4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
25 listed here are only a brief description.
26 The migration guide contains more detailed information related to new features,
27 breaking changes, and mappings for the large list of deprecated functions.
29 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
31 ### Changes between 3.0.0 and 3.0.1 [xx XXX xxxx]
33 * Corrected a few file name and file reference bugs in the build,
34 installation and setup scripts, which lead to installation verification
35 failures. Slightly enhanced the installation verification script.
39 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
41 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
46 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
47 S390X capability vector to zero. This simplifies testing of different code
48 paths on S390X architecture.
52 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
53 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
54 SP 800-38D". The communication will fail at this point.
58 * The EC_GROUP_clear_free() function is deprecated as there is nothing
59 confidential in EC_GROUP data.
63 * The byte order mark (BOM) character is ignored if encountered at the
64 beginning of a PEM-formatted file.
68 * Added CMS support for the Russian GOST algorithms.
72 * Due to move of the implementation of cryptographic operations
73 to the providers, validation of various operation parameters can
74 be postponed until the actual operation is executed where previously
75 it happened immediately when an operation parameter was set.
77 For example when setting an unsupported curve with
78 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
79 fail but later keygen operations with the EVP_PKEY_CTX will fail.
81 *OpenSSL team members and many third party contributors*
83 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
84 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
85 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
86 instead to retrieve these algorithms from a provider.
90 * On build targets where the multilib postfix is set in the build
91 configuration the libdir directory was changing based on whether
92 the lib directory with the multilib postfix exists on the system
93 or not. This unpredictable behavior was removed and eventual
94 multilib postfix is now always added to the default libdir. Use
95 `--libdir=lib` to override the libdir if adding the postfix is
100 * The triple DES key wrap functionality now conforms to RFC 3217 but is
101 no longer interoperable with OpenSSL 1.1.1.
105 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
106 function codes, this function can only cause problems for calling
111 * Add a configurable flag to output date formats as ISO 8601. Does not
112 change the default date format.
116 * Version of MSVC earlier than 1300 could get link warnings, which could
117 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
118 Support for this flag has been removed.
122 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
123 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
124 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
125 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
126 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
130 * The signatures of the functions to get and set options on SSL and
131 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
132 Some source code changes may be required.
136 * The public definitions of conf_method_st and conf_st have been
137 deprecated. They will be made opaque in a future release.
139 *Rich Salz and Tomáš Mráz*
141 * Client-initiated renegotiation is disabled by default. To allow it, use
142 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
143 flag, or the "ClientRenegotiation" config parameter as appropriate.
147 * Add "abspath" and "includedir" pragma's to config files, to prevent,
148 or modify relative pathname inclusion.
152 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
153 validated. Please consult the README-FIPS and
154 README-PROVIDERS files, as well as the migration guide.
156 *OpenSSL team members and many third party contributors*
158 * For the key types DH and DHX the allowed settable parameters are now different.
162 * The openssl commands that read keys, certificates, and CRLs now
163 automatically detect the PEM or DER format of the input files.
165 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
167 * Added enhanced PKCS#12 APIs which accept a library context.
171 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
175 * Added support for Kernel TLS (KTLS).
177 *Boris Pismenny, John Baldwin and Andrew Gallatin*
179 * Support for RFC 5746 secure renegotiation is now required by default for
180 SSL or TLS connections to succeed.
184 * The signature of the `copy` functional parameter of the
185 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
186 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
187 the signature of the `pub_decode` functional parameter of the
188 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
189 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
193 * The error return values from some control calls (ctrl) have changed.
197 * A public key check is now performed during EVP_PKEY_derive_set_peer().
201 * Many functions in the EVP_ namespace that are getters of values from
202 implementations or contexts were renamed to include get or get0 in their
203 names. Old names are provided as macro aliases for compatibility and
208 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
209 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
210 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
215 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
218 * The output from the command line applications may have minor
223 * The output from numerous "printing" may have minor changes.
227 * Windows thread synchronization uses read/write primitives (SRWLock) when
228 supported by the OS, otherwise CriticalSection continues to be used.
232 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
233 work on read only BIO source/sinks that do not support these functions.
234 This allows piping or redirection of a file BIO using stdin to be buffered
235 into memory. This is used internally in OSSL_DECODER_from_bio().
239 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
240 this function would return one of the values OSSL_STORE_INFO_NAME,
241 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
242 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
243 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
244 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
245 using this function should be amended to handle the changed return value.
249 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
250 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
251 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
252 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
253 contains more than one certificate identifier: This means that all
254 certificates referenced there MUST be part of the validation chain.
258 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
259 RC5, DESX and DES have been moved to the legacy provider.
263 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
264 RIPEMD-160 have been moved to the legacy provider.
268 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
273 * The deprecated functions EVP_PKEY_get0_RSA(),
274 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
275 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
276 well as the similarly named "get1" functions behave differently in
281 * A number of functions handling low-level keys or engines were deprecated
282 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
283 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
284 EVP_PKEY_get0_siphash().
288 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
289 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
290 will need to load the legacy crypto provider. This includes these PBE
291 algorithms which use this KDF:
292 - NID_pbeWithMD2AndDES_CBC
293 - NID_pbeWithMD5AndDES_CBC
294 - NID_pbeWithSHA1AndRC2_CBC
295 - NID_pbeWithMD2AndRC2_CBC
296 - NID_pbeWithMD5AndRC2_CBC
297 - NID_pbeWithSHA1AndDES_CBC
301 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
302 BIO_debug_callback() functions.
306 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
307 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
311 * The RAND_METHOD APIs have been deprecated.
315 * The SRP APIs have been deprecated.
319 * Add a compile time option to prevent the caching of provider fetched
320 algorithms. This is enabled by including the no-cached-fetch option
321 at configuration time.
325 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
326 count of PKCS12_DEFAULT_ITER.
328 *Tomáš Mráz and Sahana Prasad*
330 * The openssl speed command does not use low-level API calls anymore.
334 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
337 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
339 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
343 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
344 providers may supply their own group implementations (using either the "key
345 exchange" or the "key encapsulation" methods) which will automatically be
346 detected and used by libssl.
348 *Matt Caswell, Nicola Tuveri*
350 * The undocumented function X509_certificate_type() has been deprecated;
354 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
358 * Removed RSA padding mode for SSLv23 (which was only used for
359 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
360 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
365 * Deprecated the obsolete X9.31 RSA key generation related functions.
367 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
368 is not allowed to return a value > 1, this is no more taken as failure.
370 *Viktor Dukhovni and David von Oheimb*
372 * Deprecated the obsolete X9.31 RSA key generation related functions
373 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
374 BN_X931_generate_prime_ex().
378 * The default key generation method for the regular 2-prime RSA keys was
379 changed to the FIPS 186-4 B.3.6 method.
383 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
387 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
391 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
392 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
394 *Rich Salz, Richard Levitte, and David von Oheimb*
396 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
400 * Deprecated `OCSP_parse_url()`.
404 * Validation of SM2 keys has been separated from the validation of regular EC
409 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
410 switches: a validation failure triggers an early exit, returning a failure
411 exit status to the parent process.
415 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
416 to ignore unknown ciphers.
420 * The `-cipher-commands` and `-digest-commands` options
421 of the command line utility `list` have been deprecated.
422 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
426 * Added convenience functions for generating asymmetric key pairs:
427 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
428 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
432 * All of the low level EC_KEY functions have been deprecated.
434 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
436 * Deprecated all the libcrypto and libssl error string loading
441 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
442 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
447 * The `-crypt` option to the `passwd` command line tool has been removed.
451 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
456 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
460 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
461 EVP_PKEY_get1_tls_encodedpoint().
465 * The security callback, which can be customised by application code, supports
466 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
467 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
471 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
472 interface. Their functionality remains unchanged.
476 * Added new option for 'openssl list', '-providers', which will display the
477 list of loaded providers, their names, version and status. It optionally
478 displays their gettable parameters.
482 * Removed EVP_PKEY_set_alias_type().
486 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
487 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
491 * Changed all "STACK" functions to be macros instead of inline functions. Macro
492 parameters are still checked for type safety at compile time via helper
497 * Remove the RAND_DRBG API
499 *Paul Dale and Matthias St. Pierre*
501 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
502 as well as actual hostnames.
506 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
507 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
508 conversely, silently ignore DTLS protocol version bounds when configuring
509 TLS-based contexts. The commands can be repeated to set bounds of both
510 types. The same applies with the corresponding "min_protocol" and
511 "max_protocol" command-line switches, in case some application uses both TLS
514 SSL_CTX instances that are created for a fixed protocol version (e.g.
515 `TLSv1_server_method()`) also silently ignore version bounds. Previously
516 attempts to apply bounds to these protocol versions would result in an
517 error. Now only the "version-flexible" SSL_CTX instances are subject to
518 limits in configuration files in command-line options.
522 * Deprecated the `ENGINE` API. Engines should be replaced with providers
527 * Reworked the recorded ERR codes to make better space for system errors.
528 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
529 given code is a system error (true) or an OpenSSL error (false).
533 * Reworked the test perl framework to better allow parallel testing.
535 *Nicola Tuveri and David von Oheimb*
537 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
538 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
542 * 'Configure' has been changed to figure out the configuration target if
543 none is given on the command line. Consequently, the 'config' script is
544 now only a mere wrapper. All documentation is changed to only mention
547 *Rich Salz and Richard Levitte*
549 * Added a library context `OSSL_LIB_CTX` that applications as well as
550 other libraries can use to form a separate context within which
551 libcrypto operations are performed.
555 * Added various `_ex` functions to the OpenSSL API that support using
556 a non-default `OSSL_LIB_CTX`.
560 * Handshake now fails if Extended Master Secret extension is dropped
565 * Dropped interactive mode from the `openssl` program.
569 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
571 *David von Oheimb and Shane Lontis*
573 * Deprecated `EC_METHOD_get_field_type()`.
577 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
578 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
579 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
583 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
587 * Add CAdES-BES signature verification support, mostly derived
588 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
590 *Filipe Raimundo da Silva*
592 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
596 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
597 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
601 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
605 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
606 EC_KEY_precompute_mult().
610 * Deprecated EC_POINTs_mul().
614 * Removed FIPS_mode() and FIPS_mode_set().
618 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
622 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
623 EC_POINT_get_Jprojective_coordinates_GFp().
627 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
628 arrays to be more easily constructed via a series of utility functions.
629 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
630 the various push functions and finally convert to a passable OSSL_PARAM
631 array using OSSL_PARAM_BLD_to_param().
635 * The security strength of SHA1 and MD5 based signatures in TLS has been
640 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
641 contain a provider side internal key.
645 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
649 * Project text documents not yet having a proper file name extension
650 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
651 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
655 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
656 have been converted to Markdown with the goal to produce documents
657 which not only look pretty when viewed online in the browser, but
658 remain well readable inside a plain text editor.
660 To achieve this goal, a 'minimalistic' Markdown style has been applied
661 which avoids formatting elements that interfere too much with the
662 reading flow in the text file. For example, it
664 * avoids [ATX headings][] and uses [setext headings][] instead
665 (which works for `<h1>` and `<h2>` headings only).
666 * avoids [inline links][] and uses [reference links][] instead.
667 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
669 [ATX headings]: https://github.github.com/gfm/#atx-headings
670 [setext headings]: https://github.github.com/gfm/#setext-headings
671 [inline links]: https://github.github.com/gfm/#inline-link
672 [reference links]: https://github.github.com/gfm/#reference-link
673 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
674 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
676 *Matthias St. Pierre*
678 * The test suite is changed to preserve results of each test recipe.
679 A new directory test-runs/ with subdirectories named like the
680 test recipes are created in the build tree for this purpose.
684 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
685 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
686 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
688 *David von Oheimb, Martin Peylo*
690 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
691 It supports arbitrary request and response content types, GET redirection,
692 TLS, connections via HTTP(S) proxies, connections and exchange via
693 user-defined BIOs (allowing implicit connections), persistent connections,
694 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
695 The legacy OCSP-focused (and only partly documented) API
696 is retained for backward compatibility, while most of it is deprecated.
700 * Added `util/check-format.pl`, a tool for checking adherence to the
701 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
702 The checks performed are incomplete and yield some false positives.
703 Still the tool should be useful for detecting most typical glitches.
707 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
708 If domain name resolution yields multiple IP addresses all of them are tried
709 after `connect()` failures.
713 * All of the low level RSA functions have been deprecated.
717 * X509 certificates signed using SHA1 are no longer allowed at security
722 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
723 modified to use PKEY APIs. These commands are now in maintenance mode
724 and no new features will be added to them.
728 * The command line utility rsautl has been deprecated.
732 * The command line utilities genrsa and rsa have been modified to use PKEY
733 APIs. They now write PKCS#8 keys by default. These commands are now in
734 maintenance mode and no new features will be added to them.
738 * All of the low level DH functions have been deprecated.
740 *Paul Dale and Matt Caswell*
742 * All of the low level DSA functions have been deprecated.
746 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
747 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
751 * Deprecated low level ECDH and ECDSA functions.
755 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
759 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
760 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
761 a new formulation to include all the things it can be used for,
762 as well as words of caution.
766 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
770 * All of the low level HMAC functions have been deprecated.
772 *Paul Dale and David von Oheimb*
774 * Over two thousand fixes were made to the documentation, including:
775 - Common options (such as -rand/-writerand, TLS version control, etc)
776 were refactored and point to newly-enhanced descriptions in openssl.pod.
777 - Added style conformance for all options (with help from Richard Levitte),
778 documented all reported missing options, added a CI build to check
779 that all options are documented and that no unimplemented options
781 - Documented some internals, such as all use of environment variables.
782 - Addressed all internal broken L<> references.
786 * All of the low level CMAC functions have been deprecated.
790 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
791 functions have been deprecated.
793 *Paul Dale and David von Oheimb*
795 * Corrected the documentation of the return values from the `EVP_DigestSign*`
796 set of functions. The documentation mentioned negative values for some
797 errors, but this was never the case, so the mention of negative values
800 Code that followed the documentation and thereby check with something
801 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
805 * All of the low level cipher functions have been deprecated.
807 *Matt Caswell and Paul Dale*
809 * Removed include/openssl/opensslconf.h.in and replaced it with
810 include/openssl/configuration.h.in, which differs in not including
811 <openssl/macros.h>. A short header include/openssl/opensslconf.h
812 was added to include both.
814 This allows internal hacks where one might need to modify the set
815 of configured macros, for example this if deprecated symbols are
816 still supposed to be available internally:
818 #include <openssl/configuration.h>
820 #undef OPENSSL_NO_DEPRECATED
821 #define OPENSSL_SUPPRESS_DEPRECATED
823 #include <openssl/macros.h>
825 This should not be used by applications that use the exported
826 symbols, as that will lead to linking errors.
830 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
831 used in exponentiation with 512-bit moduli. No EC algorithms are
832 affected. Analysis suggests that attacks against 2-prime RSA1024,
833 3-prime RSA1536, and DSA1024 as a result of this defect would be very
834 difficult to perform and are not believed likely. Attacks against DH512
835 are considered just feasible. However, for an attack the target would
836 have to re-use the DH512 private key, which is not recommended anyway.
837 Also applications directly using the low-level API BN_mod_exp may be
838 affected if they use BN_FLG_CONSTTIME.
843 * Most memory-debug features have been deprecated, and the functionality
844 replaced with no-ops.
848 * Added documentation for the STACK API.
852 * Introduced a new method type and API, OSSL_ENCODER, to represent
853 generic encoders. These do the same sort of job that PEM writers
854 and d2i functions do, but with support for methods supplied by
855 providers, and the possibility for providers to support other
860 * Introduced a new method type and API, OSSL_DECODER, to represent
861 generic decoders. These do the same sort of job that PEM readers
862 and i2d functions do, but with support for methods supplied by
863 providers, and the possibility for providers to support other
868 * Added a .pragma directive to the syntax of configuration files, to
869 allow varying behavior in a supported and predictable manner.
870 Currently added pragma:
874 This allows dollar signs to be a keyword character unless it's
875 followed by a opening brace or parenthesis. This is useful for
876 platforms where dollar signs are commonly used in names, such as
877 volume names and system directory names on VMS.
881 * Added functionality to create an EVP_PKEY from user data.
885 * Change the interpretation of the '--api' configuration option to
886 mean that this is a desired API compatibility level with no
887 further meaning. The previous interpretation, that this would
888 also mean to remove all deprecated symbols up to and including
889 the given version, no requires that 'no-deprecated' is also used
890 in the configuration.
892 When building applications, the desired API compatibility level
893 can be set with the OPENSSL_API_COMPAT macro like before. For
894 API compatibility version below 3.0, the old style numerical
895 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
896 For version 3.0 and on, the value is expected to be the decimal
897 value calculated from the major and minor version like this:
899 MAJOR * 10000 + MINOR * 100
903 -DOPENSSL_API_COMPAT=30000 For 3.0
904 -DOPENSSL_API_COMPAT=30200 For 3.2
906 To hide declarations that are deprecated up to and including the
907 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
908 given when building the application as well.
912 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
913 access to certificate and CRL stores via URIs and OSSL_STORE
916 This adds the following functions:
918 - X509_LOOKUP_store()
919 - X509_STORE_load_file()
920 - X509_STORE_load_path()
921 - X509_STORE_load_store()
922 - SSL_add_store_cert_subjects_to_stack()
923 - SSL_CTX_set_default_verify_store()
924 - SSL_CTX_load_verify_file()
925 - SSL_CTX_load_verify_dir()
926 - SSL_CTX_load_verify_store()
930 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
931 The presence of this system service is determined at run-time.
935 * Added functionality to create an EVP_PKEY context based on data
936 for methods from providers. This takes an algorithm name and a
937 property query string and simply stores them, with the intent
938 that any operation that uses this context will use those strings
939 to fetch the needed methods implicitly, thereby making the port
940 of application written for pre-3.0 OpenSSL easier.
944 * The undocumented function NCONF_WIN32() has been deprecated; for
945 conversion details see the HISTORY section of doc/man5/config.pod
949 * Introduced the new functions EVP_DigestSignInit_ex() and
950 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
951 EVP_DigestVerifyUpdate() have been converted to functions. See the man
952 pages for further details.
956 * Over two thousand fixes were made to the documentation, including:
957 adding missing command flags, better style conformance, documentation
960 *Rich Salz, Richard Levitte*
962 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
963 X25519, X448, Ed25519 and Ed448.
967 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
972 * Deprecated the public definition of `ERR_STATE` as well as the function
973 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
978 * Added ERR functionality to give callers access to the stored function
979 names that have replaced the older function code based functions.
981 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
982 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
983 ERR_peek_error_all() and ERR_peek_last_error_all().
985 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
986 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
987 ERR_func_error_string().
991 * Extended testing to be verbose for failing tests only. The make variables
992 VERBOSE_FAILURE or VF can be used to enable this:
994 $ make VF=1 test # Unix
995 $ mms /macro=(VF=1) test ! OpenVMS
996 $ nmake VF=1 test # Windows
1000 * Added the `-copy_extensions` option to the `x509` command for use with
1001 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1002 all extensions in the request are copied to the certificate or vice versa.
1004 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1006 * Added the `-copy_extensions` option to the `req` command for use with
1007 `-x509`. When given with the `copy` or `copyall` argument,
1008 all extensions in the certification request are copied to the certificate.
1012 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1013 they generate are by default RFC 5280 compliant in the following sense:
1014 There is a subjectKeyIdentifier extension with a hash value of the public key
1015 and for not self-signed certs there is an authorityKeyIdentifier extension
1016 with a keyIdentifier field or issuer information identifying the signing key.
1017 This is done unless some configuration overrides the new default behavior,
1018 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1022 * Added several checks to `X509_verify_cert()` according to requirements in
1023 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1024 (which may be done by using the CLI option `-x509_strict`):
1025 * The basicConstraints of CA certificates must be marked critical.
1026 * CA certificates must explicitly include the keyUsage extension.
1027 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1028 * The issuer name of any certificate must not be empty.
1029 * The subject name of CA certs, certs with keyUsage crlSign,
1030 and certs without subjectAlternativeName must not be empty.
1031 * If a subjectAlternativeName extension is given it must not be empty.
1032 * The signatureAlgorithm field and the cert signature must be consistent.
1033 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1034 must not be marked critical.
1035 * The authorityKeyIdentifier must be given for X.509v3 certs
1036 unless they are self-signed.
1037 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1041 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1042 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1046 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1047 used even when parsing explicit parameters, when loading a encoded key
1048 or calling `EC_GROUP_new_from_ecpkparameters()`/
1049 `EC_GROUP_new_from_ecparameters()`.
1050 This prevents bypass of security hardening and performance gains,
1051 especially for curves with specialized EC_METHODs.
1052 By default, if a key encoded with explicit parameters is loaded and later
1053 encoded, the output is still encoded with explicit parameters, even if
1054 internally a "named" EC_GROUP is used for computation.
