378 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9379 values. This is useful for key rollover for example where several key
9380 sets may exist with different names.
9384 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9385 This was broken until now in 0.9.8 releases, such that the only way
9386 a registered ENGINE could be used (assuming it initialises
9387 successfully on the host) was to explicitly set it as the default
9388 for the relevant algorithms. This is in contradiction with 0.9.7
9389 behaviour and the documentation. With this fix, when an ENGINE is
9390 registered into a given algorithm's table of implementations, the
9391 'uptodate' flag is reset so that auto-discovery will be used next
9392 time a new context for that algorithm attempts to select an
9395 *Ian Lister (tweaked by Geoff Thorpe)*
9397 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9398 implementation in the following ways:
9400 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9403 Lack of BER streaming support means one pass streaming processing is
9404 only supported if data is detached: setting the streaming flag is
9405 ignored for embedded content.
9407 CMS support is disabled by default and must be explicitly enabled
9408 with the enable-cms configuration option.
9412 * Update the GMP engine glue to do direct copies between BIGNUM and
9413 mpz_t when openssl and GMP use the same limb size. Otherwise the
9414 existing "conversion via a text string export" trick is still used.
9416 *Paul Sheer <paulsheer@gmail.com>*
9418 * Zlib compression BIO. This is a filter BIO which compressed and
9419 uncompresses any data passed through it.
9423 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9424 RFC3394 compatible AES key wrapping.
9428 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9429 sets string data without copying. X509_ALGOR_set0() and
9430 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9431 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9432 from an X509_ATTRIBUTE structure optionally checking it occurs only
9433 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9438 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9439 to get the expected BN_FLG_CONSTTIME behavior.
9441 *Bodo Moeller (Google)*
9445 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9446 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9447 - added some more tests to do_tests.pl
9448 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9449 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9450 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9451 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9452 - various changes to netware.pl to enable gcc-cross builds on Win32
9454 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9455 - various changes to fix missing prototype warnings
9456 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9457 - added AES, WHIRLPOOL and CPUID assembler code to build files
9458 - added missing AES assembler make rules to mk1mf.pl
9459 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9461 *Guenter Knauf <eflash@gmx.net>*
9463 * Implement certificate status request TLS extension defined in RFC3546.
9464 A client can set the appropriate parameters and receive the encoded
9465 OCSP response via a callback. A server can query the supplied parameters
9466 and set the encoded OCSP response in the callback. Add simplified examples
9467 to s_client and s_server.
9471 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9474 + Binary incompatibility of ssl_ctx_st structure
9475 + DTLS interoperation with non-compliant servers
9476 + Don't call get_session_cb() without proposed session
9477 + Fix ia64 assembler code
9479 *Andy Polyakov, Steve Henson*
9481 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9483 * DTLS Handshake overhaul. There were longstanding issues with
9484 OpenSSL DTLS implementation, which were making it impossible for
9485 RFC 4347 compliant client to communicate with OpenSSL server.
9486 Unfortunately just fixing these incompatibilities would "cut off"
9487 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9488 server keeps tolerating non RFC compliant syntax. The opposite is
9489 not true, 0.9.8f client can not communicate with earlier server.
9490 This update even addresses CVE-2007-4995.
9494 * Changes to avoid need for function casts in OpenSSL: some compilers
9495 (gcc 4.2 and later) reject their use.
9496 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9499 * Add RFC4507 support to OpenSSL. This includes the corrections in
9500 RFC4507bis. The encrypted ticket format is an encrypted encoded
9501 SSL_SESSION structure, that way new session features are automatically
9504 If a client application caches session in an SSL_SESSION structure
9505 support is transparent because tickets are now stored in the encoded
9508 The SSL_CTX structure automatically generates keys for ticket
9509 protection in servers so again support should be possible
9510 with no application modification.
9512 If a client or server wishes to disable RFC4507 support then the option
9513 SSL_OP_NO_TICKET can be set.
9515 Add a TLS extension debugging callback to allow the contents of any client
9516 or server extensions to be examined.
9518 This work was sponsored by Google.
9522 * Add initial support for TLS extensions, specifically for the server_name
9523 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9524 have new members for a host name. The SSL data structure has an
9525 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9526 stored in that context to allow for session resumption, even after the
9527 SSL has been switched to a new SSL_CTX in reaction to a client's
9528 server_name extension.
9530 New functions (subject to change):
9532 SSL_get_servername()
9533 SSL_get_servername_type()
9536 New CTRL codes and macros (subject to change):
9538 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9539 - SSL_CTX_set_tlsext_servername_callback()
9540 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9541 - SSL_CTX_set_tlsext_servername_arg()
9542 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9544 openssl s_client has a new '-servername ...' option.
9546 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9547 '-key2 ...', '-servername_fatal' (subject to change). This allows
9548 testing the HostName extension for a specific single host name ('-cert'
9549 and '-key' remain fallbacks for handshakes without HostName
9550 negotiation). If the unrecognized_name alert has to be sent, this by
9551 default is a warning; it becomes fatal with the '-servername_fatal'
9554 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9556 * Add AES and SSE2 assembly language support to VC++ build.
9560 * Mitigate attack on final subtraction in Montgomery reduction.
9564 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9565 (which previously caused an internal error).
9569 * Squeeze another 10% out of IGE mode when in != out.
9573 * AES IGE mode speedup.
9575 *Dean Gaudet (Google)*
9577 * Add the Korean symmetric 128-bit cipher SEED (see
9578 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9579 add SEED ciphersuites from RFC 4162:
9581 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9582 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9583 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9584 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9586 To minimize changes between patchlevels in the OpenSSL 0.9.8
9587 series, SEED remains excluded from compilation unless OpenSSL
9588 is configured with 'enable-seed'.
9590 *KISA, Bodo Moeller*
9592 * Mitigate branch prediction attacks, which can be practical if a
9593 single processor is shared, allowing a spy process to extract
9594 information. For detailed background information, see
9595 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9596 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9597 and Necessary Software Countermeasures"). The core of the change
9598 are new versions BN_div_no_branch() and
9599 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9600 respectively, which are slower, but avoid the security-relevant
9601 conditional branches. These are automatically called by BN_div()
9602 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9603 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9604 remove a conditional branch.
9606 BN_FLG_CONSTTIME is the new name for the previous
9607 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9608 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9609 in the exponent causes BN_mod_exp_mont() to use the alternative
9610 implementation in BN_mod_exp_mont_consttime().) The old name
9611 remains as a deprecated alias.
9613 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9614 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9615 constant-time implementations for more than just exponentiation.
9616 Here too the old name is kept as a deprecated alias.
9618 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9619 the BN_BLINDING structure gets an independent copy of the
9620 modulus. This means that the previous `BIGNUM *m` argument to
9621 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9622 essentially becomes `const BIGNUM *m`, although we can't actually
9623 change this in the header file before 0.9.9. It allows
9624 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9625 enable BN_FLG_CONSTTIME.
9627 *Matthew D Wood (Intel Corp)*
9629 * In the SSL/TLS server implementation, be strict about session ID
9630 context matching (which matters if an application uses a single
9631 external cache for different purposes). Previously,
9632 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9633 set. This did ensure strict client verification, but meant that,
9634 with applications using a single external cache for quite
9635 different requirements, clients could circumvent ciphersuite
9636 restrictions for a given session ID context by starting a session
9637 in a different context.
9641 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9642 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9643 authentication-only ciphersuites.
9647 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9648 not complete and could lead to a possible single byte overflow
9649 ([CVE-2007-5135]) [Ben Laurie]
9651 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9653 * Since AES128 and AES256 (and similarly Camellia128 and
9654 Camellia256) share a single mask bit in the logic of
9655 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9656 kludge to work properly if AES128 is available and AES256 isn't
9657 (or if Camellia128 is available and Camellia256 isn't).
9661 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9662 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9663 When a point or a seed is encoded in a BIT STRING, we need to
9664 prevent the removal of trailing zero bits to get the proper DER
9665 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9666 of a NamedBitList, for which trailing 0 bits need to be removed.)
9670 * Have SSL/TLS server implementation tolerate "mismatched" record
9671 protocol version while receiving ClientHello even if the
9672 ClientHello is fragmented. (The server can't insist on the
9673 particular protocol version it has chosen before the ServerHello
9674 message has informed the client about his choice.)
9678 * Add RFC 3779 support.
9680 *Rob Austein for ARIN, Ben Laurie*
9682 * Load error codes if they are not already present instead of using a
9683 static variable. This allows them to be cleanly unloaded and reloaded.
9684 Improve header file function name parsing.
9688 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9689 or CAPABILITY handshake as required by RFCs.
9693 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9695 * Introduce limits to prevent malicious keys being able to
9696 cause a denial of service. ([CVE-2006-2940])
9698 *Steve Henson, Bodo Moeller*
9700 * Fix ASN.1 parsing of certain invalid structures that can result
9701 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9703 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9704 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9706 * Fix SSL client code which could crash if connecting to a
9707 malicious SSLv2 server. ([CVE-2006-4343])
9709 *Tavis Ormandy and Will Drewry, Google Security Team*
9711 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9712 match only those. Before that, "AES256-SHA" would be interpreted
9713 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9714 the same strength classification in 0.9.7h) as we currently only
9715 have a single AES bit in the ciphersuite description bitmap.
9716 That change, however, also applied to ciphersuite strings such as
9717 "RC4-MD5" that intentionally matched multiple ciphersuites --
9718 namely, SSL 2.0 ciphersuites in addition to the more common ones
9719 from SSL 3.0/TLS 1.0.
9721 So we change the selection algorithm again: Naming an explicit
9722 ciphersuite selects this one ciphersuite, and any other similar
9723 ciphersuite (same bitmap) from *other* protocol versions.
9724 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9725 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9727 Since SSL 2.0 does not have any ciphersuites for which the
9728 128/256 bit distinction would be relevant, this works for now.
9729 The proper fix will be to use different bits for AES128 and
9730 AES256, which would have avoided the problems from the beginning;
9731 however, bits are scarce, so we can only do this in a new release
9732 (not just a patchlevel) when we can change the SSL_CIPHER
9733 definition to split the single 'unsigned long mask' bitmap into
9734 multiple values to extend the available space.
9738 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9740 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9741 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9743 * Add AES IGE and biIGE modes.
9747 * Change the Unix randomness entropy gathering to use poll() when
9748 possible instead of select(), since the latter has some
9749 undesirable limitations.
9751 *Darryl Miles via Richard Levitte and Bodo Moeller*
9753 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9754 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9755 cannot be implicitly activated as part of, e.g., the "AES" alias.
9756 However, please upgrade to OpenSSL 0.9.9[-dev] for
9757 non-experimental use of the ECC ciphersuites to get TLS extension
9758 support, which is required for curve and point format negotiation
9759 to avoid potential handshake problems.
9763 * Disable rogue ciphersuites:
9765 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9766 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9767 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9769 The latter two were purportedly from
9770 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9773 Also deactivate the remaining ciphersuites from
9774 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9775 unofficial, and the ID has long expired.
9779 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9780 dual-core machines) and other potential thread-safety issues.
9784 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9785 versions), which is now available for royalty-free use
9786 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9787 Also, add Camellia TLS ciphersuites from RFC 4132.
9789 To minimize changes between patchlevels in the OpenSSL 0.9.8
9790 series, Camellia remains excluded from compilation unless OpenSSL
9791 is configured with 'enable-camellia'.
9795 * Disable the padding bug check when compression is in use. The padding
9796 bug check assumes the first packet is of even length, this is not
9797 necessarily true if compression is enabled and can result in false
9798 positives causing handshake failure. The actual bug test is ancient
9799 code so it is hoped that implementations will either have fixed it by
9800 now or any which still have the bug do not support compression.
9804 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9806 * When applying a cipher rule check to see if string match is an explicit
9807 cipher suite and only match that one cipher suite if it is.
9811 * Link in manifests for VC++ if needed.
9813 *Austin Ziegler <halostatue@gmail.com>*
9815 * Update support for ECC-based TLS ciphersuites according to
9816 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9817 TLS extensions, which are supported starting with the 0.9.9
9818 branch, not in the OpenSSL 0.9.8 branch).
9822 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9823 opaque EVP_CIPHER_CTX handling.
9827 * Fixes and enhancements to zlib compression code. We now only use
9828 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9829 to conform with the standards mentioned here:
9830 <http://www.zlib.net/DLL_FAQ.txt>
9831 Static zlib linking now works on Windows and the new --with-zlib-include
9832 --with-zlib-lib options to Configure can be used to supply the location
9833 of the headers and library. Gracefully handle case where zlib library
9838 * Several fixes and enhancements to the OID generation code. The old code
9839 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9840 handle numbers larger than ULONG_MAX, truncated printing and had a
9841 non standard OBJ_obj2txt() behaviour.
9845 * Add support for building of engines under engine/ as shared libraries
9846 under VC++ build system.
9850 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9851 Hopefully, we will not see any false combination of paths any more.
9855 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9857 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9858 (part of SSL_OP_ALL). This option used to disable the
9859 countermeasure against man-in-the-middle protocol-version
9860 rollback in the SSL 2.0 server implementation, which is a bad
9861 idea. ([CVE-2005-2969])
9863 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9864 for Information Security, National Institute of Advanced Industrial
9865 Science and Technology [AIST], Japan)*
9867 * Add two function to clear and return the verify parameter flags.
9871 * Keep cipherlists sorted in the source instead of sorting them at
9872 runtime, thus removing the need for a lock.
9876 * Avoid some small subgroup attacks in Diffie-Hellman.
9878 *Nick Mathewson and Ben Laurie*
9880 * Add functions for well-known primes.
9884 * Extended Windows CE support.
9886 *Satoshi Nakamura and Andy Polyakov*
9888 * Initialize SSL_METHOD structures at compile time instead of during
9889 runtime, thus removing the need for a lock.
9893 * Make PKCS7_decrypt() work even if no certificate is supplied by
9894 attempting to decrypt each encrypted key in turn. Add support to
9899 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9901 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9904 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9908 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9909 key into the same file any more.
9913 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9917 * Add -utf8 command line and config file option to 'ca'.
9919 *Stefan <stf@udoma.org*
9921 * Removed the macro des_crypt(), as it seems to conflict with some
9922 libraries. Use DES_crypt().
9926 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9927 involves renaming the source and generated shared-libs for
9928 both. The engines will accept the corrected or legacy ids
9929 ('ncipher' and '4758_cca' respectively) when binding. NB,
9930 this only applies when building 'shared'.
9932 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9934 * Add attribute functions to EVP_PKEY structure. Modify
9935 PKCS12_create() to recognize a CSP name attribute and
9936 use it. Make -CSP option work again in pkcs12 utility.
9940 * Add new functionality to the bn blinding code:
9941 - automatic re-creation of the BN_BLINDING parameters after
9942 a fixed number of uses (currently 32)
9943 - add new function for parameter creation
9944 - introduce flags to control the update behaviour of the
9945 BN_BLINDING parameters
9946 - hide BN_BLINDING structure
9947 Add a second BN_BLINDING slot to the RSA structure to improve
9948 performance when a single RSA object is shared among several
9953 * Add support for DTLS.
9955 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9957 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9958 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9962 * Remove buggy and incomplete DH cert support from
9963 ssl/ssl_rsa.c and ssl/s3_both.c
9967 * Use SHA-1 instead of MD5 as the default digest algorithm for
9968 the `apps/openssl` commands.
9972 * Compile clean with "-Wall -Wmissing-prototypes
9973 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9974 DEBUG_SAFESTACK must also be set.
9978 * Change ./Configure so that certain algorithms can be disabled by default.
9979 The new counterpiece to "no-xxx" is "enable-xxx".
9981 The patented RC5 and MDC2 algorithms will now be disabled unless
9982 "enable-rc5" and "enable-mdc2", respectively, are specified.
9984 (IDEA remains enabled despite being patented. This is because IDEA
9985 is frequently required for interoperability, and there is no license
9986 fee for non-commercial use. As before, "no-idea" can be used to
9987 avoid this algorithm.)
9991 * Add processing of proxy certificates (see RFC 3820). This work was
9992 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9993 EGEE (Enabling Grids for E-science in Europe).
9997 * RC4 performance overhaul on modern architectures/implementations, such
9998 as Intel P4, IA-64 and AMD64.
10002 * New utility extract-section.pl. This can be used specify an alternative
10003 section number in a pod file instead of having to treat each file as
10004 a separate case in Makefile. This can be done by adding two lines to the
10007 =for comment openssl_section:XXX
10009 The blank line is mandatory.
10013 * New arguments -certform, -keyform and -pass for s_client and s_server
10014 to allow alternative format key and certificate files and passphrase
10019 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10020 update associated structures and add various utility functions.
10022 Add new policy related verify parameters, include policy checking in
10023 standard verify code. Enhance 'smime' application with extra parameters
10024 to support policy checking and print out.
10028 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10029 Nehemiah processors. These extensions support AES encryption in hardware
10030 as well as RNG (though RNG support is currently disabled).
10032 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10034 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10038 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10040 *Andy Polyakov and a number of other people*
10042 * Improved PowerPC platform support. Most notably BIGNUM assembler
10043 implementation contributed by IBM.
10045 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10047 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10048 exponent rather than 'unsigned long'. There is a corresponding change to
10049 the new 'rsa_keygen' element of the RSA_METHOD structure.
10051 *Jelte Jansen, Geoff Thorpe*
10053 * Functionality for creating the initial serial number file is now
10054 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10056 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10057 number file to 1, which is bound to cause problems. To avoid
10058 the problems while respecting compatibility between different 0.9.7
10059 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10060 CA.pl for serial number initialization. With the new release 0.9.8,
10061 we can fix the problem directly in the 'ca' utility.)
10065 * Reduced header interdependencies by declaring more opaque objects in
10066 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10067 give fewer recursive includes, which could break lazy source code - so
10068 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10069 developers should define this symbol when building and using openssl to
10070 ensure they track the recommended behaviour, interfaces, [etc], but
10071 backwards-compatible behaviour prevails when this isn't defined.
10075 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10079 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10080 This will generate a random key of the appropriate length based on the
10081 cipher context. The EVP_CIPHER can provide its own random key generation
10082 routine to support keys of a specific form. This is used in the des and
10083 3des routines to generate a key of the correct parity. Update S/MIME
10084 code to use new functions and hence generate correct parity DES keys.
10085 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10086 valid (weak or incorrect parity).
10090 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10091 as looking them up. This is useful when the verified structure may contain
10092 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10093 present unless the new PKCS7_NO_CRL flag is asserted.
10097 * Extend ASN1 oid configuration module. It now additionally accepts the
10100 shortName = some long name, 1.2.3.4
10104 * Reimplemented the BN_CTX implementation. There is now no more static
10105 limitation on the number of variables it can handle nor the depth of the
10106 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10107 information can now expand as required, and rather than having a single
10108 static array of bignums, BN_CTX now uses a linked-list of such arrays
10109 allowing it to expand on demand whilst maintaining the usefulness of
10110 BN_CTX's "bundling".
10114 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10115 to allow all RSA operations to function using a single BN_CTX.
10119 * Preliminary support for certificate policy evaluation and checking. This
10120 is initially intended to pass the tests outlined in "Conformance Testing
10121 of Relying Party Client Certificate Path Processing Logic" v1.07.
10125 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10126 remained unused and not that useful. A variety of other little bignum
10127 tweaks and fixes have also been made continuing on from the audit (see
10132 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10133 associated ASN1, EVP and SSL functions and old ASN1 macros.
10137 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10138 and this should never fail. So the return value from the use of
10139 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10140 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10144 * BN_CTX_get() should return zero-valued bignums, providing the same
10145 initialised value as BN_new().
10147 *Geoff Thorpe, suggested by Ulf Möller*
10149 * Support for inhibitAnyPolicy certificate extension.
10153 * An audit of the BIGNUM code is underway, for which debugging code is
10154 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10155 is considered valid when processing BIGNUMs, and causes execution to
10156 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10157 further steps are taken to deliberately pollute unused data in BIGNUM
10158 structures to try and expose faulty code further on. For now, openssl will
10159 (in its default mode of operation) continue to tolerate the inconsistent
10160 forms that it has tolerated in the past, but authors and packagers should
10161 consider trying openssl and their own applications when compiled with
10162 these debugging symbols defined. It will help highlight potential bugs in
10163 their own code, and will improve the test coverage for OpenSSL itself. At
10164 some point, these tighter rules will become openssl's default to improve
10165 maintainability, though the assert()s and other overheads will remain only
10166 in debugging configurations. See bn.h for more details.
10168 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10170 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10171 that can only be obtained through BN_CTX_new() (which implicitly
10172 initialises it). The presence of this function only made it possible
10173 to overwrite an existing structure (and cause memory leaks).
10177 * Because of the callback-based approach for implementing LHASH as a
10178 template type, lh_insert() adds opaque objects to hash-tables and
10179 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10180 to clean up those corresponding objects before destroying the hash table
10181 (and losing the object pointers). So some over-zealous constifications in
10182 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10183 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10184 prototyped to have "const" restrictions on the object pointers they are
10185 given (and so aren't required to cast them away any more).
10189 * The tmdiff.h API was so ugly and minimal that our own timing utility
10190 (speed) prefers to use its own implementation. The two implementations
10191 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10192 its object type properly exposed (MS_TM) instead of casting to/from
10193 `char *`. This may still change yet if someone realises MS_TM and
10195 aren't necessarily the greatest nomenclatures - but this is what was used
10196 internally to the implementation so I've used that for now.
10200 * Ensure that deprecated functions do not get compiled when
10201 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10202 the self-tests were still using deprecated key-generation functions so
10203 these have been updated also.
10207 * Reorganise PKCS#7 code to separate the digest location functionality
10208 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10209 New function PKCS7_set_digest() to set the digest type for PKCS#7
10210 digestedData type. Add additional code to correctly generate the
10211 digestedData type and add support for this type in PKCS7 initialization
10216 * New function PKCS7_set0_type_other() this initializes a PKCS7
10217 structure of type "other".
10221 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10222 sure the loop does correctly stop and breaking ("division by zero")
10223 modulus operations are not performed. The (pre-generated) prime
10224 table crypto/bn/bn_prime.h was already correct, but it could not be
10225 re-generated on some platforms because of the "division by zero"
10226 situation in the script.
10228 *Ralf S. Engelschall*
10230 * Update support for ECC-based TLS ciphersuites according to
10231 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10232 SHA-1 now is only used for "small" curves (where the
10233 representation of a field element takes up to 24 bytes); for
10234 larger curves, the field element resulting from ECDH is directly
10235 used as premaster secret.
10237 *Douglas Stebila (Sun Microsystems Laboratories)*
10239 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10240 curve secp160r1 to the tests.
10242 *Douglas Stebila (Sun Microsystems Laboratories)*
10244 * Add the possibility to load symbols globally with DSO.
10246 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10248 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10249 control of the error stack.
10253 * Add support for STORE in ENGINE.
10257 * Add the STORE type. The intention is to provide a common interface
10258 to certificate and key stores, be they simple file-based stores, or
10259 HSM-type store, or LDAP stores, or...
10260 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10264 * Add a generic structure called OPENSSL_ITEM. This can be used to
10265 pass a list of arguments to any function as well as provide a way
10266 for a function to pass data back to the caller.
10270 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10271 works like BUF_strdup() but can be used to duplicate a portion of
10272 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10277 * Add the function sk_find_ex() which works like sk_find(), but will
10278 return an index to an element even if an exact match couldn't be
10279 found. The index is guaranteed to point at the element where the
10280 searched-for key would be inserted to preserve sorting order.
10284 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10285 takes an extra flags argument for optional functionality. Currently,
10286 the following flags are defined:
10288 OBJ_BSEARCH_VALUE_ON_NOMATCH
10289 This one gets OBJ_bsearch_ex() to return a pointer to the first
10290 element where the comparing function returns a negative or zero
10293 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10294 This one gets OBJ_bsearch_ex() to return a pointer to the first
10295 element where the comparing function returns zero. This is useful
10296 if there are more than one element where the comparing function
10301 * Make it possible to create self-signed certificates with 'openssl ca'
10302 in such a way that the self-signed certificate becomes part of the
10303 CA database and uses the same mechanisms for serial number generation
10304 as all other certificate signing. The new flag '-selfsign' enables
10305 this functionality. Adapt CA.sh and CA.pl.in.
10309 * Add functionality to check the public key of a certificate request
10310 against a given private. This is useful to check that a certificate
10311 request can be signed by that key (self-signing).
10315 * Make it possible to have multiple active certificates with the same
10316 subject in the CA index file. This is done only if the keyword
10317 'unique_subject' is set to 'no' in the main CA section (default
10318 if 'CA_default') of the configuration file. The value is saved
10319 with the database itself in a separate index attribute file,
10320 named like the index file with '.attr' appended to the name.
10324 * Generate multi-valued AVAs using '+' notation in config files for
10329 * Support for nameConstraints certificate extension.
10333 * Support for policyConstraints certificate extension.
10337 * Support for policyMappings certificate extension.
10341 * Make sure the default DSA_METHOD implementation only uses its
10342 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10343 and change its own handlers to be NULL so as to remove unnecessary
10344 indirection. This lets alternative implementations fallback to the
10345 default implementation more easily.
10349 * Support for directoryName in GeneralName related extensions
10354 * Make it possible to link applications using Makefile.shared.
10355 Make that possible even when linking against static libraries!
10359 * Support for single pass processing for S/MIME signing. This now
10360 means that S/MIME signing can be done from a pipe, in addition
10361 cleartext signing (multipart/signed type) is effectively streaming
10362 and the signed data does not need to be all held in memory.
10364 This is done with a new flag PKCS7_STREAM. When this flag is set
10365 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10366 is done after the data is output (and digests calculated) in
10367 SMIME_write_PKCS7().
10371 * Add full support for -rpath/-R, both in shared libraries and
10372 applications, at least on the platforms where it's known how
10377 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10378 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10379 will now compute a table of multiples of the generator that
10380 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10381 faster (notably in the case of a single point multiplication,
10382 scalar * generator).
10384 *Nils Larsch, Bodo Moeller*
10386 * IPv6 support for certificate extensions. The various extensions
10387 which use the IP:a.b.c.d can now take IPv6 addresses using the
10388 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10393 * Added an ENGINE that implements RSA by performing private key
10394 exponentiations with the GMP library. The conversions to and from
10395 GMP's mpz_t format aren't optimised nor are any montgomery forms
10396 cached, and on x86 it appears OpenSSL's own performance has caught up.
10397 However there are likely to be other architectures where GMP could
10398 provide a boost. This ENGINE is not built in by default, but it can be
10399 specified at Configure time and should be accompanied by the necessary
10400 linker additions, eg;
10401 ./config -DOPENSSL_USE_GMP -lgmp
10405 * "openssl engine" will not display ENGINE/DSO load failure errors when
10406 testing availability of engines with "-t" - the old behaviour is
10407 produced by increasing the feature's verbosity with "-tt".
10411 * ECDSA routines: under certain error conditions uninitialized BN objects
10412 could be freed. Solution: make sure initialization is performed early
10413 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10418 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10419 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10420 software implementations. For DSA and DH, parameter generation can
10421 also be overridden by providing the appropriate method callbacks.
10425 * Change the "progress" mechanism used in key-generation and
10426 primality testing to functions that take a new BN_GENCB pointer in
10427 place of callback/argument pairs. The new API functions have `_ex`
10428 postfixes and the older functions are reimplemented as wrappers for
10429 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10430 declarations of the old functions to help (graceful) attempts to
10431 migrate to the new functions. Also, the new key-generation API
10432 functions operate on a caller-supplied key-structure and return
10433 success/failure rather than returning a key or NULL - this is to
10434 help make "keygen" another member function of RSA_METHOD etc.
10436 Example for using the new callback interface:
10438 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10439 void *my_arg = ...;
10442 BN_GENCB_set(&my_cb, my_callback, my_arg);
10444 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10445 /* For the meaning of a, b in calls to my_callback(), see the
10446 * documentation of the function that calls the callback.
10447 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10448 * my_callback should return 1 if it wants BN_is_prime_ex()
10449 * to continue, or 0 to stop.
10454 * Change the ZLIB compression method to be stateful, and make it
10455 available to TLS with the number defined in
10456 draft-ietf-tls-compression-04.txt.
10460 * Add the ASN.1 structures and functions for CertificatePair, which
10461 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10463 CertificatePair ::= SEQUENCE {
10464 forward [0] Certificate OPTIONAL,
10465 reverse [1] Certificate OPTIONAL,
10466 -- at least one of the pair shall be present -- }
10468 Also implement the PEM functions to read and write certificate
10469 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10471 This needed to be defined, mostly for the sake of the LDAP
10472 attribute crossCertificatePair, but may prove useful elsewhere as
10477 * Make it possible to inhibit symlinking of shared libraries in
10478 Makefile.shared, for Cygwin's sake.
10482 * Extend the BIGNUM API by creating a function
10483 void BN_set_negative(BIGNUM *a, int neg);
10484 and a macro that behave like
10485 int BN_is_negative(const BIGNUM *a);
10487 to avoid the need to access 'a->neg' directly in applications.
10491 * Implement fast modular reduction for pseudo-Mersenne primes
10492 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10493 EC_GROUP_new_curve_GFp() will now automatically use this
10496 *Nils Larsch <nla@trustcenter.de>*
10498 * Add new lock type (CRYPTO_LOCK_BN).