1058 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1059 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1060 NULL. After this change, only the cofactor parameter can be NULL. It also
1061 does some minimal sanity checks on the passed order.
1066 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1067 An attack is simple, if the first CMS_recipientInfo is valid but the
1068 second CMS_recipientInfo is chosen ciphertext. If the second
1069 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1070 encryption key will be replaced by garbage, and the message cannot be
1071 decoded, but if the RSA decryption fails, the correct encryption key is
1072 used and the recipient will not notice the attack.
1073 As a work around for this potential attack the length of the decrypted
1074 key must be equal to the cipher default key length, in case the
1075 certifiate is not given and all recipientInfo are tried out.
1076 The old behaviour can be re-enabled in the CMS code by setting the
1077 CMS_DEBUG_DECRYPT flag.
1081 * Early start up entropy quality from the DEVRANDOM seed source has been
1082 improved for older Linux systems. The RAND subsystem will wait for
1083 /dev/random to be producing output before seeding from /dev/urandom.
1084 The seeded state is stored for future library initialisations using
1085 a system global shared memory segment. The shared memory identifier
1086 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1087 the desired value. The default identifier is 114.
1091 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1092 when primes for RSA keys are computed.
1093 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1094 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1095 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1096 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1097 This avoids possible fingerprinting of newly generated RSA modules.
1101 * Correct the extended master secret constant on EBCDIC systems. Without this
1102 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1103 negotiate EMS will fail. Unfortunately this also means that TLS connections
1104 between EBCDIC systems with this fix, and EBCDIC systems without this
1105 fix will fail if they negotiate EMS.
1109 * Changed the library initialisation so that the config file is now loaded
1110 by default. This was already the case for libssl. It now occurs for both
1111 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1112 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1116 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1117 where the former acts as a replacement for `ERR_put_error()`, and the
1118 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1119 `ERR_raise_data()` adds more flexibility by taking a format string and
1120 an arbitrary number of arguments following it, to be processed with
1125 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1126 to check if a named provider is loaded and available. When called, it
1127 will also activate all fallback providers if such are still present.
1131 * Enforce a minimum DH modulus size of 512 bits.
1135 * Changed DH parameters to generate the order q subgroup instead of 2q.
1136 Previously generated DH parameters are still accepted by DH_check
1137 but DH_generate_key works around that by clearing bit 0 of the
1138 private key for those. This avoids leaking bit 0 of the private key.
1142 * Significantly reduce secure memory usage by the randomness pools.
1146 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1151 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1152 algorithms. An implementation of a key exchange algorithm can be obtained
1153 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1154 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1155 the older EVP_PKEY_derive_init() function. See the man pages for the new
1156 functions for further details.
1160 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1164 * Removed the function names from error messages and deprecated the
1169 * Removed NextStep support and the macro OPENSSL_UNISTD
1173 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1174 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1175 Also removed "export var as function" capability; we do not export
1176 variables, only functions.
1180 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1181 an error and 1 indicating success. In previous versions of OpenSSL this
1182 was a void type. If a key was set longer than the maximum possible this
1187 * Support SM2 signing and verification schemes with X509 certificate.
1191 * Use SHA256 as the default digest for TS query in the `ts` app.
1195 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1199 * Default cipher lists/suites are now available via a function, the
1200 #defines are deprecated.
1204 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1205 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1206 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1210 * Join the directories crypto/x509 and crypto/x509v3
1214 * Added command 'openssl kdf' that uses the EVP_KDF API.
1218 * Added command 'openssl mac' that uses the EVP_MAC API.
1222 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1223 as default directories. Also added the command 'openssl info'
1224 for scripting purposes.
1228 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1233 * Add prediction resistance to the DRBG reseeding process.
1237 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1238 mandated by IEEE Std 1619-2018.
1242 * Added newline escaping functionality to a filename when using openssl dgst.
1243 This output format is to replicate the output format found in the `*sum`
1244 checksum programs. This aims to preserve backward compatibility.
1246 *Matt Eaton, Richard Levitte, and Paul Dale*
1248 * Removed the heartbeat message in DTLS feature, as it has very
1249 little usage and doesn't seem to fulfill a valuable purpose.
1250 The configuration option is now deprecated.
1254 * Changed the output of 'openssl {digestname} < file' to display the
1255 digest name in its output.
1259 * Added a new generic trace API which provides support for enabling
1260 instrumentation through trace output.
1262 *Richard Levitte & Matthias St. Pierre*
1264 * Added build tests for C++. These are generated files that only do one
1265 thing, to include one public OpenSSL head file each. This tests that
1266 the public header files can be usefully included in a C++ application.
1268 This test isn't enabled by default. It can be enabled with the option
1269 'enable-buildtest-c++'.
1273 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1277 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1281 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1285 * Added KMAC to EVP_MAC.
1289 * Added property based algorithm implementation selection framework to
1294 * Added SCA hardening for modular field inversion in EC_GROUP through
1295 a new dedicated field_inv() pointer in EC_METHOD.
1296 This also addresses a leakage affecting conversions from projective
1297 to affine coordinates.
1299 *Billy Bob Brumley, Nicola Tuveri*
1301 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1302 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1303 those algorithms that were already supported through the EVP_PKEY API
1304 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1305 and scrypt are now wrappers that call EVP_KDF.
1309 * Build devcrypto engine as a dynamic engine.
1311 *Eneas U de Queiroz*
1313 * Add keyed BLAKE2 to EVP_MAC.
1317 * Fix a bug in the computation of the endpoint-pair shared secret used
1318 by DTLS over SCTP. This breaks interoperability with older versions
1319 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1320 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1321 interoperability with such broken implementations. However, enabling
1322 this switch breaks interoperability with correct implementations.
1324 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1325 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1329 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1333 * Changed the license to the Apache License v2.0.
1337 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1339 - Major releases (indicated by incrementing the MAJOR release number)
1340 may introduce incompatible API/ABI changes.
1341 - Minor releases (indicated by incrementing the MINOR release number)
1342 may introduce new features but retain API/ABI compatibility.
1343 - Patch releases (indicated by incrementing the PATCH number)
1344 are intended for bug fixes and other improvements of existing
1345 features only (like improving performance or adding documentation)
1346 and retain API/ABI compatibility.
1350 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1354 * Remove the 'dist' target and add a tarball building script. The
1355 'dist' target has fallen out of use, and it shouldn't be
1356 necessary to configure just to create a source distribution.
1360 * Recreate the OS390-Unix config target. It no longer relies on a
1361 special script like it did for OpenSSL pre-1.1.0.
1365 * Instead of having the source directories listed in Configure, add
1366 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1371 * Add GMAC to EVP_MAC.
1375 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1379 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1380 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1381 to facilitate the continued use of MACs through raw private keys in
1382 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1386 * Deprecate ECDH_KDF_X9_62().
1390 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1391 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1392 are retained for backwards compatibility.
1396 * AES-XTS mode now enforces that its two keys are different to mitigate
1397 the attacked described in "Efficient Instantiations of Tweakable
1398 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1399 Details of this attack can be obtained from:
1400 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1404 * Rename the object files, i.e. give them other names than in previous
1405 versions. Their names now include the name of the final product, as
1406 well as its type mnemonic (bin, lib, shlib).
1410 * Added new option for 'openssl list', '-objects', which will display the
1411 list of built in objects, i.e. OIDs with names.
1415 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1416 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1421 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1422 improves application performance by removing data copies and providing
1423 applications with zero-copy system calls such as sendfile and splice.
1427 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1431 * `PKCS12_parse` now maintains the order of the parsed certificates
1432 when outputting them via `*ca` (rather than reversing it).
1436 * Deprecated pthread fork support methods.
1440 * Added support for FFDHE key exchange in TLS 1.3.
1444 * Added a new concept for OpenSSL plugability: providers. This
1445 functionality is designed to replace the ENGINE API and ENGINE
1446 implementations, and to be much more dynamic, allowing provider
1447 authors to introduce new algorithms among other things, as long as
1448 there's an API that supports the algorithm type.
1450 With this concept comes a new core API for interaction between
1451 libcrypto and provider implementations. Public libcrypto functions
1452 that want to use providers do so through this core API.
1454 The main documentation for this core API is found in
1455 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1456 refer to other manuals describing the API specific for supported
1457 algorithm types (also called operations).
1464 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1466 * Fixed an SM2 Decryption Buffer Overflow.
1468 In order to decrypt SM2 encrypted data an application is expected to
1469 call the API function EVP_PKEY_decrypt(). Typically an application will
1470 call this function twice. The first time, on entry, the "out" parameter
1471 can be NULL and, on exit, the "outlen" parameter is populated with the
1472 buffer size required to hold the decrypted plaintext. The application
1473 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1474 again, but this time passing a non-NULL value for the "out" parameter.
1476 A bug in the implementation of the SM2 decryption code means that the
1477 calculation of the buffer size required to hold the plaintext returned
1478 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1479 size required by the second call. This can lead to a buffer overflow
1480 when EVP_PKEY_decrypt() is called by the application a second time with
1481 a buffer that is too small.
1483 A malicious attacker who is able present SM2 content for decryption to
1484 an application could cause attacker chosen data to overflow the buffer
1485 by up to a maximum of 62 bytes altering the contents of other data held
1486 after the buffer, possibly changing application behaviour or causing
1487 the application to crash. The location of the buffer is application
1488 dependent but is typically heap allocated.
1493 * Fixed various read buffer overruns processing ASN.1 strings
1495 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1496 structure which contains a buffer holding the string data and a field
1497 holding the buffer length. This contrasts with normal C strings which
1498 are repesented as a buffer for the string data which is terminated
1499 with a NUL (0) byte.
1501 Although not a strict requirement, ASN.1 strings that are parsed using
1502 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1503 well as any string whose value has been set with the ASN1_STRING_set()
1504 function will additionally NUL terminate the byte array in the
1505 ASN1_STRING structure.
1507 However, it is possible for applications to directly construct valid
1508 ASN1_STRING structures which do not NUL terminate the byte array by
1509 directly setting the "data" and "length" fields in the ASN1_STRING
1510 array. This can also happen by using the ASN1_STRING_set0() function.
1512 Numerous OpenSSL functions that print ASN.1 data have been found to
1513 assume that the ASN1_STRING byte array will be NUL terminated, even
1514 though this is not guaranteed for strings that have been directly
1515 constructed. Where an application requests an ASN.1 structure to be
1516 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1517 been directly constructed by the application without NUL terminating
1518 the "data" field, then a read buffer overrun can occur.
1520 The same thing can also occur during name constraints processing
1521 of certificates (for example if a certificate has been directly
1522 constructed by the application instead of loading it via the OpenSSL
1523 parsing functions, and the certificate contains non NUL terminated
1524 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1525 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1527 If a malicious actor can cause an application to directly construct an
1528 ASN1_STRING and then process it through one of the affected OpenSSL
1529 functions then this issue could be hit. This might result in a crash
1530 (causing a Denial of Service attack). It could also result in the
1531 disclosure of private memory contents (such as private keys, or
1532 sensitive plaintext).
1537 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1539 * Fixed a problem with verifying a certificate chain when using the
1540 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1541 the certificates present in a certificate chain. It is not set by default.
1543 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1544 the chain that have explicitly encoded elliptic curve parameters was added
1545 as an additional strict check.
1547 An error in the implementation of this check meant that the result of a
1548 previous check to confirm that certificates in the chain are valid CA
1549 certificates was overwritten. This effectively bypasses the check
1550 that non-CA certificates must not be able to issue other certificates.
1552 If a "purpose" has been configured then there is a subsequent opportunity
1553 for checks that the certificate is a valid CA. All of the named "purpose"
1554 values implemented in libcrypto perform this check. Therefore, where
1555 a purpose is set the certificate chain will still be rejected even when the
1556 strict flag has been used. A purpose is set by default in libssl client and
1557 server certificate verification routines, but it can be overridden or
1558 removed by an application.
1560 In order to be affected, an application must explicitly set the
1561 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1562 for the certificate verification or, in the case of TLS client or server
1563 applications, override the default purpose.
1568 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1569 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1570 renegotiation ClientHello omits the signature_algorithms extension (where it
1571 was present in the initial ClientHello), but includes a
1572 signature_algorithms_cert extension then a NULL pointer dereference will
1573 result, leading to a crash and a denial of service attack.
1575 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1576 (which is the default configuration). OpenSSL TLS clients are not impacted by
1580 *Peter Kästle and Samuel Sapalski*
1582 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1584 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1585 create a unique hash value based on the issuer and serial number data
1586 contained within an X509 certificate. However it was failing to correctly
1587 handle any errors that may occur while parsing the issuer field (which might
1588 occur if the issuer field is maliciously constructed). This may subsequently
1589 result in a NULL pointer deref and a crash leading to a potential denial of
1595 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1596 padding mode to correctly check for rollback attacks. This is considered a
1597 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1602 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1603 functions. Previously they could overflow the output length argument in some
1604 cases where the input length is close to the maximum permissable length for
1605 an integer on the platform. In such cases the return value from the function
1606 call would be 1 (indicating success), but the output length value would be
1607 negative. This could cause applications to behave incorrectly or crash.
1612 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1613 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1614 could be exploited in a side channel attack to recover the password. Since
1615 the attack is local host only this is outside of the current OpenSSL
1616 threat model and therefore no CVE is assigned.
1618 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1623 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1625 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1626 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1627 If an attacker can control both items being compared then this could lead
1628 to a possible denial of service attack. OpenSSL itself uses the
1629 GENERAL_NAME_cmp function for two purposes:
1630 1) Comparing CRL distribution point names between an available CRL and a
1631 CRL distribution point embedded in an X509 certificate
1632 2) When verifying that a timestamp response token signer matches the
1633 timestamp authority name (exposed via the API functions
1634 TS_RESP_verify_response and TS_RESP_verify_token)
1639 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1641 * Certificates with explicit curve parameters are now disallowed in
1642 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1646 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1647 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1648 conversely, silently ignore DTLS protocol version bounds when configuring
1649 TLS-based contexts. The commands can be repeated to set bounds of both
1650 types. The same applies with the corresponding "min_protocol" and
1651 "max_protocol" command-line switches, in case some application uses both TLS
1654 SSL_CTX instances that are created for a fixed protocol version (e.g.
1655 TLSv1_server_method()) also silently ignore version bounds. Previously
1656 attempts to apply bounds to these protocol versions would result in an
1657 error. Now only the "version-flexible" SSL_CTX instances are subject to
1658 limits in configuration files in command-line options.
1662 * Handshake now fails if Extended Master Secret extension is dropped
1667 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1669 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1671 * Fixed segmentation fault in SSL_check_chain()
1672 Server or client applications that call the SSL_check_chain() function
1673 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1674 dereference as a result of incorrect handling of the
1675 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1676 or unrecognised signature algorithm is received from the peer. This could
1677 be exploited by a malicious peer in a Denial of Service attack.
1682 * Added AES consttime code for no-asm configurations
1683 an optional constant time support for AES was added
1684 when building openssl for no-asm.
1685 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1686 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1687 At this time this feature is by default disabled.
1688 It will be enabled by default in 3.0.
1692 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1694 * Revert the change of EOF detection while reading in libssl to avoid
1695 regressions in applications depending on the current way of reporting
1696 the EOF. As the existing method is not fully accurate the change to
1697 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1698 branch and will be present in the 3.0 release.
1702 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1703 when primes for RSA keys are computed.
1704 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1705 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1706 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1707 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1708 This avoids possible fingerprinting of newly generated RSA modules.
1712 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1714 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1715 while reading in libssl then we would report an error back to the
1716 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1717 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1718 therefore give a hint as to what went wrong.
1722 * Check that ed25519 and ed448 are allowed by the security level. Previously
1723 signature algorithms not using an MD were not being checked that they were
1724 allowed by the security level.
1728 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1729 was not quite right. The behaviour was not consistent between resumption
1730 and normal handshakes, and also not quite consistent with historical
1731 behaviour. The behaviour in various scenarios has been clarified and
1732 it has been updated to make it match historical behaviour as closely as
1737 * *[VMS only]* The header files that the VMS compilers include automatically,
1738 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1739 that the C++ compiler doesn't understand. This is a shortcoming in the
1740 compiler, but can be worked around with `__cplusplus` guards.
1742 C++ applications that use OpenSSL libraries must be compiled using the
1743 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1744 functions. Otherwise, only functions with symbols of less than 31
1745 characters can be used, as the linker will not be able to successfully
1746 resolve symbols with longer names.
1750 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1751 The presence of this system service is determined at run-time.
1755 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1760 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1762 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1763 number generator (RNG). This was intended to include protection in the
1764 event of a fork() system call in order to ensure that the parent and child
1765 processes did not share the same RNG state. However this protection was not
1766 being used in the default case.
1768 A partial mitigation for this issue is that the output from a high
1769 precision timer is mixed into the RNG state so the likelihood of a parent
1770 and child process sharing state is significantly reduced.
1772 If an application already calls OPENSSL_init_crypto() explicitly using
1773 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1776 *Matthias St. Pierre*
1778 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1779 used even when parsing explicit parameters, when loading a encoded key
1780 or calling `EC_GROUP_new_from_ecpkparameters()`/
1781 `EC_GROUP_new_from_ecparameters()`.
1782 This prevents bypass of security hardening and performance gains,
1783 especially for curves with specialized EC_METHODs.
1784 By default, if a key encoded with explicit parameters is loaded and later
1785 encoded, the output is still encoded with explicit parameters, even if
1786 internally a "named" EC_GROUP is used for computation.
1790 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1791 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1792 NULL. After this change, only the cofactor parameter can be NULL. It also
1793 does some minimal sanity checks on the passed order.
1798 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1799 An attack is simple, if the first CMS_recipientInfo is valid but the
1800 second CMS_recipientInfo is chosen ciphertext. If the second
1801 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1802 encryption key will be replaced by garbage, and the message cannot be
1803 decoded, but if the RSA decryption fails, the correct encryption key is
1804 used and the recipient will not notice the attack.
1805 As a work around for this potential attack the length of the decrypted
1806 key must be equal to the cipher default key length, in case the
1807 certifiate is not given and all recipientInfo are tried out.
1808 The old behaviour can be re-enabled in the CMS code by setting the
1809 CMS_DEBUG_DECRYPT flag.
1814 * Early start up entropy quality from the DEVRANDOM seed source has been
1815 improved for older Linux systems. The RAND subsystem will wait for
1816 /dev/random to be producing output before seeding from /dev/urandom.
1817 The seeded state is stored for future library initialisations using
1818 a system global shared memory segment. The shared memory identifier
1819 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1820 the desired value. The default identifier is 114.
1824 * Correct the extended master secret constant on EBCDIC systems. Without this
1825 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1826 negotiate EMS will fail. Unfortunately this also means that TLS connections
1827 between EBCDIC systems with this fix, and EBCDIC systems without this
1828 fix will fail if they negotiate EMS.
1832 * Use Windows installation paths in the mingw builds
1834 Mingw isn't a POSIX environment per se, which means that Windows
1835 paths should be used for installation.
1840 * Changed DH_check to accept parameters with order q and 2q subgroups.
1841 With order 2q subgroups the bit 0 of the private key is not secret
1842 but DH_generate_key works around that by clearing bit 0 of the
1843 private key for those. This avoids leaking bit 0 of the private key.
1847 * Significantly reduce secure memory usage by the randomness pools.
1851 * Revert the DEVRANDOM_WAIT feature for Linux systems
1853 The DEVRANDOM_WAIT feature added a select() call to wait for the
1854 /dev/random device to become readable before reading from the
1855 /dev/urandom device.
1857 It turned out that this change had negative side effects on
1858 performance which were not acceptable. After some discussion it
1859 was decided to revert this feature and leave it up to the OS
1860 resp. the platform maintainer to ensure a proper initialization
1861 during early boot time.
1863 *Matthias St. Pierre*
1865 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1867 * Add build tests for C++. These are generated files that only do one
1868 thing, to include one public OpenSSL head file each. This tests that
1869 the public header files can be usefully included in a C++ application.
1871 This test isn't enabled by default. It can be enabled with the option
1872 'enable-buildtest-c++'.
1876 * Enable SHA3 pre-hashing for ECDSA and DSA.
1880 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1881 This changes the size when using the `genpkey` command when no size is given.
1882 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1883 generation commands to use 2048 bits by default.
1887 * Reorganize the manual pages to consistently have RETURN VALUES,
1888 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1889 util/fix-doc-nits accordingly.
1891 *Paul Yang, Joshua Lock*
1893 * Add the missing accessor EVP_PKEY_get0_engine()
1897 * Have commands like `s_client` and `s_server` output the signature scheme
1898 along with other cipher suite parameters when debugging.
1902 * Make OPENSSL_config() error agnostic again.
1906 * Do the error handling in RSA decryption constant time.
1910 * Prevent over long nonces in ChaCha20-Poly1305.
1912 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1913 for every encryption operation. RFC 7539 specifies that the nonce value
1914 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1915 and front pads the nonce with 0 bytes if it is less than 12
1916 bytes. However it also incorrectly allows a nonce to be set of up to 16
1917 bytes. In this case only the last 12 bytes are significant and any
1918 additional leading bytes are ignored.
1920 It is a requirement of using this cipher that nonce values are
1921 unique. Messages encrypted using a reused nonce value are susceptible to
1922 serious confidentiality and integrity attacks. If an application changes
1923 the default nonce length to be longer than 12 bytes and then makes a
1924 change to the leading bytes of the nonce expecting the new value to be a
1925 new unique nonce then such an application could inadvertently encrypt
1926 messages with a reused nonce.
1928 Additionally the ignored bytes in a long nonce are not covered by the
1929 integrity guarantee of this cipher. Any application that relies on the
1930 integrity of these ignored leading bytes of a long nonce may be further
1931 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1932 is safe because no such use sets such a long nonce value. However user
1933 applications that use this cipher directly and set a non-default nonce
1934 length to be longer than 12 bytes may be vulnerable.
1936 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1942 * Add DEVRANDOM_WAIT feature for Linux systems
1944 On older Linux systems where the getrandom() system call is not available,
1945 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1946 Contrary to getrandom(), the /dev/urandom device will not block during
1947 early boot when the kernel CSPRNG has not been seeded yet.
1949 To mitigate this known weakness, use select() to wait for /dev/random to
1950 become readable before reading from /dev/urandom.
1952 * Ensure that SM2 only uses SM3 as digest algorithm
1956 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1958 * Change the info callback signals for the start and end of a post-handshake
1959 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1960 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1961 confused by this and assume that a TLSv1.2 renegotiation has started. This
1962 can break KeyUpdate handling. Instead we no longer signal the start and end
1963 of a post handshake message exchange (although the messages themselves are
1964 still signalled). This could break some applications that were expecting
1965 the old signals. However without this KeyUpdate is not usable for many
1970 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1972 * Timing vulnerability in DSA signature generation
1974 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1975 timing side channel attack. An attacker could use variations in the signing
1976 algorithm to recover the private key.
1978 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1983 * Timing vulnerability in ECDSA signature generation
1985 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1986 timing side channel attack. An attacker could use variations in the signing
1987 algorithm to recover the private key.
1989 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1994 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1995 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1996 of two gigabytes and the error handling improved.
1998 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1999 categorized as a normal bug, not a security issue, because the DRBG reseeds
2000 automatically and is fully functional even without additional randomness
2001 provided by the application.
2003 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2005 * Add a new ClientHello callback. Provides a callback interface that gives
2006 the application the ability to adjust the nascent SSL object at the
2007 earliest stage of ClientHello processing, immediately after extensions have
2008 been collected but before they have been processed. In particular, this
2009 callback can adjust the supported TLS versions in response to the contents
2014 * Add SM2 base algorithm support.
2018 * s390x assembly pack: add (improved) hardware-support for the following
2019 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2020 aes-cfb/cfb8, aes-ecb.
2024 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2025 parameter is no longer accepted, as it leads to a corrupt table. NULL
2026 pem_str is reserved for alias entries only.
2030 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2031 step for prime curves. The new implementation is based on formulae from
2032 differential addition-and-doubling in homogeneous projective coordinates
2033 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2034 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2035 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2036 to work in projective coordinates.
2038 *Billy Bob Brumley, Nicola Tuveri*
2040 * Change generating and checking of primes so that the error rate of not
2041 being prime depends on the intended use based on the size of the input.
2042 For larger primes this will result in more rounds of Miller-Rabin.
2043 The maximal error rate for primes with more than 1080 bits is lowered
2046 *Kurt Roeckx, Annie Yousar*
2048 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2052 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2053 moving between systems, and to avoid confusion when a Windows build is
2054 done with mingw vs with MSVC. For POSIX installs, there's still a
2055 symlink or copy named 'tsget' to avoid that confusion as well.
2059 * Revert blinding in ECDSA sign and instead make problematic addition
2060 length-invariant. Switch even to fixed-length Montgomery multiplication.
2064 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2065 step for binary curves. The new implementation is based on formulae from
2066 differential addition-and-doubling in mixed Lopez-Dahab projective
2067 coordinates, modified to independently blind the operands.
2069 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2071 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2072 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2073 EC_METHODs to implement their own specialized "ladder step", to take
2074 advantage of more favorable coordinate systems or more efficient
2075 differential addition-and-doubling algorithms.
2077 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2079 * Modified the random device based seed sources to keep the relevant
2080 file descriptors open rather than reopening them on each access.
2081 This allows such sources to operate in a chroot() jail without
2082 the associated device nodes being available. This behaviour can be
2083 controlled using RAND_keep_random_devices_open().
2087 * Numerous side-channel attack mitigations have been applied. This may have
2088 performance impacts for some algorithms for the benefit of improved
2089 security. Specific changes are noted in this change log by their respective
2094 * AIX shared library support overhaul. Switch to AIX "natural" way of
2095 handling shared libraries, which means collecting shared objects of
2096 different versions and bitnesses in one common archive. This allows to
2097 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2098 doesn't affect the way 3rd party applications are linked, only how
2099 multi-version installation is managed.
2103 * Make ec_group_do_inverse_ord() more robust and available to other
2104 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2105 mitigations are applied to the fallback BN_mod_inverse().
2106 When using this function rather than BN_mod_inverse() directly, new
2107 EC cryptosystem implementations are then safer-by-default.
2111 * Add coordinate blinding for EC_POINT and implement projective
2112 coordinate blinding for generic prime curves as a countermeasure to
2113 chosen point SCA attacks.
2115 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2117 * Add blinding to ECDSA and DSA signatures to protect against side channel
2118 attacks discovered by Keegan Ryan (NCC Group).
2122 * Enforce checking in the `pkeyutl` command to ensure that the input
2123 length does not exceed the maximum supported digest length when performing
2124 a sign, verify or verifyrecover operation.
2128 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2129 I/O in combination with something like select() or poll() will hang. This
2130 can be turned off again using SSL_CTX_clear_mode().
2131 Many applications do not properly handle non-application data records, and
2132 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2133 around the problems in those applications, but can also break some.
2134 It's recommended to read the manpages about SSL_read(), SSL_write(),
2135 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2136 SSL_CTX_set_read_ahead() again.
2140 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2141 now allow empty (zero character) pass phrases.
2145 * Apply blinding to binary field modular inversion and remove patent
2146 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2150 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2151 binary and prime elliptic curves.
2155 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2156 constant time fixed point multiplication.
2160 * Revise elliptic curve scalar multiplication with timing attack
2161 defenses: ec_wNAF_mul redirects to a constant time implementation
2162 when computing fixed point and variable point multiplication (which
2163 in OpenSSL are mostly used with secret scalars in keygen, sign,
2164 ECDH derive operations).
2165 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2168 * Updated CONTRIBUTING
2172 * Updated DRBG / RAND to request nonce and additional low entropy
2173 randomness from the system.
2175 *Matthias St. Pierre*
2177 * Updated 'openssl rehash' to use OpenSSL consistent default.
2181 * Moved the load of the ssl_conf module to libcrypto, which helps
2182 loading engines that libssl uses before libssl is initialised.
2186 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2190 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2192 *Ingo Schwarze, Rich Salz*
2194 * Added output of accepting IP address and port for 'openssl s_server'
2198 * Added a new API for TLSv1.3 ciphersuites:
2199 SSL_CTX_set_ciphersuites()
2200 SSL_set_ciphersuites()
2204 * Memory allocation failures consistently add an error to the error
2209 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2210 in libcrypto when run as setuid/setgid.
2214 * Load any config file by default when libssl is used.
2218 * Added new public header file <openssl/rand_drbg.h> and documentation
2219 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2221 *Matthias St. Pierre*
2223 * QNX support removed (cannot find contributors to get their approval
2224 for the license change).
2228 * TLSv1.3 replay protection for early data has been implemented. See the
2229 SSL_read_early_data() man page for further details.
2233 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2234 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2235 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2236 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2237 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2238 configuration has been separated out. See the ciphers man page or the
2239 SSL_CTX_set_ciphersuites() man page for more information.
2243 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2244 in responder mode now supports the new "-multi" option, which
2245 spawns the specified number of child processes to handle OCSP
2246 requests. The "-timeout" option now also limits the OCSP
2247 responder's patience to wait to receive the full client request
2248 on a newly accepted connection. Child processes are respawned
2249 as needed, and the CA index file is automatically reloaded
2250 when changed. This makes it possible to run the "ocsp" responder
2251 as a long-running service, making the OpenSSL CA somewhat more
2252 feature-complete. In this mode, most diagnostic messages logged
2253 after entering the event loop are logged via syslog(3) rather than
2258 * Added support for X448 and Ed448. Heavily based on original work by
2263 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2264 objects loaded. This adds the functions OSSL_STORE_expect() and
2265 OSSL_STORE_find() as well as needed tools to construct searches and
2266 get the search data out of them.
2270 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2271 version of OpenSSL should review their configuration settings to ensure
2272 that they are still appropriate for TLSv1.3. For further information see:
2273 <https://wiki.openssl.org/index.php/TLS1.3>
2277 * Grand redesign of the OpenSSL random generator
2279 The default RAND method now utilizes an AES-CTR DRBG according to
2280 NIST standard SP 800-90Ar1. The new random generator is essentially
2281 a port of the default random generator from the OpenSSL FIPS 2.0
2282 object module. It is a hybrid deterministic random bit generator
2283 using an AES-CTR bit stream and which seeds and reseeds itself
2284 automatically using trusted system entropy sources.
2286 Some of its new features are:
2287 - Support for multiple DRBG instances with seed chaining.
2288 - The default RAND method makes use of a DRBG.
2289 - There is a public and private DRBG instance.
2290 - The DRBG instances are fork-safe.
2291 - Keep all global DRBG instances on the secure heap if it is enabled.
2292 - The public and private DRBG instance are per thread for lock free
2295 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2297 * Changed Configure so it only says what it does and doesn't dump
2298 so much data. Instead, ./configdata.pm should be used as a script
2299 to display all sorts of configuration data.
2303 * Added processing of "make variables" to Configure.
2307 * Added SHA512/224 and SHA512/256 algorithm support.
2311 * The last traces of Netware support, first removed in 1.1.0, have
2316 * Get rid of Makefile.shared, and in the process, make the processing
2317 of certain files (rc.obj, or the .def/.map/.opt files produced from
2318 the ordinal files) more visible and hopefully easier to trace and
2319 debug (or make silent).
2323 * Make it possible to have environment variable assignments as
2324 arguments to config / Configure.
2328 * Add multi-prime RSA (RFC 8017) support.
2332 * Add SM3 implemented according to GB/T 32905-2016
2333 *Jack Lloyd <jack.lloyd@ribose.com>,*
2334 *Ronald Tse <ronald.tse@ribose.com>,*
2335 *Erick Borsboom <erick.borsboom@ribose.com>*
2337 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2338 as documented in RFC6066.
2339 Based on a patch from Tomasz Moń
2341 *Filipe Raimundo da Silva*
2343 * Add SM4 implemented according to GB/T 32907-2016.
2344 *Jack Lloyd <jack.lloyd@ribose.com>,*
2345 *Ronald Tse <ronald.tse@ribose.com>,*
2346 *Erick Borsboom <erick.borsboom@ribose.com>*
2348 * Reimplement -newreq-nodes and ERR_error_string_n; the
2349 original author does not agree with the license change.
2353 * Add ARIA AEAD TLS support.
2357 * Some macro definitions to support VS6 have been removed. Visual
2358 Studio 6 has not worked since 1.1.0
2362 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2363 without clearing the errors.
2367 * Add "atfork" functions. If building on a system that without
2368 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2369 requirements. The RAND facility now uses/requires this.
2377 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2378 not possible to disable entirely. However, it's still possible to
2379 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2382 To disable, configure with 'no-ui-console'. 'no-ui' is still
2383 possible to use as an alias. Check at compile time with the
2384 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2385 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2389 * Add a STORE module, which implements a uniform and URI based reader of
2390 stores that can contain keys, certificates, CRLs and numerous other
2391 objects. The main API is loosely based on a few stdio functions,
2392 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2393 OSSL_STORE_error and OSSL_STORE_close.
2394 The implementation uses backends called "loaders" to implement arbitrary
2395 URI schemes. There is one built in "loader" for the 'file' scheme.
2399 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2400 then adjusted to work on FreeBSD 8.4 as well.
2401 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2402 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2406 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2407 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2408 error code calls like this:
2410 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2412 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2413 that can be encoded in C. For the foreseeable future, this will only
2416 *Richard Levitte and Tim Hudson*
2418 * Removed BSD cryptodev engine.
2422 * Add a build target 'build_all_generated', to build all generated files
2423 and only that. This can be used to prepare everything that requires
2424 things like perl for a system that lacks perl and then move everything
2425 to that system and do the rest of the build there.
2429 * In the UI interface, make it possible to duplicate the user data. This
2430 can be used by engines that need to retain the data for a longer time
2431 than just the call where this user data is passed.
2435 * Ignore the '-named_curve auto' value for compatibility of applications
2438 *Tomáš Mráz <tmraz@fedoraproject.org>*
2440 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2441 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2442 alerts across multiple records (some of which could be empty). In practice
2443 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2444 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2445 support this at all. Supporting it adds significant complexity to the
2446 record layer, and its removal is unlikely to cause interoperability
2451 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2452 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2453 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2458 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2459 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2461 *Richard Levitte, Andy Polyakov*
2463 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2468 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2469 platform rather than 'mingw'.
2473 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2474 success if they are asked to add an object which already exists
2475 in the store. This change cascades to other functions which load
2476 certificates and CRLs.
2480 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2481 facilitate stack unwinding even from assembly subroutines.
2485 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2486 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2490 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2491 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2492 which is the minimum version we support.
2496 * Certificate time validation (X509_cmp_time) enforces stricter
2497 compliance with RFC 5280. Fractional seconds and timezone offsets
2498 are no longer allowed.
2502 * Add support for ARIA
2506 * s_client will now send the Server Name Indication (SNI) extension by
2507 default unless the new "-noservername" option is used. The server name is
2508 based on the host provided to the "-connect" option unless overridden by
2509 using "-servername".
2513 * Add support for SipHash
2517 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2518 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2519 prevent issues where no progress is being made and the peer continually
2520 sends unrecognised record types, using up resources processing them.
2524 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2525 using the algorithm defined in
2526 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2530 * Heartbeat support has been removed; the ABI is changed for now.
2532 *Richard Levitte, Rich Salz*
2534 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2538 * The RSA "null" method, which was partially supported to avoid patent
2539 issues, has been replaced to always returns NULL.
2546 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2548 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2549 used even when parsing explicit parameters, when loading a encoded key
2550 or calling `EC_GROUP_new_from_ecpkparameters()`/
2551 `EC_GROUP_new_from_ecparameters()`.
2552 This prevents bypass of security hardening and performance gains,
2553 especially for curves with specialized EC_METHODs.
2554 By default, if a key encoded with explicit parameters is loaded and later
2555 encoded, the output is still encoded with explicit parameters, even if
2556 internally a "named" EC_GROUP is used for computation.
2560 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2561 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2562 NULL. After this change, only the cofactor parameter can be NULL. It also
2563 does some minimal sanity checks on the passed order.
2568 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2569 An attack is simple, if the first CMS_recipientInfo is valid but the
2570 second CMS_recipientInfo is chosen ciphertext. If the second
2571 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2572 encryption key will be replaced by garbage, and the message cannot be
2573 decoded, but if the RSA decryption fails, the correct encryption key is
2574 used and the recipient will not notice the attack.
2575 As a work around for this potential attack the length of the decrypted
2576 key must be equal to the cipher default key length, in case the
2577 certifiate is not given and all recipientInfo are tried out.
2578 The old behaviour can be re-enabled in the CMS code by setting the
2579 CMS_DEBUG_DECRYPT flag.
2584 * Use Windows installation paths in the mingw builds
2586 Mingw isn't a POSIX environment per se, which means that Windows
2587 paths should be used for installation.
2592 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2594 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2595 This changes the size when using the `genpkey` command when no size is given.
2596 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2597 generation commands to use 2048 bits by default.
2601 * Prevent over long nonces in ChaCha20-Poly1305.
2603 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2604 for every encryption operation. RFC 7539 specifies that the nonce value
2605 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2606 and front pads the nonce with 0 bytes if it is less than 12
2607 bytes. However it also incorrectly allows a nonce to be set of up to 16
2608 bytes. In this case only the last 12 bytes are significant and any
2609 additional leading bytes are ignored.
2611 It is a requirement of using this cipher that nonce values are
2612 unique. Messages encrypted using a reused nonce value are susceptible to
2613 serious confidentiality and integrity attacks. If an application changes
2614 the default nonce length to be longer than 12 bytes and then makes a
2615 change to the leading bytes of the nonce expecting the new value to be a
2616 new unique nonce then such an application could inadvertently encrypt
2617 messages with a reused nonce.
2619 Additionally the ignored bytes in a long nonce are not covered by the
2620 integrity guarantee of this cipher. Any application that relies on the
2621 integrity of these ignored leading bytes of a long nonce may be further
2622 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2623 is safe because no such use sets such a long nonce value. However user
2624 applications that use this cipher directly and set a non-default nonce
2625 length to be longer than 12 bytes may be vulnerable.
2627 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2633 * Added SCA hardening for modular field inversion in EC_GROUP through
2634 a new dedicated field_inv() pointer in EC_METHOD.
2635 This also addresses a leakage affecting conversions from projective
2636 to affine coordinates.
2638 *Billy Bob Brumley, Nicola Tuveri*
2640 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2641 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2645 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2649 * Remove the 'dist' target and add a tarball building script. The
2650 'dist' target has fallen out of use, and it shouldn't be
2651 necessary to configure just to create a source distribution.
2655 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2657 * Timing vulnerability in DSA signature generation
2659 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2660 timing side channel attack. An attacker could use variations in the signing
2661 algorithm to recover the private key.
2663 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2668 * Timing vulnerability in ECDSA signature generation
2670 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2671 timing side channel attack. An attacker could use variations in the signing
2672 algorithm to recover the private key.
2674 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2679 * Add coordinate blinding for EC_POINT and implement projective
2680 coordinate blinding for generic prime curves as a countermeasure to
2681 chosen point SCA attacks.
2683 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2685 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2687 * Client DoS due to large DH parameter
2689 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2690 malicious server can send a very large prime value to the client. This will
2691 cause the client to spend an unreasonably long period of time generating a
2692 key for this prime resulting in a hang until the client has finished. This
2693 could be exploited in a Denial Of Service attack.
2695 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2700 * Cache timing vulnerability in RSA Key Generation
2702 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2703 a cache timing side channel attack. An attacker with sufficient access to
2704 mount cache timing attacks during the RSA key generation process could
2705 recover the private key.
2707 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2708 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2713 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2714 parameter is no longer accepted, as it leads to a corrupt table. NULL
2715 pem_str is reserved for alias entries only.
2719 * Revert blinding in ECDSA sign and instead make problematic addition
2720 length-invariant. Switch even to fixed-length Montgomery multiplication.
2724 * Change generating and checking of primes so that the error rate of not
2725 being prime depends on the intended use based on the size of the input.
2726 For larger primes this will result in more rounds of Miller-Rabin.
2727 The maximal error rate for primes with more than 1080 bits is lowered
2730 *Kurt Roeckx, Annie Yousar*
2732 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2736 * Add blinding to ECDSA and DSA signatures to protect against side channel
2737 attacks discovered by Keegan Ryan (NCC Group).
2741 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2742 now allow empty (zero character) pass phrases.
2746 * Certificate time validation (X509_cmp_time) enforces stricter
2747 compliance with RFC 5280. Fractional seconds and timezone offsets
2748 are no longer allowed.
2752 * Fixed a text canonicalisation bug in CMS
2754 Where a CMS detached signature is used with text content the text goes
2755 through a canonicalisation process first prior to signing or verifying a
2756 signature. This process strips trailing space at the end of lines, converts
2757 line terminators to CRLF and removes additional trailing line terminators
2758 at the end of a file. A bug in the canonicalisation process meant that
2759 some characters, such as form-feed, were incorrectly treated as whitespace
2760 and removed. This is contrary to the specification (RFC5485). This fix
2761 could mean that detached text data signed with an earlier version of
2762 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2763 signed with a fixed OpenSSL may fail to verify with an earlier version of
2764 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2765 and use the "-binary" flag (for the "cms" command line application) or set
2766 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2770 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2772 * Constructed ASN.1 types with a recursive definition could exceed the stack
2774 Constructed ASN.1 types with a recursive definition (such as can be found
2775 in PKCS7) could eventually exceed the stack given malicious input with
2776 excessive recursion. This could result in a Denial Of Service attack. There
2777 are no such structures used within SSL/TLS that come from untrusted sources
2778 so this is considered safe.