10502 * Change the ENGINE framework to automatically load engines
10503 dynamically from specific directories unless they could be
10504 found to already be built in or loaded. Move all the
10505 current engines except for the cryptodev one to a new
10506 directory engines/.
10507 The engines in engines/ are built as shared libraries if
10508 the "shared" options was given to ./Configure or ./config.
10509 Otherwise, they are inserted in libcrypto.a.
10510 /usr/local/ssl/engines is the default directory for dynamic
10511 engines, but that can be overridden at configure time through
10512 the usual use of --prefix and/or --openssldir, and at run
10513 time with the environment variable OPENSSL_ENGINES.
10515 *Geoff Thorpe and Richard Levitte*
10517 * Add Makefile.shared, a helper makefile to build shared
10518 libraries. Adapt Makefile.org.
10522 * Add version info to Win32 DLLs.
10524 *Peter 'Luna' Runestig" <peter@runestig.com>*
10526 * Add new 'medium level' PKCS#12 API. Certificates and keys
10527 can be added using this API to created arbitrary PKCS#12
10528 files while avoiding the low-level API.
10530 New options to PKCS12_create(), key or cert can be NULL and
10531 will then be omitted from the output file. The encryption
10532 algorithm NIDs can be set to -1 for no encryption, the mac
10533 iteration count can be set to 0 to omit the mac.
10535 Enhance pkcs12 utility by making the -nokeys and -nocerts
10536 options work when creating a PKCS#12 file. New option -nomac
10537 to omit the mac, NONE can be set for an encryption algorithm.
10538 New code is modified to use the enhanced PKCS12_create()
10539 instead of the low-level API.
10543 * Extend ASN1 encoder to support indefinite length constructed
10544 encoding. This can output sequences tags and octet strings in
10545 this form. Modify pk7_asn1.c to support indefinite length
10546 encoding. This is experimental and needs additional code to
10547 be useful, such as an ASN1 bio and some enhanced streaming
10550 Extend template encode functionality so that tagging is passed
10551 down to the template encoder.
10555 * Let 'openssl req' fail if an argument to '-newkey' is not
10556 recognized instead of using RSA as a default.
10560 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10561 As these are not official, they are not included in "ALL";
10562 the "ECCdraft" ciphersuite group alias can be used to select them.
10564 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10566 * Add ECDH engine support.
10568 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10570 * Add ECDH in new directory crypto/ecdh/.
10572 *Douglas Stebila (Sun Microsystems Laboratories)*
10574 * Let BN_rand_range() abort with an error after 100 iterations
10575 without success (which indicates a broken PRNG).
10579 * Change BN_mod_sqrt() so that it verifies that the input value
10580 is really the square of the return value. (Previously,
10581 BN_mod_sqrt would show GIGO behaviour.)
10585 * Add named elliptic curves over binary fields from X9.62, SECG,
10586 and WAP/WTLS; add OIDs that were still missing.
10588 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10590 * Extend the EC library for elliptic curves over binary fields
10591 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10594 EC_GF2m_simple_method
10598 EC_GROUP_new_curve_GF2m
10599 EC_GROUP_set_curve_GF2m
10600 EC_GROUP_get_curve_GF2m
10601 EC_POINT_set_affine_coordinates_GF2m
10602 EC_POINT_get_affine_coordinates_GF2m
10603 EC_POINT_set_compressed_coordinates_GF2m
10605 Point compression for binary fields is disabled by default for
10606 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10609 As binary polynomials are represented as BIGNUMs, various members
10610 of the EC_GROUP and EC_POINT data structures can be shared
10611 between the implementations for prime fields and binary fields;
10612 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10613 are essentially identical to their `..._GFp` counterparts.
10614 (For simplicity, the `..._GFp` prefix has been dropped from
10615 various internal method names.)
10617 An internal 'field_div' method (similar to 'field_mul' and
10618 'field_sqr') has been added; this is used only for binary fields.
10620 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10622 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10623 through methods ('mul', 'precompute_mult').
10625 The generic implementations (now internally called 'ec_wNAF_mul'
10626 and 'ec_wNAF_precomputed_mult') remain the default if these
10627 methods are undefined.
10629 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10631 * New function EC_GROUP_get_degree, which is defined through
10632 EC_METHOD. For curves over prime fields, this returns the bit
10633 length of the modulus.
10635 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10637 * New functions EC_GROUP_dup, EC_POINT_dup.
10638 (These simply call ..._new and ..._copy).
10640 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10642 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10643 Polynomials are represented as BIGNUMs (where the sign bit is not
10644 used) in the following functions [macros]:
10647 BN_GF2m_sub [= BN_GF2m_add]
10648 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10649 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10650 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10652 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10653 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10654 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10655 BN_GF2m_cmp [= BN_ucmp]
10657 (Note that only the 'mod' functions are actually for fields GF(2^m).
10658 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10660 For some functions, an the irreducible polynomial defining a
10661 field can be given as an 'unsigned int[]' with strictly
10662 decreasing elements giving the indices of those bits that are set;
10663 i.e., p[] represents the polynomial
10664 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10666 p[0] > p[1] > ... > p[k] = 0.
10667 This applies to the following functions:
10670 BN_GF2m_mod_mul_arr
10671 BN_GF2m_mod_sqr_arr
10672 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10673 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10674 BN_GF2m_mod_exp_arr
10675 BN_GF2m_mod_sqrt_arr
10676 BN_GF2m_mod_solve_quad_arr
10680 Conversion can be performed by the following functions:
10685 bntest.c has additional tests for binary polynomial arithmetic.
10687 Two implementations for BN_GF2m_mod_div() are available.
10688 The default algorithm simply uses BN_GF2m_mod_inv() and
10689 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10690 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10691 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10693 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10695 * Add new error code 'ERR_R_DISABLED' that can be used when some
10696 functionality is disabled at compile-time.
10698 *Douglas Stebila <douglas.stebila@sun.com>*
10700 * Change default behaviour of 'openssl asn1parse' so that more
10701 information is visible when viewing, e.g., a certificate:
10703 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10704 mode the content of non-printable OCTET STRINGs is output in a
10705 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10706 avoid the appearance of a printable string.
10708 *Nils Larsch <nla@trustcenter.de>*
10710 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10712 EC_GROUP_set_asn1_flag()
10713 EC_GROUP_get_asn1_flag()
10714 EC_GROUP_set_point_conversion_form()
10715 EC_GROUP_get_point_conversion_form()
10716 These control ASN1 encoding details:
10717 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10718 has been set to OPENSSL_EC_NAMED_CURVE.
10719 - Points are encoded in uncompressed form by default; options for
10720 asn1_for are as for point2oct, namely
10721 POINT_CONVERSION_COMPRESSED
10722 POINT_CONVERSION_UNCOMPRESSED
10723 POINT_CONVERSION_HYBRID
10725 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10727 EC_GROUP_set_seed()
10728 EC_GROUP_get0_seed()
10729 EC_GROUP_get_seed_len()
10730 This is used only for ASN1 purposes (so far).
10732 *Nils Larsch <nla@trustcenter.de>*
10734 * Add 'field_type' member to EC_METHOD, which holds the NID
10735 of the appropriate field type OID. The new function
10736 EC_METHOD_get_field_type() returns this value.
10738 *Nils Larsch <nla@trustcenter.de>*
10741 EC_POINT_point2bn()
10742 EC_POINT_bn2point()
10743 EC_POINT_point2hex()
10744 EC_POINT_hex2point()
10745 providing useful interfaces to EC_POINT_point2oct() and
10746 EC_POINT_oct2point().
10748 *Nils Larsch <nla@trustcenter.de>*
10750 * Change internals of the EC library so that the functions
10751 EC_GROUP_set_generator()
10752 EC_GROUP_get_generator()
10753 EC_GROUP_get_order()
10754 EC_GROUP_get_cofactor()
10755 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10756 to methods, which would lead to unnecessary code duplication when
10757 adding different types of curves.
10759 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10761 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10762 arithmetic, and such that modified wNAFs are generated
10763 (which avoid length expansion in many cases).
10767 * Add a function EC_GROUP_check_discriminant() (defined via
10768 EC_METHOD) that verifies that the curve discriminant is non-zero.
10770 Add a function EC_GROUP_check() that makes some sanity tests
10771 on a EC_GROUP, its generator and order. This includes
10772 EC_GROUP_check_discriminant().
10774 *Nils Larsch <nla@trustcenter.de>*
10776 * Add ECDSA in new directory crypto/ecdsa/.
10778 Add applications 'openssl ecparam' and 'openssl ecdsa'
10779 (these are based on 'openssl dsaparam' and 'openssl dsa').
10781 ECDSA support is also included in various other files across the
10782 library. Most notably,
10783 - 'openssl req' now has a '-newkey ecdsa:file' option;
10784 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10785 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10786 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10787 them suitable for ECDSA where domain parameters must be
10788 extracted before the specific public key;
10789 - ECDSA engine support has been added.
10791 *Nils Larsch <nla@trustcenter.de>*
10793 * Include some named elliptic curves, and add OIDs from X9.62,
10794 SECG, and WAP/WTLS. Each curve can be obtained from the new
10796 EC_GROUP_new_by_curve_name(),
10797 and the list of available named curves can be obtained with
10798 EC_get_builtin_curves().
10799 Also add a 'curve_name' member to EC_GROUP objects, which can be
10801 EC_GROUP_set_curve_name()
10802 EC_GROUP_get_curve_name()
10804 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10806 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10807 was actually never needed) and in BN_mul(). The removal in BN_mul()
10808 required a small change in bn_mul_part_recursive() and the addition
10809 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10810 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10811 bn_sub_words() and bn_add_words() except they take arrays with
10816 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10818 * Cleanse PEM buffers before freeing them since they may contain
10821 *Benjamin Bennett <ben@psc.edu>*
10823 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10824 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10825 authentication-only ciphersuites.
10829 * Since AES128 and AES256 share a single mask bit in the logic of
10830 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10831 kludge to work properly if AES128 is available and AES256 isn't.
10835 * Expand security boundary to match 1.1.1 module.
10839 * Remove redundant features: hash file source, editing of test vectors
10840 modify fipsld to use external fips_premain.c signature.
10844 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10845 run algorithm test programs.
10849 * Make algorithm test programs more tolerant of whitespace.
10853 * Have SSL/TLS server implementation tolerate "mismatched" record
10854 protocol version while receiving ClientHello even if the
10855 ClientHello is fragmented. (The server can't insist on the
10856 particular protocol version it has chosen before the ServerHello
10857 message has informed the client about his choice.)
10861 * Load error codes if they are not already present instead of using a
10862 static variable. This allows them to be cleanly unloaded and reloaded.
10866 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10868 * Introduce limits to prevent malicious keys being able to
10869 cause a denial of service. ([CVE-2006-2940])
10871 *Steve Henson, Bodo Moeller*
10873 * Fix ASN.1 parsing of certain invalid structures that can result
10874 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10876 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10877 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10879 * Fix SSL client code which could crash if connecting to a
10880 malicious SSLv2 server. ([CVE-2006-4343])
10882 *Tavis Ormandy and Will Drewry, Google Security Team*
10884 * Change ciphersuite string processing so that an explicit
10885 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10886 will no longer include "AES128-SHA"), and any other similar
10887 ciphersuite (same bitmap) from *other* protocol versions (so that
10888 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10889 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10890 changes from 0.9.8b and 0.9.8d.
10894 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10896 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10897 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10899 * Change the Unix randomness entropy gathering to use poll() when
10900 possible instead of select(), since the latter has some
10901 undesirable limitations.
10903 *Darryl Miles via Richard Levitte and Bodo Moeller*
10905 * Disable rogue ciphersuites:
10907 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10908 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10909 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10911 The latter two were purportedly from
10912 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10915 Also deactivate the remaining ciphersuites from
10916 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10917 unofficial, and the ID has long expired.
10921 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10922 dual-core machines) and other potential thread-safety issues.
10926 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10928 * Adapt fipsld and the build system to link against the validated FIPS
10929 module in FIPS mode.
10933 * Fixes for VC++ 2005 build under Windows.
10937 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10938 from a Windows bash shell such as MSYS. It is autodetected from the
10939 "config" script when run from a VC++ environment. Modify standard VC++
10940 build to use fipscanister.o from the GNU make build.
10944 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10946 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10947 The value now differs depending on if you build for FIPS or not.
10948 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10949 safely run with a non-FIPSed libcrypto, as it may crash because of
10950 the difference induced by this change.
10954 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10956 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10957 (part of SSL_OP_ALL). This option used to disable the
10958 countermeasure against man-in-the-middle protocol-version
10959 rollback in the SSL 2.0 server implementation, which is a bad
10960 idea. ([CVE-2005-2969])
10962 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10963 for Information Security, National Institute of Advanced Industrial
10964 Science and Technology [AIST, Japan)]*
10966 * Minimal support for X9.31 signatures and PSS padding modes. This is
10967 mainly for FIPS compliance and not fully integrated at this stage.
10971 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10972 the exponentiation using a fixed-length exponent. (Otherwise,
10973 the information leaked through timing could expose the secret key
10974 after many signatures; cf. Bleichenbacher's attack on DSA with
10979 * Make a new fixed-window mod_exp implementation the default for
10980 RSA, DSA, and DH private-key operations so that the sequence of
10981 squares and multiplies and the memory access pattern are
10982 independent of the particular secret key. This will mitigate
10983 cache-timing and potential related attacks.
10985 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10986 and this is automatically used by BN_mod_exp_mont() if the new flag
10987 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10988 will use this BN flag for private exponents unless the flag
10989 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10990 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10992 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10994 * Change the client implementation for SSLv23_method() and
10995 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10996 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10997 (Previously, the SSL 2.0 backwards compatible Client Hello
10998 message format would be used even with SSL_OP_NO_SSLv2.)
11002 * Add support for smime-type MIME parameter in S/MIME messages which some
11007 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11008 a threadsafe manner. Modify rsa code to use new function and add calls
11009 to dsa and dh code (which had race conditions before).
11013 * Include the fixed error library code in the C error file definitions
11014 instead of fixing them up at runtime. This keeps the error code
11015 structures constant.
11019 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11021 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11024 * Fixes for newer kerberos headers. NB: the casts are needed because
11025 the 'length' field is signed on one version and unsigned on another
11026 with no (?) obvious way to tell the difference, without these VC++
11027 complains. Also the "definition" of FAR (blank) is no longer included
11028 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11029 some needed definitions.
11033 * Undo Cygwin change.
11037 * Added support for proxy certificates according to RFC 3820.
11038 Because they may be a security thread to unaware applications,
11039 they must be explicitly allowed in run-time. See
11040 docs/HOWTO/proxy_certificates.txt for further information.
11044 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11046 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11047 server and client random values. Previously
11048 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11049 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11051 This change has negligible security impact because:
11053 1. Server and client random values still have 24 bytes of pseudo random
11056 2. Server and client random values are sent in the clear in the initial
11059 3. The master secret is derived using the premaster secret (48 bytes in
11060 size for static RSA ciphersuites) as well as client server and random
11063 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11066 *Stephen Henson, reported by UK NISCC*
11068 * Use Windows randomness collection on Cygwin.
11072 * Fix hang in EGD/PRNGD query when communication socket is closed
11073 prematurely by EGD/PRNGD.
11075 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11077 * Prompt for pass phrases when appropriate for PKCS12 input format.
11081 * Back-port of selected performance improvements from development
11082 branch, as well as improved support for PowerPC platforms.
11086 * Add lots of checks for memory allocation failure, error codes to indicate
11087 failure and freeing up memory if a failure occurs.
11089 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11091 * Add new -passin argument to dgst.
11095 * Perform some character comparisons of different types in X509_NAME_cmp:
11096 this is needed for some certificates that re-encode DNs into UTF8Strings
11097 (in violation of RFC3280) and can't or won't issue name rollover
11102 * Make an explicit check during certificate validation to see that
11103 the CA setting in each certificate on the chain is correct. As a
11104 side effect always do the following basic checks on extensions,
11105 not just when there's an associated purpose to the check:
11107 - if there is an unhandled critical extension (unless the user
11108 has chosen to ignore this fault)
11109 - if the path length has been exceeded (if one is set at all)
11110 - that certain extensions fit the associated purpose (if one has
11115 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11117 * Avoid a race condition when CRLs are checked in a multi threaded
11118 environment. This would happen due to the reordering of the revoked
11119 entries during signature checking and serial number lookup. Now the
11120 encoding is cached and the serial number sort performed under a lock.
11121 Add new STACK function sk_is_sorted().
11125 * Add Delta CRL to the extension code.
11129 * Various fixes to s3_pkt.c so alerts are sent properly.
11131 *David Holmes <d.holmes@f5.com>*
11133 * Reduce the chances of duplicate issuer name and serial numbers (in
11134 violation of RFC3280) using the OpenSSL certificate creation utilities.
11135 This is done by creating a random 64 bit value for the initial serial
11136 number when a serial number file is created or when a self signed
11137 certificate is created using 'openssl req -x509'. The initial serial
11138 number file is created using 'openssl x509 -next_serial' in CA.pl
11139 rather than being initialized to 1.
11143 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11145 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11146 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11148 *Joe Orton, Steve Henson*
11150 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11153 *Joe Orton, Steve Henson*
11155 * Make it possible to have multiple active certificates with the same
11156 subject in the CA index file. This is done only if the keyword
11157 'unique_subject' is set to 'no' in the main CA section (default
11158 if 'CA_default') of the configuration file. The value is saved
11159 with the database itself in a separate index attribute file,
11160 named like the index file with '.attr' appended to the name.
11164 * X509 verify fixes. Disable broken certificate workarounds when
11165 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11166 keyUsage extension present. Don't accept CRLs with unhandled critical
11167 extensions: since verify currently doesn't process CRL extensions this
11168 rejects a CRL with *any* critical extensions. Add new verify error codes
11173 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11174 A clarification of RFC2560 will require the use of OCTET STRINGs and
11175 some implementations cannot handle the current raw format. Since OpenSSL
11176 copies and compares OCSP nonces as opaque blobs without any attempt at
11177 parsing them this should not create any compatibility issues.
11181 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11182 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11183 this HMAC (and other) operations are several times slower than OpenSSL
11188 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11190 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11192 * Use the correct content when signing type "other".
11196 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11198 * Fix various bugs revealed by running the NISCC test suite:
11200 Stop out of bounds reads in the ASN1 code when presented with
11201 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11203 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11205 If verify callback ignores invalid public key errors don't try to check
11206 certificate signature with the NULL public key.
11210 * New -ignore_err option in ocsp application to stop the server
11211 exiting on the first error in a request.
11215 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11216 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11221 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11222 extra data after the compression methods not only for TLS 1.0
11223 but also for SSL 3.0 (as required by the specification).
11225 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11227 * Change X509_certificate_type() to mark the key as exported/exportable
11228 when it's 512 *bits* long, not 512 bytes.
11232 * Change AES_cbc_encrypt() so it outputs exact multiple of
11233 blocks during encryption.
11237 * Various fixes to base64 BIO and non blocking I/O. On write
11238 flushes were not handled properly if the BIO retried. On read
11239 data was not being buffered properly and had various logic bugs.
11240 This also affects blocking I/O when the data being decoded is a
11245 * Various S/MIME bugfixes and compatibility changes:
11246 output correct application/pkcs7 MIME type if
11247 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11248 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11249 of files as .eml work). Correctly handle very long lines in MIME
11254 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11256 * Countermeasure against the Klima-Pokorny-Rosa extension of
11257 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11258 a protocol version number mismatch like a decryption error
11259 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11263 * Turn on RSA blinding by default in the default implementation
11264 to avoid a timing attack. Applications that don't want it can call
11265 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11266 They would be ill-advised to do so in most cases.
11268 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11270 * Change RSA blinding code so that it works when the PRNG is not
11271 seeded (in this case, the secret RSA exponent is abused as
11272 an unpredictable seed -- if it is not unpredictable, there
11273 is no point in blinding anyway). Make RSA blinding thread-safe
11274 by remembering the creator's thread ID in rsa->blinding and
11275 having all other threads use local one-time blinding factors
11276 (this requires more computation than sharing rsa->blinding, but
11277 avoids excessive locking; and if an RSA object is not shared
11278 between threads, blinding will still be very fast).
11282 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11283 ENGINE as defaults for all supported algorithms irrespective of
11284 the 'flags' parameter. 'flags' is now honoured, so applications
11285 should make sure they are passing it correctly.
11289 * Target "mingw" now allows native Windows code to be generated in
11290 the Cygwin environment as well as with the MinGW compiler.
11294 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11296 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11297 via timing by performing a MAC computation even if incorrect
11298 block cipher padding has been found. This is a countermeasure
11299 against active attacks where the attacker has to distinguish
11300 between bad padding and a MAC verification error. ([CVE-2003-0078])
11302 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11303 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11304 Martin Vuagnoux (EPFL, Ilion)*
11306 * Make the no-err option work as intended. The intention with no-err
11307 is not to have the whole error stack handling routines removed from
11308 libcrypto, it's only intended to remove all the function name and
11309 reason texts, thereby removing some of the footprint that may not
11310 be interesting if those errors aren't displayed anyway.
11312 NOTE: it's still possible for any application or module to have its
11313 own set of error texts inserted. The routines are there, just not
11314 used by default when no-err is given.
11318 * Add support for FreeBSD on IA64.
11320 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11322 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11323 Kerberos function mit_des_cbc_cksum(). Before this change,
11324 the value returned by DES_cbc_cksum() was like the one from
11325 mit_des_cbc_cksum(), except the bytes were swapped.
11327 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11329 * Allow an application to disable the automatic SSL chain building.
11330 Before this a rather primitive chain build was always performed in
11331 ssl3_output_cert_chain(): an application had no way to send the
11332 correct chain if the automatic operation produced an incorrect result.
11334 Now the chain builder is disabled if either:
11336 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11338 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11340 The reasoning behind this is that an application would not want the
11341 auto chain building to take place if extra chain certificates are
11342 present and it might also want a means of sending no additional
11343 certificates (for example the chain has two certificates and the
11348 * Add the possibility to build without the ENGINE framework.
11350 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11352 * Under Win32 gmtime() can return NULL: check return value in
11353 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11357 * DSA routines: under certain error conditions uninitialized BN objects
11358 could be freed. Solution: make sure initialization is performed early
11359 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11360 Nils Larsch <nla@trustcenter.de> via PR#459)
11364 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11365 checked on reconnect on the client side, therefore session resumption
11366 could still fail with a "ssl session id is different" error. This
11367 behaviour is masked when SSL_OP_ALL is used due to
11368 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11369 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11370 followup to PR #377.
11374 * IA-32 assembler support enhancements: unified ELF targets, support
11375 for SCO/Caldera platforms, fix for Cygwin shared build.
11379 * Add support for FreeBSD on sparc64. As a consequence, support for
11380 FreeBSD on non-x86 processors is separate from x86 processors on
11381 the config script, much like the NetBSD support.
11383 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11385 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11387 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11390 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11391 code (06) was taken as the first octet of the session ID and the last
11392 octet was ignored consequently. As a result SSLv2 client side session
11393 caching could not have worked due to the session ID mismatch between
11395 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11400 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11401 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11406 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11407 seems that in spite of existing for more than a year, many application
11408 author have done nothing to provide the necessary callbacks, which
11409 means that this particular engine will not work properly anywhere.
11410 This is a very unfortunate situation which forces us, in the name
11411 of usability, to give the hw_ncipher.c a static lock, which is part
11413 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11414 appear in 0.9.8 or later. We EXPECT application authors to have
11415 dealt properly with this when 0.9.8 is released (unless we actually
11416 make such changes in the libcrypto locking code that changes will
11417 have to be made anyway).
11421 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11422 octets have been read, EOF or an error occurs. Without this change
11423 some truncated ASN1 structures will not produce an error.
11427 * Disable Heimdal support, since it hasn't been fully implemented.
11428 Still give the possibility to force the use of Heimdal, but with
11429 warnings and a request that patches get sent to openssl-dev.
11433 * Add the VC-CE target, introduce the WINCE sysname, and add
11434 INSTALL.WCE and appropriate conditionals to make it build.
11436 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11438 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11439 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11440 edit numbers of the version.
11442 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11444 * Introduce safe string copy and catenation functions
11445 (BUF_strlcpy() and BUF_strlcat()).
11447 *Ben Laurie (CHATS) and Richard Levitte*
11449 * Avoid using fixed-size buffers for one-line DNs.
11451 *Ben Laurie (CHATS)*
11453 * Add BUF_MEM_grow_clean() to avoid information leakage when
11454 resizing buffers containing secrets, and use where appropriate.
11456 *Ben Laurie (CHATS)*
11458 * Avoid using fixed size buffers for configuration file location.
11460 *Ben Laurie (CHATS)*
11462 * Avoid filename truncation for various CA files.
11464 *Ben Laurie (CHATS)*
11466 * Use sizeof in preference to magic numbers.
11468 *Ben Laurie (CHATS)*
11470 * Avoid filename truncation in cert requests.
11472 *Ben Laurie (CHATS)*
11474 * Add assertions to check for (supposedly impossible) buffer
11477 *Ben Laurie (CHATS)*
11479 * Don't cache truncated DNS entries in the local cache (this could
11480 potentially lead to a spoofing attack).
11482 *Ben Laurie (CHATS)*
11484 * Fix various buffers to be large enough for hex/decimal
11485 representations in a platform independent manner.
11487 *Ben Laurie (CHATS)*
11489 * Add CRYPTO_realloc_clean() to avoid information leakage when
11490 resizing buffers containing secrets, and use where appropriate.
11492 *Ben Laurie (CHATS)*
11494 * Add BIO_indent() to avoid much slightly worrying code to do
11497 *Ben Laurie (CHATS)*
11499 * Convert sprintf()/BIO_puts() to BIO_printf().
11501 *Ben Laurie (CHATS)*
11503 * buffer_gets() could terminate with the buffer only half
11506 *Ben Laurie (CHATS)*
11508 * Add assertions to prevent user-supplied crypto functions from
11509 overflowing internal buffers by having large block sizes, etc.
11511 *Ben Laurie (CHATS)*
11513 * New OPENSSL_assert() macro (similar to assert(), but enabled
11516 *Ben Laurie (CHATS)*
11518 * Eliminate unused copy of key in RC4.
11520 *Ben Laurie (CHATS)*
11522 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11524 *Ben Laurie (CHATS)*
11526 * Fix off-by-one error in EGD path.
11528 *Ben Laurie (CHATS)*
11530 * If RANDFILE path is too long, ignore instead of truncating.
11532 *Ben Laurie (CHATS)*
11534 * Eliminate unused and incorrectly sized X.509 structure
11537 *Ben Laurie (CHATS)*
11539 * Eliminate unused and dangerous function knumber().
11541 *Ben Laurie (CHATS)*
11543 * Eliminate unused and dangerous structure, KSSL_ERR.
11545 *Ben Laurie (CHATS)*
11547 * Protect against overlong session ID context length in an encoded
11548 session object. Since these are local, this does not appear to be
11551 *Ben Laurie (CHATS)*
11553 * Change from security patch (see 0.9.6e below) that did not affect
11554 the 0.9.6 release series:
11556 Remote buffer overflow in SSL3 protocol - an attacker could
11557 supply an oversized master key in Kerberos-enabled versions.
11560 *Ben Laurie (CHATS)*
11562 * Change the SSL kerb5 codes to match RFC 2712.
11566 * Make -nameopt work fully for req and add -reqopt switch.
11568 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11570 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11572 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11574 * Make sure tests can be performed even if the corresponding algorithms
11575 have been removed entirely. This was also the last step to make
11576 OpenSSL compilable with DJGPP under all reasonable conditions.
11578 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11580 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11581 to allow version independent disabling of normally unselected ciphers,
11582 which may be activated as a side-effect of selecting a single cipher.
11584 (E.g., cipher list string "RSA" enables ciphersuites that are left
11585 out of "ALL" because they do not provide symmetric encryption.
11586 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11588 *Lutz Jaenicke, Bodo Moeller*
11590 * Add appropriate support for separate platform-dependent build
11591 directories. The recommended way to make a platform-dependent
11592 build directory is the following (tested on Linux), maybe with
11595 # Place yourself outside of the OpenSSL source tree. In
11596 # this example, the environment variable OPENSSL_SOURCE
11597 # is assumed to contain the absolute OpenSSL source directory.
11598 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11599 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11600 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11601 mkdir -p `dirname $F`
11602 ln -s $OPENSSL_SOURCE/$F $F
11605 To be absolutely sure not to disturb the source tree, a "make clean"
11606 is a good thing. If it isn't successful, don't worry about it,
11607 it probably means the source directory is very clean.
11611 * Make sure any ENGINE control commands make local copies of string
11612 pointers passed to them whenever necessary. Otherwise it is possible
11613 the caller may have overwritten (or deallocated) the original string
11614 data when a later ENGINE operation tries to use the stored values.
11616 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11618 * Improve diagnostics in file reading and command-line digests.
11620 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11622 * Add AES modes CFB and OFB to the object database. Correct an
11623 error in AES-CFB decryption.
11627 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11628 allows existing EVP_CIPHER_CTX structures to be reused after
11629 calling `EVP_*Final()`. This behaviour is used by encryption
11630 BIOs and some applications. This has the side effect that
11631 applications must explicitly clean up cipher contexts with
11632 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11636 * Check the values of dna and dnb in bn_mul_recursive before calling
11637 bn_mul_comba (a non zero value means the a or b arrays do not contain
11638 n2 elements) and fallback to bn_mul_normal if either is not zero.