2780 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2786 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2788 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2789 effectively reduced to only comparing the least significant bit of each
2790 byte. This allows an attacker to forge messages that would be considered as
2791 authenticated in an amount of tries lower than that guaranteed by the
2792 security claims of the scheme. The module can only be compiled by the
2793 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2795 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2801 * Add a build target 'build_all_generated', to build all generated files
2802 and only that. This can be used to prepare everything that requires
2803 things like perl for a system that lacks perl and then move everything
2804 to that system and do the rest of the build there.
2808 * Backport SSL_OP_NO_RENGOTIATION
2810 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2811 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2812 changes this is no longer possible in 1.1.0. Therefore the new
2813 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2814 1.1.0 to provide equivalent functionality.
2816 Note that if an application built against 1.1.0h headers (or above) is run
2817 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2818 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2822 * Removed the OS390-Unix config target. It relied on a script that doesn't
2827 * rsaz_1024_mul_avx2 overflow bug on x86_64
2829 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2830 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2831 Analysis suggests that attacks against RSA and DSA as a result of this
2832 defect would be very difficult to perform and are not believed likely.
2833 Attacks against DH1024 are considered just feasible, because most of the
2834 work necessary to deduce information about a private key may be performed
2835 offline. The amount of resources required for such an attack would be
2836 significant. However, for an attack on TLS to be meaningful, the server
2837 would have to share the DH1024 private key among multiple clients, which is
2838 no longer an option since CVE-2016-0701.
2840 This only affects processors that support the AVX2 but not ADX extensions
2841 like Intel Haswell (4th generation).
2843 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2844 was originally found via the OSS-Fuzz project.
2849 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2851 * bn_sqrx8x_internal carry bug on x86_64
2853 There is a carry propagating bug in the x86_64 Montgomery squaring
2854 procedure. No EC algorithms are affected. Analysis suggests that attacks
2855 against RSA and DSA as a result of this defect would be very difficult to
2856 perform and are not believed likely. Attacks against DH are considered just
2857 feasible (although very difficult) because most of the work necessary to
2858 deduce information about a private key may be performed offline. The amount
2859 of resources required for such an attack would be very significant and
2860 likely only accessible to a limited number of attackers. An attacker would
2861 additionally need online access to an unpatched system using the target
2862 private key in a scenario with persistent DH parameters and a private
2863 key that is shared between multiple clients.
2865 This only affects processors that support the BMI1, BMI2 and ADX extensions
2866 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2868 This issue was reported to OpenSSL by the OSS-Fuzz project.
2873 * Malformed X.509 IPAddressFamily could cause OOB read
2875 If an X.509 certificate has a malformed IPAddressFamily extension,
2876 OpenSSL could do a one-byte buffer overread. The most likely result
2877 would be an erroneous display of the certificate in text format.
2879 This issue was reported to OpenSSL by the OSS-Fuzz project.
2884 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2886 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2887 platform rather than 'mingw'.
2891 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2892 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2893 which is the minimum version we support.
2897 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2899 * Encrypt-Then-Mac renegotiation crash
2901 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2902 negotiated where it was not in the original handshake (or vice-versa) then
2903 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2904 and servers are affected.
2906 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2911 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2913 * Truncated packet could crash via OOB read
2915 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2916 cipher is being used, then a truncated packet can cause that host to
2917 perform an out-of-bounds read, usually resulting in a crash.
2919 This issue was reported to OpenSSL by Robert Święcki of Google.
2924 * Bad (EC)DHE parameters cause a client crash
2926 If a malicious server supplies bad parameters for a DHE or ECDHE key
2927 exchange then this can result in the client attempting to dereference a
2928 NULL pointer leading to a client crash. This could be exploited in a Denial
2931 This issue was reported to OpenSSL by Guido Vranken.
2936 * BN_mod_exp may produce incorrect results on x86_64
2938 There is a carry propagating bug in the x86_64 Montgomery squaring
2939 procedure. No EC algorithms are affected. Analysis suggests that attacks
2940 against RSA and DSA as a result of this defect would be very difficult to
2941 perform and are not believed likely. Attacks against DH are considered just
2942 feasible (although very difficult) because most of the work necessary to
2943 deduce information about a private key may be performed offline. The amount
2944 of resources required for such an attack would be very significant and
2945 likely only accessible to a limited number of attackers. An attacker would
2946 additionally need online access to an unpatched system using the target
2947 private key in a scenario with persistent DH parameters and a private
2948 key that is shared between multiple clients. For example this can occur by
2949 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2950 similar to CVE-2015-3193 but must be treated as a separate problem.
2952 This issue was reported to OpenSSL by the OSS-Fuzz project.
2957 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2959 * ChaCha20/Poly1305 heap-buffer-overflow
2961 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2962 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2963 crash. This issue is not considered to be exploitable beyond a DoS.
2965 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2970 * CMS Null dereference
2972 Applications parsing invalid CMS structures can crash with a NULL pointer
2973 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2974 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2975 structure callback if an attempt is made to free certain invalid encodings.
2976 Only CHOICE structures using a callback which do not handle NULL value are
2979 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2984 * Montgomery multiplication may produce incorrect results
2986 There is a carry propagating bug in the Broadwell-specific Montgomery
2987 multiplication procedure that handles input lengths divisible by, but
2988 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2989 and DH private keys are impossible. This is because the subroutine in
2990 question is not used in operations with the private key itself and an input
2991 of the attacker's direct choice. Otherwise the bug can manifest itself as
2992 transient authentication and key negotiation failures or reproducible
2993 erroneous outcome of public-key operations with specially crafted input.
2994 Among EC algorithms only Brainpool P-512 curves are affected and one
2995 presumably can attack ECDH key negotiation. Impact was not analyzed in
2996 detail, because pre-requisites for attack are considered unlikely. Namely
2997 multiple clients have to choose the curve in question and the server has to
2998 share the private key among them, neither of which is default behaviour.
2999 Even then only clients that chose the curve will be affected.
3001 This issue was publicly reported as transient failures and was not
3002 initially recognized as a security issue. Thanks to Richard Morgan for
3003 providing reproducible case.
3008 * Removed automatic addition of RPATH in shared libraries and executables,
3009 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3013 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3015 * Fix Use After Free for large message sizes
3017 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3018 message larger than approx 16k is received then the underlying buffer to
3019 store the incoming message is reallocated and moved. Unfortunately a
3020 dangling pointer to the old location is left which results in an attempt to
3021 write to the previously freed location. This is likely to result in a
3022 crash, however it could potentially lead to execution of arbitrary code.
3024 This issue only affects OpenSSL 1.1.0a.
3026 This issue was reported to OpenSSL by Robert Święcki.
3031 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3033 * OCSP Status Request extension unbounded memory growth
3035 A malicious client can send an excessively large OCSP Status Request
3036 extension. If that client continually requests renegotiation, sending a
3037 large OCSP Status Request extension each time, then there will be unbounded
3038 memory growth on the server. This will eventually lead to a Denial Of
3039 Service attack through memory exhaustion. Servers with a default
3040 configuration are vulnerable even if they do not support OCSP. Builds using
3041 the "no-ocsp" build time option are not affected.
3043 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3048 * SSL_peek() hang on empty record
3050 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3051 sends an empty record. This could be exploited by a malicious peer in a
3052 Denial Of Service attack.
3054 This issue was reported to OpenSSL by Alex Gaynor.
3059 * Excessive allocation of memory in tls_get_message_header() and
3060 dtls1_preprocess_fragment()
3062 A (D)TLS message includes 3 bytes for its length in the header for the
3063 message. This would allow for messages up to 16Mb in length. Messages of
3064 this length are excessive and OpenSSL includes a check to ensure that a
3065 peer is sending reasonably sized messages in order to avoid too much memory
3066 being consumed to service a connection. A flaw in the logic of version
3067 1.1.0 means that memory for the message is allocated too early, prior to
3068 the excessive message length check. Due to way memory is allocated in
3069 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3070 to service a connection. This could lead to a Denial of Service through
3071 memory exhaustion. However, the excessive message length check still takes
3072 place, and this would cause the connection to immediately fail. Assuming
3073 that the application calls SSL_free() on the failed connection in a timely
3074 manner then the 21Mb of allocated memory will then be immediately freed
3075 again. Therefore the excessive memory allocation will be transitory in
3076 nature. This then means that there is only a security impact if:
3078 1) The application does not call SSL_free() in a timely manner in the event
3079 that the connection fails
3081 2) The application is working in a constrained environment where there is
3082 very little free memory
3084 3) The attacker initiates multiple connection attempts such that there are
3085 multiple connections in a state where memory has been allocated for the
3086 connection; SSL_free() has not yet been called; and there is insufficient
3087 memory to service the multiple requests.
3089 Except in the instance of (1) above any Denial Of Service is likely to be
3090 transitory because as soon as the connection fails the memory is
3091 subsequently freed again in the SSL_free() call. However there is an
3092 increased risk during this period of application crashes due to the lack of
3093 memory - which would then mean a more serious Denial of Service.
3095 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3096 (CVE-2016-6307 and CVE-2016-6308)
3100 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3101 had to be removed. Primary reason is that vendor assembler can't
3102 assemble our modules with -KPIC flag. As result it, assembly
3103 support, was not even available as option. But its lack means
3104 lack of side-channel resistant code, which is incompatible with
3105 security by todays standards. Fortunately gcc is readily available
3106 prepackaged option, which we firmly point at...
3110 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3112 * Windows command-line tool supports UTF-8 opt-in option for arguments
3113 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3114 (to any value) allows Windows user to access PKCS#12 file generated
3115 with Windows CryptoAPI and protected with non-ASCII password, as well
3116 as files generated under UTF-8 locale on Linux also protected with
3121 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3122 have been disabled by default and removed from DEFAULT, just like RC4.
3123 See the RC4 item below to re-enable both.
3127 * The method for finding the storage location for the Windows RAND seed file
3128 has changed. First we check %RANDFILE%. If that is not set then we check
3129 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3130 all else fails we fall back to C:\.
3134 * The EVP_EncryptUpdate() function has had its return type changed from void
3135 to int. A return of 0 indicates and error while a return of 1 indicates
3140 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3141 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3142 off the constant time implementation for RSA, DSA and DH have been made
3143 no-ops and deprecated.
3147 * Windows RAND implementation was simplified to only get entropy by
3148 calling CryptGenRandom(). Various other RAND-related tickets
3151 *Joseph Wylie Yandle, Rich Salz*
3153 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3154 and `OPENSSL_LH_`, respectively. The old names are available
3155 with API compatibility. They new names are now completely documented.
3159 * Unify TYPE_up_ref(obj) methods signature.
3160 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3161 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3162 int (instead of void) like all others TYPE_up_ref() methods.
3163 So now these methods also check the return value of CRYPTO_atomic_add(),
3164 and the validity of object reference counter.
3166 *fdasilvayy@gmail.com*
3168 * With Windows Visual Studio builds, the .pdb files are installed
3169 alongside the installed libraries and executables. For a static
3170 library installation, ossl_static.pdb is the associate compiler
3171 generated .pdb file to be used when linking programs.
3175 * Remove openssl.spec. Packaging files belong with the packagers.
3179 * Automatic Darwin/OSX configuration has had a refresh, it will now
3180 recognise x86_64 architectures automatically. You can still decide
3181 to build for a different bitness with the environment variable
3182 KERNEL_BITS (can be 32 or 64), for example:
3184 KERNEL_BITS=32 ./config
3188 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3189 256 bit AES and HMAC with SHA256.
3193 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3197 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3201 * To enable users to have their own config files and build file templates,
3202 Configure looks in the directory indicated by the environment variable
3203 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3204 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3205 name and is used as is.
3209 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3210 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3211 X509_CERT_FILE_CTX was removed.
3215 * "shared" builds are now the default. To create only static libraries use
3216 the "no-shared" Configure option.
3220 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3221 All of these option have not worked for some while and are fundamental
3226 * Make various cleanup routines no-ops and mark them as deprecated. Most
3227 global cleanup functions are no longer required because they are handled
3228 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3229 Explicitly de-initing can cause problems (e.g. where a library that uses
3230 OpenSSL de-inits, but an application is still using it). The affected
3231 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3232 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3233 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3234 COMP_zlib_cleanup().
3238 * --strict-warnings no longer enables runtime debugging options
3239 such as REF_DEBUG. Instead, debug options are automatically
3240 enabled with '--debug' builds.
3242 *Andy Polyakov, Emilia Käsper*
3244 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3245 have been moved out of the public header files. New functions for managing
3246 these have been added.
3250 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3251 objects have been moved out of the public header files. New
3252 functions for managing these have been added.
3256 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3257 have been moved out of the public header files. New functions for managing
3258 these have been added.
3262 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3263 moved out of the public header files. New functions for managing these
3268 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3272 * Removed the mk1mf build scripts.
3276 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3277 it is always safe to #include a header now.
3281 * Removed the aged BC-32 config and all its supporting scripts
3285 * Removed support for Ultrix, Netware, and OS/2.
3289 * Add support for HKDF.
3291 *Alessandro Ghedini*
3293 * Add support for blake2b and blake2s
3297 * Added support for "pipelining". Ciphers that have the
3298 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3299 encryptions/decryptions simultaneously. There are currently no built-in
3300 ciphers with this property but the expectation is that engines will be able
3301 to offer it to significantly improve throughput. Support has been extended
3302 into libssl so that multiple records for a single connection can be
3303 processed in one go (for >=TLS 1.1).
3307 * Added the AFALG engine. This is an async capable engine which is able to
3308 offload work to the Linux kernel. In this initial version it only supports
3309 AES128-CBC. The kernel must be version 4.1.0 or greater.
3313 * OpenSSL now uses a new threading API. It is no longer necessary to
3314 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3315 are two supported threading models: pthreads and windows threads. It is
3316 also possible to configure OpenSSL at compile time for "no-threads". The
3317 old threading API should no longer be used. The functions have been
3318 replaced with "no-op" compatibility macros.
3320 *Alessandro Ghedini, Matt Caswell*
3322 * Modify behavior of ALPN to invoke callback after SNI/servername
3323 callback, such that updates to the SSL_CTX affect ALPN.
3327 * Add SSL_CIPHER queries for authentication and key-exchange.
3331 * Changes to the DEFAULT cipherlist:
3332 - Prefer (EC)DHE handshakes over plain RSA.
3333 - Prefer AEAD ciphers over legacy ciphers.
3334 - Prefer ECDSA over RSA when both certificates are available.
3335 - Prefer TLSv1.2 ciphers/PRF.
3336 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3341 * Change the ECC default curve list to be this, in order: x25519,
3342 secp256r1, secp521r1, secp384r1.
3346 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3347 disabled by default. They can be re-enabled using the
3348 enable-weak-ssl-ciphers option to Configure.
3352 * If the server has ALPN configured, but supports no protocols that the
3353 client advertises, send a fatal "no_application_protocol" alert.
3354 This behaviour is SHALL in RFC 7301, though it isn't universally
3355 implemented by other servers.
3359 * Add X25519 support.
3360 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3361 for public and private key encoding using the format documented in
3362 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3363 key generation and key derivation.
3365 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3370 * Deprecate SRP_VBASE_get_by_user.
3371 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3372 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3373 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3374 seed, even if the seed is configured.
3376 Users should use SRP_VBASE_get1_by_user instead. Note that in
3377 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3378 also that even though configuring the SRP seed attempts to hide
3379 invalid usernames by continuing the handshake with fake
3380 credentials, this behaviour is not constant time and no strong
3381 guarantees are made that the handshake is indistinguishable from
3382 that of a valid user.
3386 * Configuration change; it's now possible to build dynamic engines
3387 without having to build shared libraries and vice versa. This
3388 only applies to the engines in `engines/`, those in `crypto/engine/`
3389 will always be built into libcrypto (i.e. "static").
3391 Building dynamic engines is enabled by default; to disable, use
3392 the configuration option "disable-dynamic-engine".
3394 The only requirements for building dynamic engines are the
3395 presence of the DSO module and building with position independent
3396 code, so they will also automatically be disabled if configuring
3397 with "disable-dso" or "disable-pic".
3399 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3400 are also taken away from openssl/opensslconf.h, as they are
3405 * Configuration change; if there is a known flag to compile
3406 position independent code, it will always be applied on the
3407 libcrypto and libssl object files, and never on the application
3408 object files. This means other libraries that use routines from
3409 libcrypto / libssl can be made into shared libraries regardless
3410 of how OpenSSL was configured.
3412 If this isn't desirable, the configuration options "disable-pic"
3413 or "no-pic" can be used to disable the use of PIC. This will
3414 also disable building shared libraries and dynamic engines.
3418 * Removed JPAKE code. It was experimental and has no wide use.
3422 * The INSTALL_PREFIX Makefile variable has been renamed to
3423 DESTDIR. That makes for less confusion on what this variable
3424 is for. Also, the configuration option --install_prefix is
3429 * Heartbeat for TLS has been removed and is disabled by default
3430 for DTLS; configure with enable-heartbeats. Code that uses the
3431 old #define's might need to be updated.
3433 *Emilia Käsper, Rich Salz*
3435 * Rename REF_CHECK to REF_DEBUG.
3439 * New "unified" build system
3441 The "unified" build system is aimed to be a common system for all
3442 platforms we support. With it comes new support for VMS.
3444 This system builds supports building in a different directory tree
3445 than the source tree. It produces one Makefile (for unix family
3446 or lookalikes), or one descrip.mms (for VMS).
3448 The source of information to make the Makefile / descrip.mms is
3449 small files called 'build.info', holding the necessary
3450 information for each directory with source to compile, and a
3451 template in Configurations, like unix-Makefile.tmpl or
3454 With this change, the library names were also renamed on Windows
3455 and on VMS. They now have names that are closer to the standard
3456 on Unix, and include the major version number, and in certain
3457 cases, the architecture they are built for. See "Notes on shared
3458 libraries" in INSTALL.
3460 We rely heavily on the perl module Text::Template.
3464 * Added support for auto-initialisation and de-initialisation of the library.
3465 OpenSSL no longer requires explicit init or deinit routines to be called,
3466 except in certain circumstances. See the OPENSSL_init_crypto() and
3467 OPENSSL_init_ssl() man pages for further information.
3471 * The arguments to the DTLSv1_listen function have changed. Specifically the
3472 "peer" argument is now expected to be a BIO_ADDR object.
3474 * Rewrite of BIO networking library. The BIO library lacked consistent
3475 support of IPv6, and adding it required some more extensive
3476 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3477 which hold all types of addresses and chains of address information.
3478 It also introduces a new API, with functions like BIO_socket,
3479 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3480 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3481 have been adapted accordingly.
3485 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3490 * CRIME protection: disable compression by default, even if OpenSSL is
3491 compiled with zlib enabled. Applications can still enable compression
3492 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3493 using the SSL_CONF library to configure compression.
3497 * The signature of the session callback configured with
3498 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3499 was explicitly marked as `const unsigned char*` instead of
3504 * Always DPURIFY. Remove the use of uninitialized memory in the
3505 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3509 * Removed many obsolete configuration items, including
3510 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3511 MD2_CHAR, MD2_INT, MD2_LONG
3513 IDEA_SHORT, IDEA_LONG
3514 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3516 *Rich Salz, with advice from Andy Polyakov*
3518 * Many BN internals have been moved to an internal header file.
3520 *Rich Salz with help from Andy Polyakov*
3522 * Configuration and writing out the results from it has changed.
3523 Files such as Makefile include/openssl/opensslconf.h and are now
3524 produced through general templates, such as Makefile.in and
3525 crypto/opensslconf.h.in and some help from the perl module
3528 Also, the center of configuration information is no longer
3529 Makefile. Instead, Configure produces a perl module in
3530 configdata.pm which holds most of the config data (in the hash
3531 table %config), the target data that comes from the target
3532 configuration in one of the `Configurations/*.conf` files (in
3537 * To clarify their intended purposes, the Configure options
3538 --prefix and --openssldir change their semantics, and become more
3539 straightforward and less interdependent.
3541 --prefix shall be used exclusively to give the location INSTALLTOP
3542 where programs, scripts, libraries, include files and manuals are
3543 going to be installed. The default is now /usr/local.
3545 --openssldir shall be used exclusively to give the default
3546 location OPENSSLDIR where certificates, private keys, CRLs are
3547 managed. This is also where the default openssl.cnf gets
3549 If the directory given with this option is a relative path, the
3550 values of both the --prefix value and the --openssldir value will
3551 be combined to become OPENSSLDIR.
3552 The default for --openssldir is INSTALLTOP/ssl.
3554 Anyone who uses --openssldir to specify where OpenSSL is to be
3555 installed MUST change to use --prefix instead.
3559 * The GOST engine was out of date and therefore it has been removed. An up
3560 to date GOST engine is now being maintained in an external repository.
3561 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3562 support for GOST ciphersuites (these are only activated if a GOST engine
3567 * EGD is no longer supported by default; use enable-egd when
3570 *Ben Kaduk and Rich Salz*
3572 * The distribution now has Makefile.in files, which are used to
3573 create Makefile's when Configure is run. *Configure must be run
3574 before trying to build now.*
3578 * The return value for SSL_CIPHER_description() for error conditions
3583 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3585 Obtaining and performing DNSSEC validation of TLSA records is
3586 the application's responsibility. The application provides
3587 the TLSA records of its choice to OpenSSL, and these are then
3588 used to authenticate the peer.
3590 The TLSA records need not even come from DNS. They can, for
3591 example, be used to implement local end-entity certificate or
3592 trust-anchor "pinning", where the "pin" data takes the form
3593 of TLSA records, which can augment or replace verification
3594 based on the usual WebPKI public certification authorities.
3598 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3599 continues to support deprecated interfaces in default builds.
3600 However, applications are strongly advised to compile their
3601 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3602 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3603 or the 1.1.0 releases.
3605 In environments in which all applications have been ported to
3606 not use any deprecated interfaces OpenSSL's Configure script
3607 should be used with the --api=1.1.0 option to entirely remove
3608 support for the deprecated features from the library and
3609 unconditionally disable them in the installed headers.
3610 Essentially the same effect can be achieved with the "no-deprecated"
3611 argument to Configure, except that this will always restrict
3612 the build to just the latest API, rather than a fixed API
3615 As applications are ported to future revisions of the API,
3616 they should update their compile-time OPENSSL_API_COMPAT define
3617 accordingly, but in most cases should be able to continue to
3618 compile with later releases.
3620 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3621 0x10000000L and 0x00908000L, respectively. However those
3622 versions did not support the OPENSSL_API_COMPAT feature, and
3623 so applications are not typically tested for explicit support
3624 of just the undeprecated features of either release.
3628 * Add support for setting the minimum and maximum supported protocol.
3629 It can bet set via the SSL_set_min_proto_version() and
3630 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3631 MaxProtocol. It's recommended to use the new APIs to disable
3632 protocols instead of disabling individual protocols using
3633 SSL_set_options() or SSL_CONF's Protocol. This change also
3634 removes support for disabling TLS 1.2 in the OpenSSL TLS
3635 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3639 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3643 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3644 and integrates ECDSA and ECDH functionality into EC. Implementations can
3645 now redirect key generation and no longer need to convert to or from
3648 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3649 include the ec.h header file instead.
3653 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3654 ciphers who are no longer supported and drops support the ephemeral RSA key
3655 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3659 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3660 opaque. For HMAC_CTX, the following constructors and destructors
3663 HMAC_CTX *HMAC_CTX_new(void);
3664 void HMAC_CTX_free(HMAC_CTX *ctx);
3666 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3667 destroy such methods has been added. See EVP_MD_meth_new(3) and
3668 EVP_CIPHER_meth_new(3) for documentation.
3671 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3672 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3673 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3674 an already created structure.
3675 2) For consistency with the majority of our object creators and
3676 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3677 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3678 for deprecated builds.
3682 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3683 cryptographic operations to be performed asynchronously as long as an
3684 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3685 further details. Libssl has also had this capability integrated with the
3686 introduction of the new mode SSL_MODE_ASYNC and associated error
3687 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3688 pages. This work was developed in partnership with Intel Corp.
3692 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3693 always enabled now. If you want to disable the support you should
3694 exclude it using the list of supported ciphers. This also means that the
3695 "-no_ecdhe" option has been removed from s_server.
3699 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3700 SSL_{CTX_}set1_curves() which can set a list.
3704 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3705 curve you want to support using SSL_{CTX_}set1_curves().
3709 * State machine rewrite. The state machine code has been significantly
3710 refactored in order to remove much duplication of code and solve issues
3711 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3712 further details). This change does have some associated API changes.
3713 Notably the SSL_state() function has been removed and replaced by
3714 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3715 SSL_set_state() has been removed altogether. The previous handshake states
3716 defined in ssl.h and ssl3.h have also been removed.
3720 * All instances of the string "ssleay" in the public API were replaced
3721 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3722 Some error codes related to internal RSA_eay API's were renamed.
3726 * The demo files in crypto/threads were moved to demo/threads.
3730 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3733 *Matt Caswell, Rich Salz*
3735 * New ASN.1 embed macro.
3737 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3738 structure is not allocated: it is part of the parent. That is instead of
3746 This reduces memory fragmentation and make it impossible to accidentally
3747 set a mandatory field to NULL.
3749 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3750 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3751 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3756 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3760 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3761 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3762 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3763 DES and RC4 ciphersuites.
3767 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3768 This changes the decoding behaviour for some invalid messages,
3769 though the change is mostly in the more lenient direction, and
3770 legacy behaviour is preserved as much as possible.
3774 * Fix no-stdio build.
3775 *David Woodhouse <David.Woodhouse@intel.com> and also*
3776 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3778 * New testing framework
3779 The testing framework has been largely rewritten and is now using
3780 perl and the perl modules Test::Harness and an extended variant of
3781 Test::More called OpenSSL::Test to do its work. All test scripts in
3782 test/ have been rewritten into test recipes, and all direct calls to
3783 executables in test/Makefile have become individual recipes using the
3784 simplified testing OpenSSL::Test::Simple.
3786 For documentation on our testing modules, do:
3788 perldoc test/testlib/OpenSSL/Test/Simple.pm
3789 perldoc test/testlib/OpenSSL/Test.pm
3793 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3794 are used; the latter aborts on memory leaks (usually checked on exit).
3795 Some undocumented "set malloc, etc., hooks" functions were removed
3796 and others were changed. All are now documented.
3800 * In DSA_generate_parameters_ex, if the provided seed is too short,
3803 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3805 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3806 from RFC4279, RFC4785, RFC5487, RFC5489.
3808 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3809 original RSA_PSK patch.
3813 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3814 era flag was never set throughout the codebase (only read). Also removed
3815 SSL3_FLAGS_POP_BUFFER which was only used if
3816 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3820 * Changed the default name options in the "ca", "crl", "req" and "x509"
3821 to be "oneline" instead of "compat".
3825 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3826 not aware of clients that still exhibit this bug, and the workaround
3827 hasn't been working properly for a while.
3831 * The return type of BIO_number_read() and BIO_number_written() as well as
3832 the corresponding num_read and num_write members in the BIO structure has
3833 changed from unsigned long to uint64_t. On platforms where an unsigned
3834 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3839 * Given the pervasive nature of TLS extensions it is inadvisable to run
3840 OpenSSL without support for them. It also means that maintaining
3841 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3842 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3846 * Removed support for the two export grade static DH ciphersuites
3847 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3848 were newly added (along with a number of other static DH ciphersuites) to
3849 1.0.2. However the two export ones have *never* worked since they were
3850 introduced. It seems strange in any case to be adding new export
3851 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3855 * Version negotiation has been rewritten. In particular SSLv23_method(),
3856 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3857 and turned into macros which simply call the new preferred function names
3858 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3859 should use the new names instead. Also as part of this change the ssl23.h
3860 header file has been removed.
3864 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3865 code and the associated standard is no longer considered fit-for-purpose.
3869 * RT2547 was closed. When generating a private key, try to make the
3870 output file readable only by the owner. This behavior change might
3871 be noticeable when interacting with other software.
3873 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3878 * Added HTTP GET support to the ocsp command.
3882 * Changed default digest for the dgst and enc commands from MD5 to
3887 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3891 * Added support for TLS extended master secret from
3892 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3893 initial patch which was a great help during development.
3897 * All libssl internal structures have been removed from the public header
3898 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3899 now redundant). Users should not attempt to access internal structures
3900 directly. Instead they should use the provided API functions.
3904 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3905 Access to deprecated functions can be re-enabled by running config with
3906 "enable-deprecated". In addition applications wishing to use deprecated
3907 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3908 will, by default, disable some transitive includes that previously existed
3909 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3913 * Added support for OCB mode. OpenSSL has been granted a patent license
3914 compatible with the OpenSSL license for use of OCB. Details are available
3915 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3916 for OCB can be removed by calling config with no-ocb.
3920 * SSLv2 support has been removed. It still supports receiving a SSLv2
3921 compatible client hello.
3925 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3926 done while fixing the error code for the key-too-small case.
3928 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3930 * CA.sh has been removed; use CA.pl instead.
3934 * Removed old DES API.
3938 * Remove various unsupported platforms:
3944 Sinix/ReliantUNIX RM400
3949 16-bit platforms such as WIN16
3953 * Clean up OPENSSL_NO_xxx #define's
3954 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3955 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3956 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3957 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3958 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3959 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3960 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3961 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3962 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3963 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3967 * Cleaned up dead code
3968 Remove all but one '#ifdef undef' which is to be looked at.
3972 * Clean up calling of xxx_free routines.
3973 Just like free(), fix most of the xxx_free routines to accept
3974 NULL. Remove the non-null checks from callers. Save much code.
3978 * Add secure heap for storage of private keys (when possible).
3979 Add BIO_s_secmem(), CBIGNUM, etc.
3980 Contributed by Akamai Technologies under our Corporate CLA.
3984 * Experimental support for a new, fast, unbiased prime candidate generator,
3985 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3987 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3989 * New output format NSS in the sess_id command line tool. This allows
3990 exporting the session id and the master key in NSS keylog format.
3992 *Martin Kaiser <martin@kaiser.cx>*
3994 * Harmonize version and its documentation. -f flag is used to display
3997 *mancha <mancha1@zoho.com>*
3999 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4000 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4002 *mancha <mancha1@zoho.com>*
4004 * Fix some double frees. These are not thought to be exploitable.
4006 *mancha <mancha1@zoho.com>*
4008 * A missing bounds check in the handling of the TLS heartbeat extension
4009 can be used to reveal up to 64k of memory to a connected client or
4012 Thanks for Neel Mehta of Google Security for discovering this bug and to
4013 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4014 preparing the fix ([CVE-2014-0160])
4016 *Adam Langley, Bodo Moeller*
4018 * Fix for the attack described in the paper "Recovering OpenSSL
4019 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4020 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4021 <http://eprint.iacr.org/2014/140>
4023 Thanks to Yuval Yarom and Naomi Benger for discovering this
4024 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4026 *Yuval Yarom and Naomi Benger*
4028 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4029 this fixes a limitation in previous versions of OpenSSL.
4033 * Experimental encrypt-then-mac support.
4035 Experimental support for encrypt then mac from
4036 draft-gutmann-tls-encrypt-then-mac-02.txt
4038 To enable it set the appropriate extension number (0x42 for the test
4039 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4041 For non-compliant peers (i.e. just about everything) this should have no
4044 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4048 * Add EVP support for key wrapping algorithms, to avoid problems with
4049 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4050 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4051 algorithms and include tests cases.
4055 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4060 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4061 MGF1 digest and OAEP label.
4065 * Make openssl verify return errors.
4067 *Chris Palmer <palmer@google.com> and Ben Laurie*
4069 * New function ASN1_TIME_diff to calculate the difference between two
4070 ASN1_TIME structures or one structure and the current time.
4074 * Update fips_test_suite to support multiple command line options. New
4075 test to induce all self test errors in sequence and check expected
4080 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4081 sign or verify all in one operation.
4085 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4086 test programs and fips_test_suite. Includes functionality to parse
4087 the minimal script output of fipsalgest.pl directly.
4091 * Add authorisation parameter to FIPS_module_mode_set().
4095 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4099 * Use separate DRBG fields for internal and external flags. New function
4100 FIPS_drbg_health_check() to perform on demand health checking. Add
4101 generation tests to fips_test_suite with reduced health check interval to
4102 demonstrate periodic health checking. Add "nodh" option to
4103 fips_test_suite to skip very slow DH test.
4107 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4112 * More extensive health check for DRBG checking many more failure modes.
4113 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4114 combination: call this in fips_test_suite.
4118 * Add support for canonical generation of DSA parameter 'g'. See
4121 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4122 POST to handle HMAC cases.
4126 * Add functions FIPS_module_version() and FIPS_module_version_text()
4127 to return numerical and string versions of the FIPS module number.
4131 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4132 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4133 outside the validated module in the FIPS capable OpenSSL.
4137 * Minor change to DRBG entropy callback semantics. In some cases
4138 there is no multiple of the block length between min_len and
4139 max_len. Allow the callback to return more than max_len bytes
4140 of entropy but discard any extra: it is the callback's responsibility
4141 to ensure that the extra data discarded does not impact the
4142 requested amount of entropy.
4146 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4147 information in FIPS186-3, SP800-57 and SP800-131A.
4151 * CCM support via EVP. Interface is very similar to GCM case except we
4152 must supply all data in one chunk (i.e. no update, final) and the
4153 message length must be supplied if AAD is used. Add algorithm test
4158 * Initial version of POST overhaul. Add POST callback to allow the status
4159 of POST to be monitored and/or failures induced. Modify fips_test_suite
4160 to use callback. Always run all selftests even if one fails.
4164 * XTS support including algorithm test driver in the fips_gcmtest program.
4165 Note: this does increase the maximum key length from 32 to 64 bytes but
4166 there should be no binary compatibility issues as existing applications
4167 will never use XTS mode.
4171 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4172 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4173 performs algorithm blocking for unapproved PRNG types. Also do not
4174 set PRNG type in FIPS_mode_set(): leave this to the application.
4175 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4176 the standard OpenSSL PRNG: set additional data to a date time vector.
4180 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4181 This shouldn't present any incompatibility problems because applications
4182 shouldn't be using these directly and any that are will need to rethink
4183 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4187 * Extensive self tests and health checking required by SP800-90 DRBG.
4188 Remove strength parameter from FIPS_drbg_instantiate and always
4189 instantiate at maximum supported strength.
4193 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4197 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4201 * New function DH_compute_key_padded() to compute a DH key and pad with
4202 leading zeroes if needed: this complies with SP800-56A et al.
4206 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4207 anything, incomplete, subject to change and largely untested at present.
4211 * Modify fipscanisteronly build option to only build the necessary object
4212 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4216 * Add experimental option FIPSSYMS to give all symbols in
4217 fipscanister.o and FIPS or fips prefix. This will avoid
4218 conflicts with future versions of OpenSSL. Add perl script
4219 util/fipsas.pl to preprocess assembly language source files
4220 and rename any affected symbols.
4224 * Add selftest checks and algorithm block of non-fips algorithms in
4225 FIPS mode. Remove DES2 from selftests.
4229 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4230 return internal method without any ENGINE dependencies. Add new
4231 tiny fips sign and verify functions.
4235 * New build option no-ec2m to disable characteristic 2 code.
4239 * New build option "fipscanisteronly". This only builds fipscanister.o
4240 and (currently) associated fips utilities. Uses the file Makefile.fips
4241 instead of Makefile.org as the prototype.
4245 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4246 Update fips_gcmtest to use IV generator.
4250 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4251 setting output buffer to NULL. The `*Final` function must be
4252 called although it will not retrieve any additional data. The tag
4253 can be set or retrieved with a ctrl. The IV length is by default 12
4254 bytes (96 bits) but can be set to an alternative value. If the IV
4255 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4260 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4261 underlying do_cipher function handles all cipher semantics itself
4262 including padding and finalisation. This is useful if (for example)
4263 an ENGINE cipher handles block padding itself. The behaviour of
4264 do_cipher is subtly changed if this flag is set: the return value
4265 is the number of characters written to the output buffer (zero is
4266 no longer an error code) or a negative error code. Also if the
4267 input buffer is NULL and length 0 finalisation should be performed.
4271 * If a candidate issuer certificate is already part of the constructed
4272 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4276 * Improve forward-security support: add functions
4278 void SSL_CTX_set_not_resumable_session_callback(
4279 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4280 void SSL_set_not_resumable_session_callback(
4281 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4283 for use by SSL/TLS servers; the callback function will be called whenever a
4284 new session is created, and gets to decide whether the session may be
4285 cached to make it resumable (return 0) or not (return 1). (As by the
4286 SSL/TLS protocol specifications, the session_id sent by the server will be
4287 empty to indicate that the session is not resumable; also, the server will
4288 not generate RFC 4507 (RFC 5077) session tickets.)
4290 A simple reasonable callback implementation is to return is_forward_secure.
4291 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4292 by the SSL/TLS server library, indicating whether it can provide forward
4295 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4297 * New -verify_name option in command line utilities to set verification
4302 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4303 Add CMAC pkey methods.
4307 * Experimental renegotiation in s_server -www mode. If the client
4308 browses /reneg connection is renegotiated. If /renegcert it is
4309 renegotiated requesting a certificate.
4313 * Add an "external" session cache for debugging purposes to s_server. This
4314 should help trace issues which normally are only apparent in deployed
4315 multi-process servers.
4319 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4320 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4321 BIO_set_cipher() and some obscure PEM functions were changed so they
4322 can now return an error. The RAND changes required a change to the
4323 RAND_METHOD structure.
4327 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4328 a gcc attribute to warn if the result of a function is ignored. This
4329 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4330 whose return value is often ignored.
4334 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4335 These allow SCTs (signed certificate timestamps) to be requested and
4336 validated when establishing a connection.
4338 *Rob Percival <robpercival@google.com>*
4343 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4345 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4346 used even when parsing explicit parameters, when loading a encoded key
4347 or calling `EC_GROUP_new_from_ecpkparameters()`/
4348 `EC_GROUP_new_from_ecparameters()`.
4349 This prevents bypass of security hardening and performance gains,
4350 especially for curves with specialized EC_METHODs.
4351 By default, if a key encoded with explicit parameters is loaded and later
4352 encoded, the output is still encoded with explicit parameters, even if
4353 internally a "named" EC_GROUP is used for computation.
4357 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4358 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4359 NULL. After this change, only the cofactor parameter can be NULL. It also
4360 does some minimal sanity checks on the passed order.
4365 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4366 An attack is simple, if the first CMS_recipientInfo is valid but the
4367 second CMS_recipientInfo is chosen ciphertext. If the second
4368 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4369 encryption key will be replaced by garbage, and the message cannot be
4370 decoded, but if the RSA decryption fails, the correct encryption key is
4371 used and the recipient will not notice the attack.
4372 As a work around for this potential attack the length of the decrypted
4373 key must be equal to the cipher default key length, in case the
4374 certifiate is not given and all recipientInfo are tried out.
4375 The old behaviour can be re-enabled in the CMS code by setting the
4376 CMS_DEBUG_DECRYPT flag.
4381 * Document issue with installation paths in diverse Windows builds
4383 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4384 binaries and run-time config file.
4389 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4391 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4392 This changes the size when using the `genpkey` command when no size is given.
4393 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4394 generation commands to use 2048 bits by default.
4398 * Add FIPS support for Android Arm 64-bit
4400 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4401 Module in Version 2.0.10. For some reason, the corresponding target
4402 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4403 built with FIPS support on Android Arm 64-bit. This omission has been
4406 *Matthias St. Pierre*
4408 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4410 * 0-byte record padding oracle
4412 If an application encounters a fatal protocol error and then calls
4413 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4414 then OpenSSL can respond differently to the calling application if a 0 byte
4415 record is received with invalid padding compared to if a 0 byte record is
4416 received with an invalid MAC. If the application then behaves differently
4417 based on that in a way that is detectable to the remote peer, then this
4418 amounts to a padding oracle that could be used to decrypt data.
4420 In order for this to be exploitable "non-stitched" ciphersuites must be in
4421 use. Stitched ciphersuites are optimised implementations of certain
4422 commonly used ciphersuites. Also the application must call SSL_shutdown()
4423 twice even if a protocol error has occurred (applications should not do
4424 this but some do anyway).
4426 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4427 Aviram, with additional investigation by Steven Collison and Andrew
4428 Hourselt. It was reported to OpenSSL on 10th December 2018.
4433 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4437 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4439 * Microarchitecture timing vulnerability in ECC scalar multiplication
4441 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4442 shown to be vulnerable to a microarchitecture timing side channel attack.
4443 An attacker with sufficient access to mount local timing attacks during
4444 ECDSA signature generation could recover the private key.
4446 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4447 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4453 * Timing vulnerability in DSA signature generation
4455 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4456 timing side channel attack. An attacker could use variations in the signing
4457 algorithm to recover the private key.
4459 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4464 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4465 Module, accidentally introduced while backporting security fixes from the
4466 development branch and hindering the use of ECC in FIPS mode.
4470 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4472 * Client DoS due to large DH parameter
4474 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4475 malicious server can send a very large prime value to the client. This will
4476 cause the client to spend an unreasonably long period of time generating a
4477 key for this prime resulting in a hang until the client has finished. This
4478 could be exploited in a Denial Of Service attack.
4480 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4485 * Cache timing vulnerability in RSA Key Generation
4487 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4488 a cache timing side channel attack. An attacker with sufficient access to
4489 mount cache timing attacks during the RSA key generation process could
4490 recover the private key.