11642 * Fix escaping of non-ASCII characters when using the -subj option
11643 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11647 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11648 form for "surname", serialNumber has no short form.
11649 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11650 therefore remove "mail" short name for "internet 7".
11651 The OID for unique identifiers in X509 certificates is
11652 x500UniqueIdentifier, not uniqueIdentifier.
11653 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11657 * Add an "init" command to the ENGINE config module and auto initialize
11658 ENGINEs. Without any "init" command the ENGINE will be initialized
11659 after all ctrl commands have been executed on it. If init=1 the
11660 ENGINE is initialized at that point (ctrls before that point are run
11661 on the uninitialized ENGINE and after on the initialized one). If
11662 init=0 then the ENGINE will not be initialized at all.
11666 * Fix the 'app_verify_callback' interface so that the user-defined
11667 argument is actually passed to the callback: In the
11668 SSL_CTX_set_cert_verify_callback() prototype, the callback
11669 declaration has been changed from
11672 int (*cb)(X509_STORE_CTX *,void *);
11673 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11674 i=s->ctx->app_verify_callback(&ctx)
11675 has been changed into
11676 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11678 To update applications using SSL_CTX_set_cert_verify_callback(),
11679 a dummy argument can be added to their callback functions.
11681 *D. K. Smetters <smetters@parc.xerox.com>*
11683 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11685 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11687 * Add and OPENSSL_LOAD_CONF define which will cause
11688 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11689 This allows older applications to transparently support certain
11690 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11691 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11692 load the config file and OPENSSL_add_all_algorithms_conf() which will
11693 always load it have also been added.
11697 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11698 Adjust NIDs and EVP layer.
11700 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11702 * Config modules support in openssl utility.
11704 Most commands now load modules from the config file,
11705 though in a few (such as version) this isn't done
11706 because it couldn't be used for anything.
11708 In the case of ca and req the config file used is
11709 the same as the utility itself: that is the -config
11710 command line option can be used to specify an
11715 * Move default behaviour from OPENSSL_config(). If appname is NULL
11716 use "openssl_conf" if filename is NULL use default openssl config file.
11720 * Add an argument to OPENSSL_config() to allow the use of an alternative
11721 config section name. Add a new flag to tolerate a missing config file
11722 and move code to CONF_modules_load_file().
11726 * Support for crypto accelerator cards from Accelerated Encryption
11727 Processing, www.aep.ie. (Use engine 'aep')
11728 The support was copied from 0.9.6c [engine] and adapted/corrected
11729 to work with the new engine framework.
11731 *AEP Inc. and Richard Levitte*
11733 * Support for SureWare crypto accelerator cards from Baltimore
11734 Technologies. (Use engine 'sureware')
11735 The support was copied from 0.9.6c [engine] and adapted
11736 to work with the new engine framework.
11740 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11741 make the newer ENGINE framework commands for the CHIL engine work.
11743 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11745 * Make it possible to produce shared libraries on ReliantUNIX.
11747 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11749 * Add the configuration target debug-linux-ppro.
11750 Make 'openssl rsa' use the general key loading routines
11751 implemented in `apps.c`, and make those routines able to
11752 handle the key format FORMAT_NETSCAPE and the variant
11755 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11757 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11759 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11761 * Add -keyform to rsautl, and document -engine.
11763 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11765 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11766 BIO_R_NO_SUCH_FILE error code rather than the generic
11767 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11771 * Add new functions
11772 ERR_peek_last_error
11773 ERR_peek_last_error_line
11774 ERR_peek_last_error_line_data.
11775 These are similar to
11777 ERR_peek_error_line
11778 ERR_peek_error_line_data,
11779 but report on the latest error recorded rather than the first one
11780 still in the error queue.
11782 *Ben Laurie, Bodo Moeller*
11784 * default_algorithms option in ENGINE config module. This allows things
11786 default_algorithms = ALL
11787 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11791 * Preliminary ENGINE config module.
11795 * New experimental application configuration code.
11799 * Change the AES code to follow the same name structure as all other
11800 symmetric ciphers, and behave the same way. Move everything to
11801 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11803 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11805 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11807 *Ben Laurie and Theo de Raadt*
11809 * Add option to output public keys in req command.
11811 *Massimiliano Pala madwolf@openca.org*
11813 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11814 (up to about 10% better than before for P-192 and P-224).
11818 * New functions/macros
11820 SSL_CTX_set_msg_callback(ctx, cb)
11821 SSL_CTX_set_msg_callback_arg(ctx, arg)
11822 SSL_set_msg_callback(ssl, cb)
11823 SSL_set_msg_callback_arg(ssl, arg)
11825 to request calling a callback function
11827 void cb(int write_p, int version, int content_type,
11828 const void *buf, size_t len, SSL *ssl, void *arg)
11830 whenever a protocol message has been completely received
11831 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11832 protocol version according to which the SSL library interprets
11833 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11834 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11835 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11836 specification (change_cipher_spec(20), alert(21), handshake(22)).
11837 'buf' and 'len' point to the actual message, 'ssl' to the
11838 SSL object, and 'arg' is the application-defined value set by
11839 SSL[_CTX]_set_msg_callback_arg().
11841 'openssl s_client' and 'openssl s_server' have new '-msg' options
11842 to enable a callback that displays all protocol messages.
11846 * Change the shared library support so shared libraries are built as
11847 soon as the corresponding static library is finished, and thereby get
11848 openssl and the test programs linked against the shared library.
11849 This still only happens when the keyword "shard" has been given to
11850 the configuration scripts.
11852 NOTE: shared library support is still an experimental thing, and
11853 backward binary compatibility is still not guaranteed.
11855 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11857 * Add support for Subject Information Access extension.
11859 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11861 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11862 additional bytes when new memory had to be allocated, not just
11863 when reusing an existing buffer.
11867 * New command line and configuration option 'utf8' for the req command.
11868 This allows field values to be specified as UTF8 strings.
11872 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11873 runs for the former and machine-readable output for the latter.
11877 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11878 of the e-mail address in the DN (i.e., it will go into a certificate
11879 extension only). The new configuration file option 'email_in_dn = no'
11880 has the same effect.
11882 *Massimiliano Pala madwolf@openca.org*
11884 * Change all functions with names starting with `des_` to be starting
11885 with `DES_` instead. Add wrappers that are compatible with libdes,
11886 but are named `_ossl_old_des_*`. Finally, add macros that map the
11887 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11888 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11889 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11892 Since we provide two compatibility mappings, the user needs to
11893 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11894 compatibility is desired. The default (i.e., when that macro
11895 isn't defined) is OpenSSL 0.9.6c compatibility.
11897 There are also macros that enable and disable the support of old
11898 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11899 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11900 are defined, the default will apply: to support the old des routines.
11902 In either case, one must include openssl/des.h to get the correct
11903 definitions. Do not try to just include openssl/des_old.h, that
11906 NOTE: This is a major break of an old API into a new one. Software
11907 authors are encouraged to switch to the `DES_` style functions. Some
11908 time in the future, des_old.h and the libdes compatibility functions
11909 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11910 default), and then completely removed.
11914 * Test for certificates which contain unsupported critical extensions.
11915 If such a certificate is found during a verify operation it is
11916 rejected by default: this behaviour can be overridden by either
11917 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11918 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11919 X509_supported_extension() has also been added which returns 1 if a
11920 particular extension is supported.
11924 * Modify the behaviour of EVP cipher functions in similar way to digests
11925 to retain compatibility with existing code.
11929 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11930 compatibility with existing code. In particular the 'ctx' parameter does
11931 not have to be to be initialized before the call to EVP_DigestInit() and
11932 it is tidied up after a call to EVP_DigestFinal(). New function
11933 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11934 EVP_MD_CTX_copy() changed to not require the destination to be
11935 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11936 requires the destination to be valid.
11938 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11939 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11943 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11944 so that complete 'Handshake' protocol structures are kept in memory
11945 instead of overwriting 'msg_type' and 'length' with 'body' data.
11949 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11951 *Massimo Santin via Richard Levitte*
11953 * Major restructuring to the underlying ENGINE code. This includes
11954 reduction of linker bloat, separation of pure "ENGINE" manipulation
11955 (initialisation, etc) from functionality dealing with implementations
11956 of specific crypto interfaces. This change also introduces integrated
11957 support for symmetric ciphers and digest implementations - so ENGINEs
11958 can now accelerate these by providing EVP_CIPHER and EVP_MD
11959 implementations of their own. This is detailed in
11960 [crypto/engine/README.md](crypto/engine/README.md)
11961 as it couldn't be adequately described here. However, there are a few
11962 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11963 were changed in the original introduction of ENGINE code have now
11964 reverted back - the hooking from this code to ENGINE is now a good
11965 deal more passive and at run-time, operations deal directly with
11966 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11967 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11968 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11969 they were not being used by the framework as there is no concept of a
11970 BIGNUM_METHOD and they could not be generalised to the new
11971 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11972 ENGINE_cpy() has been removed as it cannot be consistently defined in
11977 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11981 * Change mkdef.pl to sort symbols that get the same entry number,
11982 and make sure the automatically generated functions `ERR_load_*`
11983 become part of libeay.num as well.
11987 * New function SSL_renegotiate_pending(). This returns true once
11988 renegotiation has been requested (either SSL_renegotiate() call
11989 or HelloRequest/ClientHello received from the peer) and becomes
11990 false once a handshake has been completed.
11991 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11992 sends a HelloRequest, but does not ensure that a handshake takes
11993 place. SSL_renegotiate_pending() is useful for checking if the
11994 client has followed the request.)
11998 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11999 By default, clients may request session resumption even during
12000 renegotiation (if session ID contexts permit); with this option,
12001 session resumption is possible only in the first handshake.
12003 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12004 more bits available for options that should not be part of
12005 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12009 * Add some demos for certificate and certificate request creation.
12013 * Make maximum certificate chain size accepted from the peer application
12014 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12015 "Douglas E. Engert" <deengert@anl.gov>.
12019 * Add support for shared libraries for Unixware-7
12020 (Boyd Lynn Gerber <gerberb@zenez.com>).
12024 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12025 be done prior to destruction. Use this to unload error strings from
12026 ENGINEs that load their own error strings. NB: This adds two new API
12027 functions to "get" and "set" this destroy handler in an ENGINE.
12031 * Alter all existing ENGINE implementations (except "openssl" and
12032 "openbsd") to dynamically instantiate their own error strings. This
12033 makes them more flexible to be built both as statically-linked ENGINEs
12034 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12035 Also, add stub code to each that makes building them as self-contained
12036 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12040 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12041 implementations into applications that are completely implemented in
12042 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12043 commands that can be used to configure what shared-library to load and
12044 to control aspects of the way it is handled. Also, made an update to
12045 the [README-Engine.md](README-Engine.md) file
12046 that brings its information up-to-date and
12047 provides some information and instructions on the "dynamic" ENGINE
12048 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12052 * Make it possible to unload ranges of ERR strings with a new
12053 "ERR_unload_strings" function.
12057 * Add a copy() function to EVP_MD.
12061 * Make EVP_MD routines take a context pointer instead of just the
12062 md_data void pointer.
12066 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12067 that the digest can only process a single chunk of data
12068 (typically because it is provided by a piece of
12069 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12070 is only going to provide a single chunk of data, and hence the
12071 framework needn't accumulate the data for oneshot drivers.
12075 * As with "ERR", make it possible to replace the underlying "ex_data"
12076 functions. This change also alters the storage and management of global
12077 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12078 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12079 index counters. The API functions that use this state have been changed
12080 to take a "class_index" rather than pointers to the class's local STACK
12081 and counter, and there is now an API function to dynamically create new
12082 classes. This centralisation allows us to (a) plug a lot of the
12083 thread-safety problems that existed, and (b) makes it possible to clean
12084 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12085 such data would previously have always leaked in application code and
12086 workarounds were in place to make the memory debugging turn a blind eye
12087 to it. Application code that doesn't use this new function will still
12088 leak as before, but their memory debugging output will announce it now
12089 rather than letting it slide.
12091 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12092 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12093 has a return value to indicate success or failure.
12097 * Make it possible to replace the underlying "ERR" functions such that the
12098 global state (2 LHASH tables and 2 locks) is only used by the "default"
12099 implementation. This change also adds two functions to "get" and "set"
12100 the implementation prior to it being automatically set the first time
12101 any other ERR function takes place. Ie. an application can call "get",
12102 pass the return value to a module it has just loaded, and that module
12103 can call its own "set" function using that value. This means the
12104 module's "ERR" operations will use (and modify) the error state in the
12105 application and not in its own statically linked copy of OpenSSL code.
12109 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12110 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12111 the operation, and provides a more encapsulated way for external code
12112 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12113 to use these functions rather than manually incrementing the counts.
12115 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12119 * Add EVP test program.
12123 * Add symmetric cipher support to ENGINE. Expect the API to change!
12127 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12128 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12129 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12130 These allow a CRL to be built without having to access X509_CRL fields
12131 directly. Modify 'ca' application to use new functions.
12135 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12136 bug workarounds. Rollback attack detection is a security feature.
12137 The problem will only arise on OpenSSL servers when TLSv1 is not
12138 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12139 Software authors not wanting to support TLSv1 will have special reasons
12140 for their choice and can explicitly enable this option.
12142 *Bodo Moeller, Lutz Jaenicke*
12144 * Rationalise EVP so it can be extended: don't include a union of
12145 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12146 (similar to those existing for EVP_CIPHER_CTX).
12151 EVP_MD_CTX_init(&md); /* new function call */
12152 EVP_DigestInit(&md, EVP_sha1());
12153 EVP_DigestUpdate(&md, in, len);
12154 EVP_DigestFinal(&md, out, NULL);
12155 EVP_MD_CTX_cleanup(&md); /* new function call */
12159 * Make DES key schedule conform to the usual scheme, as well as
12160 correcting its structure. This means that calls to DES functions
12161 now have to pass a pointer to a des_key_schedule instead of a
12162 plain des_key_schedule (which was actually always a pointer
12165 des_key_schedule ks;
12167 des_set_key_checked(..., &ks);
12168 des_ncbc_encrypt(..., &ks, ...);
12170 (Note that a later change renames 'des_...' into 'DES_...'.)
12174 * Initial reduction of linker bloat: the use of some functions, such as
12175 PEM causes large amounts of unused functions to be linked in due to
12176 poor organisation. For example pem_all.c contains every PEM function
12177 which has a knock on effect of linking in large amounts of (unused)
12178 ASN1 code. Grouping together similar functions and splitting unrelated
12179 functions prevents this.
12183 * Cleanup of EVP macros.
12187 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12188 correct `_ecb suffix`.
12192 * Add initial OCSP responder support to ocsp application. The
12193 revocation information is handled using the text based index
12194 use by the ca application. The responder can either handle
12195 requests generated internally, supplied in files (for example
12196 via a CGI script) or using an internal minimal server.
12200 * Add configuration choices to get zlib compression for TLS.
12204 * Changes to Kerberos SSL for RFC 2712 compliance:
12205 1. Implemented real KerberosWrapper, instead of just using
12206 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12207 2. Implemented optional authenticator field of KerberosWrapper.
12209 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12210 and authenticator structs; see crypto/krb5/.
12212 Generalized Kerberos calls to support multiple Kerberos libraries.
12213 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12214 via Richard Levitte*
12216 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12217 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12218 values for each of the key sizes rather than having just
12219 parameters (and 'speed' generating keys each time).
12223 * Speed up EVP routines.
12226 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12227 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12228 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12229 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12231 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12232 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12233 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12236 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12238 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12242 * Added the OS2-EMX target.
12244 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12246 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12247 New functions to support `NCONF `routines in extension code.
12248 New function `CONF_set_nconf()`
12249 to allow functions which take an `NCONF` to also handle the old `LHASH`
12250 structure: this means that the old `CONF` compatible routines can be
12251 retained (in particular w.rt. extensions) without having to duplicate the
12252 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12256 * Enhance the general user interface with mechanisms for inner control
12257 and with possibilities to have yes/no kind of prompts.
12261 * Change all calls to low-level digest routines in the library and
12262 applications to use EVP. Add missing calls to HMAC_cleanup() and
12263 don't assume HMAC_CTX can be copied using memcpy().
12265 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12267 * Add the possibility to control engines through control names but with
12268 arbitrary arguments instead of just a string.
12269 Change the key loaders to take a UI_METHOD instead of a callback
12270 function pointer. NOTE: this breaks binary compatibility with earlier
12271 versions of OpenSSL [engine].
12272 Adapt the nCipher code for these new conditions and add a card insertion
12277 * Enhance the general user interface with mechanisms to better support
12278 dialog box interfaces, application-defined prompts, the possibility
12279 to use defaults (for example default passwords from somewhere else)
12280 and interrupts/cancellations.
12284 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12285 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12289 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12290 tidy up some unnecessarily weird code in 'sk_new()').
12292 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12294 * Change the key loading routines for ENGINEs to use the same kind
12295 callback (pem_password_cb) as all other routines that need this
12300 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12301 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12302 than this minimum value is recommended.
12306 * New random seeder for OpenVMS, using the system process statistics
12307 that are easily reachable.
12311 * Windows apparently can't transparently handle global
12312 variables defined in DLLs. Initialisations such as:
12314 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12316 won't compile. This is used by the any applications that need to
12317 declare their own ASN1 modules. This was fixed by adding the option
12318 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12319 needed for static libraries under Win32.
12323 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12324 setting of purpose and trust fields. New X509_STORE trust and
12325 purpose functions and tidy up setting in other SSL functions.
12329 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12330 structure. These are inherited by X509_STORE_CTX when it is
12331 initialised. This allows various defaults to be set in the
12332 X509_STORE structure (such as flags for CRL checking and custom
12333 purpose or trust settings) for functions which only use X509_STORE_CTX
12334 internally such as S/MIME.
12336 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12337 trust settings if they are not set in X509_STORE. This allows X509_STORE
12338 purposes and trust (in S/MIME for example) to override any set by default.
12340 Add command line options for CRL checking to smime, s_client and s_server
12345 * Initial CRL based revocation checking. If the CRL checking flag(s)
12346 are set then the CRL is looked up in the X509_STORE structure and
12347 its validity and signature checked, then if the certificate is found
12348 in the CRL the verify fails with a revoked error.
12350 Various new CRL related callbacks added to X509_STORE_CTX structure.
12352 Command line options added to 'verify' application to support this.
12354 This needs some additional work, such as being able to handle multiple
12355 CRLs with different times, extension based lookup (rather than just
12356 by subject name) and ultimately more complete V2 CRL extension
12361 * Add a general user interface API (crypto/ui/). This is designed
12362 to replace things like des_read_password and friends (backward
12363 compatibility functions using this new API are provided).
12364 The purpose is to remove prompting functions from the DES code
12365 section as well as provide for prompting through dialog boxes in
12366 a window system and the like.
12370 * Add "ex_data" support to ENGINE so implementations can add state at a
12371 per-structure level rather than having to store it globally.
12375 * Make it possible for ENGINE structures to be copied when retrieved by
12376 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12377 This causes the "original" ENGINE structure to act like a template,
12378 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12379 operational state can be localised to each ENGINE structure, despite the
12380 fact they all share the same "methods". New ENGINE structures returned in
12381 this case have no functional references and the return value is the single
12382 structural reference. This matches the single structural reference returned
12383 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12388 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12389 needs to match any other type at all we need to manually clear the
12394 * Changes to the "openssl engine" utility to include;
12395 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12396 about an ENGINE's available control commands.
12397 - executing control commands from command line arguments using the
12398 '-pre' and '-post' switches. '-post' is only used if '-t' is
12399 specified and the ENGINE is successfully initialised. The syntax for
12400 the individual commands are colon-separated, for example;
12401 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12405 * New dynamic control command support for ENGINEs. ENGINEs can now
12406 declare their own commands (numbers), names (strings), descriptions,
12407 and input types for run-time discovery by calling applications. A
12408 subset of these commands are implicitly classed as "executable"
12409 depending on their input type, and only these can be invoked through
12410 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12411 can be based on user input, config files, etc). The distinction is
12412 that "executable" commands cannot return anything other than a boolean
12413 result and can only support numeric or string input, whereas some
12414 discoverable commands may only be for direct use through
12415 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12416 pointers, or other custom uses. The "executable" commands are to
12417 support parameterisations of ENGINE behaviour that can be
12418 unambiguously defined by ENGINEs and used consistently across any
12419 OpenSSL-based application. Commands have been added to all the
12420 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12421 control over shared-library paths without source code alterations.
12425 * Changed all ENGINE implementations to dynamically allocate their
12426 ENGINEs rather than declaring them statically. Apart from this being
12427 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12428 this also allows the implementations to compile without using the
12429 internal engine_int.h header.
12433 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12434 'const' value. Any code that should be able to modify a RAND_METHOD
12435 should already have non-const pointers to it (ie. they should only
12436 modify their own ones).
12440 * Made a variety of little tweaks to the ENGINE code.
12441 - "atalla" and "ubsec" string definitions were moved from header files
12442 to C code. "nuron" string definitions were placed in variables
12443 rather than hard-coded - allowing parameterisation of these values
12444 later on via ctrl() commands.
12445 - Removed unused "#if 0"'d code.
12446 - Fixed engine list iteration code so it uses ENGINE_free() to release
12447 structural references.
12448 - Constified the RAND_METHOD element of ENGINE structures.
12449 - Constified various get/set functions as appropriate and added
12450 missing functions (including a catch-all ENGINE_cpy that duplicates
12451 all ENGINE values onto a new ENGINE except reference counts/state).
12452 - Removed NULL parameter checks in get/set functions. Setting a method
12453 or function to NULL is a way of cancelling out a previously set
12454 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12455 and doesn't justify the extra error symbols and code.
12456 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12457 flags from engine_int.h to engine.h.
12458 - Changed prototypes for ENGINE handler functions (init(), finish(),
12459 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12463 * Implement binary inversion algorithm for BN_mod_inverse in addition
12464 to the algorithm using long division. The binary algorithm can be
12465 used only if the modulus is odd. On 32-bit systems, it is faster
12466 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12467 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12468 up to 450 bits. In 64-bit environments, the binary algorithm
12469 appears to be advantageous for much longer moduli; here we use it
12470 for moduli up to 2048 bits.
12474 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12475 could not support the combine flag in choice fields.
12479 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12480 extensions from a certificate request to the certificate.
12484 * Allow multiple 'certopt' and 'nameopt' options to be separated
12485 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12486 file: this allows the display of the certificate about to be
12487 signed to be customised, to allow certain fields to be included
12488 or excluded and extension details. The old system didn't display
12489 multicharacter strings properly, omitted fields not in the policy
12490 and couldn't display additional details such as extensions.
12494 * Function EC_POINTs_mul for multiple scalar multiplication
12495 of an arbitrary number of elliptic curve points
12496 \sum scalars[i]*points[i],
12497 optionally including the generator defined for the EC_GROUP:
12498 scalar*generator + \sum scalars[i]*points[i].
12500 EC_POINT_mul is a simple wrapper function for the typical case
12501 that the point list has just one item (besides the optional
12506 * First EC_METHODs for curves over GF(p):
12508 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12509 operations and provides various method functions that can also
12510 operate with faster implementations of modular arithmetic.
12512 EC_GFp_mont_method() reuses most functions that are part of
12513 EC_GFp_simple_method, but uses Montgomery arithmetic.
12515 *Bodo Moeller; point addition and point doubling
12516 implementation directly derived from source code provided by
12517 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12519 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12520 crypto/ec/ec_lib.c):
12522 Curves are EC_GROUP objects (with an optional group generator)
12523 based on EC_METHODs that are built into the library.
12525 Points are EC_POINT objects based on EC_GROUP objects.
12527 Most of the framework would be able to handle curves over arbitrary
12528 finite fields, but as there are no obvious types for fields other
12529 than GF(p), some functions are limited to that for now.
12533 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12534 that the file contains a complete HTTP response.
12538 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12539 change the def and num file printf format specifier from "%-40sXXX"
12540 to "%-39s XXX". The latter will always guarantee a space after the
12541 field while the former will cause them to run together if the field
12542 is 40 of more characters long.
12546 * Constify the cipher and digest 'method' functions and structures
12547 and modify related functions to take constant EVP_MD and EVP_CIPHER
12552 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12553 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12557 * Modify `EVP_Digest*()` routines so they now return values. Although the
12558 internal software routines can never fail additional hardware versions
12563 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12565 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12566 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12569 ERR_R_NESTED_ASN1_ERROR
12571 ERR_R_MISSING_ASN1_EOS
12572 were 4 .. 9, conflicting with
12573 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12575 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12576 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12578 Add new error code 'ERR_R_INTERNAL_ERROR'.
12582 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12587 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12588 sets the subject name for a new request or supersedes the
12589 subject name in a given request. Formats that can be parsed are
12590 'CN=Some Name, OU=myOU, C=IT'
12592 'CN=Some Name/OU=myOU/C=IT'.
12594 Add options '-batch' and '-verbose' to 'openssl req'.
12596 *Massimiliano Pala <madwolf@hackmasters.net>*
12598 * Introduce the possibility to access global variables through
12599 functions on platform were that's the best way to handle exporting
12600 global variables in shared libraries. To enable this functionality,
12601 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12602 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12603 is normally done by Configure or something similar).
12605 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12606 in the source file (foo.c) like this:
12608 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12609 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12611 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12612 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12614 OPENSSL_DECLARE_GLOBAL(int,foo);
12615 #define foo OPENSSL_GLOBAL_REF(foo)
12616 OPENSSL_DECLARE_GLOBAL(double,bar);
12617 #define bar OPENSSL_GLOBAL_REF(bar)
12619 The #defines are very important, and therefore so is including the
12620 header file everywhere where the defined globals are used.
12622 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12623 of ASN.1 items, but that structure is a bit different.
12625 The largest change is in util/mkdef.pl which has been enhanced with
12626 better and easier to understand logic to choose which symbols should
12627 go into the Windows .def files as well as a number of fixes and code
12628 cleanup (among others, algorithm keywords are now sorted
12629 lexicographically to avoid constant rewrites).
12633 * In BN_div() keep a copy of the sign of 'num' before writing the
12634 result to 'rm' because if rm==num the value will be overwritten
12635 and produce the wrong result if 'num' is negative: this caused
12636 problems with BN_mod() and BN_nnmod().
12640 * Function OCSP_request_verify(). This checks the signature on an
12641 OCSP request and verifies the signer certificate. The signer
12642 certificate is just checked for a generic purpose and OCSP request
12647 * Add OCSP_check_validity() function to check the validity of OCSP
12648 responses. OCSP responses are prepared in real time and may only
12649 be a few seconds old. Simply checking that the current time lies
12650 between thisUpdate and nextUpdate max reject otherwise valid responses
12651 caused by either OCSP responder or client clock inaccuracy. Instead
12652 we allow thisUpdate and nextUpdate to fall within a certain period of
12653 the current time. The age of the response can also optionally be
12654 checked. Two new options -validity_period and -status_age added to
12659 * If signature or public key algorithm is unrecognized print out its
12660 OID rather that just UNKNOWN.
12664 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12665 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12666 ID to be generated from the issuer certificate alone which can then be
12667 passed to OCSP_id_issuer_cmp().
12671 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12672 ASN1 modules to export functions returning ASN1_ITEM pointers
12673 instead of the ASN1_ITEM structures themselves. This adds several
12674 new macros which allow the underlying ASN1 function/structure to
12675 be accessed transparently. As a result code should not use ASN1_ITEM
12676 references directly (such as &X509_it) but instead use the relevant
12677 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12678 use of the new ASN1 code on platforms where exporting structures
12679 is problematical (for example in shared libraries) but exporting
12680 functions returning pointers to structures is not.
12684 * Add support for overriding the generation of SSL/TLS session IDs.
12685 These callbacks can be registered either in an SSL_CTX or per SSL.
12686 The purpose of this is to allow applications to control, if they wish,
12687 the arbitrary values chosen for use as session IDs, particularly as it
12688 can be useful for session caching in multiple-server environments. A
12689 command-line switch for testing this (and any client code that wishes
12690 to use such a feature) has been added to "s_server".
12692 *Geoff Thorpe, Lutz Jaenicke*
12694 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12695 of the form `#if defined(...) || defined(...) || ...` and
12696 `#if !defined(...) && !defined(...) && ...`. This also avoids
12697 the growing number of special cases it was previously handling.
12701 * Make all configuration macros available for application by making
12702 sure they are available in opensslconf.h, by giving them names starting
12703 with `OPENSSL_` to avoid conflicts with other packages and by making
12704 sure e_os2.h will cover all platform-specific cases together with
12706 Additionally, it is now possible to define configuration/platform-
12707 specific names (called "system identities"). In the C code, these
12708 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12709 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12710 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12715 * New option -set_serial to 'req' and 'x509' this allows the serial
12716 number to use to be specified on the command line. Previously self
12717 signed certificates were hard coded with serial number 0 and the
12718 CA options of 'x509' had to use a serial number in a file which was
12723 * New options to 'ca' utility to support V2 CRL entry extensions.
12724 Currently CRL reason, invalidity date and hold instruction are
12725 supported. Add new CRL extensions to V3 code and some new objects.