4492 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4493 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4498 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4499 parameter is no longer accepted, as it leads to a corrupt table. NULL
4500 pem_str is reserved for alias entries only.
4504 * Revert blinding in ECDSA sign and instead make problematic addition
4505 length-invariant. Switch even to fixed-length Montgomery multiplication.
4509 * Change generating and checking of primes so that the error rate of not
4510 being prime depends on the intended use based on the size of the input.
4511 For larger primes this will result in more rounds of Miller-Rabin.
4512 The maximal error rate for primes with more than 1080 bits is lowered
4515 *Kurt Roeckx, Annie Yousar*
4517 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4521 * Add blinding to ECDSA and DSA signatures to protect against side channel
4522 attacks discovered by Keegan Ryan (NCC Group).
4526 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4527 now allow empty (zero character) pass phrases.
4531 * Certificate time validation (X509_cmp_time) enforces stricter
4532 compliance with RFC 5280. Fractional seconds and timezone offsets
4533 are no longer allowed.
4537 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4539 * Constructed ASN.1 types with a recursive definition could exceed the stack
4541 Constructed ASN.1 types with a recursive definition (such as can be found
4542 in PKCS7) could eventually exceed the stack given malicious input with
4543 excessive recursion. This could result in a Denial Of Service attack. There
4544 are no such structures used within SSL/TLS that come from untrusted sources
4545 so this is considered safe.
4547 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4553 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4555 * Read/write after SSL object in error state
4557 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4558 mechanism. The intent was that if a fatal error occurred during a handshake
4559 then OpenSSL would move into the error state and would immediately fail if
4560 you attempted to continue the handshake. This works as designed for the
4561 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4562 SSL_connect()), however due to a bug it does not work correctly if
4563 SSL_read() or SSL_write() is called directly. In that scenario, if the
4564 handshake fails then a fatal error will be returned in the initial function
4565 call. If SSL_read()/SSL_write() is subsequently called by the application
4566 for the same SSL object then it will succeed and the data is passed without
4567 being decrypted/encrypted directly from the SSL/TLS record layer.
4569 In order to exploit this issue an application bug would have to be present
4570 that resulted in a call to SSL_read()/SSL_write() being issued after having
4571 already received a fatal error.
4573 This issue was reported to OpenSSL by David Benjamin (Google).
4578 * rsaz_1024_mul_avx2 overflow bug on x86_64
4580 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4581 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4582 Analysis suggests that attacks against RSA and DSA as a result of this
4583 defect would be very difficult to perform and are not believed likely.
4584 Attacks against DH1024 are considered just feasible, because most of the
4585 work necessary to deduce information about a private key may be performed
4586 offline. The amount of resources required for such an attack would be
4587 significant. However, for an attack on TLS to be meaningful, the server
4588 would have to share the DH1024 private key among multiple clients, which is
4589 no longer an option since CVE-2016-0701.
4591 This only affects processors that support the AVX2 but not ADX extensions
4592 like Intel Haswell (4th generation).
4594 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4595 was originally found via the OSS-Fuzz project.
4600 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4602 * bn_sqrx8x_internal carry bug on x86_64
4604 There is a carry propagating bug in the x86_64 Montgomery squaring
4605 procedure. No EC algorithms are affected. Analysis suggests that attacks
4606 against RSA and DSA as a result of this defect would be very difficult to
4607 perform and are not believed likely. Attacks against DH are considered just
4608 feasible (although very difficult) because most of the work necessary to
4609 deduce information about a private key may be performed offline. The amount
4610 of resources required for such an attack would be very significant and
4611 likely only accessible to a limited number of attackers. An attacker would
4612 additionally need online access to an unpatched system using the target
4613 private key in a scenario with persistent DH parameters and a private
4614 key that is shared between multiple clients.
4616 This only affects processors that support the BMI1, BMI2 and ADX extensions
4617 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4619 This issue was reported to OpenSSL by the OSS-Fuzz project.
4624 * Malformed X.509 IPAddressFamily could cause OOB read
4626 If an X.509 certificate has a malformed IPAddressFamily extension,
4627 OpenSSL could do a one-byte buffer overread. The most likely result
4628 would be an erroneous display of the certificate in text format.
4630 This issue was reported to OpenSSL by the OSS-Fuzz project.
4634 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4636 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4637 platform rather than 'mingw'.
4641 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4643 * Truncated packet could crash via OOB read
4645 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4646 cipher is being used, then a truncated packet can cause that host to
4647 perform an out-of-bounds read, usually resulting in a crash.
4649 This issue was reported to OpenSSL by Robert Święcki of Google.
4654 * BN_mod_exp may produce incorrect results on x86_64
4656 There is a carry propagating bug in the x86_64 Montgomery squaring
4657 procedure. No EC algorithms are affected. Analysis suggests that attacks
4658 against RSA and DSA as a result of this defect would be very difficult to
4659 perform and are not believed likely. Attacks against DH are considered just
4660 feasible (although very difficult) because most of the work necessary to
4661 deduce information about a private key may be performed offline. The amount
4662 of resources required for such an attack would be very significant and
4663 likely only accessible to a limited number of attackers. An attacker would
4664 additionally need online access to an unpatched system using the target
4665 private key in a scenario with persistent DH parameters and a private
4666 key that is shared between multiple clients. For example this can occur by
4667 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4668 similar to CVE-2015-3193 but must be treated as a separate problem.
4670 This issue was reported to OpenSSL by the OSS-Fuzz project.
4675 * Montgomery multiplication may produce incorrect results
4677 There is a carry propagating bug in the Broadwell-specific Montgomery
4678 multiplication procedure that handles input lengths divisible by, but
4679 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4680 and DH private keys are impossible. This is because the subroutine in
4681 question is not used in operations with the private key itself and an input
4682 of the attacker's direct choice. Otherwise the bug can manifest itself as
4683 transient authentication and key negotiation failures or reproducible
4684 erroneous outcome of public-key operations with specially crafted input.
4685 Among EC algorithms only Brainpool P-512 curves are affected and one
4686 presumably can attack ECDH key negotiation. Impact was not analyzed in
4687 detail, because pre-requisites for attack are considered unlikely. Namely
4688 multiple clients have to choose the curve in question and the server has to
4689 share the private key among them, neither of which is default behaviour.
4690 Even then only clients that chose the curve will be affected.
4692 This issue was publicly reported as transient failures and was not
4693 initially recognized as a security issue. Thanks to Richard Morgan for
4694 providing reproducible case.
4699 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4700 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4701 prevent issues where no progress is being made and the peer continually
4702 sends unrecognised record types, using up resources processing them.
4706 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4708 * Missing CRL sanity check
4710 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4711 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4712 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4714 This issue only affects the OpenSSL 1.0.2i
4719 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4721 * OCSP Status Request extension unbounded memory growth
4723 A malicious client can send an excessively large OCSP Status Request
4724 extension. If that client continually requests renegotiation, sending a
4725 large OCSP Status Request extension each time, then there will be unbounded
4726 memory growth on the server. This will eventually lead to a Denial Of
4727 Service attack through memory exhaustion. Servers with a default
4728 configuration are vulnerable even if they do not support OCSP. Builds using
4729 the "no-ocsp" build time option are not affected.
4731 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4736 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4739 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4745 * OOB write in MDC2_Update()
4747 An overflow can occur in MDC2_Update() either if called directly or
4748 through the EVP_DigestUpdate() function using MDC2. If an attacker
4749 is able to supply very large amounts of input data after a previous
4750 call to EVP_EncryptUpdate() with a partial block then a length check
4751 can overflow resulting in a heap corruption.
4753 The amount of data needed is comparable to SIZE_MAX which is impractical
4756 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4761 * Malformed SHA512 ticket DoS
4763 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4764 DoS attack where a malformed ticket will result in an OOB read which will
4767 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4768 a custom server callback and ticket lookup mechanism.
4770 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4775 * OOB write in BN_bn2dec()
4777 The function BN_bn2dec() does not check the return value of BN_div_word().
4778 This can cause an OOB write if an application uses this function with an
4779 overly large BIGNUM. This could be a problem if an overly large certificate
4780 or CRL is printed out from an untrusted source. TLS is not affected because
4781 record limits will reject an oversized certificate before it is parsed.
4783 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4788 * OOB read in TS_OBJ_print_bio()
4790 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4791 the total length the OID text representation would use and not the amount
4792 of data written. This will result in OOB reads when large OIDs are
4795 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4800 * Pointer arithmetic undefined behaviour
4802 Avoid some undefined pointer arithmetic
4804 A common idiom in the codebase is to check limits in the following manner:
4807 Where "p" points to some malloc'd data of SIZE bytes and
4810 "len" here could be from some externally supplied data (e.g. from a TLS
4813 The rules of C pointer arithmetic are such that "p + len" is only well
4814 defined where len <= SIZE. Therefore the above idiom is actually
4815 undefined behaviour.
4817 For example this could cause problems if some malloc implementation
4818 provides an address for "p" such that "p + len" actually overflows for
4819 values of len that are too big and therefore p + len < limit.
4821 This issue was reported to OpenSSL by Guido Vranken
4826 * Constant time flag not preserved in DSA signing
4828 Operations in the DSA signing algorithm should run in constant time in
4829 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4830 implementation means that a non-constant time codepath is followed for
4831 certain operations. This has been demonstrated through a cache-timing
4832 attack to be sufficient for an attacker to recover the private DSA key.
4834 This issue was reported by César Pereida (Aalto University), Billy Brumley
4835 (Tampere University of Technology), and Yuval Yarom (The University of
4836 Adelaide and NICTA).
4841 * DTLS buffered message DoS
4843 In a DTLS connection where handshake messages are delivered out-of-order
4844 those messages that OpenSSL is not yet ready to process will be buffered
4845 for later use. Under certain circumstances, a flaw in the logic means that
4846 those messages do not get removed from the buffer even though the handshake
4847 has been completed. An attacker could force up to approx. 15 messages to
4848 remain in the buffer when they are no longer required. These messages will
4849 be cleared when the DTLS connection is closed. The default maximum size for
4850 a message is 100k. Therefore the attacker could force an additional 1500k
4851 to be consumed per connection. By opening many simulataneous connections an
4852 attacker could cause a DoS attack through memory exhaustion.
4854 This issue was reported to OpenSSL by Quan Luo.
4859 * DTLS replay protection DoS
4861 A flaw in the DTLS replay attack protection mechanism means that records
4862 that arrive for future epochs update the replay protection "window" before
4863 the MAC for the record has been validated. This could be exploited by an
4864 attacker by sending a record for the next epoch (which does not have to
4865 decrypt or have a valid MAC), with a very large sequence number. This means
4866 that all subsequent legitimate packets are dropped causing a denial of
4867 service for a specific DTLS connection.
4869 This issue was reported to OpenSSL by the OCAP audit team.
4874 * Certificate message OOB reads
4876 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4877 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4878 theoretical DoS risk but this has not been observed in practice on common
4881 The messages affected are client certificate, client certificate request
4882 and server certificate. As a result the attack can only be performed
4883 against a client or a server which enables client authentication.
4885 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4890 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4892 * Prevent padding oracle in AES-NI CBC MAC check
4894 A MITM attacker can use a padding oracle attack to decrypt traffic
4895 when the connection uses an AES CBC cipher and the server support
4898 This issue was introduced as part of the fix for Lucky 13 padding
4899 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4900 constant time by making sure that always the same bytes are read and
4901 compared against either the MAC or padding bytes. But it no longer
4902 checked that there was enough data to have both the MAC and padding
4905 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4909 * Fix EVP_EncodeUpdate overflow
4911 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4912 Base64 encoding of binary data. If an attacker is able to supply very large
4913 amounts of input data then a length check can overflow resulting in a heap
4916 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4917 the `PEM_write_bio*` family of functions. These are mainly used within the
4918 OpenSSL command line applications, so any application which processes data
4919 from an untrusted source and outputs it as a PEM file should be considered
4920 vulnerable to this issue. User applications that call these APIs directly
4921 with large amounts of untrusted data may also be vulnerable.
4923 This issue was reported by Guido Vranken.
4928 * Fix EVP_EncryptUpdate overflow
4930 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4931 is able to supply very large amounts of input data after a previous call to
4932 EVP_EncryptUpdate() with a partial block then a length check can overflow
4933 resulting in a heap corruption. Following an analysis of all OpenSSL
4934 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4935 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4936 the first called function after an EVP_EncryptInit(), and therefore that
4937 specific call must be safe. The second form is where the length passed to
4938 EVP_EncryptUpdate() can be seen from the code to be some small value and
4939 therefore there is no possibility of an overflow. Since all instances are
4940 one of these two forms, it is believed that there can be no overflows in
4941 internal code due to this problem. It should be noted that
4942 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4943 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4944 of these calls have also been analysed too and it is believed there are no
4945 instances in internal usage where an overflow could occur.
4947 This issue was reported by Guido Vranken.
4952 * Prevent ASN.1 BIO excessive memory allocation
4954 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4955 a short invalid encoding can cause allocation of large amounts of memory
4956 potentially consuming excessive resources or exhausting memory.
4958 Any application parsing untrusted data through d2i BIO functions is
4959 affected. The memory based functions such as d2i_X509() are *not* affected.
4960 Since the memory based functions are used by the TLS library, TLS
4961 applications are not affected.
4963 This issue was reported by Brian Carpenter.
4970 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4971 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4972 in arbitrary stack data being returned in the buffer.
4974 This issue was reported by Guido Vranken.
4979 * Modify behavior of ALPN to invoke callback after SNI/servername
4980 callback, such that updates to the SSL_CTX affect ALPN.
4984 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4989 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4990 methods are enabled and ssl2 is disabled the methods return NULL.
4994 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4996 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4997 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4998 provide any "EXPORT" or "LOW" strength ciphers.
5002 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5003 is by default disabled at build-time. Builds that are not configured with
5004 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5005 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5006 will need to explicitly call either of:
5008 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5010 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5012 as appropriate. Even if either of those is used, or the application
5013 explicitly uses the version-specific SSLv2_method() or its client and
5014 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5015 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5016 ciphers, and SSLv2 56-bit DES are no longer available.
5021 * Fix a double-free in DSA code
5023 A double free bug was discovered when OpenSSL parses malformed DSA private
5024 keys and could lead to a DoS attack or memory corruption for applications
5025 that receive DSA private keys from untrusted sources. This scenario is
5028 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5034 * Disable SRP fake user seed to address a server memory leak.
5036 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5038 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5039 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5040 was changed to ignore the "fake user" SRP seed, even if the seed
5043 Users should use SRP_VBASE_get1_by_user instead. Note that in
5044 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5045 also that even though configuring the SRP seed attempts to hide
5046 invalid usernames by continuing the handshake with fake
5047 credentials, this behaviour is not constant time and no strong
5048 guarantees are made that the handshake is indistinguishable from
5049 that of a valid user.
5054 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5056 In the BN_hex2bn function the number of hex digits is calculated using an
5057 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5058 large values of `i` this can result in `bn_expand` not allocating any
5059 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5060 field as NULL leading to a subsequent NULL ptr deref. For very large values
5061 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5062 In this case memory is allocated to the internal BIGNUM data field, but it
5063 is insufficiently sized leading to heap corruption. A similar issue exists
5064 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5065 is ever called by user applications with very large untrusted hex/dec data.
5066 This is anticipated to be a rare occurrence.
5068 All OpenSSL internal usage of these functions use data that is not expected
5069 to be untrusted, e.g. config file data or application command line
5070 arguments. If user developed applications generate config file data based
5071 on untrusted data then it is possible that this could also lead to security
5072 consequences. This is also anticipated to be rare.
5074 This issue was reported to OpenSSL by Guido Vranken.
5079 * Fix memory issues in `BIO_*printf` functions
5081 The internal `fmtstr` function used in processing a "%s" format string in
5082 the `BIO_*printf` functions could overflow while calculating the length of a
5083 string and cause an OOB read when printing very long strings.
5085 Additionally the internal `doapr_outch` function can attempt to write to an
5086 OOB memory location (at an offset from the NULL pointer) in the event of a
5087 memory allocation failure. In 1.0.2 and below this could be caused where
5088 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5089 could be in processing a very long "%s" format string. Memory leaks can
5092 The first issue may mask the second issue dependent on compiler behaviour.
5093 These problems could enable attacks where large amounts of untrusted data
5094 is passed to the `BIO_*printf` functions. If applications use these functions
5095 in this way then they could be vulnerable. OpenSSL itself uses these
5096 functions when printing out human-readable dumps of ASN.1 data. Therefore
5097 applications that print this data could be vulnerable if the data is from
5098 untrusted sources. OpenSSL command line applications could also be
5099 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5100 as command line arguments.
5102 Libssl is not considered directly vulnerable. Additionally certificates etc
5103 received via remote connections via libssl are also unlikely to be able to
5104 trigger these issues because of message size limits enforced within libssl.
5106 This issue was reported to OpenSSL Guido Vranken.
5111 * Side channel attack on modular exponentiation
5113 A side-channel attack was found which makes use of cache-bank conflicts on
5114 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5115 of RSA keys. The ability to exploit this issue is limited as it relies on
5116 an attacker who has control of code in a thread running on the same
5117 hyper-threaded core as the victim thread which is performing decryptions.
5119 This issue was reported to OpenSSL by Yuval Yarom, The University of
5120 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5121 Nadia Heninger, University of Pennsylvania with more information at
5122 <http://cachebleed.info>.
5127 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5128 if no keysize is specified with default_bits. This fixes an
5129 omission in an earlier change that changed all RSA/DSA key generation
5130 commands to use 2048 bits by default.
5134 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5136 * DH small subgroups
5138 Historically OpenSSL only ever generated DH parameters based on "safe"
5139 primes. More recently (in version 1.0.2) support was provided for
5140 generating X9.42 style parameter files such as those required for RFC 5114
5141 support. The primes used in such files may not be "safe". Where an
5142 application is using DH configured with parameters based on primes that are
5143 not "safe" then an attacker could use this fact to find a peer's private
5144 DH exponent. This attack requires that the attacker complete multiple
5145 handshakes in which the peer uses the same private DH exponent. For example
5146 this could be used to discover a TLS server's private DH exponent if it's
5147 reusing the private DH exponent or it's using a static DH ciphersuite.
5149 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5150 TLS. It is not on by default. If the option is not set then the server
5151 reuses the same private DH exponent for the life of the server process and
5152 would be vulnerable to this attack. It is believed that many popular
5153 applications do set this option and would therefore not be at risk.
5155 The fix for this issue adds an additional check where a "q" parameter is
5156 available (as is the case in X9.42 based parameters). This detects the
5157 only known attack, and is the only possible defense for static DH
5158 ciphersuites. This could have some performance impact.
5160 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5161 default and cannot be disabled. This could have some performance impact.
5163 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5168 * SSLv2 doesn't block disabled ciphers
5170 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5171 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5172 been disabled, provided that the SSLv2 protocol was not also disabled via
5175 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5176 and Sebastian Schinzel.
5181 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5183 * BN_mod_exp may produce incorrect results on x86_64
5185 There is a carry propagating bug in the x86_64 Montgomery squaring
5186 procedure. No EC algorithms are affected. Analysis suggests that attacks
5187 against RSA and DSA as a result of this defect would be very difficult to
5188 perform and are not believed likely. Attacks against DH are considered just
5189 feasible (although very difficult) because most of the work necessary to
5190 deduce information about a private key may be performed offline. The amount
5191 of resources required for such an attack would be very significant and
5192 likely only accessible to a limited number of attackers. An attacker would
5193 additionally need online access to an unpatched system using the target
5194 private key in a scenario with persistent DH parameters and a private
5195 key that is shared between multiple clients. For example this can occur by
5196 default in OpenSSL DHE based SSL/TLS ciphersuites.
5198 This issue was reported to OpenSSL by Hanno Böck.
5203 * Certificate verify crash with missing PSS parameter
5205 The signature verification routines will crash with a NULL pointer
5206 dereference if presented with an ASN.1 signature using the RSA PSS
5207 algorithm and absent mask generation function parameter. Since these
5208 routines are used to verify certificate signature algorithms this can be
5209 used to crash any certificate verification operation and exploited in a
5210 DoS attack. Any application which performs certificate verification is
5211 vulnerable including OpenSSL clients and servers which enable client
5214 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5219 * X509_ATTRIBUTE memory leak
5221 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5222 memory. This structure is used by the PKCS#7 and CMS routines so any
5223 application which reads PKCS#7 or CMS data from untrusted sources is
5224 affected. SSL/TLS is not affected.
5226 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5232 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5233 This changes the decoding behaviour for some invalid messages,
5234 though the change is mostly in the more lenient direction, and
5235 legacy behaviour is preserved as much as possible.
5239 * In DSA_generate_parameters_ex, if the provided seed is too short,
5242 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5244 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5246 * Alternate chains certificate forgery
5248 During certificate verification, OpenSSL will attempt to find an
5249 alternative certificate chain if the first attempt to build such a chain
5250 fails. An error in the implementation of this logic can mean that an
5251 attacker could cause certain checks on untrusted certificates to be
5252 bypassed, such as the CA flag, enabling them to use a valid leaf
5253 certificate to act as a CA and "issue" an invalid certificate.