12729 * New function EVP_CIPHER_CTX_set_padding() this is used to
12730 disable standard block padding (aka PKCS#5 padding) in the EVP
12731 API, which was previously mandatory. This means that the data is
12732 not padded in any way and so the total length much be a multiple
12733 of the block size, otherwise an error occurs.
12737 * Initial (incomplete) OCSP SSL support.
12741 * New function OCSP_parse_url(). This splits up a URL into its host,
12742 port and path components: primarily to parse OCSP URLs. New -url
12743 option to ocsp utility.
12747 * New nonce behavior. The return value of OCSP_check_nonce() now
12748 reflects the various checks performed. Applications can decide
12749 whether to tolerate certain situations such as an absent nonce
12750 in a response when one was present in a request: the ocsp application
12751 just prints out a warning. New function OCSP_add1_basic_nonce()
12752 this is to allow responders to include a nonce in a response even if
12753 the request is nonce-less.
12757 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12758 skipped when using openssl x509 multiple times on a single input file,
12759 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12763 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12764 set string type: to handle setting ASN1_TIME structures. Fix ca
12765 utility to correctly initialize revocation date of CRLs.
12769 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12770 the clients preferred ciphersuites and rather use its own preferences.
12771 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12772 Internet Explorer by ensuring unchanged hash method during stepup.
12773 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12777 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12778 to aes and add a new 'exist' option to print out symbols that don't
12783 * Additional options to ocsp utility to allow flags to be set and
12784 additional certificates supplied.
12788 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12789 OCSP client a number of certificate to only verify the response
12794 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12795 handle the new API. Currently only ECB, CBC modes supported. Add new
12798 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12799 Encryption Standard (AES) Ciphersuites for Transport Layer
12800 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12801 not enabled by default and were not part of the "ALL" ciphersuite
12802 alias because they were not yet official; they could be
12803 explicitly requested by specifying the "AESdraft" ciphersuite
12804 group alias. In the final release of OpenSSL 0.9.7, the group
12805 alias is called "AES" and is part of "ALL".)
12807 *Ben Laurie, Steve Henson, Bodo Moeller*
12809 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12810 request to response.
12814 * Functions for OCSP responders. OCSP_request_onereq_count(),
12815 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12816 extract information from a certificate request. OCSP_response_create()
12817 creates a response and optionally adds a basic response structure.
12818 OCSP_basic_add1_status() adds a complete single response to a basic
12819 response and returns the OCSP_SINGLERESP structure just added (to allow
12820 extensions to be included for example). OCSP_basic_add1_cert() adds a
12821 certificate to a basic response and OCSP_basic_sign() signs a basic
12822 response with various flags. New helper functions ASN1_TIME_check()
12823 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12824 (converts ASN1_TIME to GeneralizedTime).
12828 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12829 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12830 structure from a certificate. X509_pubkey_digest() digests the public_key
12831 contents: this is used in various key identifiers.
12835 * Make sk_sort() tolerate a NULL argument.
12837 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12839 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12840 passed by the function are trusted implicitly. If any of them signed the
12841 response then it is assumed to be valid and is not verified.
12845 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12846 to data. This was previously part of the PKCS7 ASN1 code. This
12847 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12848 *Steve Henson, reported by Kenneth R. Robinette
12849 <support@securenetterm.com>*
12851 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12852 routines: without these tracing memory leaks is very painful.
12853 Fix leaks in PKCS12 and PKCS7 routines.
12857 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12858 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12859 effectively meant GeneralizedTime would never be used. Now it
12860 is initialised to -1 but X509_time_adj() now has to check the value
12861 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12862 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12863 *Steve Henson, reported by Kenneth R. Robinette
12864 <support@securenetterm.com>*
12866 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12867 result in a zero length in the ASN1_INTEGER structure which was
12868 not consistent with the structure when d2i_ASN1_INTEGER() was used
12869 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12870 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12871 where it did not print out a minus for negative ASN1_INTEGER.
12875 * Add summary printout to ocsp utility. The various functions which
12876 convert status values to strings have been renamed to:
12877 OCSP_response_status_str(), OCSP_cert_status_str() and
12878 OCSP_crl_reason_str() and are no longer static. New options
12879 to verify nonce values and to disable verification. OCSP response
12880 printout format cleaned up.
12884 * Add additional OCSP certificate checks. These are those specified
12885 in RFC2560. This consists of two separate checks: the CA of the
12886 certificate being checked must either be the OCSP signer certificate
12887 or the issuer of the OCSP signer certificate. In the latter case the
12888 OCSP signer certificate must contain the OCSP signing extended key
12889 usage. This check is performed by attempting to match the OCSP
12890 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12891 in the OCSP_CERTID structures of the response.
12895 * Initial OCSP certificate verification added to OCSP_basic_verify()
12896 and related routines. This uses the standard OpenSSL certificate
12897 verify routines to perform initial checks (just CA validity) and
12898 to obtain the certificate chain. Then additional checks will be
12899 performed on the chain. Currently the root CA is checked to see
12900 if it is explicitly trusted for OCSP signing. This is used to set
12901 a root CA as a global signing root: that is any certificate that
12902 chains to that CA is an acceptable OCSP signing certificate.
12906 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12907 extensions from a separate configuration file.
12908 As when reading extensions from the main configuration file,
12909 the '-extensions ...' option may be used for specifying the
12912 *Massimiliano Pala <madwolf@comune.modena.it>*
12914 * New OCSP utility. Allows OCSP requests to be generated or
12915 read. The request can be sent to a responder and the output
12916 parsed, outputted or printed in text form. Not complete yet:
12917 still needs to check the OCSP response validity.
12921 * New subcommands for 'openssl ca':
12922 `openssl ca -status <serial>` prints the status of the cert with
12923 the given serial number (according to the index file).
12924 `openssl ca -updatedb` updates the expiry status of certificates
12927 *Massimiliano Pala <madwolf@comune.modena.it>*
12929 * New '-newreq-nodes' command option to CA.pl. This is like
12930 '-newreq', but calls 'openssl req' with the '-nodes' option
12931 so that the resulting key is not encrypted.
12933 *Damien Miller <djm@mindrot.org>*
12935 * New configuration for the GNU Hurd.
12937 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12939 * Initial code to implement OCSP basic response verify. This
12940 is currently incomplete. Currently just finds the signer's
12941 certificate and verifies the signature on the response.
12945 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12946 value of OPENSSLDIR. This is available via the new '-d' option
12947 to 'openssl version', and is also included in 'openssl version -a'.
12951 * Allowing defining memory allocation callbacks that will be given
12952 file name and line number information in additional arguments
12953 (a `const char*` and an int). The basic functionality remains, as
12954 well as the original possibility to just replace malloc(),
12955 realloc() and free() by functions that do not know about these
12956 additional arguments. To register and find out the current
12957 settings for extended allocation functions, the following
12958 functions are provided:
12960 CRYPTO_set_mem_ex_functions
12961 CRYPTO_set_locked_mem_ex_functions
12962 CRYPTO_get_mem_ex_functions
12963 CRYPTO_get_locked_mem_ex_functions
12965 These work the same way as CRYPTO_set_mem_functions and friends.
12966 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12967 extended allocation function is enabled.
12968 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12969 a conventional allocation function is enabled.
12971 *Richard Levitte, Bodo Moeller*
12973 * Finish off removing the remaining LHASH function pointer casts.
12974 There should no longer be any prototype-casting required when using
12975 the LHASH abstraction, and any casts that remain are "bugs". See
12976 the callback types and macros at the head of lhash.h for details
12977 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12981 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12982 If /dev/[u]random devices are not available or do not return enough
12983 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12985 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12986 /etc/entropy will be queried once each in this sequence, querying stops
12987 when enough entropy was collected without querying more sockets.
12991 * Change the Unix RAND_poll() variant to be able to poll several
12992 random devices, as specified by DEVRANDOM, until a sufficient amount
12993 of data has been collected. We spend at most 10 ms on each file
12994 (select timeout) and read in non-blocking mode. DEVRANDOM now
12995 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12996 (previously it was just the string "/dev/urandom"), so on typical
12997 platforms the 10 ms delay will never occur.
12998 Also separate out the Unix variant to its own file, rand_unix.c.
12999 For VMS, there's a currently-empty rand_vms.c.
13003 * Move OCSP client related routines to ocsp_cl.c. These
13004 provide utility functions which an application needing
13005 to issue a request to an OCSP responder and analyse the
13006 response will typically need: as opposed to those which an
13007 OCSP responder itself would need which will be added later.
13009 OCSP_request_sign() signs an OCSP request with an API similar
13010 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13011 response. OCSP_response_get1_basic() extracts basic response
13012 from response. OCSP_resp_find_status(): finds and extracts status
13013 information from an OCSP_CERTID structure (which will be created
13014 when the request structure is built). These are built from lower
13015 level functions which work on OCSP_SINGLERESP structures but
13016 won't normally be used unless the application wishes to examine
13017 extensions in the OCSP response for example.
13019 Replace nonce routines with a pair of functions.
13020 OCSP_request_add1_nonce() adds a nonce value and optionally
13021 generates a random value. OCSP_check_nonce() checks the
13022 validity of the nonce in an OCSP response.
13026 * Change function OCSP_request_add() to OCSP_request_add0_id().
13027 This doesn't copy the supplied OCSP_CERTID and avoids the
13028 need to free up the newly created id. Change return type
13029 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13030 This can then be used to add extensions to the request.
13031 Deleted OCSP_request_new(), since most of its functionality
13032 is now in OCSP_REQUEST_new() (and the case insensitive name
13033 clash) apart from the ability to set the request name which
13034 will be added elsewhere.
13038 * Update OCSP API. Remove obsolete extensions argument from
13039 various functions. Extensions are now handled using the new
13040 OCSP extension code. New simple OCSP HTTP function which
13041 can be used to send requests and parse the response.
13045 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13046 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13047 uses the special reorder version of SET OF to sort the attributes
13048 and reorder them to match the encoded order. This resolves a long
13049 standing problem: a verify on a PKCS7 structure just after signing
13050 it used to fail because the attribute order did not match the
13051 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13052 it uses the received order. This is necessary to tolerate some broken
13053 software that does not order SET OF. This is handled by encoding
13054 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13055 to produce the required SET OF.
13059 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13060 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13061 files to get correct declarations of the ASN.1 item variables.
13065 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13066 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13067 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13068 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13069 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13070 ASN1_ITEM and no wrapper functions.
13074 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13075 replace the old function pointer based I/O routines. Change most of
13076 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13080 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13081 lines, recognize more "algorithms" that can be deselected, and make
13082 it complain about algorithm deselection that isn't recognised.
13086 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13087 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13088 to use new functions. Add NO_ASN1_OLD which can be set to remove
13089 some old style ASN1 functions: this can be used to determine if old
13090 code will still work when these eventually go away.
13094 * New extension functions for OCSP structures, these follow the
13095 same conventions as certificates and CRLs.
13099 * New function X509V3_add1_i2d(). This automatically encodes and
13100 adds an extension. Its behaviour can be customised with various
13101 flags to append, replace or delete. Various wrappers added for
13102 certificates and CRLs.
13106 * Fix to avoid calling the underlying ASN1 print routine when
13107 an extension cannot be parsed. Correct a typo in the
13108 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13112 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13113 entries for variables.
13117 * Add functionality to `apps/openssl.c` for detecting locking
13118 problems: As the program is single-threaded, all we have
13119 to do is register a locking callback using an array for
13120 storing which locks are currently held by the program.
13124 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13125 SSL_get_ex_data_X509_STORE_idx(), which is used in
13126 ssl_verify_cert_chain() and thus can be called at any time
13127 during TLS/SSL handshakes so that thread-safety is essential.
13128 Unfortunately, the ex_data design is not at all suited
13129 for multi-threaded use, so it probably should be abolished.
13133 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13135 *Broadcom, tweaked and integrated by Geoff Thorpe*
13137 * Move common extension printing code to new function
13138 X509V3_print_extensions(). Reorganise OCSP print routines and
13139 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13143 * New function X509_signature_print() to remove duplication in some
13148 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13149 set (this was treated exactly the same as SET OF previously). This
13150 is used to reorder the STACK representing the structure to match the
13151 encoding. This will be used to get round a problem where a PKCS7
13152 structure which was signed could not be verified because the STACK
13153 order did not reflect the encoded order.
13157 * Reimplement the OCSP ASN1 module using the new code.
13161 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13162 for its ASN1 operations. The old style function pointers still exist
13163 for now but they will eventually go away.
13167 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13168 completely replaces the old ASN1 functionality with a table driven
13169 encoder and decoder which interprets an ASN1_ITEM structure describing
13170 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13171 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13172 has also been converted to the new form.
13176 * Change BN_mod_exp_recp so that negative moduli are tolerated
13177 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13178 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13179 for negative moduli.
13183 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13184 of not touching the result's sign bit.
13188 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13193 * Changed the LHASH code to use prototypes for callbacks, and created
13194 macros to declare and implement thin (optionally static) functions
13195 that provide type-safety and avoid function pointer casting for the
13196 type-specific callbacks.
13200 * Added Kerberos Cipher Suites to be used with TLS, as written in
13202 *Veers Staats <staatsvr@asc.hpc.mil>,
13203 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13205 * Reformat the FAQ so the different questions and answers can be divided
13206 in sections depending on the subject.
13210 * Have the zlib compression code load ZLIB.DLL dynamically under
13215 * New function BN_mod_sqrt for computing square roots modulo a prime
13216 (using the probabilistic Tonelli-Shanks algorithm unless
13217 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13218 be handled deterministically).
13220 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13222 * Make BN_mod_inverse faster by explicitly handling small quotients
13223 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13224 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13228 * New function BN_kronecker.
13232 * Fix BN_gcd so that it works on negative inputs; the result is
13233 positive unless both parameters are zero.
13234 Previously something reasonably close to an infinite loop was
13235 possible because numbers could be growing instead of shrinking
13236 in the implementation of Euclid's algorithm.
13240 * Fix BN_is_word() and BN_is_one() macros to take into account the
13241 sign of the number in question.
13243 Fix BN_is_word(a,w) to work correctly for w == 0.
13245 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13246 because its test if the absolute value of 'a' equals 'w'.
13247 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13248 it exists mostly for use in the implementations of BN_is_zero(),
13249 BN_is_one(), and BN_is_word().
13253 * New function BN_swap.
13257 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13258 the exponentiation functions are more likely to produce reasonable
13259 results on negative inputs.
13263 * Change BN_mod_mul so that the result is always non-negative.
13264 Previously, it could be negative if one of the factors was negative;
13265 I don't think anyone really wanted that behaviour.
13269 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13270 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13271 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13272 and add new functions:
13281 BN_mod_lshift1_quick
13283 BN_mod_lshift_quick
13285 These functions always generate non-negative results.
13287 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13288 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13290 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13291 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13292 be reduced modulo `m`.
13294 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13297 The following entry accidentally appeared in the CHANGES file
13298 distributed with OpenSSL 0.9.7. The modifications described in
13299 it do *not* apply to OpenSSL 0.9.7.
13301 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13302 was actually never needed) and in BN_mul(). The removal in BN_mul()
13303 required a small change in bn_mul_part_recursive() and the addition
13304 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13305 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13306 bn_sub_words() and bn_add_words() except they take arrays with
13312 * In 'openssl passwd', verify passwords read from the terminal
13313 unless the '-salt' option is used (which usually means that
13314 verification would just waste user's time since the resulting
13315 hash is going to be compared with some given password hash)
13316 or the new '-noverify' option is used.
13318 This is an incompatible change, but it does not affect
13319 non-interactive use of 'openssl passwd' (passwords on the command
13320 line, '-stdin' option, '-in ...' option) and thus should not
13321 cause any problems.
13325 * Remove all references to RSAref, since there's no more need for it.
13329 * Make DSO load along a path given through an environment variable
13330 (SHLIB_PATH) with shl_load().
13334 * Constify the ENGINE code as a result of BIGNUM constification.
13335 Also constify the RSA code and most things related to it. In a
13336 few places, most notable in the depth of the ASN.1 code, ugly
13337 casts back to non-const were required (to be solved at a later
13342 * Make it so the openssl application has all engines loaded by default.
13346 * Constify the BIGNUM routines a little more.
13350 * Add the following functions:
13352 ENGINE_load_cswift()
13354 ENGINE_load_atalla()
13355 ENGINE_load_nuron()
13356 ENGINE_load_builtin_engines()
13358 That way, an application can itself choose if external engines that
13359 are built-in in OpenSSL shall ever be used or not. The benefit is
13360 that applications won't have to be linked with libdl or other dso
13361 libraries unless it's really needed.
13363 Changed 'openssl engine' to load all engines on demand.
13364 Changed the engine header files to avoid the duplication of some
13365 declarations (they differed!).
13369 * 'openssl engine' can now list capabilities.
13373 * Better error reporting in 'openssl engine'.
13377 * Never call load_dh_param(NULL) in s_server.
13381 * Add engine application. It can currently list engines by name and
13382 identity, and test if they are actually available.
13386 * Improve RPM specification file by forcing symbolic linking and making
13387 sure the installed documentation is also owned by root.root.
13389 *Damien Miller <djm@mindrot.org>*
13391 * Give the OpenSSL applications more possibilities to make use of
13392 keys (public as well as private) handled by engines.
13396 * Add OCSP code that comes from CertCo.
13400 * Add VMS support for the Rijndael code.
13404 * Added untested support for Nuron crypto accelerator.
13408 * Add support for external cryptographic devices. This code was
13409 previously distributed separately as the "engine" branch.
13411 *Geoff Thorpe, Richard Levitte*
13413 * Rework the filename-translation in the DSO code. It is now possible to
13414 have far greater control over how a "name" is turned into a filename
13415 depending on the operating environment and any oddities about the
13416 different shared library filenames on each system.
13420 * Support threads on FreeBSD-elf in Configure.
13424 * Fix for SHA1 assembly problem with MASM: it produces
13425 warnings about corrupt line number information when assembling
13426 with debugging information. This is caused by the overlapping
13429 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13432 NCONF_get_number() has no error checking at all. As a replacement,
13433 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13434 promoted strongly. The old NCONF_get_number is kept around for
13435 binary backward compatibility.
13436 Make it possible for methods to load from something other than a BIO,
13437 by providing a function pointer that is given a name instead of a BIO.
13438 For example, this could be used to load configuration data from an
13443 * Fix for non blocking accept BIOs. Added new I/O special reason
13444 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13445 with non blocking I/O was not possible because no retry code was
13446 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13451 * Added the beginnings of Rijndael support.
13455 * Fix for bug in DirectoryString mask setting. Add support for
13456 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13457 to allow certificate printing to more controllable, additional
13458 'certopt' option to 'x509' to allow new printing options to be
13463 * Clean old EAY MD5 hack from e_os.h.
13467 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13469 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13470 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13472 *Joe Orton, Steve Henson*
13474 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13476 * Fix additional bug revealed by the NISCC test suite:
13478 Stop bug triggering large recursion when presented with
13479 certain ASN.1 tags ([CVE-2003-0851])
13483 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13485 * Fix various bugs revealed by running the NISCC test suite:
13487 Stop out of bounds reads in the ASN1 code when presented with
13488 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13490 If verify callback ignores invalid public key errors don't try to check
13491 certificate signature with the NULL public key.
13495 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13496 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13501 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13502 extra data after the compression methods not only for TLS 1.0
13503 but also for SSL 3.0 (as required by the specification).
13505 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13507 * Change X509_certificate_type() to mark the key as exported/exportable
13508 when it's 512 *bits* long, not 512 bytes.
13512 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13514 * Countermeasure against the Klima-Pokorny-Rosa extension of
13515 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13516 a protocol version number mismatch like a decryption error
13517 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13521 * Turn on RSA blinding by default in the default implementation
13522 to avoid a timing attack. Applications that don't want it can call
13523 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13524 They would be ill-advised to do so in most cases.
13526 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13528 * Change RSA blinding code so that it works when the PRNG is not
13529 seeded (in this case, the secret RSA exponent is abused as
13530 an unpredictable seed -- if it is not unpredictable, there
13531 is no point in blinding anyway). Make RSA blinding thread-safe
13532 by remembering the creator's thread ID in rsa->blinding and
13533 having all other threads use local one-time blinding factors
13534 (this requires more computation than sharing rsa->blinding, but
13535 avoids excessive locking; and if an RSA object is not shared
13536 between threads, blinding will still be very fast).
13540 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13542 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13543 via timing by performing a MAC computation even if incorrect
13544 block cipher padding has been found. This is a countermeasure
13545 against active attacks where the attacker has to distinguish
13546 between bad padding and a MAC verification error. ([CVE-2003-0078])
13548 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13549 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13550 Martin Vuagnoux (EPFL, Ilion)*
13552 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13554 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13555 memory from its contents. This is done with a counter that will
13556 place alternating values in each byte. This can be used to solve
13557 two issues: 1) the removal of calls to memset() by highly optimizing
13558 compilers, and 2) cleansing with other values than 0, since those can
13559 be read through on certain media, for example a swap space on disk.
13563 * Bugfix: client side session caching did not work with external caching,
13564 because the session->cipher setting was not restored when reloading
13565 from the external cache. This problem was masked, when
13566 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13567 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13571 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13572 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13574 *Zeev Lieber <zeev-l@yahoo.com>*
13576 * Undo an undocumented change introduced in 0.9.6e which caused
13577 repeated calls to OpenSSL_add_all_ciphers() and
13578 OpenSSL_add_all_digests() to be ignored, even after calling
13583 * Change the default configuration reader to deal with last line not
13584 being properly terminated.
13588 * Change X509_NAME_cmp() so it applies the special rules on handling
13589 DN values that are of type PrintableString, as well as RDNs of type
13590 emailAddress where the value has the type ia5String.
13592 *stefank@valicert.com via Richard Levitte*
13594 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13595 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13596 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13597 the bitwise-OR of the two for use by the majority of applications
13598 wanting this behaviour, and update the docs. The documented
13599 behaviour and actual behaviour were inconsistent and had been
13600 changing anyway, so this is more a bug-fix than a behavioural
13603 *Geoff Thorpe, diagnosed by Nadav Har'El*
13605 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13606 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13610 * Fix initialization code race conditions in
13611 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13612 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13613 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13614 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13615 ssl2_get_cipher_by_char(),
13616 ssl3_get_cipher_by_char().
13618 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13620 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13621 the cached sessions are flushed, as the remove_cb() might use ex_data
13622 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13623 (see [openssl.org #212]).
13625 *Geoff Thorpe, Lutz Jaenicke*
13627 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13628 length, instead of the encoding length to d2i_ASN1_OBJECT.
13632 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13634 * [In 0.9.6g-engine release:]
13635 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13637 *Lynn Gazis <lgazis@rainbow.com>*
13639 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13641 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13642 and get fix the header length calculation.
13643 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13644 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13646 * Use proper error handling instead of 'assertions' in buffer
13647 overflow checks added in 0.9.6e. This prevents DoS (the
13648 assertions could call abort()).
13650 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13652 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13654 * Add various sanity checks to asn1_get_length() to reject
13655 the ASN1 length bytes if they exceed sizeof(long), will appear
13656 negative or the content length exceeds the length of the
13659 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13661 * Fix cipher selection routines: ciphers without encryption had no flags
13662 for the cipher strength set and where therefore not handled correctly
13663 by the selection routines (PR #130).
13667 * Fix EVP_dsa_sha macro.
13672 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13673 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13674 that was added in OpenSSL 0.9.6d.
13676 As the countermeasure turned out to be incompatible with some
13677 broken SSL implementations, the new option is part of SSL_OP_ALL.
13678 SSL_OP_ALL is usually employed when compatibility with weird SSL
13679 implementations is desired (e.g. '-bugs' option to 's_client' and
13680 's_server'), so the new option is automatically set in many
13685 * Changes in security patch:
13687 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13688 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13689 Air Force Materiel Command, USAF, under agreement number
13692 * Add various sanity checks to asn1_get_length() to reject
13693 the ASN1 length bytes if they exceed sizeof(long), will appear
13694 negative or the content length exceeds the length of the
13695 supplied buffer. ([CVE-2002-0659])
13697 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13699 * Assertions for various potential buffer overflows, not known to
13700 happen in practice.
13702 *Ben Laurie (CHATS)*
13704 * Various temporary buffers to hold ASCII versions of integers were
13705 too small for 64 bit platforms. ([CVE-2002-0655])
13706 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13708 * Remote buffer overflow in SSL3 protocol - an attacker could
13709 supply an oversized session ID to a client. ([CVE-2002-0656])
13711 *Ben Laurie (CHATS)*
13713 * Remote buffer overflow in SSL2 protocol - an attacker could
13714 supply an oversized client master key. ([CVE-2002-0656])
13716 *Ben Laurie (CHATS)*
13718 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13720 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13721 encoded as NULL) with id-dsa-with-sha1.
13723 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13725 * Check various `X509_...()` return values in `apps/req.c`.
13727 *Nils Larsch <nla@trustcenter.de>*
13729 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13730 an end-of-file condition would erroneously be flagged, when the CRLF
13731 was just at the end of a processed block. The bug was discovered when
13732 processing data through a buffering memory BIO handing the data to a
13733 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13734 <ptsekov@syntrex.com> and Nedelcho Stanev.
13738 * Implement a countermeasure against a vulnerability recently found
13739 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13740 before application data chunks to avoid the use of known IVs
13741 with data potentially chosen by the attacker.
13745 * Fix length checks in ssl3_get_client_hello().
13749 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13750 to prevent ssl3_read_internal() from incorrectly assuming that
13751 ssl3_read_bytes() found application data while handshake
13752 processing was enabled when in fact s->s3->in_read_app_data was
13753 merely automatically cleared during the initial handshake.
13755 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13757 * Fix object definitions for Private and Enterprise: they were not
13758 recognized in their shortname (=lowercase) representation. Extend
13759 obj_dat.pl to issue an error when using undefined keywords instead
13760 of silently ignoring the problem (Svenning Sorensen
13761 <sss@sss.dnsalias.net>).
13765 * Fix DH_generate_parameters() so that it works for 'non-standard'
13766 generators, i.e. generators other than 2 and 5. (Previously, the
13767 code did not properly initialise the 'add' and 'rem' values to
13768 BN_generate_prime().)
13770 In the new general case, we do not insist that 'generator' is
13771 actually a primitive root: This requirement is rather pointless;
13772 a generator of the order-q subgroup is just as good, if not
13777 * Map new X509 verification errors to alerts. Discovered and submitted by
13778 Tom Wu <tom@arcot.com>.
13782 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13783 returning non-zero before the data has been completely received
13784 when using non-blocking I/O.
13786 *Bodo Moeller; problem pointed out by John Hughes*
13788 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13790 *Ben Laurie, Lutz Jaenicke*
13792 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13793 Yoram Zahavi <YoramZ@gilian.com>).
13797 * Add information about CygWin 1.3 and on, and preserve proper
13798 configuration for the versions before that.
13800 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13802 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13803 check whether we deal with a copy of a session and do not delete from
13804 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13805 <izhar@checkpoint.com>.
13809 * Do not store session data into the internal session cache, if it
13810 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13811 flag is set). Proposed by Aslam <aslam@funk.com>.
13815 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13820 * [In 0.9.6d-engine release:]
13821 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13823 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13825 * Add the configuration target linux-s390x.
13827 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13829 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13830 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13831 variable as an indication that a ClientHello message has been
13832 received. As the flag value will be lost between multiple
13833 invocations of ssl3_accept when using non-blocking I/O, the
13834 function may not be aware that a handshake has actually taken
13835 place, thus preventing a new session from being added to the
13838 To avoid this problem, we now set s->new_session to 2 instead of
13839 using a local variable.
13841 *Lutz Jaenicke, Bodo Moeller*
13843 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13844 if the SSL_R_LENGTH_MISMATCH error is detected.
13846 *Geoff Thorpe, Bodo Moeller*
13848 * New 'shared_ldflag' column in Configure platform table.
13852 * Fix EVP_CIPHER_mode macro.
13854 *"Dan S. Camper" <dan@bti.net>*
13856 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13857 type, we must throw them away by setting rr->length to 0.
13859 *D P Chang <dpc@qualys.com>*
13861 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13863 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13864 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13865 worked incorrectly for those cases where range = `10..._2` and
13866 `3*range` is two bits longer than range.)
13870 * Only add signing time to PKCS7 structures if it is not already
13875 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13876 OBJ_ld_ce should be OBJ_id_ce.
13877 Also some ip-pda OIDs in crypto/objects/objects.txt were
13878 incorrect (cf. RFC 3039).
13880 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13882 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13883 returns early because it has nothing to do.
13885 *Andy Schneider <andy.schneider@bjss.co.uk>*
13887 * [In 0.9.6c-engine release:]
13888 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13890 *Andy Schneider <andy.schneider@bjss.co.uk>*
13892 * [In 0.9.6c-engine release:]
13893 Add support for Cryptographic Appliance's keyserver technology.
13894 (Use engine 'keyclient')
13896 *Cryptographic Appliances and Geoff Thorpe*
13898 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13899 is called via tools/c89.sh because arguments have to be
13900 rearranged (all '-L' options must appear before the first object
13903 *Richard Shapiro <rshapiro@abinitio.com>*
13905 * [In 0.9.6c-engine release:]
13906 Add support for Broadcom crypto accelerator cards, backported
13909 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13911 * [In 0.9.6c-engine release:]
13912 Add support for SureWare crypto accelerator cards from
13913 Baltimore Technologies. (Use engine 'sureware')
13915 *Baltimore Technologies and Mark Cox*
13917 * [In 0.9.6c-engine release:]
13918 Add support for crypto accelerator cards from Accelerated
13919 Encryption Processing, www.aep.ie. (Use engine 'aep')
13921 *AEP Inc. and Mark Cox*
13923 * Add a configuration entry for gcc on UnixWare.
13925 *Gary Benson <gbenson@redhat.com>*
13927 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13928 messages are stored in a single piece (fixed-length part and
13929 variable-length part combined) and fix various bugs found on the way.
13933 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13934 instead. BIO_gethostbyname() does not know what timeouts are
13935 appropriate, so entries would stay in cache even when they have
13937 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13939 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13940 faced with a pathologically small ClientHello fragment that does
13941 not contain client_version: Instead of aborting with an error,
13942 simply choose the highest available protocol version (i.e.,
13943 TLS 1.0 unless it is disabled). In practice, ClientHello
13944 messages are never sent like this, but this change gives us
13945 strictly correct behaviour at least for TLS.
13949 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13950 never resets s->method to s->ctx->method when called from within
13951 one of the SSL handshake functions.
13953 *Bodo Moeller; problem pointed out by Niko Baric*
13955 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13956 (sent using the client's version number) if client_version is
13957 smaller than the protocol version in use. Also change
13958 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13959 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13960 the client will at least see that alert.
13964 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13969 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13970 client receives HelloRequest while in a handshake.
13972 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13974 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13975 should end in 'break', not 'goto end' which circumvents various
13976 cleanups done in state SSL_ST_OK. But session related stuff
13977 must be disabled for SSL_ST_OK in the case that we just sent a
13980 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13981 before just sending a HelloRequest.
13983 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13985 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13986 reveal whether illegal block cipher padding was found or a MAC
13987 verification error occurred. (Neither SSLerr() codes nor alerts
13988 are directly visible to potential attackers, but the information
13989 may leak via logfiles.)
13991 Similar changes are not required for the SSL 2.0 implementation
13992 because the number of padding bytes is sent in clear for SSL 2.0,
13993 and the extra bytes are just ignored. However ssl/s2_pkt.c
13994 failed to verify that the purported number of padding bytes is in
13999 * Add OpenUNIX-8 support including shared libraries
14000 (Boyd Lynn Gerber <gerberb@zenez.com>).
14004 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14005 'wristwatch attack' using huge encoding parameters (cf.
14006 James H. Manger's CRYPTO 2001 paper). Note that the
14007 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14008 encoding parameters and hence was not vulnerable.
14012 * BN_sqr() bug fix.
14014 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14016 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14017 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14018 followed by modular reduction.
14020 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14022 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14023 equivalent based on BN_pseudo_rand() instead of BN_rand().
14027 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14028 This function was broken, as the check for a new client hello message
14029 to handle SGC did not allow these large messages.
14030 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14034 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14038 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14039 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14043 * Rework the configuration and shared library support for Tru64 Unix.
14044 The configuration part makes use of modern compiler features and
14045 still retains old compiler behavior for those that run older versions
14046 of the OS. The shared library support part includes a variant that
14047 uses the RPATH feature, and is available through the special
14048 configuration target "alpha-cc-rpath", which will never be selected
14051 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14053 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14054 with the same message size as in ssl3_get_certificate_request().
14055 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14056 messages might inadvertently be reject as too long.
14058 *Petr Lampa <lampa@fee.vutbr.cz>*
14060 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14064 * Modified SSL library such that the verify_callback that has been set
14065 specifically for an SSL object with SSL_set_verify() is actually being
14066 used. Before the change, a verify_callback set with this function was
14067 ignored and the verify_callback() set in the SSL_CTX at the time of
14068 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14069 to allow the necessary settings.
14073 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14074 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14075 done automatically (in contradiction to the requirements of the C
14076 standard). This made problems when used from OpenSSH.
14080 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14081 dh->length and always used
14083 BN_rand_range(priv_key, dh->p).
14085 BN_rand_range() is not necessary for Diffie-Hellman, and this
14086 specific range makes Diffie-Hellman unnecessarily inefficient if
14087 dh->length (recommended exponent length) is much smaller than the
14088 length of dh->p. We could use BN_rand_range() if the order of
14089 the subgroup was stored in the DH structure, but we only have
14094 BN_rand(priv_key, l, ...)
14096 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14103 RSA_eay_public_encrypt
14104 RSA_eay_private_decrypt
14105 RSA_eay_private_encrypt (signing)
14106 RSA_eay_public_decrypt (signature verification)
14108 (default implementations for RSA_public_encrypt,
14109 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14110 always reject numbers >= n.
14114 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14115 to synchronize access to 'locking_thread'. This is necessary on
14116 systems where access to 'locking_thread' (an 'unsigned long'
14117 variable) is not atomic.
14121 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14122 *before* setting the 'crypto_lock_rand' flag. The previous code had
14123 a race condition if 0 is a valid thread ID.
14125 *Travis Vitek <vitek@roguewave.com>*
14127 * Add support for shared libraries under Irix.
14129 *Albert Chin-A-Young <china@thewrittenword.com>*
14131 * Add configuration option to build on Linux on both big-endian and
14132 little-endian MIPS.
14134 *Ralf Baechle <ralf@uni-koblenz.de>*
14136 * Add the possibility to create shared libraries on HP-UX.
14140 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14142 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14143 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14144 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14145 PRNG state recovery was possible based on the output of
14146 one PRNG request appropriately sized to gain knowledge on
14147 'md' followed by enough consecutive 1-byte PRNG requests
14148 to traverse all of 'state'.
14150 1. When updating 'md_local' (the current thread's copy of 'md')
14151 during PRNG output generation, hash all of the previous
14152 'md_local' value, not just the half used for PRNG output.
14154 2. Make the number of bytes from 'state' included into the hash
14155 independent from the number of PRNG bytes requested.
14157 The first measure alone would be sufficient to avoid
14158 Markku-Juhani's attack. (Actually it had never occurred
14159 to me that the half of 'md_local' used for chaining was the
14160 half from which PRNG output bytes were taken -- I had always
14161 assumed that the secret half would be used.) The second
14162 measure makes sure that additional data from 'state' is never
14163 mixed into 'md_local' in small portions; this heuristically
14164 further strengthens the PRNG.
14168 * Fix crypto/bn/asm/mips3.s.
14172 * When only the key is given to "enc", the IV is undefined. Print out
14173 an error message in this case.
14177 * Handle special case when X509_NAME is empty in X509 printing routines.
14181 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14182 positive and less than q.
14186 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14187 used: it isn't thread safe and the add_lock_callback should handle
14190 *Paul Rose <Paul.Rose@bridge.com>*
14192 * Verify that incoming data obeys the block size in
14193 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14199 *Ulf Möller, Bodo Möller*
14201 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14202 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14203 when fixing the server behaviour for backwards-compatible 'client
14204 hello' messages. (Note that the attack is impractical against
14205 SSL 3.0 and TLS 1.0 anyway because length and version checking
14206 means that the probability of guessing a valid ciphertext is
14207 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14210 Before 0.9.5, the countermeasure (hide the error by generating a
14211 random 'decryption result') did not work properly because
14212 ERR_clear_error() was missing, meaning that SSL_get_error() would
14213 detect the supposedly ignored error.
14215 Both problems are now fixed.
14219 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14220 (previously it was 1024).
14224 * Fix for compatibility mode trust settings: ignore trust settings
14225 unless some valid trust or reject settings are present.
14229 * Fix for blowfish EVP: its a variable length cipher.
14233 * Fix various bugs related to DSA S/MIME verification. Handle missing
14234 parameters in DSA public key structures and return an error in the
14235 DSA routines if parameters are absent.
14239 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14240 in the current directory if neither $RANDFILE nor $HOME was set.
14241 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14242 caused some confusion to Windows users who haven't defined $HOME.
14243 Thus RAND_file_name() is changed again: e_os.h can define a
14244 DEFAULT_HOME, which will be used if $HOME is not set.
14245 For Windows, we use "C:"; on other platforms, we still require
14246 environment variables.
14248 * Move 'if (!initialized) RAND_poll()' into regions protected by
14249 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14250 having multiple threads call RAND_poll() concurrently.
14254 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14255 combination of a flag and a thread ID variable.
14256 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14257 flag), *other* threads can enter ssleay_add_bytes without obeying
14258 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14259 that they do not hold after the first thread unsets add_do_not_lock).
14263 * Change bctest again: '-x' expressions are not available in all
14264 versions of 'test'.
14268 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14270 * Fix a couple of memory leaks in PKCS7_dataDecode()
14272 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14274 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14275 the default extension for executables, if any. Also, make the perl
14276 scripts that use symlink() to test if it really exists and use "cp"
14277 if it doesn't. All this made OpenSSL compilable and installable in
14282 * Fix for asn1_GetSequence() for indefinite length constructed data.
14283 If SEQUENCE is length is indefinite just set c->slen to the total
14284 amount of data available.
14286 *Steve Henson, reported by shige@FreeBSD.org*
14288 *This change does not apply to 0.9.7.*
14290 * Change bctest to avoid here-documents inside command substitution
14291 (workaround for FreeBSD /bin/sh bug).
14292 For compatibility with Ultrix, avoid shell functions (introduced
14293 in the bctest version that searches along $PATH).
14297 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14298 with des_encrypt() defined on some operating systems, like Solaris
14303 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14304 On the Importance of Eliminating Errors in Cryptographic
14305 Computations, J. Cryptology 14 (2001) 2, 101-119,
14306 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14310 * MIPS assembler BIGNUM division bug fix.
14314 * Disabled incorrect Alpha assembler code.
14318 * Fix PKCS#7 decode routines so they correctly update the length
14319 after reading an EOC for the EXPLICIT tag.
14323 *This change does not apply to 0.9.7.*
14325 * Fix bug in PKCS#12 key generation routines. This was triggered
14326 if a 3DES key was generated with a 0 initial byte. Include
14327 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14328 (but broken) behaviour.
14332 * Enhance bctest to search for a working bc along $PATH and print
14335 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14337 * Fix memory leaks in err.c: free err_data string if necessary;
14338 don't write to the wrong index in ERR_set_error_data.
14342 * Implement ssl23_peek (analogous to ssl23_read), which previously
14347 * Replace rdtsc with `_emit` statements for VC++ version 5.
14349 *Jeremy Cooper <jeremy@baymoo.org>*
14351 * Make it possible to reuse SSLv2 sessions.
14355 * In copy_email() check for >= 0 as a return value for
14356 X509_NAME_get_index_by_NID() since 0 is a valid index.
14358 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14360 * Avoid coredump with unsupported or invalid public keys by checking if
14361 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14362 PKCS7_verify() fails with non detached data.
14366 * Don't use getenv in library functions when run as setuid/setgid.
14367 New function OPENSSL_issetugid().
14371 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14372 due to incorrect handling of multi-threading:
14374 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14376 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14378 3. Count how many times MemCheck_off() has been called so that
14379 nested use can be treated correctly. This also avoids
14380 inband-signalling in the previous code (which relied on the
14381 assumption that thread ID 0 is impossible).
14385 * Add "-rand" option also to s_client and s_server.
14389 * Fix CPU detection on Irix 6.x.
14390 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14391 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14393 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14398 *This change does not apply to 0.9.7.*
14400 * Use the cached encoding of an X509_NAME structure rather than
14401 copying it. This is apparently the reason for the libsafe "errors"
14402 but the code is actually correct.
14406 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14407 Bleichenbacher's DSA attack.
14408 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14409 to be set and top=0 forces the highest bit to be set; top=-1 is new
14410 and leaves the highest bit random.
14412 *Ulf Moeller, Bodo Moeller*
14414 * In the `NCONF_...`-based implementations for `CONF_...` queries
14415 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14416 a temporary CONF structure with the data component set to NULL
14417 (which gives segmentation faults in lh_retrieve).
14418 Instead, use NULL for the CONF pointer in CONF_get_string and
14419 CONF_get_number (which may use environment variables) and directly
14420 return NULL from CONF_get_section.
14424 * Fix potential buffer overrun for EBCDIC.
14428 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14429 keyUsage if basicConstraints absent for a CA.
14433 * Make SMIME_write_PKCS7() write mail header values with a format that
14434 is more generally accepted (no spaces before the semicolon), since
14435 some programs can't parse those values properly otherwise. Also make
14436 sure BIO's that break lines after each write do not create invalid
14441 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14442 macros previously used would not encode an empty SEQUENCE OF
14443 and break the signature.
14447 *This change does not apply to 0.9.7.*
14449 * Zero the premaster secret after deriving the master secret in
14454 * Add some EVP_add_digest_alias registrations (as found in
14455 OpenSSL_add_all_digests()) to SSL_library_init()
14456 aka OpenSSL_add_ssl_algorithms(). This provides improved
14457 compatibility with peers using X.509 certificates
14458 with unconventional AlgorithmIdentifier OIDs.
14462 * Fix for Irix with NO_ASM.
14464 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14466 * ./config script fixes.
14468 *Ulf Moeller, Richard Levitte*
14470 * Fix 'openssl passwd -1'.
14474 * Change PKCS12_key_gen_asc() so it can cope with non null
14475 terminated strings whose length is passed in the passlen
14476 parameter, for example from PEM callbacks. This was done
14477 by adding an extra length parameter to asc2uni().
14479 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14481 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14482 call failed, free the DSA structure.
14486 * Fix to uni2asc() to cope with zero length Unicode strings.
14487 These are present in some PKCS#12 files.
14491 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14492 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14493 when writing a 32767 byte record.
14495 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14497 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14498 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14500 (RSA objects have a reference count access to which is protected
14501 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14502 so they are meant to be shared between threads.)
14503 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14504 "Reddie, Steven" <Steven.Reddie@ca.com>*
14506 * Fix a deadlock in CRYPTO_mem_leaks().
14510 * Use better test patterns in bntest.
14514 * rand_win.c fix for Borland C.
14518 * BN_rshift bugfix for n == 0.
14522 * Add a 'bctest' script that checks for some known 'bc' bugs
14523 so that 'make test' does not abort just because 'bc' is broken.
14527 * Store verify_result within SSL_SESSION also for client side to
14528 avoid potential security hole. (Re-used sessions on the client side
14529 always resulted in verify_result==X509_V_OK, not using the original
14530 result of the server certificate verification.)
14534 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14535 SSL3_RT_APPLICATION_DATA, return 0.
14536 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14541 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14542 releases, have been re-implemented by renaming the previous
14543 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14544 and ssl3_read_internal, respectively, and adding 'peek' parameters
14545 to them. The new ssl[23]_{read,peek} functions are calls to
14546 ssl[23]_read_internal with the 'peek' flag set appropriately.
14547 A 'peek' parameter has also been added to ssl3_read_bytes, which
14548 does the actual work for ssl3_read_internal.
14552 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14553 the method-specific "init()" handler. Also clean up ex_data after
14554 calling the method-specific "finish()" handler. Previously, this was
14555 happening the other way round.
14559 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14560 The previous value, 12, was not always sufficient for BN_mod_exp().
14564 * Make sure that shared libraries get the internal name engine with
14565 the full version number and not just 0. This should mark the
14566 shared libraries as not backward compatible. Of course, this should
14567 be changed again when we can guarantee backward binary compatibility.
14571 * Fix typo in get_cert_by_subject() in by_dir.c
14573 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14575 * Rework the system to generate shared libraries:
14577 - Make note of the expected extension for the shared libraries and
14578 if there is a need for symbolic links from for example libcrypto.so.0
14579 to libcrypto.so.0.9.7. There is extended info in Configure for
14582 - Make as few rebuilds of the shared libraries as possible.
14584 - Still avoid linking the OpenSSL programs with the shared libraries.
14586 - When installing, install the shared libraries separately from the
14591 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14593 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14594 and not in SSL_clear because the latter is also used by the
14595 accept/connect functions; previously, the settings made by
14596 SSL_set_read_ahead would be lost during the handshake.
14598 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14600 * Correct util/mkdef.pl to be selective about disabled algorithms.
14601 Previously, it would create entries for disabled algorithms no
14606 * Added several new manual pages for SSL_* function.
14610 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14612 * In ssl23_get_client_hello, generate an error message when faced
14613 with an initial SSL 3.0/TLS record that is too small to contain the
14614 first two bytes of the ClientHello message, i.e. client_version.
14615 (Note that this is a pathologic case that probably has never happened
14616 in real life.) The previous approach was to use the version number
14617 from the record header as a substitute; but our protocol choice
14618 should not depend on that one because it is not authenticated
14619 by the Finished messages.
14623 * More robust randomness gathering functions for Windows.
14625 *Jeffrey Altman <jaltman@columbia.edu>*
14627 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14628 not set then we don't setup the error code for issuer check errors
14629 to avoid possibly overwriting other errors which the callback does
14630 handle. If an application does set the flag then we assume it knows
14631 what it is doing and can handle the new informational codes
14636 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14637 a general "ANY" type, as such it should be able to decode anything
14638 including tagged types. However it didn't check the class so it would
14639 wrongly interpret tagged types in the same way as their universal
14640 counterpart and unknown types were just rejected. Changed so that the
14641 tagged and unknown types are handled in the same way as a SEQUENCE:
14642 that is the encoding is stored intact. There is also a new type
14643 "V_ASN1_OTHER" which is used when the class is not universal, in this
14644 case we have no idea what the actual type is so we just lump them all
14649 * On VMS, stdout may very well lead to a file that is written to
14650 in a record-oriented fashion. That means that every write() will
14651 write a separate record, which will be read separately by the
14652 programs trying to read from it. This can be very confusing.
14654 The solution is to put a BIO filter in the way that will buffer
14655 text until a linefeed is reached, and then write everything a
14656 line at a time, so every record written will be an actual line,
14657 not chunks of lines and not (usually doesn't happen, but I've
14658 seen it once) several lines in one record. BIO_f_linebuffer() is
14661 Currently, it's a VMS-only method, because that's where it has
14662 been tested well enough.
14666 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14667 it can return incorrect results.
14668 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14669 but it was in 0.9.6-beta[12].)
14673 * Disable the check for content being present when verifying detached
14674 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14675 include zero length content when signing messages.
14679 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14680 BIO_ctrl (for BIO pairs).
14684 * Add DSO method for VMS.
14688 * Bug fix: Montgomery multiplication could produce results with the
14693 * Add RPM specification openssl.spec and modify it to build three
14694 packages. The default package contains applications, application
14695 documentation and run-time libraries. The devel package contains
14696 include files, static libraries and function documentation. The
14697 doc package contains the contents of the doc directory. The original
14698 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14702 * Add a large number of documentation files for many SSL routines.
14704 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14706 * Add a configuration entry for Sony News 4.
14708 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14710 * Don't set the two most significant bits to one when generating a
14711 random number < q in the DSA library.
14715 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14716 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14717 the underlying transport is blocking) if a handshake took place.
14718 (The default behaviour is needed by applications such as s_client
14719 and s_server that use select() to determine when to use SSL_read;
14720 but for applications that know in advance when to expect data, it
14721 just makes things more complicated.)
14725 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14730 * Add a few more EBCDIC conditionals that make `req` and `x509`
14731 work better on such systems.
14733 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14735 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14736 Update PKCS12_parse() so it copies the friendlyName and the
14737 keyid to the certificates aux info.
14741 * Fix bug in PKCS7_verify() which caused an infinite loop
14742 if there was more than one signature.
14744 *Sven Uszpelkat <su@celocom.de>*
14746 * Major change in util/mkdef.pl to include extra information
14747 about each symbol, as well as presenting variables as well
14748 as functions. This change means that there's n more need
14749 to rebuild the .num files when some algorithms are excluded.
14753 * Allow the verify time to be set by an application,
14754 rather than always using the current time.
14758 * Phase 2 verify code reorganisation. The certificate
14759 verify code now looks up an issuer certificate by a
14760 number of criteria: subject name, authority key id
14761 and key usage. It also verifies self signed certificates
14762 by the same criteria. The main comparison function is
14763 X509_check_issued() which performs these checks.
14765 Lot of changes were necessary in order to support this
14766 without completely rewriting the lookup code.
14768 Authority and subject key identifier are now cached.
14770 The LHASH 'certs' is X509_STORE has now been replaced
14771 by a STACK_OF(X509_OBJECT). This is mainly because an
14772 LHASH can't store or retrieve multiple objects with
14773 the same hash value.
14775 As a result various functions (which were all internal
14776 use only) have changed to handle the new X509_STORE
14777 structure. This will break anything that messed round
14778 with X509_STORE internally.
14780 The functions X509_STORE_add_cert() now checks for an
14781 exact match, rather than just subject name.
14783 The X509_STORE API doesn't directly support the retrieval
14784 of multiple certificates matching a given criteria, however
14785 this can be worked round by performing a lookup first
14786 (which will fill the cache with candidate certificates)
14787 and then examining the cache for matches. This is probably
14788 the best we can do without throwing out X509_LOOKUP
14789 entirely (maybe later...).
14791 The X509_VERIFY_CTX structure has been enhanced considerably.
14793 All certificate lookup operations now go via a get_issuer()
14794 callback. Although this currently uses an X509_STORE it
14795 can be replaced by custom lookups. This is a simple way
14796 to bypass the X509_STORE hackery necessary to make this
14797 work and makes it possible to use more efficient techniques
14798 in future. A very simple version which uses a simple
14799 STACK for its trusted certificate store is also provided
14800 using X509_STORE_CTX_trusted_stack().
14802 The verify_cb() and verify() callbacks now have equivalents
14803 in the X509_STORE_CTX structure.
14805 X509_STORE_CTX also has a 'flags' field which can be used
14806 to customise the verify behaviour.
14810 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14811 excludes S/MIME capabilities.
14815 * When a certificate request is read in keep a copy of the
14816 original encoding of the signed data and use it when outputting
14817 again. Signatures then use the original encoding rather than
14818 a decoded, encoded version which may cause problems if the
14819 request is improperly encoded.
14823 * For consistency with other BIO_puts implementations, call
14824 buffer_write(b, ...) directly in buffer_puts instead of calling
14827 In BIO_puts, increment b->num_write as in BIO_write.
14829 *Peter.Sylvester@EdelWeb.fr*
14831 * Fix BN_mul_word for the case where the word is 0. (We have to use
14832 BN_zero, we may not return a BIGNUM with an array consisting of
14833 words set to zero.)
14837 * Avoid calling abort() from within the library when problems are
14838 detected, except if preprocessor symbols have been defined
14839 (such as REF_CHECK, BN_DEBUG etc.).
14843 * New openssl application 'rsautl'. This utility can be
14844 used for low-level RSA operations. DER public key
14845 BIO/fp routines also added.
14849 * New Configure entry and patches for compiling on QNX 4.
14851 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14853 * A demo state-machine implementation was sponsored by
14854 Nuron (<http://www.nuron.com/>) and is now available in
14855 demos/state_machine.
14859 * New options added to the 'dgst' utility for signature
14860 generation and verification.
14864 * Unrecognized PKCS#7 content types are now handled via a
14865 catch all ASN1_TYPE structure. This allows unsupported
14866 types to be stored as a "blob" and an application can
14867 encode and decode it manually.
14871 * Fix various signed/unsigned issues to make a_strex.c
14872 compile under VC++.
14874 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14876 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14877 length if passed a buffer. ASN1_INTEGER_to_BN failed
14878 if passed a NULL BN and its argument was negative.
14880 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14882 * Modification to PKCS#7 encoding routines to output definite
14883 length encoding. Since currently the whole structures are in
14884 memory there's not real point in using indefinite length
14885 constructed encoding. However if OpenSSL is compiled with
14886 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14890 * Added BIO_vprintf() and BIO_vsnprintf().
14894 * Added more prefixes to parse for in the strings written
14895 through a logging bio, to cover all the levels that are available
14896 through syslog. The prefixes are now:
14898 PANIC, EMERG, EMR => LOG_EMERG
14899 ALERT, ALR => LOG_ALERT
14900 CRIT, CRI => LOG_CRIT
14901 ERROR, ERR => LOG_ERR
14902 WARNING, WARN, WAR => LOG_WARNING
14903 NOTICE, NOTE, NOT => LOG_NOTICE
14904 INFO, INF => LOG_INFO
14905 DEBUG, DBG => LOG_DEBUG
14907 and as before, if none of those prefixes are present at the
14908 beginning of the string, LOG_ERR is chosen.
14910 On Win32, the `LOG_*` levels are mapped according to this:
14912 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14913 LOG_WARNING => EVENTLOG_WARNING_TYPE
14914 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14918 * Made it possible to reconfigure with just the configuration
14919 argument "reconf" or "reconfigure". The command line arguments
14920 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14921 and are retrieved from there when reconfiguring.
14927 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14929 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14933 * The obj_dat.pl script was messing up the sorting of object
14934 names. The reason was that it compared the quoted version
14935 of strings as a result "OCSP" > "OCSP Signing" because
14936 " > SPACE. Changed script to store unquoted versions of
14937 names and add quotes on output. It was also omitting some
14938 names from the lookup table if they were given a default
14939 value (that is if SN is missing it is given the same
14940 value as LN and vice versa), these are now added on the
14941 grounds that if an object has a name we should be able to
14942 look it up. Finally added warning output when duplicate
14943 short or long names are found.
14947 * Changes needed for Tandem NSK.
14949 *Scott Uroff <scott@xypro.com>*
14951 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14952 RSA_padding_check_SSLv23(), special padding was never detected
14953 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14954 version rollback attacks was not effective.
14956 In s23_clnt.c, don't use special rollback-attack detection padding
14957 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14958 client; similarly, in s23_srvr.c, don't do the rollback check if
14959 SSL 2.0 is the only protocol enabled in the server.
14963 * Make it possible to get hexdumps of unprintable data with 'openssl
14964 asn1parse'. By implication, the functions ASN1_parse_dump() and
14965 BIO_dump_indent() are added.
14969 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14970 these print out strings and name structures based on various
14971 flags including RFC2253 support and proper handling of
14972 multibyte characters. Added options to the 'x509' utility
14973 to allow the various flags to be set.
14977 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14978 Also change the functions X509_cmp_current_time() and
14979 X509_gmtime_adj() work with an ASN1_TIME structure,
14980 this will enable certificates using GeneralizedTime in validity
14981 dates to be checked.
14985 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14986 negative public key encodings) on by default,
14987 NO_NEG_PUBKEY_BUG can be set to disable it.
14991 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14992 content octets. An i2c_ASN1_OBJECT is unnecessary because
14993 the encoding can be trivially obtained from the structure.
14997 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14998 not read locks (`CRYPTO_r_[un]lock`).
15002 * A first attempt at creating official support for shared
15003 libraries through configuration. I've kept it so the
15004 default is static libraries only, and the OpenSSL programs
15005 are always statically linked for now, but there are
15006 preparations for dynamic linking in place.
15007 This has been tested on Linux and Tru64.
15011 * Randomness polling function for Win9x, as described in:
15012 Peter Gutmann, Software Generation of Practically Strong
15017 * Fix so PRNG is seeded in req if using an already existing
15022 * New options to smime application. -inform and -outform
15023 allow alternative formats for the S/MIME message including
15024 PEM and DER. The -content option allows the content to be
15025 specified separately. This should allow things like Netscape
15026 form signing output easier to verify.
15030 * Fix the ASN1 encoding of tags using the 'long form'.
15034 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15035 STRING types. These convert content octets to and from the
15036 underlying type. The actual tag and length octets are
15037 already assumed to have been read in and checked. These
15038 are needed because all other string types have virtually
15039 identical handling apart from the tag. By having versions
15040 of the ASN1 functions that just operate on content octets
15041 IMPLICIT tagging can be handled properly. It also allows
15042 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15043 and ASN1_INTEGER are identical apart from the tag.
15047 * Change the handling of OID objects as follows:
15049 - New object identifiers are inserted in objects.txt, following
15050 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15051 - objects.pl is used to process obj_mac.num and create a new
15053 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15056 This is currently kind of a hack, and the perl code in objects.pl
15057 isn't very elegant, but it works as I intended. The simplest way
15058 to check that it worked correctly is to look in obj_dat.h and
15059 check the array nid_objs and make sure the objects haven't moved
15060 around (this is important!). Additions are OK, as well as
15061 consistent name changes.
15065 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15069 * Addition of the command line parameter '-rand file' to 'openssl req'.
15070 The given file adds to whatever has already been seeded into the
15071 random pool through the RANDFILE configuration file option or
15072 environment variable, or the default random state file.
15076 * mkstack.pl now sorts each macro group into lexical order.
15077 Previously the output order depended on the order the files
15078 appeared in the directory, resulting in needless rewriting
15083 * Patches to make OpenSSL compile under Win32 again. Mostly
15084 work arounds for the VC++ problem that it treats func() as
15085 func(void). Also stripped out the parts of mkdef.pl that
15086 added extra typesafe functions: these no longer exist.
15090 * Reorganisation of the stack code. The macros are now all
15091 collected in safestack.h . Each macro is defined in terms of
15092 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15093 DEBUG_SAFESTACK is now handled in terms of function casts,
15094 this has the advantage of retaining type safety without the
15095 use of additional functions. If DEBUG_SAFESTACK is not defined
15096 then the non typesafe macros are used instead. Also modified the
15097 mkstack.pl script to handle the new form. Needs testing to see
15098 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15099 the default if no major problems. Similar behaviour for ASN1_SET_OF
15100 and PKCS12_STACK_OF.