5255 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5260 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5262 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5263 incompatibility in the handling of HMAC. The previous ABI has now been
5268 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5270 * Malformed ECParameters causes infinite loop
5272 When processing an ECParameters structure OpenSSL enters an infinite loop
5273 if the curve specified is over a specially malformed binary polynomial
5276 This can be used to perform denial of service against any
5277 system which processes public keys, certificate requests or
5278 certificates. This includes TLS clients and TLS servers with
5279 client authentication enabled.
5281 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5286 * Exploitable out-of-bounds read in X509_cmp_time
5288 X509_cmp_time does not properly check the length of the ASN1_TIME
5289 string and can read a few bytes out of bounds. In addition,
5290 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5293 An attacker can use this to craft malformed certificates and CRLs of
5294 various sizes and potentially cause a segmentation fault, resulting in
5295 a DoS on applications that verify certificates or CRLs. TLS clients
5296 that verify CRLs are affected. TLS clients and servers with client
5297 authentication enabled may be affected if they use custom verification
5300 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5301 independently by Hanno Böck.
5306 * PKCS7 crash with missing EnvelopedContent
5308 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5309 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5310 with missing content and trigger a NULL pointer dereference on parsing.
5312 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5313 structures from untrusted sources are affected. OpenSSL clients and
5314 servers are not affected.
5316 This issue was reported to OpenSSL by Michal Zalewski (Google).
5321 * CMS verify infinite loop with unknown hash function
5323 When verifying a signedData message the CMS code can enter an infinite loop
5324 if presented with an unknown hash function OID. This can be used to perform
5325 denial of service against any system which verifies signedData messages using
5327 This issue was reported to OpenSSL by Johannes Bauer.
5332 * Race condition handling NewSessionTicket
5334 If a NewSessionTicket is received by a multi-threaded client when attempting to
5335 reuse a previous ticket then a race condition can occur potentially leading to
5336 a double free of the ticket data.
5341 * Only support 256-bit or stronger elliptic curves with the
5342 'ecdh_auto' setting (server) or by default (client). Of supported
5343 curves, prefer P-256 (both).
5347 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5349 * ClientHello sigalgs DoS fix
5351 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5352 invalid signature algorithms extension a NULL pointer dereference will
5353 occur. This can be exploited in a DoS attack against the server.
5355 This issue was was reported to OpenSSL by David Ramos of Stanford
5359 *Stephen Henson and Matt Caswell*
5361 * Multiblock corrupted pointer fix
5363 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5364 feature only applies on 64 bit x86 architecture platforms that support AES
5365 NI instructions. A defect in the implementation of "multiblock" can cause
5366 OpenSSL's internal write buffer to become incorrectly set to NULL when
5367 using non-blocking IO. Typically, when the user application is using a
5368 socket BIO for writing, this will only result in a failed connection.
5369 However if some other BIO is used then it is likely that a segmentation
5370 fault will be triggered, thus enabling a potential DoS attack.
5372 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5377 * Segmentation fault in DTLSv1_listen fix
5379 The DTLSv1_listen function is intended to be stateless and processes the
5380 initial ClientHello from many peers. It is common for user code to loop
5381 over the call to DTLSv1_listen until a valid ClientHello is received with
5382 an associated cookie. A defect in the implementation of DTLSv1_listen means
5383 that state is preserved in the SSL object from one invocation to the next
5384 that can lead to a segmentation fault. Errors processing the initial
5385 ClientHello can trigger this scenario. An example of such an error could be
5386 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5389 This issue was reported to OpenSSL by Per Allansson.
5394 * Segmentation fault in ASN1_TYPE_cmp fix
5396 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5397 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5398 certificate signature algorithm consistency this can be used to crash any
5399 certificate verification operation and exploited in a DoS attack. Any
5400 application which performs certificate verification is vulnerable including
5401 OpenSSL clients and servers which enable client authentication.
5406 * Segmentation fault for invalid PSS parameters fix
5408 The signature verification routines will crash with a NULL pointer
5409 dereference if presented with an ASN.1 signature using the RSA PSS
5410 algorithm and invalid parameters. Since these routines are used to verify
5411 certificate signature algorithms this can be used to crash any
5412 certificate verification operation and exploited in a DoS attack. Any
5413 application which performs certificate verification is vulnerable including
5414 OpenSSL clients and servers which enable client authentication.
5416 This issue was was reported to OpenSSL by Brian Carpenter.
5421 * ASN.1 structure reuse memory corruption fix
5423 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5424 memory corruption via an invalid write. Such reuse is and has been
5425 strongly discouraged and is believed to be rare.
5427 Applications that parse structures containing CHOICE or ANY DEFINED BY
5428 components may be affected. Certificate parsing (d2i_X509 and related
5429 functions) are however not affected. OpenSSL clients and servers are
5435 * PKCS7 NULL pointer dereferences fix
5437 The PKCS#7 parsing code does not handle missing outer ContentInfo
5438 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5439 missing content and trigger a NULL pointer dereference on parsing.
5441 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5442 otherwise parse PKCS#7 structures from untrusted sources are
5443 affected. OpenSSL clients and servers are not affected.
5445 This issue was reported to OpenSSL by Michal Zalewski (Google).
5450 * DoS via reachable assert in SSLv2 servers fix
5452 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5453 servers that both support SSLv2 and enable export cipher suites by sending
5454 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5456 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5457 (OpenSSL development team).
5462 * Empty CKE with client auth and DHE fix
5464 If client auth is used then a server can seg fault in the event of a DHE
5465 ciphersuite being selected and a zero length ClientKeyExchange message
5466 being sent by the client. This could be exploited in a DoS attack.
5471 * Handshake with unseeded PRNG fix
5473 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5474 with an unseeded PRNG. The conditions are:
5475 - The client is on a platform where the PRNG has not been seeded
5476 automatically, and the user has not seeded manually
5477 - A protocol specific client method version has been used (i.e. not
5478 SSL_client_methodv23)
5479 - A ciphersuite is used that does not require additional random data from
5480 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5482 If the handshake succeeds then the client random that has been used will
5483 have been generated from a PRNG with insufficient entropy and therefore the
5484 output may be predictable.
5486 For example using the following command with an unseeded openssl will
5487 succeed on an unpatched platform:
5489 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5494 * Use After Free following d2i_ECPrivatekey error fix
5496 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5497 could cause a use after free condition. This, in turn, could cause a double
5498 free in several private key parsing functions (such as d2i_PrivateKey
5499 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5500 for applications that receive EC private keys from untrusted
5501 sources. This scenario is considered rare.
5503 This issue was discovered by the BoringSSL project and fixed in their
5509 * X509_to_X509_REQ NULL pointer deref fix
5511 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5512 the certificate key is invalid. This function is rarely used in practice.
5514 This issue was discovered by Brian Carpenter.
5519 * Removed the export ciphers from the DEFAULT ciphers
5523 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5525 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5526 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5527 So far those who have to target multiple platforms would compromise
5528 and argue that binary targeting say ARMv5 would still execute on
5529 ARMv8. "Universal" build resolves this compromise by providing
5530 near-optimal performance even on newer platforms.
5534 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5535 (other platforms pending).
5537 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5539 * Add support for the SignedCertificateTimestampList certificate and
5540 OCSP response extensions from RFC6962.
5544 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5545 for corner cases. (Certain input points at infinity could lead to
5546 bogus results, with non-infinity inputs mapped to infinity too.)
5550 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5551 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5552 common cases are optimized and there still is room for further
5553 improvements. Vector Permutation AES for Altivec is also added.
5557 * Add support for little-endian ppc64 Linux target.
5559 *Marcelo Cerri (IBM)*
5561 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5562 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5563 are optimized and there still is room for further improvements.
5564 Both 32- and 64-bit modes are supported.
5566 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5568 * Improved ARMv7 NEON support.
5572 * Support for SPARC Architecture 2011 crypto extensions, first
5573 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5574 SHA256/512, MD5, GHASH and modular exponentiation.
5576 *Andy Polyakov, David Miller*
5578 * Accelerated modular exponentiation for Intel processors, a.k.a.
5581 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5583 * Support for new and upcoming Intel processors, including AVX2,
5584 BMI and SHA ISA extensions. This includes additional "stitched"
5585 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5588 This work was sponsored by Intel Corp.
5592 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5593 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5594 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5598 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5599 this fixes a limitation in previous versions of OpenSSL.
5603 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5604 MGF1 digest and OAEP label.
5608 * Add EVP support for key wrapping algorithms, to avoid problems with
5609 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5610 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5611 algorithms and include tests cases.
5615 * Add functions to allocate and set the fields of an ECDSA_METHOD
5618 *Douglas E. Engert, Steve Henson*
5620 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5621 difference in days and seconds between two tm or ASN1_TIME structures.
5625 * Add -rev test option to s_server to just reverse order of characters
5626 received by client and send back to server. Also prints an abbreviated
5627 summary of the connection parameters.
5631 * New option -brief for s_client and s_server to print out a brief summary
5632 of connection parameters.
5636 * Add callbacks for arbitrary TLS extensions.
5638 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5640 * New option -crl_download in several openssl utilities to download CRLs
5641 from CRLDP extension in certificates.
5645 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5649 * New function X509_CRL_diff to generate a delta CRL from the difference
5650 of two full CRLs. Add support to "crl" utility.
5654 * New functions to set lookup_crls function and to retrieve
5655 X509_STORE from X509_STORE_CTX.
5659 * Print out deprecated issuer and subject unique ID fields in
5664 * Extend OCSP I/O functions so they can be used for simple general purpose
5665 HTTP as well as OCSP. New wrapper function which can be used to download
5666 CRLs using the OCSP API.
5670 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5674 * `SSL_CONF*` functions. These provide a common framework for application
5675 configuration using configuration files or command lines.
5679 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5680 message callback and prints the results. Needs compile time option
5681 "enable-ssl-trace". New options to s_client and s_server to enable
5686 * New ctrl and macro to retrieve supported points extensions.
5687 Print out extension in s_server and s_client.
5691 * New functions to retrieve certificate signature and signature
5696 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5701 * New Suite B modes for TLS code. These use and enforce the requirements
5702 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5703 only use Suite B curves. The Suite B modes can be set by using the
5704 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5708 * New chain verification flags for Suite B levels of security. Check
5709 algorithms are acceptable when flags are set in X509_verify_cert.
5713 * Make tls1_check_chain return a set of flags indicating checks passed
5714 by a certificate chain. Add additional tests to handle client
5715 certificates: checks for matching certificate type and issuer name
5720 * If an attempt is made to use a signature algorithm not in the peer
5721 preference list abort the handshake. If client has no suitable
5722 signature algorithms in response to a certificate request do not
5723 use the certificate.
5727 * If server EC tmp key is not in client preference list abort handshake.
5731 * Add support for certificate stores in CERT structure. This makes it
5732 possible to have different stores per SSL structure or one store in
5733 the parent SSL_CTX. Include distinct stores for certificate chain
5734 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5735 to build and store a certificate chain in CERT structure: returning
5736 an error if the chain cannot be built: this will allow applications
5737 to test if a chain is correctly configured.
5739 Note: if the CERT based stores are not set then the parent SSL_CTX
5740 store is used to retain compatibility with existing behaviour.
5744 * New function ssl_set_client_disabled to set a ciphersuite disabled
5745 mask based on the current session, check mask when sending client
5746 hello and checking the requested ciphersuite.
5750 * New ctrls to retrieve and set certificate types in a certificate
5751 request message. Print out received values in s_client. If certificate
5752 types is not set with custom values set sensible values based on
5753 supported signature algorithms.
5757 * Support for distinct client and server supported signature algorithms.
5761 * Add certificate callback. If set this is called whenever a certificate
5762 is required by client or server. An application can decide which
5763 certificate chain to present based on arbitrary criteria: for example
5764 supported signature algorithms. Add very simple example to s_server.
5765 This fixes many of the problems and restrictions of the existing client
5766 certificate callback: for example you can now clear an existing
5767 certificate and specify the whole chain.
5771 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5772 the certificate can be used for (if anything). Set valid_flags field
5773 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5774 to have similar checks in it.
5776 Add new "cert_flags" field to CERT structure and include a "strict mode".
5777 This enforces some TLS certificate requirements (such as only permitting
5778 certificate signature algorithms contained in the supported algorithms
5779 extension) which some implementations ignore: this option should be used
5780 with caution as it could cause interoperability issues.
5784 * Update and tidy signature algorithm extension processing. Work out
5785 shared signature algorithms based on preferences and peer algorithms
5786 and print them out in s_client and s_server. Abort handshake if no
5787 shared signature algorithms.
5791 * Add new functions to allow customised supported signature algorithms
5792 for SSL and SSL_CTX structures. Add options to s_client and s_server
5797 * New function SSL_certs_clear() to delete all references to certificates
5798 from an SSL structure. Before this once a certificate had been added
5799 it couldn't be removed.
5803 * Integrate hostname, email address and IP address checking with certificate
5804 verification. New verify options supporting checking in openssl utility.
5808 * Fixes and wildcard matching support to hostname and email checking
5809 functions. Add manual page.
5811 *Florian Weimer (Red Hat Product Security Team)*
5813 * New functions to check a hostname email or IP address against a
5814 certificate. Add options x509 utility to print results of checks against
5819 * Fix OCSP checking.
5821 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5823 * Initial experimental support for explicitly trusted non-root CAs.
5824 OpenSSL still tries to build a complete chain to a root but if an
5825 intermediate CA has a trust setting included that is used. The first
5826 setting is used: whether to trust (e.g., -addtrust option to the x509
5831 * Add -trusted_first option which attempts to find certificates in the
5832 trusted store even if an untrusted chain is also supplied.
5836 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5837 platform support for Linux and Android.
5841 * Support for linux-x32, ILP32 environment in x86_64 framework.
5845 * Experimental multi-implementation support for FIPS capable OpenSSL.
5846 When in FIPS mode the approved implementations are used as normal,
5847 when not in FIPS mode the internal unapproved versions are used instead.
5848 This means that the FIPS capable OpenSSL isn't forced to use the
5849 (often lower performance) FIPS implementations outside FIPS mode.
5853 * Transparently support X9.42 DH parameters when calling
5854 PEM_read_bio_DHparameters. This means existing applications can handle
5855 the new parameter format automatically.
5859 * Initial experimental support for X9.42 DH parameter format: mainly
5860 to support use of 'q' parameter for RFC5114 parameters.
5864 * Add DH parameters from RFC5114 including test data to dhtest.
5868 * Support for automatic EC temporary key parameter selection. If enabled
5869 the most preferred EC parameters are automatically used instead of
5870 hardcoded fixed parameters. Now a server just has to call:
5871 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5872 support ECDH and use the most appropriate parameters.
5876 * Enhance and tidy EC curve and point format TLS extension code. Use
5877 static structures instead of allocation if default values are used.
5878 New ctrls to set curves we wish to support and to retrieve shared curves.
5879 Print out shared curves in s_server. New options to s_server and s_client
5880 to set list of supported curves.
5884 * New ctrls to retrieve supported signature algorithms and
5885 supported curve values as an array of NIDs. Extend openssl utility
5886 to print out received values.
5890 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5891 between NIDs and the more common NIST names such as "P-256". Enhance
5892 ecparam utility and ECC method to recognise the NIST names for curves.
5896 * Enhance SSL/TLS certificate chain handling to support different
5897 chains for each certificate instead of one chain in the parent SSL_CTX.
5901 * Support for fixed DH ciphersuite client authentication: where both
5902 server and client use DH certificates with common parameters.
5906 * Support for fixed DH ciphersuites: those requiring DH server
5911 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5913 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5914 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5915 X509_CINF_get_signature were reverted post internal team review.
5920 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5922 * OCSP Status Request extension unbounded memory growth
5924 A malicious client can send an excessively large OCSP Status Request
5925 extension. If that client continually requests renegotiation, sending a
5926 large OCSP Status Request extension each time, then there will be unbounded
5927 memory growth on the server. This will eventually lead to a Denial Of
5928 Service attack through memory exhaustion. Servers with a default
5929 configuration are vulnerable even if they do not support OCSP. Builds using
5930 the "no-ocsp" build time option are not affected.
5932 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5937 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5940 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5946 * OOB write in MDC2_Update()
5948 An overflow can occur in MDC2_Update() either if called directly or
5949 through the EVP_DigestUpdate() function using MDC2. If an attacker
5950 is able to supply very large amounts of input data after a previous
5951 call to EVP_EncryptUpdate() with a partial block then a length check
5952 can overflow resulting in a heap corruption.
5954 The amount of data needed is comparable to SIZE_MAX which is impractical
5957 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5962 * Malformed SHA512 ticket DoS
5964 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5965 DoS attack where a malformed ticket will result in an OOB read which will
5968 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5969 a custom server callback and ticket lookup mechanism.
5971 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5976 * OOB write in BN_bn2dec()
5978 The function BN_bn2dec() does not check the return value of BN_div_word().
5979 This can cause an OOB write if an application uses this function with an
5980 overly large BIGNUM. This could be a problem if an overly large certificate
5981 or CRL is printed out from an untrusted source. TLS is not affected because
5982 record limits will reject an oversized certificate before it is parsed.
5984 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5989 * OOB read in TS_OBJ_print_bio()
5991 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5992 the total length the OID text representation would use and not the amount
5993 of data written. This will result in OOB reads when large OIDs are
5996 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6001 * Pointer arithmetic undefined behaviour
6003 Avoid some undefined pointer arithmetic
6005 A common idiom in the codebase is to check limits in the following manner:
6008 Where "p" points to some malloc'd data of SIZE bytes and
6011 "len" here could be from some externally supplied data (e.g. from a TLS
6014 The rules of C pointer arithmetic are such that "p + len" is only well
6015 defined where len <= SIZE. Therefore the above idiom is actually
6016 undefined behaviour.
6018 For example this could cause problems if some malloc implementation
6019 provides an address for "p" such that "p + len" actually overflows for
6020 values of len that are too big and therefore p + len < limit.
6022 This issue was reported to OpenSSL by Guido Vranken
6027 * Constant time flag not preserved in DSA signing
6029 Operations in the DSA signing algorithm should run in constant time in
6030 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6031 implementation means that a non-constant time codepath is followed for
6032 certain operations. This has been demonstrated through a cache-timing
6033 attack to be sufficient for an attacker to recover the private DSA key.
6035 This issue was reported by César Pereida (Aalto University), Billy Brumley
6036 (Tampere University of Technology), and Yuval Yarom (The University of
6037 Adelaide and NICTA).
6042 * DTLS buffered message DoS
6044 In a DTLS connection where handshake messages are delivered out-of-order
6045 those messages that OpenSSL is not yet ready to process will be buffered
6046 for later use. Under certain circumstances, a flaw in the logic means that
6047 those messages do not get removed from the buffer even though the handshake
6048 has been completed. An attacker could force up to approx. 15 messages to
6049 remain in the buffer when they are no longer required. These messages will
6050 be cleared when the DTLS connection is closed. The default maximum size for
6051 a message is 100k. Therefore the attacker could force an additional 1500k
6052 to be consumed per connection. By opening many simulataneous connections an
6053 attacker could cause a DoS attack through memory exhaustion.
6055 This issue was reported to OpenSSL by Quan Luo.
6060 * DTLS replay protection DoS
6062 A flaw in the DTLS replay attack protection mechanism means that records
6063 that arrive for future epochs update the replay protection "window" before
6064 the MAC for the record has been validated. This could be exploited by an
6065 attacker by sending a record for the next epoch (which does not have to
6066 decrypt or have a valid MAC), with a very large sequence number. This means
6067 that all subsequent legitimate packets are dropped causing a denial of
6068 service for a specific DTLS connection.
6070 This issue was reported to OpenSSL by the OCAP audit team.
6075 * Certificate message OOB reads
6077 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6078 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6079 theoretical DoS risk but this has not been observed in practice on common
6082 The messages affected are client certificate, client certificate request
6083 and server certificate. As a result the attack can only be performed
6084 against a client or a server which enables client authentication.
6086 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6091 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6093 * Prevent padding oracle in AES-NI CBC MAC check
6095 A MITM attacker can use a padding oracle attack to decrypt traffic
6096 when the connection uses an AES CBC cipher and the server support
6099 This issue was introduced as part of the fix for Lucky 13 padding
6100 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6101 constant time by making sure that always the same bytes are read and
6102 compared against either the MAC or padding bytes. But it no longer
6103 checked that there was enough data to have both the MAC and padding
6106 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6111 * Fix EVP_EncodeUpdate overflow
6113 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6114 Base64 encoding of binary data. If an attacker is able to supply very large
6115 amounts of input data then a length check can overflow resulting in a heap
6118 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6119 the `PEM_write_bio*` family of functions. These are mainly used within the
6120 OpenSSL command line applications, so any application which processes data
6121 from an untrusted source and outputs it as a PEM file should be considered
6122 vulnerable to this issue. User applications that call these APIs directly
6123 with large amounts of untrusted data may also be vulnerable.