15104 * When some versions of IIS use the 'NET' form of private key the
15105 key derivation algorithm is different. Normally MD5(password) is
15106 used as a 128 bit RC4 key. In the modified case
15107 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15108 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15109 as the old Netscape_RSA functions except they have an additional
15110 'sgckey' parameter which uses the modified algorithm. Also added
15111 an -sgckey command line option to the rsa utility. Thanks to
15112 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15113 algorithm to openssl-dev.
15117 * The evp_local.h macros were using 'c.##kname' which resulted in
15118 invalid expansion on some systems (SCO 5.0.5 for example).
15119 Corrected to 'c.kname'.
15121 *Phillip Porch <root@theporch.com>*
15123 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15124 a STACK of email addresses from a certificate or request, these look
15125 in the subject name and the subject alternative name extensions and
15126 omit any duplicate addresses.
15130 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15131 This makes DSA verification about 2 % faster.
15135 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15136 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15137 plus overhead for 1024 bit moduli).
15138 This makes exponentiations about 0.5 % faster for 1024 bit
15139 exponents (as measured by "openssl speed rsa2048").
15143 * Rename memory handling macros to avoid conflicts with other
15145 Malloc => OPENSSL_malloc
15146 Malloc_locked => OPENSSL_malloc_locked
15147 Realloc => OPENSSL_realloc
15148 Free => OPENSSL_free
15152 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15153 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15157 * CygWin32 support.
15159 *John Jarvie <jjarvie@newsguy.com>*
15161 * The type-safe stack code has been rejigged. It is now only compiled
15162 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15163 by default all type-specific stack functions are "#define"d back to
15164 standard stack functions. This results in more streamlined output
15165 but retains the type-safety checking possibilities of the original
15170 * The STACK code has been cleaned up, and certain type declarations
15171 that didn't make a lot of sense have been brought in line. This has
15172 also involved a cleanup of sorts in safestack.h to more correctly
15173 map type-safe stack functions onto their plain stack counterparts.
15174 This work has also resulted in a variety of "const"ifications of
15175 lots of the code, especially `_cmp` operations which should normally
15176 be prototyped with "const" parameters anyway.
15180 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15181 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15182 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15183 where all of 'md' is used each time the PRNG is used, but 'state'
15184 is used only indexed by a cyclic counter. As entropy may not be
15185 well distributed from the beginning, 'md' is important as a
15186 chaining variable. However, the output function chains only half
15187 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15188 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15189 in all of 'state' being rewritten, with the new values depending
15190 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15194 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15195 the handshake is continued after ssl_verify_cert_chain();
15196 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15197 can lead to 'unexplainable' connection aborts later.
15199 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15201 * Major EVP API cipher revision.
15202 Add hooks for extra EVP features. This allows various cipher
15203 parameters to be set in the EVP interface. Support added for variable
15204 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15205 setting of RC2 and RC5 parameters.
15207 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15210 Remove lots of duplicated code from the EVP library. For example *every*
15211 cipher init() function handles the 'iv' in the same way according to the
15212 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15213 for CFB and OFB modes they zero ctx->num.
15215 New functionality allows removal of S/MIME code RC2 hack.
15217 Most of the routines have the same form and so can be declared in terms
15220 By shifting this to the top level EVP_CipherInit() it can be removed from
15221 all individual ciphers. If the cipher wants to handle IVs or keys
15222 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15225 Change lots of functions like EVP_EncryptUpdate() to now return a
15226 value: although software versions of the algorithms cannot fail
15227 any installed hardware versions can.
15231 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15232 this option is set, tolerate broken clients that send the negotiated
15233 protocol version number instead of the requested protocol version
15238 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15239 i.e. non-zero for export ciphersuites, zero otherwise.
15240 Previous versions had this flag inverted, inconsistent with
15241 rsa_tmp_cb (..._TMP_RSA_CB).
15243 *Bodo Moeller; problem reported by Amit Chopra*
15245 * Add missing DSA library text string. Work around for some IIS
15246 key files with invalid SEQUENCE encoding.
15250 * Add a document (doc/standards.txt) that list all kinds of standards
15251 and so on that are implemented in OpenSSL.
15255 * Enhance c_rehash script. Old version would mishandle certificates
15256 with the same subject name hash and wouldn't handle CRLs at all.
15257 Added -fingerprint option to crl utility, to support new c_rehash
15262 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15266 * Fix for SSL server purpose checking. Server checking was
15267 rejecting certificates which had extended key usage present
15268 but no ssl client purpose.
15270 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15272 * Make PKCS#12 code work with no password. The PKCS#12 spec
15273 is a little unclear about how a blank password is handled.
15274 Since the password in encoded as a BMPString with terminating
15275 double NULL a zero length password would end up as just the
15276 double NULL. However no password at all is different and is
15277 handled differently in the PKCS#12 key generation code. NS
15278 treats a blank password as zero length. MSIE treats it as no
15279 password on export: but it will try both on import. We now do
15280 the same: PKCS12_parse() tries zero length and no password if
15281 the password is set to "" or NULL (NULL is now a valid password:
15282 it wasn't before) as does the pkcs12 application.
15286 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15287 perror when PEM_read_bio_X509_REQ fails, the error message must
15288 be obtained from the error queue.
15292 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15293 it in ERR_remove_state if appropriate, and change ERR_get_state
15294 accordingly to avoid race conditions (this is necessary because
15295 thread_hash is no longer constant once set).
15299 * Bugfix for linux-elf makefile.one.
15303 * RSA_get_default_method() will now cause a default
15304 RSA_METHOD to be chosen if one doesn't exist already.
15305 Previously this was only set during a call to RSA_new()
15306 or RSA_new_method(NULL) meaning it was possible for
15307 RSA_get_default_method() to return NULL.
15311 * Added native name translation to the existing DSO code
15312 that will convert (if the flag to do so is set) filenames
15313 that are sufficiently small and have no path information
15314 into a canonical native form. Eg. "blah" converted to
15315 "libblah.so" or "blah.dll" etc.
15319 * New function ERR_error_string_n(e, buf, len) which is like
15320 ERR_error_string(e, buf), but writes at most 'len' bytes
15321 including the 0 terminator. For ERR_error_string_n, 'buf'
15324 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15326 * CONF library reworked to become more general. A new CONF
15327 configuration file reader "class" is implemented as well as a
15328 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15329 old `CONF_*` functions are still there, but are reimplemented to
15330 work in terms of the new functions. Also, a set of functions
15331 to handle the internal storage of the configuration data is
15332 provided to make it easier to write new configuration file
15333 reader "classes" (I can definitely see something reading a
15334 configuration file in XML format, for example), called `_CONF_*`,
15335 or "the configuration storage API"...
15337 The new configuration file reading functions are:
15339 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15340 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15342 NCONF_default, NCONF_WIN32
15344 NCONF_dump_fp, NCONF_dump_bio
15346 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15347 NCONF_new creates a new CONF object. This works in the same way
15348 as other interfaces in OpenSSL, like the BIO interface.
15349 `NCONF_dump_*` dump the internal storage of the configuration file,
15350 which is useful for debugging. All other functions take the same
15351 arguments as the old `CONF_*` functions with the exception of the
15352 first that must be a `CONF *` instead of a `LHASH *`.
15354 To make it easier to use the new classes with the old `CONF_*` functions,
15355 the function CONF_set_default_method is provided.
15359 * Add '-tls1' option to 'openssl ciphers', which was already
15360 mentioned in the documentation but had not been implemented.
15361 (This option is not yet really useful because even the additional
15362 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15366 * Initial DSO code added into libcrypto for letting OpenSSL (and
15367 OpenSSL-based applications) load shared libraries and bind to
15368 them in a portable way.
15370 *Geoff Thorpe, with contributions from Richard Levitte*
15372 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15374 * Make sure _lrotl and _lrotr are only used with MSVC.
15376 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15377 (the default implementation of RAND_status).
15379 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15380 to '-clrext' (= clear extensions), as intended and documented.
15381 *Bodo Moeller; inconsistency pointed out by Michael Attili
15382 <attili@amaxo.com>*
15384 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15385 was larger than the MD block size.
15387 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15389 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15390 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15391 using the passed key: if the passed key was a private key the result
15392 of X509_print(), for example, would be to print out all the private key
15397 * des_quad_cksum() byte order bug fix.
15398 *Ulf Möller, using the problem description in krb4-0.9.7, where
15399 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15401 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15404 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15406 * For easily testing in shell scripts whether some command
15407 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15408 returns with exit code 0 iff no command of the given name is available.
15409 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15410 the output goes to stdout and nothing is printed to stderr.
15411 Additional arguments are always ignored.
15413 Since for each cipher there is a command of the same name,
15414 the 'no-cipher' compilation switches can be tested this way.
15416 ('openssl no-XXX' is not able to detect pseudo-commands such
15417 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15421 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15425 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15426 is set; it will be thrown away anyway because each handshake creates
15428 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15429 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15430 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15431 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15435 * New s_client option -ign_eof: EOF at stdin is ignored, and
15436 'Q' and 'R' lose their special meanings (quit/renegotiate).
15437 This is part of what -quiet does; unlike -quiet, -ign_eof
15438 does not suppress any output.
15442 * Add compatibility options to the purpose and trust code. The
15443 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15444 accepts a certificate or CA, this was the previous behaviour,
15445 with all the associated security issues.
15447 X509_TRUST_COMPAT is the old trust behaviour: only and
15448 automatically trust self signed roots in certificate store. A
15449 new trust setting X509_TRUST_DEFAULT is used to specify that
15450 a purpose has no associated trust setting and it should instead
15451 use the value in the default purpose.
15455 * Fix the PKCS#8 DSA private key code so it decodes keys again
15456 and fix a memory leak.
15460 * In util/mkerr.pl (which implements 'make errors'), preserve
15461 reason strings from the previous version of the .c file, as
15462 the default to have only downcase letters (and digits) in
15463 automatically generated reasons codes is not always appropriate.
15467 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15468 using strerror. Previously, ERR_reason_error_string() returned
15469 library names as reason strings for SYSerr; but SYSerr is a special
15470 case where small numbers are errno values, not library numbers.
15474 * Add '-dsaparam' option to 'openssl dhparam' application. This
15475 converts DSA parameters into DH parameters. (When creating parameters,
15476 DSA_generate_parameters is used.)
15480 * Include 'length' (recommended exponent length) in C code generated
15481 by 'openssl dhparam -C'.
15485 * The second argument to set_label in perlasm was already being used
15486 so couldn't be used as a "file scope" flag. Moved to third argument
15491 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15492 instead of RAND_bytes for encryption IVs and salts.
15496 * Include RAND_status() into RAND_METHOD instead of implementing
15497 it only for md_rand.c Otherwise replacing the PRNG by calling
15498 RAND_set_rand_method would be impossible.
15502 * Don't let DSA_generate_key() enter an infinite loop if the random
15503 number generation fails.
15507 * New 'rand' application for creating pseudo-random output.
15511 * Added configuration support for Linux/IA64
15513 *Rolf Haberrecker <rolf@suse.de>*
15515 * Assembler module support for Mingw32.
15519 * Shared library support for HPUX (in shlib/).
15521 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15523 * Shared library support for Solaris gcc.
15525 *Lutz Behnke <behnke@trustcenter.de>*
15527 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15529 * PKCS7_encrypt() was adding text MIME headers twice because they
15530 were added manually and by SMIME_crlf_copy().
15534 * In bntest.c don't call BN_rand with zero bits argument.
15536 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15538 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15539 case was implemented. This caused BN_div_recp() to fail occasionally.
15543 * Add an optional second argument to the set_label() in the perl
15544 assembly language builder. If this argument exists and is set
15545 to 1 it signals that the assembler should use a symbol whose
15546 scope is the entire file, not just the current function. This
15547 is needed with MASM which uses the format label:: for this scope.
15549 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15551 * Change the ASN1 types so they are typedefs by default. Before
15552 almost all types were #define'd to ASN1_STRING which was causing
15553 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15558 * Change names of new functions to the new get1/get0 naming
15559 convention: After 'get1', the caller owns a reference count
15560 and has to call `..._free`; 'get0' returns a pointer to some
15561 data structure without incrementing reference counters.
15562 (Some of the existing 'get' functions increment a reference
15563 counter, some don't.)
15564 Similarly, 'set1' and 'add1' functions increase reference
15565 counters or duplicate objects.
15569 * Allow for the possibility of temp RSA key generation failure:
15570 the code used to assume it always worked and crashed on failure.
15574 * Fix potential buffer overrun problem in BIO_printf().
15575 *Ulf Möller, using public domain code by Patrick Powell; problem
15576 pointed out by David Sacerdote <das33@cornell.edu>*
15578 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15579 RAND_egd() and RAND_status(). In the command line application,
15580 the EGD socket can be specified like a seed file using RANDFILE
15585 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15586 Some CAs (e.g. Verisign) distribute certificates in this form.
15590 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15591 list to exclude them. This means that no special compilation option
15592 is needed to use anonymous DH: it just needs to be included in the
15597 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15598 EVP_MD_type. The old functionality is available in a new macro called
15599 EVP_MD_md(). Change code that uses it and update docs.
15603 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15604 where the `void *` argument is replaced by a function pointer argument.
15605 Previously `void *` was abused to point to functions, which works on
15606 many platforms, but is not correct. As these functions are usually
15607 called by macros defined in OpenSSL header files, most source code
15608 should work without changes.
15612 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15613 sections with information on -D... compiler switches used for
15614 compiling the library so that applications can see them. To enable
15615 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15616 must be defined. E.g.,
15617 #define OPENSSL_ALGORITHM_DEFINES
15618 #include <openssl/opensslconf.h>
15619 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15621 *Richard Levitte, Ulf and Bodo Möller*
15623 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15628 * Change the 'other' type in certificate aux info to a STACK_OF
15629 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15630 the required ASN1 format: arbitrary types determined by an OID.
15634 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15635 argument to 'req'. This is not because the function is newer or
15636 better than others it just uses the work 'NEW' in the certificate
15637 request header lines. Some software needs this.
15641 * Reorganise password command line arguments: now passwords can be
15642 obtained from various sources. Delete the PEM_cb function and make
15643 it the default behaviour: i.e. if the callback is NULL and the
15644 usrdata argument is not NULL interpret it as a null terminated pass
15645 phrase. If usrdata and the callback are NULL then the pass phrase
15646 is prompted for as usual.
15650 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15651 the support is automatically enabled. The resulting binaries will
15652 autodetect the card and use it if present.
15654 *Ben Laurie and Compaq Inc.*
15656 * Work around for Netscape hang bug. This sends certificate request
15657 and server done in one record. Since this is perfectly legal in the
15658 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15659 the bugs/SSLv3 entry for more info.
15663 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15667 * Add -rand argument to smime and pkcs12 applications and read/write
15672 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15676 * Add command line password options to the remaining applications.
15680 * Bug fix for BN_div_recp() for numerators with an even number of
15685 * More tests in bntest.c, and changed test_bn output.
15689 * ./config recognizes MacOS X now.
15693 * Bug fix for BN_div() when the first words of num and divisor are
15694 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15698 * Add support for various broken PKCS#8 formats, and command line
15699 options to produce them.
15703 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15704 get temporary BIGNUMs from a BN_CTX.
15708 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15713 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15714 include a #define from the old name to the new. The original intent
15715 was that statically linked binaries could for example just call
15716 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15717 link with digests. This never worked because SSLeay_add_all_digests()
15718 and SSLeay_add_all_ciphers() were in the same source file so calling
15719 one would link with the other. They are now in separate source files.
15723 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15727 * Use a less unusual form of the Miller-Rabin primality test (it used
15728 a binary algorithm for exponentiation integrated into the Miller-Rabin
15729 loop, our standard modexp algorithms are faster).
15733 * Support for the EBCDIC character set completed.
15735 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15737 * Source code cleanups: use const where appropriate, eliminate casts,
15738 use `void *` instead of `char *` in lhash.
15742 * Bugfix: ssl3_send_server_key_exchange was not restartable
15743 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15744 this the server could overwrite ephemeral keys that the client
15749 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15750 using 50 iterations of the Rabin-Miller test.
15752 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15753 iterations of the Rabin-Miller test as required by the appendix
15754 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15755 As BN_is_prime_fasttest includes trial division, DSA parameter
15756 generation becomes much faster.
15758 This implies a change for the callback functions in DSA_is_prime
15759 and DSA_generate_parameters: The callback function is called once
15760 for each positive witness in the Rabin-Miller test, not just
15761 occasionally in the inner loop; and the parameters to the
15762 callback function now provide an iteration count for the outer
15763 loop rather than for the current invocation of the inner loop.
15764 DSA_generate_parameters additionally can call the callback
15765 function with an 'iteration count' of -1, meaning that a
15766 candidate has passed the trial division test (when q is generated
15767 from an application-provided seed, trial division is skipped).
15771 * New function BN_is_prime_fasttest that optionally does trial
15772 division before starting the Rabin-Miller test and has
15773 an additional BN_CTX * argument (whereas BN_is_prime always
15774 has to allocate at least one BN_CTX).
15775 'callback(1, -1, cb_arg)' is called when a number has passed the
15776 trial division stage.
15780 * Fix for bug in CRL encoding. The validity dates weren't being handled
15785 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15789 * New function BN_pseudo_rand().
15793 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15794 bignum version of BN_from_montgomery() with the working code from
15795 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15800 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15801 made it impossible to use the same SSL_SESSION data structure in
15802 SSL2 clients in multiple threads.
15806 * The return value of RAND_load_file() no longer counts bytes obtained
15807 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15808 to seed the PRNG (previously an explicit byte count was required).
15810 *Ulf Möller, Bodo Möller*
15812 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15813 used `char *` instead of `void *` and had casts all over the place.
15817 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15821 * Retain source code compatibility for BN_prime_checks macro:
15822 BN_is_prime(..., BN_prime_checks, ...) now uses
15823 BN_prime_checks_for_size to determine the appropriate number of
15824 Rabin-Miller iterations.
15828 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15829 DH_CHECK_P_NOT_SAFE_PRIME.
15830 (Check if this is true? OpenPGP calls them "strong".)
15834 * Merge the functionality of "dh" and "gendh" programs into a new program
15835 "dhparam". The old programs are retained for now but will handle DH keys
15836 (instead of parameters) in future.
15840 * Make the ciphers, s_server and s_client programs check the return values
15841 when a new cipher list is set.
15845 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15846 ciphers. Before when the 56bit ciphers were enabled the sorting was
15849 The syntax for the cipher sorting has been extended to support sorting by
15850 cipher-strength (using the strength_bits hard coded in the tables).
15851 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
15853 Fix a bug in the cipher-command parser: when supplying a cipher command
15854 string with an "undefined" symbol (neither command nor alphanumeric
15855 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15856 an error is flagged.
15858 Due to the strength-sorting extension, the code of the
15859 ssl_create_cipher_list() function was completely rearranged. I hope that
15860 the readability was also increased :-)
15862 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15864 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15865 for the first serial number and places 2 in the serial number file. This
15866 avoids problems when the root CA is created with serial number zero and
15867 the first user certificate has the same issuer name and serial number
15872 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15873 the new code. Add documentation for this stuff.
15877 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15878 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15879 structures and behave in an analogous way to the X509v3 functions:
15880 they shouldn't be called directly but wrapper functions should be used
15883 So we also now have some wrapper functions that call the X509at functions
15884 when passed certificate requests. (TO DO: similar things can be done with
15885 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15886 things. Some of these need some d2i or i2d and print functionality
15887 because they handle more complex structures.)
15891 * Add missing #ifndefs that caused missing symbols when building libssl
15892 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15893 NO_RSA in `ssl/s2*.c`.
15895 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15897 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15898 has a return value which indicates the quality of the random data
15899 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15900 error queue. New function RAND_pseudo_bytes() generates output that is
15901 guaranteed to be unique but not unpredictable. RAND_add is like
15902 RAND_seed, but takes an extra argument for an entropy estimate
15903 (RAND_seed always assumes full entropy).
15907 * Do more iterations of Rabin-Miller probable prime test (specifically,
15908 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15909 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15910 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15911 false-positive rate of at most 2^-80 for random input.
15915 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15919 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15920 in the 0.9.5 release), this returns the chain
15921 from an X509_CTX structure with a dup of the stack and all
15922 the X509 reference counts upped: so the stack will exist
15923 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15926 Also make SSL_SESSION_print() print out the verify return
15931 * Add manpage for the pkcs12 command. Also change the default
15932 behaviour so MAC iteration counts are used unless the new
15933 -nomaciter option is used. This improves file security and
15934 only older versions of MSIE (4.0 for example) need it.
15938 * Honor the no-xxx Configure options when creating .DEF files.
15942 * Add PKCS#10 attributes to field table: challengePassword,
15943 unstructuredName and unstructuredAddress. These are taken from
15944 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15945 international characters are used.
15947 More changes to X509_ATTRIBUTE code: allow the setting of types
15948 based on strings. Remove the 'loc' parameter when adding
15949 attributes because these will be a SET OF encoding which is sorted
15954 * Initial changes to the 'req' utility to allow request generation
15955 automation. This will allow an application to just generate a template
15956 file containing all the field values and have req construct the
15959 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15960 used all over the place including certificate requests and PKCS#7
15961 structures. They are currently handled manually where necessary with
15962 some primitive wrappers for PKCS#7. The new functions behave in a
15963 manner analogous to the X509 extension functions: they allow
15964 attributes to be looked up by NID and added.
15966 Later something similar to the X509V3 code would be desirable to
15967 automatically handle the encoding, decoding and printing of the
15968 more complex types. The string types like challengePassword can
15969 be handled by the string table functions.
15971 Also modified the multi byte string table handling. Now there is
15972 a 'global mask' which masks out certain types. The table itself
15973 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15974 is useful when for example there is only one permissible type
15975 (as in countryName) and using the mask might result in no valid
15980 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15981 SSL_get_peer_finished to allow applications to obtain the latest
15982 Finished messages sent to the peer or expected from the peer,
15983 respectively. (SSL_get_peer_finished is usually the Finished message
15984 actually received from the peer, otherwise the protocol will be aborted.)
15986 As the Finished message are message digests of the complete handshake
15987 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15988 be used for external authentication procedures when the authentication
15989 provided by SSL/TLS is not desired or is not enough.
15993 * Enhanced support for Alpha Linux is added. Now ./config checks if
15994 the host supports BWX extension and if Compaq C is present on the
15995 $PATH. Just exploiting of the BWX extension results in 20-30%
15996 performance kick for some algorithms, e.g. DES and RC4 to mention
15997 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16002 * Add support for MS "fast SGC". This is arguably a violation of the
16003 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16004 weak crypto and after checking the certificate is SGC a second one
16005 with strong crypto. MS SGC stops the first handshake after receiving
16006 the server certificate message and sends a second client hello. Since
16007 a server will typically do all the time consuming operations before
16008 expecting any further messages from the client (server key exchange
16009 is the most expensive) there is little difference between the two.
16011 To get OpenSSL to support MS SGC we have to permit a second client
16012 hello message after we have sent server done. In addition we have to
16013 reset the MAC if we do get this second client hello.
16017 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16018 if a DER encoded private key is RSA or DSA traditional format. Changed
16019 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16020 format DER encoded private key. Newer code should use PKCS#8 format which
16021 has the key type encoded in the ASN1 structure. Added DER private key
16022 support to pkcs8 application.
16026 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16027 ciphersuites has been selected (as required by the SSL 3/TLS 1
16028 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16029 is set, we interpret this as a request to violate the specification
16030 (the worst that can happen is a handshake failure, and 'correct'
16031 behaviour would result in a handshake failure anyway).
16035 * In SSL_CTX_add_session, take into account that there might be multiple
16036 SSL_SESSION structures with the same session ID (e.g. when two threads
16037 concurrently obtain them from an external cache).
16038 The internal cache can handle only one SSL_SESSION with a given ID,
16039 so if there's a conflict, we now throw out the old one to achieve
16044 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16045 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16046 some routines that use cipher OIDs: some ciphers do not have OIDs
16047 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16052 * Simplify the trust setting structure and code. Now we just have
16053 two sequences of OIDs for trusted and rejected settings. These will
16054 typically have values the same as the extended key usage extension
16055 and any application specific purposes.
16057 The trust checking code now has a default behaviour: it will just
16058 check for an object with the same NID as the passed id. Functions can
16059 be provided to override either the default behaviour or the behaviour
16060 for a given id. SSL client, server and email already have functions
16061 in place for compatibility: they check the NID and also return "trusted"
16062 if the certificate is self signed.
16066 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16067 traditional format into an EVP_PKEY structure.
16071 * Add a password callback function PEM_cb() which either prompts for
16072 a password if usr_data is NULL or otherwise assumes it is a null
16073 terminated password. Allow passwords to be passed on command line
16074 environment or config files in a few more utilities.
16078 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16079 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16080 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16081 Update documentation.
16085 * Support for ASN1 "NULL" type. This could be handled before by using
16086 ASN1_TYPE but there wasn't any function that would try to read a NULL
16087 and produce an error if it couldn't. For compatibility we also have
16088 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16089 don't allocate anything because they don't need to.
16093 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16096 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16098 * Rebuild of the memory allocation routines used by OpenSSL code and
16099 possibly others as well. The purpose is to make an interface that
16100 provide hooks so anyone can build a separate set of allocation and
16101 deallocation routines to be used by OpenSSL, for example memory
16102 pool implementations, or something else, which was previously hard
16103 since Malloc(), Realloc() and Free() were defined as macros having
16104 the values malloc, realloc and free, respectively (except for Win32
16105 compilations). The same is provided for memory debugging code.
16106 OpenSSL already comes with functionality to find memory leaks, but
16107 this gives people a chance to debug other memory problems.
16109 With these changes, a new set of functions and macros have appeared:
16111 CRYPTO_set_mem_debug_functions() [F]
16112 CRYPTO_get_mem_debug_functions() [F]
16113 CRYPTO_dbg_set_options() [F]
16114 CRYPTO_dbg_get_options() [F]
16115 CRYPTO_malloc_debug_init() [M]
16117 The memory debug functions are NULL by default, unless the library
16118 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16119 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16120 gives the standard debugging functions that come with OpenSSL) or
16121 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16122 provided by the library user) must be used. When the standard
16123 debugging functions are used, CRYPTO_dbg_set_options can be used to
16124 request additional information:
16125 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16126 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16128 Also, things like CRYPTO_set_mem_functions will always give the
16129 expected result (the new set of functions is used for allocation
16130 and deallocation) at all times, regardless of platform and compiler
16133 To finish it up, some functions that were never use in any other
16134 way than through macros have a new API and new semantic:
16136 CRYPTO_dbg_malloc()
16137 CRYPTO_dbg_realloc()
16140 All macros of value have retained their old syntax.
16142 *Richard Levitte and Bodo Moeller*
16144 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16145 ordering of SMIMECapabilities wasn't in "strength order" and there
16146 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16151 * Some ASN1 types with illegal zero length encoding (INTEGER,
16152 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16154 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16156 * Merge in my S/MIME library for OpenSSL. This provides a simple
16157 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16158 functionality to handle multipart/signed properly) and a utility
16159 called 'smime' to call all this stuff. This is based on code I
16160 originally wrote for Celo who have kindly allowed it to be
16161 included in OpenSSL.
16165 * Add variants des_set_key_checked and des_set_key_unchecked of
16166 des_set_key (aka des_key_sched). Global variable des_check_key
16167 decides which of these is called by des_set_key; this way
16168 des_check_key behaves as it always did, but applications and
16169 the library itself, which was buggy for des_check_key == 1,
16170 have a cleaner way to pick the version they need.
16174 * New function PKCS12_newpass() which changes the password of a
16179 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16180 dynamic mix. In both cases the ids can be used as an index into the
16181 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16182 functions so they accept a list of the field values and the
16183 application doesn't need to directly manipulate the X509_TRUST
16188 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16193 * Modify the way the V3 extension code looks up extensions. This now
16194 works in a similar way to the object code: we have some "standard"
16195 extensions in a static table which is searched with OBJ_bsearch()
16196 and the application can add dynamic ones if needed. The file
16197 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16198 updated whenever a new extension is added to the core code and kept
16199 in ext_nid order. There is a simple program 'tabtest.c' which checks
16200 this. New extensions are not added too often so this file can readily
16201 be maintained manually.
16203 There are two big advantages in doing things this way. The extensions
16204 can be looked up immediately and no longer need to be "added" using
16205 X509V3_add_standard_extensions(): this function now does nothing.
16206 Side note: I get *lots* of email saying the extension code doesn't
16207 work because people forget to call this function.
16208 Also no dynamic allocation is done unless new extensions are added:
16209 so if we don't add custom extensions there is no need to call
16210 X509V3_EXT_cleanup().
16214 * Modify enc utility's salting as follows: make salting the default. Add a
16215 magic header, so unsalted files fail gracefully instead of just decrypting
16216 to garbage. This is because not salting is a big security hole, so people
16217 should be discouraged from doing it.
16221 * Fixes and enhancements to the 'x509' utility. It allowed a message
16222 digest to be passed on the command line but it only used this
16223 parameter when signing a certificate. Modified so all relevant
16224 operations are affected by the digest parameter including the
16225 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16226 DSA key was used because it didn't fix the digest.
16230 * Initial certificate chain verify code. Currently tests the untrusted
16231 certificates for consistency with the verify purpose (which is set
16232 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16234 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16235 this is because it will reject chains with invalid extensions whereas
16236 every previous version of OpenSSL and SSLeay made no checks at all.
16238 Trust code: checks the root CA for the relevant trust settings. Trust
16239 settings have an initial value consistent with the verify purpose: e.g.
16240 if the verify purpose is for SSL client use it expects the CA to be
16241 trusted for SSL client use. However the default value can be changed to
16242 permit custom trust settings: one example of this would be to only trust
16243 certificates from a specific "secure" set of CAs.
16245 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16246 which should be used for version portability: especially since the
16247 verify structure is likely to change more often now.
16249 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16250 to set them. If not set then assume SSL clients will verify SSL servers
16253 Two new options to the verify program: -untrusted allows a set of
16254 untrusted certificates to be passed in and -purpose which sets the
16255 intended purpose of the certificate. If a purpose is set then the
16256 new chain verify code is used to check extension consistency.
16260 * Support for the authority information access extension.
16264 * Modify RSA and DSA PEM read routines to transparently handle
16265 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16266 public keys in a format compatible with certificate
16267 SubjectPublicKeyInfo structures. Unfortunately there were already
16268 functions called *_PublicKey_* which used various odd formats so
16269 these are retained for compatibility: however the DSA variants were
16270 never in a public release so they have been deleted. Changed dsa/rsa
16271 utilities to handle the new format: note no releases ever handled public
16272 keys so we should be OK.
16274 The primary motivation for this change is to avoid the same fiasco
16275 that dogs private keys: there are several incompatible private key
16276 formats some of which are standard and some OpenSSL specific and
16277 require various evil hacks to allow partial transparent handling and
16278 even then it doesn't work with DER formats. Given the option anything
16279 other than PKCS#8 should be dumped: but the other formats have to
16280 stay in the name of compatibility.
16282 With public keys and the benefit of hindsight one standard format
16283 is used which works with EVP_PKEY, RSA or DSA structures: though
16284 it clearly returns an error if you try to read the wrong kind of key.
16286 Added a -pubkey option to the 'x509' utility to output the public key.
16287 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16288 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16289 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16290 that do the same as the `EVP_PKEY_assign_*()` except they up the
16291 reference count of the added key (they don't "swallow" the
16296 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16297 CRLs would fail if the file contained no certificates or no CRLs:
16298 added a new function to read in both types and return the number
16299 read: this means that if none are read it will be an error. The
16300 DER versions of the certificate and CRL reader would always fail
16301 because it isn't possible to mix certificates and CRLs in DER format
16302 without choking one or the other routine. Changed this to just read
16303 a certificate: this is the best we can do. Also modified the code
16304 in `apps/verify.c` to take notice of return codes: it was previously
16305 attempting to read in certificates from NULL pointers and ignoring
16306 any errors: this is one reason why the cert and CRL reader seemed
16307 to work. It doesn't check return codes from the default certificate
16308 routines: these may well fail if the certificates aren't installed.
16312 * Code to support otherName option in GeneralName.
16316 * First update to verify code. Change the verify utility
16317 so it warns if it is passed a self signed certificate:
16318 for consistency with the normal behaviour. X509_verify
16319 has been modified to it will now verify a self signed
16320 certificate if *exactly* the same certificate appears
16321 in the store: it was previously impossible to trust a
16322 single self signed certificate. This means that:
16323 openssl verify ss.pem
16324 now gives a warning about a self signed certificate but
16325 openssl verify -CAfile ss.pem ss.pem
16330 * For servers, store verify_result in SSL_SESSION data structure
16331 (and add it to external session representation).
16332 This is needed when client certificate verifications fails,
16333 but an application-provided verification callback (set by
16334 SSL_CTX_set_cert_verify_callback) allows accepting the session
16335 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16336 but returns 1): When the session is reused, we have to set
16337 ssl->verify_result to the appropriate error code to avoid
16340 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16342 * Fix a bug in the new PKCS#7 code: it didn't consider the
16343 case in PKCS7_dataInit() where the signed PKCS7 structure
16344 didn't contain any existing data because it was being created.
16346 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16348 * Add a salt to the key derivation routines in enc.c. This
16349 forms the first 8 bytes of the encrypted file. Also add a
16350 -S option to allow a salt to be input on the command line.
16354 * New function X509_cmp(). Oddly enough there wasn't a function
16355 to compare two certificates. We do this by working out the SHA1
16356 hash and comparing that. X509_cmp() will be needed by the trust
16361 * SSL_get1_session() is like SSL_get_session(), but increments
16362 the reference count in the SSL_SESSION returned.
16364 *Geoff Thorpe <geoff@eu.c2.net>*
16366 * Fix for 'req': it was adding a null to request attributes.
16367 Also change the X509_LOOKUP and X509_INFO code to handle
16368 certificate auxiliary information.
16372 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16377 * Add the possibility to add extra information to the memory leak
16378 detecting output, to form tracebacks, showing from where each
16379 allocation was originated: CRYPTO_push_info("constant string") adds
16380 the string plus current file name and line number to a per-thread
16381 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16382 is like calling CYRPTO_pop_info() until the stack is empty.
16383 Also updated memory leak detection code to be multi-thread-safe.
16387 * Add options -text and -noout to pkcs7 utility and delete the
16388 encryption options which never did anything. Update docs.
16392 * Add options to some of the utilities to allow the pass phrase
16393 to be included on either the command line (not recommended on
16394 OSes like Unix) or read from the environment. Update the
16395 manpages and fix a few bugs.
16399 * Add a few manpages for some of the openssl commands.
16403 * Fix the -revoke option in ca. It was freeing up memory twice,
16404 leaking and not finding already revoked certificates.
16408 * Extensive changes to support certificate auxiliary information.
16409 This involves the use of X509_CERT_AUX structure and X509_AUX
16410 functions. An X509_AUX function such as PEM_read_X509_AUX()
16411 can still read in a certificate file in the usual way but it
16412 will also read in any additional "auxiliary information". By
16413 doing things this way a fair degree of compatibility can be
16414 retained: existing certificates can have this information added
16415 using the new 'x509' options.
16417 Current auxiliary information includes an "alias" and some trust
16418 settings. The trust settings will ultimately be used in enhanced
16419 certificate chain verification routines: currently a certificate
16420 can only be trusted if it is self signed and then it is trusted
16425 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16426 The problem was that one of the replacement routines had not been working
16427 since SSLeay releases. For now the offending routine has been replaced
16428 with non-optimised assembler. Even so, this now gives around 95%
16429 performance improvement for 1024 bit RSA signs.
16433 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16434 handling. Most clients have the effective key size in bits equal to
16435 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16436 A few however don't do this and instead use the size of the decrypted key
16437 to determine the RC2 key length and the AlgorithmIdentifier to determine
16438 the effective key length. In this case the effective key length can still
16439 be 40 bits but the key length can be 168 bits for example. This is fixed
16440 by manually forcing an RC2 key into the EVP_PKEY structure because the
16441 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16442 the key length and effective key length are equal.
16446 * Add a bunch of functions that should simplify the creation of
16447 X509_NAME structures. Now you should be able to do:
16448 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16449 and have it automatically work out the correct field type and fill in
16450 the structures. The more adventurous can try:
16451 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16452 and it will (hopefully) work out the correct multibyte encoding.
16456 * Change the 'req' utility to use the new field handling and multibyte
16457 copy routines. Before the DN field creation was handled in an ad hoc
16458 way in req, ca, and x509 which was rather broken and didn't support
16459 BMPStrings or UTF8Strings. Since some software doesn't implement
16460 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16461 using the dirstring_type option. See the new comment in the default
16462 openssl.cnf for more info.
16466 * Make crypto/rand/md_rand.c more robust:
16467 - Assure unique random numbers after fork().
16468 - Make sure that concurrent threads access the global counter and
16469 md serializably so that we never lose entropy in them
16470 or use exactly the same state in multiple threads.
16471 Access to the large state is not always serializable because
16472 the additional locking could be a performance killer, and
16473 md should be large enough anyway.
16477 * New file `apps/app_rand.c` with commonly needed functionality
16478 for handling the random seed file.
16480 Use the random seed file in some applications that previously did not:
16482 dsaparam -genkey (which also ignored its '-rand' option),
16485 x509 (when signing).
16486 Except on systems with /dev/urandom, it is crucial to have a random
16487 seed file at least for key creation, DSA signing, and for DH exchanges;
16488 for RSA signatures we could do without one.
16490 gendh and gendsa (unlike genrsa) used to read only the first byte
16491 of each file listed in the '-rand' option. The function as previously
16492 found in genrsa is now in app_rand.c and is used by all programs
16493 that support '-rand'.
16497 * In RAND_write_file, use mode 0600 for creating files;
16498 don't just chmod when it may be too late.
16502 * Report an error from X509_STORE_load_locations
16503 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16507 * New function ASN1_mbstring_copy() this copies a string in either
16508 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16509 into an ASN1_STRING type. A mask of permissible types is passed
16510 and it chooses the "minimal" type to use or an error if not type
16515 * Add function equivalents to the various macros in asn1.h. The old
16516 macros are retained with an `M_` prefix. Code inside the library can
16517 use the `M_` macros. External code (including the openssl utility)
16518 should *NOT* in order to be "shared library friendly".
16522 * Add various functions that can check a certificate's extensions
16523 to see if it usable for various purposes such as SSL client,
16524 server or S/MIME and CAs of these types. This is currently
16525 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16526 verification. Also added a -purpose flag to x509 utility to
16527 print out all the purposes.
16531 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16536 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16537 for, obtain and decode and extension and obtain its critical flag.
16538 This allows all the necessary extension code to be handled in a
16539 single function call.
16543 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16544 platforms. See crypto/rc4/rc4_enc.c for further details.
16548 * New -noout option to asn1parse. This causes no output to be produced
16549 its main use is when combined with -strparse and -out to extract data
16550 from a file (which may not be in ASN.1 format).
16554 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16555 when producing the local key id.
16557 *Richard Levitte <levitte@stacken.kth.se>*
16559 * New option -dhparam in s_server. This allows a DH parameter file to be
16560 stated explicitly. If it is not stated then it tries the first server
16561 certificate file. The previous behaviour hard coded the filename
16566 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16567 a public key to be input or output. For example:
16568 openssl rsa -in key.pem -pubout -out pubkey.pem
16569 Also added necessary DSA public key functions to handle this.
16573 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16574 in the message. This was handled by allowing
16575 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16577 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16579 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16580 to the end of the strings whereas this didn't. This would cause problems
16581 if strings read with d2i_ASN1_bytes() were later modified.
16583 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16585 * Fix for base64 decode bug. When a base64 bio reads only one line of
16586 data and it contains EOF it will end up returning an error. This is
16587 caused by input 46 bytes long. The cause is due to the way base64
16588 BIOs find the start of base64 encoded data. They do this by trying a
16589 trial decode on each line until they find one that works. When they
16590 do a flag is set and it starts again knowing it can pass all the
16591 data directly through the decoder. Unfortunately it doesn't reset
16592 the context it uses. This means that if EOF is reached an attempt
16593 is made to pass two EOFs through the context and this causes the
16594 resulting error. This can also cause other problems as well. As is
16595 usual with these problems it takes *ages* to find and the fix is
16596 trivial: move one line.
16598 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16600 * Ugly workaround to get s_client and s_server working under Windows. The
16601 old code wouldn't work because it needed to select() on sockets and the
16602 tty (for keypresses and to see if data could be written). Win32 only
16603 supports select() on sockets so we select() with a 1s timeout on the
16604 sockets and then see if any characters are waiting to be read, if none
16605 are present then we retry, we also assume we can always write data to
16606 the tty. This isn't nice because the code then blocks until we've
16607 received a complete line of data and it is effectively polling the
16608 keyboard at 1s intervals: however it's quite a bit better than not
16609 working at all :-) A dedicated Windows application might handle this
16610 with an event loop for example.
16614 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16615 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16616 will be called when RSA_sign() and RSA_verify() are used. This is useful
16617 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16618 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16619 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16620 This necessitated the support of an extra signature type NID_md5_sha1
16621 for SSL signatures and modifications to the SSL library to use it instead
16622 of calling RSA_public_decrypt() and RSA_private_encrypt().
16626 * Add new -verify -CAfile and -CApath options to the crl program, these
16627 will lookup a CRL issuers certificate and verify the signature in a
16628 similar way to the verify program. Tidy up the crl program so it
16629 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16630 less strict. It will now permit CRL extensions even if it is not
16631 a V2 CRL: this will allow it to tolerate some broken CRLs.
16635 * Initialize all non-automatic variables each time one of the openssl
16636 sub-programs is started (this is necessary as they may be started
16637 multiple times from the "OpenSSL>" prompt).
16639 *Lennart Bang, Bodo Moeller*
16641 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16642 removing all other RSA functionality (this is what NO_RSA does). This
16643 is so (for example) those in the US can disable those operations covered
16644 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16649 * Non-copying interface to BIO pairs.
16650 (still largely untested)
16654 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16655 ASCII string. This was handled independently in various places before.
16659 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16660 UTF8 strings a character at a time.
16664 * Use client_version from client hello to select the protocol
16665 (s23_srvr.c) and for RSA client key exchange verification
16666 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16670 * Add various utility functions to handle SPKACs, these were previously
16671 handled by poking round in the structure internals. Added new function
16672 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16673 print, verify and generate SPKACs. Based on an original idea from
16674 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16678 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16682 * Allow the config file extension section to be overwritten on the
16683 command line. Based on an original idea from Massimiliano Pala
16684 <madwolf@comune.modena.it>. The new option is called -extensions
16685 and can be applied to ca, req and x509. Also -reqexts to override
16686 the request extensions in req and -crlexts to override the crl extensions
16691 * Add new feature to the SPKAC handling in ca. Now you can include
16692 the same field multiple times by preceding it by "XXXX." for example:
16695 this is the same syntax as used in the req config file.
16699 * Allow certificate extensions to be added to certificate requests. These
16700 are specified in a 'req_extensions' option of the req section of the
16701 config file. They can be printed out with the -text option to req but
16702 are otherwise ignored at present.
16706 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16707 data read consists of only the final block it would not decrypted because
16708 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16709 A misplaced 'break' also meant the decrypted final block might not be
16710 copied until the next read.
16714 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16715 a few extra parameters to the DH structure: these will be useful if
16716 for example we want the value of 'q' or implement X9.42 DH.
16720 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16721 provides hooks that allow the default DSA functions or functions on a
16722 "per key" basis to be replaced. This allows hardware acceleration and
16723 hardware key storage to be handled without major modification to the
16724 library. Also added low-level modexp hooks and CRYPTO_EX structure and
16725 associated functions.
16729 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16730 as "read only": it can't be written to and the buffer it points to will
16731 not be freed. Reading from a read only BIO is much more efficient than
16732 a normal memory BIO. This was added because there are several times when
16733 an area of memory needs to be read from a BIO. The previous method was
16734 to create a memory BIO and write the data to it, this results in two
16735 copies of the data and an O(n^2) reading algorithm. There is a new
16736 function BIO_new_mem_buf() which creates a read only memory BIO from
16737 an area of memory. Also modified the PKCS#7 routines to use read only
16742 * Bugfix: ssl23_get_client_hello did not work properly when called in
16743 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16744 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16745 but a retry condition occurred while trying to read the rest.
16749 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16750 NID_pkcs7_encrypted by default: this was wrong since this should almost
16751 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16752 the encrypted data type: this is a more sensible place to put it and it
16753 allows the PKCS#12 code to be tidied up that duplicated this
16758 * Changed obj_dat.pl script so it takes its input and output files on
16759 the command line. This should avoid shell escape redirection problems
16764 * Initial support for certificate extension requests, these are included
16765 in things like Xenroll certificate requests. Included functions to allow
16766 extensions to be obtained and added.
16770 * -crlf option to s_client and s_server for sending newlines as
16771 CRLF (as required by many protocols).
16775 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16777 * Install libRSAglue.a when OpenSSL is built with RSAref.
16779 *Ralf S. Engelschall*
16781 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16783 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16785 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16790 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16791 DH parameters/keys (q is lost during that conversion, but the resulting
16792 DH parameters contain its length).
16794 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16795 much faster than DH_generate_parameters (which creates parameters
16796 where `p = 2*q + 1`), and also the smaller q makes DH computations
16797 much more efficient (160-bit exponentiation instead of 1024-bit
16798 exponentiation); so this provides a convenient way to support DHE
16799 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16800 utter importance to use
16801 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16803 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16804 when such DH parameters are used, because otherwise small subgroup
16805 attacks may become possible!
16809 * Avoid memory leak in i2d_DHparams.
16813 * Allow the -k option to be used more than once in the enc program:
16814 this allows the same encrypted message to be read by multiple recipients.
16818 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16819 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16820 it will always use the numerical form of the OID, even if it has a short
16825 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16826 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16827 otherwise bn_mod_exp was called. In the case of hardware keys for example
16828 no private key components need be present and it might store extra data
16829 in the RSA structure, which cannot be accessed from bn_mod_exp.
16830 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16831 private key operations.
16835 * Added support for SPARC Linux.
16839 * pem_password_cb function type incompatibly changed from
16840 typedef int pem_password_cb(char *buf, int size, int rwflag);
16842 ....(char *buf, int size, int rwflag, void *userdata);
16843 so that applications can pass data to their callbacks:
16844 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16845 additional void * argument, which is just handed through whenever
16846 the password callback is called.
16848 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16850 New function SSL_CTX_set_default_passwd_cb_userdata.
16852 Compatibility note: As many C implementations push function arguments
16853 onto the stack in reverse order, the new library version is likely to
16854 interoperate with programs that have been compiled with the old
16855 pem_password_cb definition (PEM_whatever takes some data that
16856 happens to be on the stack as its last argument, and the callback
16857 just ignores this garbage); but there is no guarantee whatsoever that
16860 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16861 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16862 problems not only on Windows, but also on some Unix platforms.
16863 To avoid problematic command lines, these definitions are now in an
16864 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16865 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16869 * MIPS III/IV assembler module is reimplemented.
16873 * More DES library cleanups: remove references to srand/rand and
16874 delete an unused file.
16878 * Add support for the free Netwide assembler (NASM) under Win32,
16879 since not many people have MASM (ml) and it can be hard to obtain.
16880 This is currently experimental but it seems to work OK and pass all
16881 the tests. Check out INSTALL.W32 for info.
16885 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16886 without temporary keys kept an extra copy of the server key,
16887 and connections with temporary keys did not free everything in case
16892 * New function RSA_check_key and new openssl rsa option -check
16893 for verifying the consistency of RSA keys.
16895 *Ulf Moeller, Bodo Moeller*
16897 * Various changes to make Win32 compile work:
16898 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16899 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16900 comparison" warnings.
16901 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16905 * Add a debugging option to PKCS#5 v2 key generation function: when
16906 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16907 derived keys are printed to stderr.
16911 * Copy the flags in ASN1_STRING_dup().
16913 *Roman E. Pavlov <pre@mo.msk.ru>*
16915 * The x509 application mishandled signing requests containing DSA
16916 keys when the signing key was also DSA and the parameters didn't match.
16918 It was supposed to omit the parameters when they matched the signing key:
16919 the verifying software was then supposed to automatically use the CA's
16920 parameters if they were absent from the end user certificate.
16922 Omitting parameters is no longer recommended. The test was also
16923 the wrong way round! This was probably due to unusual behaviour in
16924 EVP_cmp_parameters() which returns 1 if the parameters match.
16925 This meant that parameters were omitted when they *didn't* match and
16926 the certificate was useless. Certificates signed with 'ca' didn't have
16929 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16931 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16932 The interface is as follows:
16933 Applications can use
16934 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16935 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16936 "off" is now the default.
16937 The library internally uses
16938 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16939 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16940 to disable memory-checking temporarily.
16942 Some inconsistent states that previously were possible (and were
16943 even the default) are now avoided.
16945 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16946 with each memory chunk allocated; this is occasionally more helpful
16947 than just having a counter.
16949 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16951 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16956 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16957 which largely parallels "options", but is for changing API behaviour,
16958 whereas "options" are about protocol behaviour.
16959 Initial "mode" flags are:
16961 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16962 a single record has been written.
16963 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16964 retries use the same buffer location.
16965 (But all of the contents must be
16970 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16973 * Fix problems with no-hmac etc.
16975 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16977 * New functions RSA_get_default_method(), RSA_set_method() and
16978 RSA_get_method(). These allows replacement of RSA_METHODs without having
16979 to mess around with the internals of an RSA structure.
16983 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16984 Also really enable memory leak checks in openssl.c and in some
16987 *Chad C. Mulligan, Bodo Moeller*
16989 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16990 up the length of negative integers. This has now been simplified to just
16991 store the length when it is first determined and use it later, rather
16992 than trying to keep track of where data is copied and updating it to
16994 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16996 * Add a new function PKCS7_signatureVerify. This allows the verification
16997 of a PKCS#7 signature but with the signing certificate passed to the
16998 function itself. This contrasts with PKCS7_dataVerify which assumes the
16999 certificate is present in the PKCS#7 structure. This isn't always the
17000 case: certificates can be omitted from a PKCS#7 structure and be
17001 distributed by "out of band" means (such as a certificate database).
17005 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17006 function prototypes in pem.h, also change util/mkdef.pl to add the
17007 necessary function names.
17011 * mk1mf.pl (used by Windows builds) did not properly read the
17012 options set by Configure in the top level Makefile, and Configure
17013 was not even able to write more than one option correctly.
17014 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17018 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17019 file to be loaded from a BIO or FILE pointer. The BIO version will
17020 for example allow memory BIOs to contain config info.
17024 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17025 Whoever hopes to achieve shared-library compatibility across versions
17026 must use this, not the compile-time macro.
17027 (Exercise 0.9.4: Which is the minimum library version required by
17029 Note: All this applies only to multi-threaded programs, others don't
17034 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17035 through a BIO pair triggered the default case, i.e.
17036 SSLerr(...,SSL_R_UNKNOWN_STATE).
17040 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17041 can use the SSL library even if none of the specific BIOs is
17046 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17047 for the encoded length.
17049 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17051 * Add initial documentation of the X509V3 functions.
17055 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17056 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17057 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17058 secure PKCS#8 private key format with a high iteration count.
17062 * Fix determination of Perl interpreter: A perl or perl5
17063 *directory* in $PATH was also accepted as the interpreter.
17065 *Ralf S. Engelschall*
17067 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17068 wrong with it but it was very old and did things like calling
17069 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17070 unusual formatting.
17074 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17075 to use the new extension code.
17079 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17080 with macros. This should make it easier to change their form, add extra
17081 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17086 * Add to configuration table a new entry that can specify an alternative
17087 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17088 according to Mark Crispin <MRC@Panda.COM>.
17092 * DES CBC did not update the IV. Weird.
17096 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17097 Changing the behaviour of the former might break existing programs --
17098 where IV updating is needed, des_ncbc_encrypt can be used.
17101 * When bntest is run from "make test" it drives bc to check its
17102 calculations, as well as internally checking them. If an internal check
17103 fails, it needs to cause bc to give a non-zero result or make test carries
17104 on without noticing the failure. Fixed.
17108 * DES library cleanups.
17112 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17113 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17114 ciphers. NOTE: although the key derivation function has been verified
17115 against some published test vectors it has not been extensively tested
17116 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17121 * Instead of "mkdir -p", which is not fully portable, use new
17122 Perl script "util/mkdir-p.pl".
17126 * Rewrite the way password based encryption (PBE) is handled. It used to
17127 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17128 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17129 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17130 the 'parameter' field of the AlgorithmIdentifier is passed to the
17131 underlying key generation function so it must do its own ASN1 parsing.
17132 This has also changed the EVP_PBE_CipherInit() function which now has a
17133 'parameter' argument instead of literal salt and iteration count values
17134 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17138 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17139 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17140 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17141 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17142 value was just used as a "magic string" and not used directly its
17143 value doesn't matter.
17147 * Introduce some semblance of const correctness to BN. Shame C doesn't
17152 * "linux-sparc64" configuration (ultrapenguin).
17154 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17155 "linux-sparc" configuration.
17157 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17159 * config now generates no-xxx options for missing ciphers.
17163 * Support the EBCDIC character set (work in progress).
17164 File ebcdic.c not yet included because it has a different license.
17166 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17168 * Support BS2000/OSD-POSIX.
17170 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17172 * Make callbacks for key generation use `void *` instead of `char *`.
17176 * Make S/MIME samples compile (not yet tested).
17180 * Additional typesafe stacks.
17184 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17188 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17190 * New configuration variant "sco5-gcc".
17192 * Updated some demos.
17194 *Sean O Riordain, Wade Scholine*
17196 * Add missing BIO_free at exit of pkcs12 application.
17200 * Fix memory leak in conf.c.
17204 * Updates for Win32 to assembler version of MD5.
17208 * Set #! path to perl in `apps/der_chop` to where we found it
17209 instead of using a fixed path.
17213 * SHA library changes for irix64-mips4-cc.
17217 * Improvements for VMS support.
17221 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17223 * Bignum library bug fix. IRIX 6 passes "make test" now!
17224 This also avoids the problems with SC4.2 and unpatched SC5.
17226 *Andy Polyakov <appro@fy.chalmers.se>*
17228 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17229 These are required because of the typesafe stack would otherwise break
17230 existing code. If old code used a structure member which used to be STACK
17231 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17232 sk_num or sk_value it would produce an error because the num, data members
17233 are not present in STACK_OF. Now it just produces a warning. sk_set
17234 replaces the old method of assigning a value to sk_value
17235 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17236 that does this will no longer work (and should use sk_set instead) but
17237 this could be regarded as a "questionable" behaviour anyway.
17241 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17242 correctly handle encrypted S/MIME data.
17246 * Change type of various DES function arguments from des_cblock
17247 (which means, in function argument declarations, pointer to char)
17248 to des_cblock * (meaning pointer to array with 8 char elements),
17249 which allows the compiler to do more typechecking; it was like
17250 that back in SSLeay, but with lots of ugly casts.
17252 Introduce new type const_des_cblock.
17256 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17257 problems: find RecipientInfo structure that matches recipient certificate
17258 and initialise the ASN1 structures properly based on passed cipher.
17262 * Belatedly make the BN tests actually check the results.
17266 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17267 to and from BNs: it was completely broken. New compilation option
17268 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17269 key elements as negative integers.
17273 * Reorganize and speed up MD5.
17275 *Andy Polyakov <appro@fy.chalmers.se>*
17279 *Richard Levitte <richard@levitte.org>*
17281 * New option -out to asn1parse to allow the parsed structure to be
17282 output to a file. This is most useful when combined with the -strparse
17283 option to examine the output of things like OCTET STRINGS.
17287 * Make SSL library a little more fool-proof by not requiring any longer
17288 that `SSL_set_{accept,connect}_state` be called before
17289 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17290 in many applications because usually everything *appeared* to work as
17291 intended anyway -- now it really works as intended).
17295 * Move openssl.cnf out of lib/.
17299 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17300 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17301 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17303 *Ralf S. Engelschall*
17305 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17306 handle PKCS#7 enveloped data properly.
17308 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17310 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17311 copying pointers. The cert_st handling is changed by this in
17312 various ways (and thus what used to be known as ctx->default_cert
17313 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17314 any longer when s->cert does not give us what we need).
17315 ssl_cert_instantiate becomes obsolete by this change.
17316 As soon as we've got the new code right (possibly it already is?),
17317 we have solved a couple of bugs of the earlier code where s->cert
17318 was used as if it could not have been shared with other SSL structures.
17320 Note that using the SSL API in certain dirty ways now will result
17321 in different behaviour than observed with earlier library versions:
17322 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17323 does not influence s as it used to.
17325 In order to clean up things more thoroughly, inside SSL_SESSION
17326 we don't use CERT any longer, but a new structure SESS_CERT
17327 that holds per-session data (if available); currently, this is
17328 the peer's certificate chain and, for clients, the server's certificate
17329 and temporary key. CERT holds only those values that can have
17330 meaningful defaults in an SSL_CTX.
17334 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17335 from the internal representation. Various PKCS#7 fixes: remove some
17336 evil casts and set the enc_dig_alg field properly based on the signing
17341 * Allow PKCS#12 password to be set from the command line or the
17342 environment. Let 'ca' get its config file name from the environment
17343 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17348 * Allow certificate policies extension to use an IA5STRING for the
17349 organization field. This is contrary to the PKIX definition but
17350 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17355 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17356 without disallowing inline assembler and the like for non-pedantic builds.
17360 * Support Borland C++ builder.
17362 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17368 * SHA-1 cleanups and performance enhancements.
17370 *Andy Polyakov <appro@fy.chalmers.se>*
17372 * Sparc v8plus assembler for the bignum library.
17374 *Andy Polyakov <appro@fy.chalmers.se>*
17376 * Accept any -xxx and +xxx compiler options in Configure.
17380 * Update HPUX configuration.
17384 * Add missing `sk_<type>_unshift()` function to safestack.h
17386 *Ralf S. Engelschall*
17388 * New function SSL_CTX_use_certificate_chain_file that sets the
17389 "extra_cert"s in addition to the certificate. (This makes sense
17390 only for "PEM" format files, as chains as a whole are not
17395 * Support verify_depth from the SSL API.
17396 x509_vfy.c had what can be considered an off-by-one-error:
17397 Its depth (which was not part of the external interface)
17398 was actually counting the number of certificates in a chain;
17399 now it really counts the depth.
17403 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17404 instead of X509err, which often resulted in confusing error
17405 messages since the error codes are not globally unique
17406 (e.g. an alleged error in ssl3_accept when a certificate
17407 didn't match the private key).
17409 * New function SSL_CTX_set_session_id_context that allows to set a default
17410 value (so that you don't need SSL_set_session_id_context for each
17411 connection using the SSL_CTX).
17415 * OAEP decoding bug fix.
17419 * Support INSTALL_PREFIX for package builders, as proposed by
17424 * New Configure options "threads" and "no-threads". For systems
17425 where the proper compiler options are known (currently Solaris
17426 and Linux), "threads" is the default.
17430 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17434 * Install various scripts to $(OPENSSLDIR)/misc, not to
17435 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17436 such as /usr/local/bin.
17440 * "make linux-shared" to build shared libraries.
17442 *Niels Poppe <niels@netbox.org>*
17444 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17448 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17449 extension adding in x509 utility.
17453 * Remove NOPROTO sections and error code comments.
17457 * Partial rewrite of the DEF file generator to now parse the ANSI
17462 * New Configure options --prefix=DIR and --openssldir=DIR.
17466 * Complete rewrite of the error code script(s). It is all now handled
17467 by one script at the top level which handles error code gathering,
17468 header rewriting and C source file generation. It should be much better
17469 than the old method: it now uses a modified version of Ulf's parser to
17470 read the ANSI prototypes in all header files (thus the old K&R definitions
17471 aren't needed for error creation any more) and do a better job of
17472 translating function codes into names. The old 'ASN1 error code embedded
17473 in a comment' is no longer necessary and it doesn't use .err files which
17474 have now been deleted. Also the error code call doesn't have to appear all
17475 on one line (which resulted in some large lines...).
17479 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17483 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17484 0 (which usually indicates a closed connection), but continue reading.
17488 * Fix some race conditions.
17492 * Add support for CRL distribution points extension. Add Certificate
17493 Policies and CRL distribution points documentation.
17497 * Move the autogenerated header file parts to crypto/opensslconf.h.
17501 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17502 8 of keying material. Merlin has also confirmed interop with this fix
17503 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17505 *Merlin Hughes <merlin@baltimore.ie>*
17507 * Fix lots of warnings.
17509 *Richard Levitte <levitte@stacken.kth.se>*
17511 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17512 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17514 *Richard Levitte <levitte@stacken.kth.se>*
17516 * Fix problems with sizeof(long) == 8.
17518 *Andy Polyakov <appro@fy.chalmers.se>*
17520 * Change functions to ANSI C.
17524 * Fix typos in error codes.
17526 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17528 * Remove defunct assembler files from Configure.
17532 * SPARC v8 assembler BIGNUM implementation.
17534 *Andy Polyakov <appro@fy.chalmers.se>*
17536 * Support for Certificate Policies extension: both print and set.
17537 Various additions to support the r2i method this uses.
17541 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17542 return a const string when you are expecting an allocated buffer.
17546 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17547 types DirectoryString and DisplayText.
17551 * Add code to allow r2i extensions to access the configuration database,
17552 add an LHASH database driver and add several ctx helper functions.
17556 * Fix an evil bug in bn_expand2() which caused various BN functions to
17557 fail when they extended the size of a BIGNUM.
17561 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17562 support typesafe stack.
17566 * Fix typo in SSL_[gs]et_options().
17568 *Nils Frostberg <nils@medcom.se>*
17570 * Delete various functions and files that belonged to the (now obsolete)
17571 old X509V3 handling code.
17575 * New Configure option "rsaref".
17579 * Don't auto-generate pem.h.
17583 * Introduce type-safe ASN.1 SETs.
17587 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17589 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17591 * Introduce type-safe STACKs. This will almost certainly break lots of code
17592 that links with OpenSSL (well at least cause lots of warnings), but fear
17593 not: the conversion is trivial, and it eliminates loads of evil casts. A
17594 few STACKed things have been converted already. Feel free to convert more.
17595 In the fullness of time, I'll do away with the STACK type altogether.
17599 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17600 specified in `<certfile>` by updating the entry in the index.txt file.
17601 This way one no longer has to edit the index.txt file manually for
17602 revoking a certificate. The -revoke option does the gory details now.
17604 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17606 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17607 `-text` option at all and this way the `-noout -text` combination was
17608 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17610 *Ralf S. Engelschall*
17612 * Make sure a corresponding plain text error message exists for the
17613 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17614 verify callback function determined that a certificate was revoked.
17616 *Ralf S. Engelschall*
17618 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17619 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17620 all available ciphers including rc5, which was forgotten until now.
17621 In order to let the testing shell script know which algorithms
17622 are available, a new (up to now undocumented) command
17623 `openssl list-cipher-commands` is used.
17627 * Bugfix: s_client occasionally would sleep in select() when
17628 it should have checked SSL_pending() first.
17632 * New functions DSA_do_sign and DSA_do_verify to provide access to
17633 the raw DSA values prior to ASN.1 encoding.
17637 * Tweaks to Configure
17639 *Niels Poppe <niels@netbox.org>*
17641 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17646 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17650 * New config option to avoid instructions that are illegal on the 80386.
17651 The default code is faster, but requires at least a 486.
17655 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17656 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17657 same as SSL2_VERSION anyway.
17661 * New "-showcerts" option for s_client.
17665 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17666 application. Various cleanups and fixes.
17670 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17671 modify error routines to work internally. Add error codes and PBE init
17672 to library startup routines.
17676 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17677 packing functions to asn1 and evp. Changed function names and error
17678 codes along the way.
17682 * PKCS12 integration: and so it begins... First of several patches to
17683 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17684 objects to objects.h
17688 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17689 and display support for Thawte strong extranet extension.
17693 * Add LinuxPPC support.
17695 *Jeff Dubrule <igor@pobox.org>*
17697 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17698 bn_div_words in alpha.s.
17700 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17702 * Make sure the RSA OAEP test is skipped under -DRSAref because
17703 OAEP isn't supported when OpenSSL is built with RSAref.
17705 *Ulf Moeller <ulf@fitug.de>*
17707 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17708 so they no longer are missing under -DNOPROTO.
17710 *Soren S. Jorvang <soren@t.dk>*
17712 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17714 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17715 doesn't work when the session is reused. Coming soon!
17719 * Fix a security hole, that allows sessions to be reused in the wrong
17720 context thus bypassing client cert protection! All software that uses
17721 client certs and session caches in multiple contexts NEEDS PATCHING to
17722 allow session reuse! A fuller solution is in the works.
17724 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17726 * Some more source tree cleanups (removed obsolete files
17727 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17728 permission on "config" script to be executable) and a fix for the INSTALL
17731 *Ulf Moeller <ulf@fitug.de>*
17733 * Remove some legacy and erroneous uses of malloc, free instead of
17736 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17738 * Make rsa_oaep_test return non-zero on error.
17740 *Ulf Moeller <ulf@fitug.de>*
17742 * Add support for native Solaris shared libraries. Configure
17743 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17744 if someone would make that last step automatic.
17746 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17748 * ctx_size was not built with the right compiler during "make links". Fixed.
17752 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17753 except NULL ciphers". This means the default cipher list will no longer
17754 enable NULL ciphers. They need to be specifically enabled e.g. with
17755 the string "DEFAULT:eNULL".
17759 * Fix to RSA private encryption routines: if p < q then it would
17760 occasionally produce an invalid result. This will only happen with
17761 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17765 * Be less restrictive and allow also `perl util/perlpath.pl
17766 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17767 because this way one can also use an interpreter named `perl5` (which is
17768 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17769 installed as `perl`).
17771 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17773 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17775 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17777 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17778 advapi32.lib to Win32 build and change the pem test comparison
17779 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17780 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17781 and crypto/des/ede_cbcm_enc.c.
17785 * DES quad checksum was broken on big-endian architectures. Fixed.
17789 * Comment out two functions in bio.h that aren't implemented. Fix up the
17790 Win32 test batch file so it (might) work again. The Win32 test batch file
17791 is horrible: I feel ill....
17795 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17796 in e_os.h. Audit of header files to check ANSI and non ANSI
17797 sections: 10 functions were absent from non ANSI section and not exported
17798 from Windows DLLs. Fixed up libeay.num for new functions.
17802 * Make `openssl version` output lines consistent.
17804 *Ralf S. Engelschall*
17806 * Fix Win32 symbol export lists for BIO functions: Added
17807 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17808 to ms/libeay{16,32}.def.
17810 *Ralf S. Engelschall*
17812 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17813 fine under Unix and passes some trivial tests I've now added. But the
17814 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17815 added to make sure no one expects that this stuff really works in the
17816 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17817 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17820 *Ralf S. Engelschall*
17822 * Fix the generation of two part addresses in perl.
17824 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17826 * Add config entry for Linux on MIPS.
17828 *John Tobey <jtobey@channel1.com>*
17830 * Make links whenever Configure is run, unless we are on Windoze.
17834 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17835 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17840 * Add a useful kludge to allow package maintainers to specify compiler and
17841 other platforms details on the command line without having to patch the
17842 Configure script every time: One now can use
17843 `perl Configure <id>:<details>`,
17844 i.e. platform ids are allowed to have details appended
17845 to them (separated by colons). This is treated as there would be a static
17846 pre-configured entry in Configure's %table under key `<id>` with value
17847 `<details>` and `perl Configure <id>` is called. So, when you want to
17848 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17849 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17850 now, which overrides the FreeBSD-elf entry on-the-fly.
17852 *Ralf S. Engelschall*
17854 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17858 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17859 on the `perl Configure ...` command line. This way one can compile
17860 OpenSSL libraries with Position Independent Code (PIC) which is needed
17861 for linking it into DSOs.
17863 *Ralf S. Engelschall*
17865 * Remarkably, export ciphers were totally broken and no-one had noticed!
17870 * Cleaned up the LICENSE document: The official contact for any license
17871 questions now is the OpenSSL core team under openssl-core@openssl.org.
17872 And add a paragraph about the dual-license situation to make sure people
17873 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17874 to the OpenSSL toolkit.
17876 *Ralf S. Engelschall*
17878 * General source tree makefile cleanups: Made `making xxx in yyy...`
17879 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17880 Additionally cleaned up the `make links` target: Remove unnecessary
17881 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17882 to speed processing and no longer clutter the display with confusing
17883 stuff. Instead only the actually done links are displayed.
17885 *Ralf S. Engelschall*
17887 * Permit null encryption ciphersuites, used for authentication only. It used
17888 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17889 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17894 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17895 signed attributes when verifying signatures (this would break them),
17896 the detached data encoding was wrong and public keys obtained using
17897 X509_get_pubkey() weren't freed.
17901 * Add text documentation for the BUFFER functions. Also added a work around
17902 to a Win95 console bug. This was triggered by the password read stuff: the
17903 last character typed gets carried over to the next fread(). If you were
17904 generating a new cert request using 'req' for example then the last
17905 character of the passphrase would be CR which would then enter the first
17910 * Added the new 'Includes OpenSSL Cryptography Software' button as
17911 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17912 button and can be used by applications based on OpenSSL to show the
17913 relationship to the OpenSSL project.
17915 *Ralf S. Engelschall*
17917 * Remove confusing variables in function signatures in files
17918 ssl/ssl_lib.c and ssl/ssl.h.
17920 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17922 * Don't install bss_file.c under PREFIX/include/
17924 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17926 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17927 functions that return function pointers and has support for NT specific
17928 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17929 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17930 unsigned to signed types: this was killing the Win32 compile.
17934 * Add new certificate file to stack functions,
17935 SSL_add_dir_cert_subjects_to_stack() and
17936 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17937 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17938 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17939 This means that Apache-SSL and similar packages don't have to mess around
17940 to add as many CAs as they want to the preferred list.
17944 * Experiment with doxygen documentation. Currently only partially applied to
17946 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17947 openssl.doxy as the configuration file.
17951 * Get rid of remaining C++-style comments which strict C compilers hate.
17953 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17955 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17956 compiled in by default: it has problems with large keys.
17960 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17961 DH private keys and/or callback functions which directly correspond to
17962 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17963 is needed for applications which have to configure certificates on a
17964 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17966 For the RSA certificate situation is makes no difference, but
17967 for the DSA certificate situation this fixes the "no shared cipher"
17968 problem where the OpenSSL cipher selection procedure failed because the
17969 temporary keys were not overtaken from the context and the API provided
17970 no way to reconfigure them.
17971 The new functions now let applications reconfigure the stuff and they
17972 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17973 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17974 non-public-API function ssl_cert_instantiate() is used as a helper
17975 function and also to reduce code redundancy inside ssl_rsa.c.
17977 *Ralf S. Engelschall*
17979 * Move s_server -dcert and -dkey options out of the undocumented feature
17980 area because they are useful for the DSA situation and should be
17981 recognized by the users.
17983 *Ralf S. Engelschall*
17985 * Fix the cipher decision scheme for export ciphers: the export bits are
17986 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17987 SSL_EXP_MASK. So, the original variable has to be used instead of the
17988 already masked variable.
17990 *Richard Levitte <levitte@stacken.kth.se>*
17992 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17994 *Richard Levitte <levitte@stacken.kth.se>*
17996 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17997 from `int` to `unsigned int` because it is a length and initialized by
17998 EVP_DigestFinal() which expects an `unsigned int *`.
18000 *Richard Levitte <levitte@stacken.kth.se>*
18002 * Don't hard-code path to Perl interpreter on shebang line of Configure
18003 script. Instead use the usual Shell->Perl transition trick.
18005 *Ralf S. Engelschall*
18007 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18008 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18009 -noout -modulus` as it's already the case for `openssl rsa -noout
18010 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18011 currently the public key is printed (a decision which was already done by
18012 `openssl dsa -modulus` in the past) which serves a similar purpose.
18013 Additionally the NO_RSA no longer completely removes the whole -modulus
18014 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18017 *Ralf S. Engelschall*
18019 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18020 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18022 *Arne Ansper <arne@ats.cyber.ee>*
18024 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18025 to be added. Now both 'req' and 'ca' can use new objects defined in the
18030 * Add cool BIO that does syslog (or event log on NT).
18032 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18034 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18035 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18036 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18037 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18041 * Add preliminary config info for new extension code.
18045 * Make RSA_NO_PADDING really use no padding.
18047 *Ulf Moeller <ulf@fitug.de>*
18049 * Generate errors when private/public key check is done.
18053 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18054 for some CRL extensions and new objects added.
18058 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18059 key usage extension and fuller support for authority key id.
18063 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18064 padding method for RSA, which is recommended for new applications in PKCS
18065 #1 v2.0 (RFC 2437, October 1998).
18066 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18067 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18068 against Bleichbacher's attack on RSA.
18069 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18072 * Updates to the new SSL compression code
18074 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18076 * Fix so that the version number in the master secret, when passed
18077 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18078 (because the server will not accept higher), that the version number
18079 is 0x03,0x01, not 0x03,0x00
18081 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18083 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18084 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18085 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18089 * Support for RAW extensions where an arbitrary extension can be
18090 created by including its DER encoding. See `apps/openssl.cnf` for
18095 * Make sure latest Perl versions don't interpret some generated C array
18096 code as Perl array code in the crypto/err/err_genc.pl script.
18098 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18100 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18101 not many people have the assembler. Various Win32 compilation fixes and
18102 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18103 build instructions.
18107 * Modify configure script 'Configure' to automatically create crypto/date.h
18108 file under Win32 and also build pem.h from pem.org. New script
18109 util/mkfiles.pl to create the MINFO file on environments that can't do a
18110 'make files': perl util/mkfiles.pl >MINFO should work.
18114 * Major rework of DES function declarations, in the pursuit of correctness
18115 and purity. As a result, many evil casts evaporated, and some weirdness,
18116 too. You may find this causes warnings in your code. Zapping your evil
18117 casts will probably fix them. Mostly.
18121 * Fix for a typo in asn1.h. Bug fix to object creation script
18122 obj_dat.pl. It considered a zero in an object definition to mean
18123 "end of object": none of the objects in objects.h have any zeros
18124 so it wasn't spotted.
18126 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18128 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18129 Masking (CBCM). In the absence of test vectors, the best I have been able
18130 to do is check that the decrypt undoes the encrypt, so far. Send me test
18131 vectors if you have them.
18135 * Correct calculation of key length for export ciphers (too much space was
18136 allocated for null ciphers). This has not been tested!
18140 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18141 message is now correct (it understands "crypto" and "ssl" on its
18142 command line). There is also now an "update" option. This will update
18143 the util/ssleay.num and util/libeay.num files with any new functions.
18145 perl util/mkdef.pl crypto ssl update
18146 it will update them.
18150 * Overhauled the Perl interface:
18151 - ported BN stuff to OpenSSL's different BN library
18152 - made the perl/ source tree CVS-aware
18153 - renamed the package from SSLeay to OpenSSL (the files still contain
18154 their history because I've copied them in the repository)
18155 - removed obsolete files (the test scripts will be replaced
18156 by better Test::Harness variants in the future)
18158 *Ralf S. Engelschall*
18160 * First cut for a very conservative source tree cleanup:
18161 1. merge various obsolete readme texts into doc/ssleay.txt
18162 where we collect the old documents and readme texts.
18163 2. remove the first part of files where I'm already sure that we no
18164 longer need them because of three reasons: either they are just temporary
18165 files which were left by Eric or they are preserved original files where
18166 I've verified that the diff is also available in the CVS via "cvs diff
18167 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18168 the crypto/md/ stuff).
18170 *Ralf S. Engelschall*
18172 * More extension code. Incomplete support for subject and issuer alt
18173 name, issuer and authority key id. Change the i2v function parameters
18174 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18175 what that's for :-) Fix to ASN1 macro which messed up
18176 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18180 * Preliminary support for ENUMERATED type. This is largely copied from the
18185 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18187 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18189 * Make sure `make rehash` target really finds the `openssl` program.
18191 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18193 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18194 like to hear about it if this slows down other processors.
18198 * Add CygWin32 platform information to Configure script.
18200 *Alan Batie <batie@aahz.jf.intel.com>*
18202 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18204 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18206 * New program nseq to manipulate netscape certificate sequences
18210 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18215 * Fixes to BN code. Previously the default was to define BN_RECURSION
18216 but the BN code had some problems that would cause failures when
18217 doing certificate verification and some other functions.
18219 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18221 * Add ASN1 and PEM code to support netscape certificate sequences.
18225 * Add ASN1 and PEM code to support netscape certificate sequences.
18229 * Add several PKIX and private extended key usage OIDs.
18233 * Modify the 'ca' program to handle the new extension code. Modify
18234 openssl.cnf for new extension format, add comments.
18238 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18239 and add a sample to openssl.cnf so req -x509 now adds appropriate
18244 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18245 error code, add initial support to X509_print() and x509 application.
18249 * Takes a deep breath and start adding X509 V3 extension support code. Add
18250 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18251 stuff is currently isolated and isn't even compiled yet.
18255 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18256 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18257 Removed the versions check from X509 routines when loading extensions:
18258 this allows certain broken certificates that don't set the version
18259 properly to be processed.
18263 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18264 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18265 can still be regenerated with "make depend".
18269 * Spelling mistake in C version of CAST-128.
18271 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18273 * Changes to the error generation code. The perl script err-code.pl
18274 now reads in the old error codes and retains the old numbers, only
18275 adding new ones if necessary. It also only changes the .err files if new
18276 codes are added. The makefiles have been modified to only insert errors
18277 when needed (to avoid needlessly modifying header files). This is done
18278 by only inserting errors if the .err file is newer than the auto generated
18279 C file. To rebuild all the error codes from scratch (the old behaviour)
18280 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18281 or delete all the .err files.
18285 * CAST-128 was incorrectly implemented for short keys. The C version has
18286 been fixed, but is untested. The assembler versions are also fixed, but
18287 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18288 to regenerate it if needed.
18289 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18290 Hagino <itojun@kame.net>*
18292 * File was opened incorrectly in randfile.c.
18294 *Ulf Möller <ulf@fitug.de>*
18296 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18297 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18298 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18299 al: it's just almost always a UTCTime. Note this patch adds new error
18300 codes so do a "make errors" if there are problems.
18304 * Correct Linux 1 recognition in config.
18306 *Ulf Möller <ulf@fitug.de>*
18308 * Remove pointless MD5 hash when using DSA keys in ca.
18310 *Anonymous <nobody@replay.com>*
18312 * Generate an error if given an empty string as a cert directory. Also
18313 generate an error if handed NULL (previously returned 0 to indicate an
18314 error, but didn't set one).
18316 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18318 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18322 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18323 parameters. This was causing a warning which killed off the Win32 compile.
18327 * Remove C++ style comments from crypto/bn/bn_local.h.
18329 *Neil Costigan <neil.costigan@celocom.com>*
18331 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18332 based on a text string, looking up short and long names and finally
18333 "dot" format. The "dot" format stuff didn't work. Added new function
18334 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18335 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18336 OID is not part of the table.
18340 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18341 X509_LOOKUP_by_alias().
18345 * Sort openssl functions by name.
18349 * Get the `gendsa` command working and add it to the `list` command. Remove
18350 encryption from sample DSA keys (in case anyone is interested the password
18355 * Make *all* `*_free` functions accept a NULL pointer.
18357 *Frans Heymans <fheymans@isaserver.be>*
18359 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18362 *Anonymous <nobody@replay.com>*
18364 * s_server should send the CAfile as acceptable CAs, not its own cert.
18366 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18368 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18370 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18372 * Temp key "for export" tests were wrong in s3_srvr.c.
18374 *Anonymous <nobody@replay.com>*
18376 * Add prototype for temp key callback functions
18377 SSL_CTX_set_tmp_{rsa,dh}_callback().
18381 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18382 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18386 * X509_name_add_entry() freed the wrong thing after an error.
18388 *Arne Ansper <arne@ats.cyber.ee>*
18390 * rsa_eay.c would attempt to free a NULL context.
18392 *Arne Ansper <arne@ats.cyber.ee>*
18394 * BIO_s_socket() had a broken should_retry() on Windoze.
18396 *Arne Ansper <arne@ats.cyber.ee>*
18398 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18400 *Arne Ansper <arne@ats.cyber.ee>*
18402 * Make sure the already existing X509_STORE->depth variable is initialized
18403 in X509_STORE_new(), but document the fact that this variable is still
18404 unused in the certificate verification process.
18406 *Ralf S. Engelschall*
18408 * Fix the various library and `apps/` files to free up pkeys obtained from
18409 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18413 * Fix reference counting in X509_PUBKEY_get(). This makes
18414 demos/maurice/example2.c work, amongst others, probably.
18416 *Steve Henson and Ben Laurie*
18418 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18419 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18420 are no longer created. This way we have a single and consistent command
18421 line interface `openssl <command>`, similar to `cvs <command>`.
18423 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18425 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18426 BIT STRING wrapper always have zero unused bits.
18430 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18434 * Make the top-level INSTALL documentation easier to understand.
18438 * Makefiles updated to exit if an error occurs in a sub-directory
18439 make (including if user presses ^C) [Paul Sutton]
18441 * Make Montgomery context stuff explicit in RSA data structure.
18445 * Fix build order of pem and err to allow for generated pem.h.
18449 * Fix renumbering bug in X509_NAME_delete_entry().
18453 * Enhanced the err-ins.pl script so it makes the error library number
18454 global and can add a library name. This is needed for external ASN1 and
18455 other error libraries.
18459 * Fixed sk_insert which never worked properly.
18463 * Fix ASN1 macros so they can handle indefinite length constructed
18464 EXPLICIT tags. Some non standard certificates use these: they can now
18469 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18470 into a single doc/ssleay.txt bundle. This way the information is still
18471 preserved but no longer messes up this directory. Now it's new room for
18472 the new set of documentation files.
18474 *Ralf S. Engelschall*
18476 * SETs were incorrectly DER encoded. This was a major pain, because they
18477 shared code with SEQUENCEs, which aren't coded the same. This means that
18478 almost everything to do with SETs or SEQUENCEs has either changed name or
18479 number of arguments.
18481 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18483 * Fix test data to work with the above.
18487 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18488 was already fixed by Eric for 0.9.1 it seems.
18490 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18492 * Autodetect FreeBSD3.
18496 * Fix various bugs in Configure. This affects the following platforms:
18500 unixware-2.0-pentium
18505 * Eliminate generated files from CVS. Reorder tests to regenerate files
18506 before they are needed.
18510 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18514 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18516 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18517 changed SSLeay to OpenSSL in version strings.
18519 *Ralf S. Engelschall*
18521 * Some fixups to the top-level documents.
18525 * Fixed the nasty bug where rsaref.h was not found under compile-time
18526 because the symlink to include/ was missing.
18528 *Ralf S. Engelschall*
18530 * Incorporated the popular no-RSA/DSA-only patches
18531 which allow to compile a RSA-free SSLeay.
18533 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18535 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18536 when "ssleay" is still not found.
18538 *Ralf S. Engelschall*
18540 * Added more platforms to Configure: Cray T3E, HPUX 11,
18542 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18544 * Updated the README file.
18546 *Ralf S. Engelschall*
18548 * Added various .cvsignore files in the CVS repository subdirs
18549 to make a "cvs update" really silent.
18551 *Ralf S. Engelschall*
18553 * Recompiled the error-definition header files and added
18554 missing symbols to the Win32 linker tables.
18556 *Ralf S. Engelschall*
18558 * Cleaned up the top-level documents;
18559 o new files: CHANGES and LICENSE
18560 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18561 o merged COPYRIGHT into LICENSE
18562 o removed obsolete TODO file
18563 o renamed MICROSOFT to INSTALL.W32
18565 *Ralf S. Engelschall*
18567 * Removed dummy files from the 0.9.1b source tree:
18569 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18570 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18571 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18572 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18573 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18576 *Ralf S. Engelschall*
18578 * Added various platform portability fixes.
18582 * The Genesis of the OpenSSL rpject:
18583 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18584 Young and Tim J. Hudson created while they were working for C2Net until
18587 *The OpenSSL Project*
18589 ### Changes between 0.9.0b and 0.9.1b [not released]
18591 * Updated a few CA certificates under certs/
18595 * Changed some BIGNUM api stuff.
18599 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18600 DGUX x86, Linux Alpha, etc.
18604 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18605 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18610 * Add -strparse option to asn1pars program which parses nested
18613 *Dr Stephen Henson <shenson@bigfoot.com>*
18615 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18619 * DSA fix for "ca" program.
18623 * Added "-genkey" option to "dsaparam" program.
18627 * Added RIPE MD160 (rmd160) message digest.
18631 * Added -a (all) option to "ssleay version" command.
18635 * Added PLATFORM define which is the id given to Configure.
18639 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18643 * Extended the ASN.1 parser routines.
18647 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18651 * Added a BN_CTX to the BN library.
18655 * Fixed the weak key values in DES library
18659 * Changed API in EVP library for cipher aliases.
18663 * Added support for RC2/64bit cipher.
18667 * Converted the lhash library to the crypto/mem.c functions.
18671 * Added more recognized ASN.1 object ids.
18675 * Added more RSA padding checks for SSL/TLS.
18679 * Added BIO proxy/filter functionality.
18683 * Added extra_certs to SSL_CTX which can be used
18684 send extra CA certificates to the client in the CA cert chain sending
18685 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18689 * Now Fortezza is denied in the authentication phase because
18690 this is key exchange mechanism is not supported by SSLeay at all.
18694 * Additional PKCS1 checks.
18698 * Support the string "TLSv1" for all TLS v1 ciphers.
18702 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18703 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18707 * Fixed a few memory leaks.
18711 * Fixed various code and comment typos.
18715 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18716 bytes sent in the client random.
18718 *Edward Bishop <ebishop@spyglass.com>*
18722 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18723 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18724 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18725 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18726 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18727 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18728 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18729 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18730 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18731 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18732 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18733 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18734 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18735 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18736 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18737 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18738 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18739 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18740 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18741 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18742 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18743 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18744 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18745 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18746 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18747 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18748 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18749 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18750 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18751 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18752 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18753 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18754 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18755 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18756 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18757 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18758 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18759 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18760 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18761 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18762 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18763 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18764 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18765 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18766 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18767 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18768 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18769 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18770 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18771 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18772 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18773 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18774 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18775 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18776 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18777 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18778 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18779 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18780 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18781 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18782 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18783 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18784 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18785 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18786 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18787 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18788 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18789 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18790 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18791 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18792 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18793 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18794 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18795 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18796 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18797 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18798 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18799 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18800 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18801 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18802 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18803 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18804 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18805 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18806 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18807 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18808 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18809 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18810 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18811 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18812 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18813 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18814 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18815 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18816 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18817 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18818 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18819 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18820 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18821 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18822 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18823 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18824 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18825 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18826 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18827 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18828 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18829 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18830 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18831 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18832 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18833 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18834 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18835 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18836 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18837 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18838 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18839 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18840 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18841 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18842 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18843 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18844 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18845 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18846 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18847 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18848 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18849 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18850 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18851 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18852 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18853 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18854 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18855 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18856 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18857 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18858 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18859 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18860 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18861 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18862 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18863 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18864 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18865 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18866 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18867 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18868 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18869 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18870 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18871 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18872 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18873 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18874 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18875 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18876 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18877 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18878 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18879 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18880 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18881 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18882 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18883 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655