6125 This issue was reported by Guido Vranken.
6130 * Fix EVP_EncryptUpdate overflow
6132 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6133 is able to supply very large amounts of input data after a previous call to
6134 EVP_EncryptUpdate() with a partial block then a length check can overflow
6135 resulting in a heap corruption. Following an analysis of all OpenSSL
6136 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6137 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6138 the first called function after an EVP_EncryptInit(), and therefore that
6139 specific call must be safe. The second form is where the length passed to
6140 EVP_EncryptUpdate() can be seen from the code to be some small value and
6141 therefore there is no possibility of an overflow. Since all instances are
6142 one of these two forms, it is believed that there can be no overflows in
6143 internal code due to this problem. It should be noted that
6144 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6145 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6146 of these calls have also been analysed too and it is believed there are no
6147 instances in internal usage where an overflow could occur.
6149 This issue was reported by Guido Vranken.
6154 * Prevent ASN.1 BIO excessive memory allocation
6156 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6157 a short invalid encoding can casuse allocation of large amounts of memory
6158 potentially consuming excessive resources or exhausting memory.
6160 Any application parsing untrusted data through d2i BIO functions is
6161 affected. The memory based functions such as d2i_X509() are *not* affected.
6162 Since the memory based functions are used by the TLS library, TLS
6163 applications are not affected.
6165 This issue was reported by Brian Carpenter.
6172 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6173 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6174 in arbitrary stack data being returned in the buffer.
6176 This issue was reported by Guido Vranken.
6181 * Modify behavior of ALPN to invoke callback after SNI/servername
6182 callback, such that updates to the SSL_CTX affect ALPN.
6186 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6191 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6192 methods are enabled and ssl2 is disabled the methods return NULL.
6196 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6198 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6199 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6200 provide any "EXPORT" or "LOW" strength ciphers.
6204 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6205 is by default disabled at build-time. Builds that are not configured with
6206 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6207 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6208 will need to explicitly call either of:
6210 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6212 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6214 as appropriate. Even if either of those is used, or the application
6215 explicitly uses the version-specific SSLv2_method() or its client and
6216 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6217 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6218 ciphers, and SSLv2 56-bit DES are no longer available.
6223 * Fix a double-free in DSA code
6225 A double free bug was discovered when OpenSSL parses malformed DSA private
6226 keys and could lead to a DoS attack or memory corruption for applications
6227 that receive DSA private keys from untrusted sources. This scenario is
6230 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6236 * Disable SRP fake user seed to address a server memory leak.
6238 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6240 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6241 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6242 was changed to ignore the "fake user" SRP seed, even if the seed
6245 Users should use SRP_VBASE_get1_by_user instead. Note that in
6246 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6247 also that even though configuring the SRP seed attempts to hide
6248 invalid usernames by continuing the handshake with fake
6249 credentials, this behaviour is not constant time and no strong
6250 guarantees are made that the handshake is indistinguishable from
6251 that of a valid user.
6256 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6258 In the BN_hex2bn function the number of hex digits is calculated using an
6259 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6260 large values of `i` this can result in `bn_expand` not allocating any
6261 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6262 field as NULL leading to a subsequent NULL ptr deref. For very large values
6263 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6264 In this case memory is allocated to the internal BIGNUM data field, but it
6265 is insufficiently sized leading to heap corruption. A similar issue exists
6266 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6267 is ever called by user applications with very large untrusted hex/dec data.
6268 This is anticipated to be a rare occurrence.
6270 All OpenSSL internal usage of these functions use data that is not expected
6271 to be untrusted, e.g. config file data or application command line
6272 arguments. If user developed applications generate config file data based
6273 on untrusted data then it is possible that this could also lead to security
6274 consequences. This is also anticipated to be rare.
6276 This issue was reported to OpenSSL by Guido Vranken.
6281 * Fix memory issues in `BIO_*printf` functions
6283 The internal `fmtstr` function used in processing a "%s" format string in
6284 the `BIO_*printf` functions could overflow while calculating the length of a
6285 string and cause an OOB read when printing very long strings.
6287 Additionally the internal `doapr_outch` function can attempt to write to an
6288 OOB memory location (at an offset from the NULL pointer) in the event of a
6289 memory allocation failure. In 1.0.2 and below this could be caused where
6290 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6291 could be in processing a very long "%s" format string. Memory leaks can
6294 The first issue may mask the second issue dependent on compiler behaviour.
6295 These problems could enable attacks where large amounts of untrusted data
6296 is passed to the `BIO_*printf` functions. If applications use these functions
6297 in this way then they could be vulnerable. OpenSSL itself uses these
6298 functions when printing out human-readable dumps of ASN.1 data. Therefore
6299 applications that print this data could be vulnerable if the data is from
6300 untrusted sources. OpenSSL command line applications could also be
6301 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6302 as command line arguments.
6304 Libssl is not considered directly vulnerable. Additionally certificates etc
6305 received via remote connections via libssl are also unlikely to be able to
6306 trigger these issues because of message size limits enforced within libssl.
6308 This issue was reported to OpenSSL Guido Vranken.
6313 * Side channel attack on modular exponentiation
6315 A side-channel attack was found which makes use of cache-bank conflicts on
6316 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6317 of RSA keys. The ability to exploit this issue is limited as it relies on
6318 an attacker who has control of code in a thread running on the same
6319 hyper-threaded core as the victim thread which is performing decryptions.
6321 This issue was reported to OpenSSL by Yuval Yarom, The University of
6322 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6323 Nadia Heninger, University of Pennsylvania with more information at
6324 <http://cachebleed.info>.
6329 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6330 if no keysize is specified with default_bits. This fixes an
6331 omission in an earlier change that changed all RSA/DSA key generation
6332 commands to use 2048 bits by default.
6336 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6338 * Protection for DH small subgroup attacks
6340 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6341 switched on by default and cannot be disabled. This could have some
6346 * SSLv2 doesn't block disabled ciphers
6348 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6349 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6350 been disabled, provided that the SSLv2 protocol was not also disabled via
6353 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6354 and Sebastian Schinzel.
6359 * Reject DH handshakes with parameters shorter than 1024 bits.
6363 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6365 * Certificate verify crash with missing PSS parameter
6367 The signature verification routines will crash with a NULL pointer
6368 dereference if presented with an ASN.1 signature using the RSA PSS
6369 algorithm and absent mask generation function parameter. Since these
6370 routines are used to verify certificate signature algorithms this can be
6371 used to crash any certificate verification operation and exploited in a
6372 DoS attack. Any application which performs certificate verification is
6373 vulnerable including OpenSSL clients and servers which enable client
6376 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6381 * X509_ATTRIBUTE memory leak
6383 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6384 memory. This structure is used by the PKCS#7 and CMS routines so any
6385 application which reads PKCS#7 or CMS data from untrusted sources is
6386 affected. SSL/TLS is not affected.
6388 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6394 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6395 This changes the decoding behaviour for some invalid messages,
6396 though the change is mostly in the more lenient direction, and
6397 legacy behaviour is preserved as much as possible.
6401 * In DSA_generate_parameters_ex, if the provided seed is too short,
6402 use a random seed, as already documented.
6404 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6406 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6408 * Alternate chains certificate forgery
6410 During certificate verfification, OpenSSL will attempt to find an
6411 alternative certificate chain if the first attempt to build such a chain
6412 fails. An error in the implementation of this logic can mean that an
6413 attacker could cause certain checks on untrusted certificates to be
6414 bypassed, such as the CA flag, enabling them to use a valid leaf
6415 certificate to act as a CA and "issue" an invalid certificate.
6417 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6423 * Race condition handling PSK identify hint
6425 If PSK identity hints are received by a multi-threaded client then
6426 the values are wrongly updated in the parent SSL_CTX structure. This can
6427 result in a race condition potentially leading to a double free of the
6433 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6435 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6436 incompatibility in the handling of HMAC. The previous ABI has now been
6439 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6441 * Malformed ECParameters causes infinite loop
6443 When processing an ECParameters structure OpenSSL enters an infinite loop
6444 if the curve specified is over a specially malformed binary polynomial
6447 This can be used to perform denial of service against any
6448 system which processes public keys, certificate requests or
6449 certificates. This includes TLS clients and TLS servers with
6450 client authentication enabled.
6452 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6457 * Exploitable out-of-bounds read in X509_cmp_time
6459 X509_cmp_time does not properly check the length of the ASN1_TIME
6460 string and can read a few bytes out of bounds. In addition,
6461 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6464 An attacker can use this to craft malformed certificates and CRLs of
6465 various sizes and potentially cause a segmentation fault, resulting in
6466 a DoS on applications that verify certificates or CRLs. TLS clients
6467 that verify CRLs are affected. TLS clients and servers with client
6468 authentication enabled may be affected if they use custom verification
6471 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6472 independently by Hanno Böck.
6477 * PKCS7 crash with missing EnvelopedContent
6479 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6480 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6481 with missing content and trigger a NULL pointer dereference on parsing.
6483 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6484 structures from untrusted sources are affected. OpenSSL clients and
6485 servers are not affected.
6487 This issue was reported to OpenSSL by Michal Zalewski (Google).
6492 * CMS verify infinite loop with unknown hash function
6494 When verifying a signedData message the CMS code can enter an infinite loop
6495 if presented with an unknown hash function OID. This can be used to perform
6496 denial of service against any system which verifies signedData messages using
6498 This issue was reported to OpenSSL by Johannes Bauer.
6503 * Race condition handling NewSessionTicket
6505 If a NewSessionTicket is received by a multi-threaded client when attempting to
6506 reuse a previous ticket then a race condition can occur potentially leading to
6507 a double free of the ticket data.
6512 * Reject DH handshakes with parameters shorter than 768 bits.
6514 *Kurt Roeckx and Emilia Kasper*
6516 * dhparam: generate 2048-bit parameters by default.
6518 *Kurt Roeckx and Emilia Kasper*
6520 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6522 * Segmentation fault in ASN1_TYPE_cmp fix
6524 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6525 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6526 certificate signature algorithm consistency this can be used to crash any
6527 certificate verification operation and exploited in a DoS attack. Any
6528 application which performs certificate verification is vulnerable including
6529 OpenSSL clients and servers which enable client authentication.
6534 * ASN.1 structure reuse memory corruption fix
6536 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6537 memory corruption via an invalid write. Such reuse is and has been
6538 strongly discouraged and is believed to be rare.
6540 Applications that parse structures containing CHOICE or ANY DEFINED BY
6541 components may be affected. Certificate parsing (d2i_X509 and related
6542 functions) are however not affected. OpenSSL clients and servers are
6548 * PKCS7 NULL pointer dereferences fix
6550 The PKCS#7 parsing code does not handle missing outer ContentInfo
6551 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6552 missing content and trigger a NULL pointer dereference on parsing.
6554 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6555 otherwise parse PKCS#7 structures from untrusted sources are
6556 affected. OpenSSL clients and servers are not affected.
6558 This issue was reported to OpenSSL by Michal Zalewski (Google).
6563 * DoS via reachable assert in SSLv2 servers fix
6565 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6566 servers that both support SSLv2 and enable export cipher suites by sending
6567 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6569 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6570 (OpenSSL development team).
6575 * Use After Free following d2i_ECPrivatekey error fix
6577 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6578 could cause a use after free condition. This, in turn, could cause a double
6579 free in several private key parsing functions (such as d2i_PrivateKey
6580 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6581 for applications that receive EC private keys from untrusted
6582 sources. This scenario is considered rare.
6584 This issue was discovered by the BoringSSL project and fixed in their
6590 * X509_to_X509_REQ NULL pointer deref fix
6592 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6593 the certificate key is invalid. This function is rarely used in practice.
6595 This issue was discovered by Brian Carpenter.
6600 * Removed the export ciphers from the DEFAULT ciphers
6604 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6606 * Build fixes for the Windows and OpenVMS platforms
6608 *Matt Caswell and Richard Levitte*
6610 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6612 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6613 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6614 dereference. This could lead to a Denial Of Service attack. Thanks to
6615 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6620 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6621 dtls1_buffer_record function under certain conditions. In particular this
6622 could occur if an attacker sent repeated DTLS records with the same
6623 sequence number but for the next epoch. The memory leak could be exploited
6624 by an attacker in a Denial of Service attack through memory exhaustion.
6625 Thanks to Chris Mueller for reporting this issue.
6630 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6631 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6632 method would be set to NULL which could later result in a NULL pointer
6633 dereference. Thanks to Frank Schmirler for reporting this issue.
6638 * Abort handshake if server key exchange message is omitted for ephemeral
6641 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6642 reporting this issue.
6647 * Remove non-export ephemeral RSA code on client and server. This code
6648 violated the TLS standard by allowing the use of temporary RSA keys in
6649 non-export ciphersuites and could be used by a server to effectively
6650 downgrade the RSA key length used to a value smaller than the server
6651 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6652 INRIA or reporting this issue.
6657 * Fixed issue where DH client certificates are accepted without verification.
6658 An OpenSSL server will accept a DH certificate for client authentication
6659 without the certificate verify message. This effectively allows a client to
6660 authenticate without the use of a private key. This only affects servers
6661 which trust a client certificate authority which issues certificates
6662 containing DH keys: these are extremely rare and hardly ever encountered.
6663 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6669 * Ensure that the session ID context of an SSL is updated when its
6670 SSL_CTX is updated via SSL_set_SSL_CTX.
6672 The session ID context is typically set from the parent SSL_CTX,
6673 and can vary with the CTX.
6677 * Fix various certificate fingerprint issues.
6679 By using non-DER or invalid encodings outside the signed portion of a
6680 certificate the fingerprint can be changed without breaking the signature.
6681 Although no details of the signed portion of the certificate can be changed
6682 this can cause problems with some applications: e.g. those using the
6683 certificate fingerprint for blacklists.
6685 1. Reject signatures with non zero unused bits.
6687 If the BIT STRING containing the signature has non zero unused bits reject
6688 the signature. All current signature algorithms require zero unused bits.
6690 2. Check certificate algorithm consistency.
6692 Check the AlgorithmIdentifier inside TBS matches the one in the
6693 certificate signature. NB: this will result in signature failure
6694 errors for some broken certificates.
6696 Thanks to Konrad Kraszewski from Google for reporting this issue.
6698 3. Check DSA/ECDSA signatures use DER.
6700 Re-encode DSA/ECDSA signatures and compare with the original received
6701 signature. Return an error if there is a mismatch.
6703 This will reject various cases including garbage after signature
6704 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6705 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6706 (negative or with leading zeroes).
6708 Further analysis was conducted and fixes were developed by Stephen Henson
6709 of the OpenSSL core team.
6715 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6716 results on some platforms, including x86_64. This bug occurs at random
6717 with a very low probability, and is not known to be exploitable in any
6718 way, though its exact impact is difficult to determine. Thanks to Pieter
6719 Wuille (Blockstream) who reported this issue and also suggested an initial
6720 fix. Further analysis was conducted by the OpenSSL development team and
6721 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6722 the OpenSSL core team.
6727 * Do not resume sessions on the server if the negotiated protocol
6728 version does not match the session's version. Resuming with a different
6729 version, while not strictly forbidden by the RFC, is of questionable
6730 sanity and breaks all known clients.
6732 *David Benjamin, Emilia Käsper*
6734 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6735 early CCS messages during renegotiation. (Note that because
6736 renegotiation is encrypted, this early CCS was not exploitable.)
6740 * Tighten client-side session ticket handling during renegotiation:
6741 ensure that the client only accepts a session ticket if the server sends
6742 the extension anew in the ServerHello. Previously, a TLS client would
6743 reuse the old extension state and thus accept a session ticket if one was
6744 announced in the initial ServerHello.
6746 Similarly, ensure that the client requires a session ticket if one
6747 was advertised in the ServerHello. Previously, a TLS client would
6748 ignore a missing NewSessionTicket message.
6752 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6756 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6757 sends a carefully crafted handshake message, to cause OpenSSL to fail
6758 to free up to 64k of memory causing a memory leak. This could be
6759 exploited in a Denial Of Service attack. This issue affects OpenSSL
6760 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6761 whether SRTP is used or configured. Implementations of OpenSSL that
6762 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6764 The fix was developed by the OpenSSL team.
6769 * Session Ticket Memory Leak.
6771 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6772 integrity of that ticket is first verified. In the event of a session
6773 ticket integrity check failing, OpenSSL will fail to free memory
6774 causing a memory leak. By sending a large number of invalid session
6775 tickets an attacker could exploit this issue in a Denial Of Service
6781 * Build option no-ssl3 is incomplete.
6783 When OpenSSL is configured with "no-ssl3" as a build option, servers
6784 could accept and complete a SSL 3.0 handshake, and clients could be
6785 configured to send them.
6788 *Akamai and the OpenSSL team*
6790 * Add support for TLS_FALLBACK_SCSV.
6791 Client applications doing fallback retries should call
6792 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6795 *Adam Langley, Bodo Moeller*
6797 * Add additional DigestInfo checks.
6799 Re-encode DigestInto in DER and check against the original when
6800 verifying RSA signature: this will reject any improperly encoded
6801 DigestInfo structures.
6803 Note: this is a precautionary measure and no attacks are currently known.
6807 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6809 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6810 SRP code can be overrun an internal buffer. Add sanity check that
6811 g, A, B < N to SRP code.
6813 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6814 Group for discovering this issue.
6819 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6820 TLS 1.0 instead of higher protocol versions when the ClientHello message
6821 is badly fragmented. This allows a man-in-the-middle attacker to force a
6822 downgrade to TLS 1.0 even if both the server and the client support a
6823 higher protocol version, by modifying the client's TLS records.
6825 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6826 researching this issue.
6831 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6832 to a denial of service attack. A malicious server can crash the client
6833 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6834 ciphersuite and sending carefully crafted handshake messages.
6836 Thanks to Felix Gröbert (Google) for discovering and researching this
6842 * By sending carefully crafted DTLS packets an attacker could cause openssl
6843 to leak memory. This can be exploited through a Denial of Service attack.
6844 Thanks to Adam Langley for discovering and researching this issue.
6849 * An attacker can force openssl to consume large amounts of memory whilst
6850 processing DTLS handshake messages. This can be exploited through a
6851 Denial of Service attack.
6852 Thanks to Adam Langley for discovering and researching this issue.
6857 * An attacker can force an error condition which causes openssl to crash
6858 whilst processing DTLS packets due to memory being freed twice. This
6859 can be exploited through a Denial of Service attack.
6860 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6866 * If a multithreaded client connects to a malicious server using a resumed
6867 session and the server sends an ec point format extension it could write
6868 up to 255 bytes to freed memory.
6870 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6876 * A malicious server can crash an OpenSSL client with a null pointer
6877 dereference (read) by specifying an SRP ciphersuite even though it was not
6878 properly negotiated with the client. This can be exploited through a
6879 Denial of Service attack.
6881 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6882 discovering and researching this issue.
6887 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6888 X509_name_oneline, X509_name_print_ex et al. to leak some information
6889 from the stack. Applications may be affected if they echo pretty printing
6890 output to the attacker.
6892 Thanks to Ivan Fratric (Google) for discovering this issue.
6895 *Emilia Käsper, and Steve Henson*
6897 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6898 for corner cases. (Certain input points at infinity could lead to
6899 bogus results, with non-infinity inputs mapped to infinity too.)
6903 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6905 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6906 handshake can force the use of weak keying material in OpenSSL
6907 SSL/TLS clients and servers.
6909 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6910 researching this issue. ([CVE-2014-0224])
6912 *KIKUCHI Masashi, Steve Henson*
6914 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6915 OpenSSL DTLS client the code can be made to recurse eventually crashing
6918 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6921 *Imre Rad, Steve Henson*
6923 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6924 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6925 client or server. This is potentially exploitable to run arbitrary
6926 code on a vulnerable client or server.
6928 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6930 *Jüri Aedla, Steve Henson*
6932 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6933 are subject to a denial of service attack.
6935 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6936 this issue. ([CVE-2014-3470])
6938 *Felix Gröbert, Ivan Fratric, Steve Henson*
6940 * Harmonize version and its documentation. -f flag is used to display
6943 *mancha <mancha1@zoho.com>*
6945 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6946 in i2d_ECPrivateKey.
6948 *mancha <mancha1@zoho.com>*
6950 * Fix some double frees. These are not thought to be exploitable.
6952 *mancha <mancha1@zoho.com>*
6954 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6956 * A missing bounds check in the handling of the TLS heartbeat extension
6957 can be used to reveal up to 64k of memory to a connected client or
6960 Thanks for Neel Mehta of Google Security for discovering this bug and to
6961 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6962 preparing the fix ([CVE-2014-0160])
6964 *Adam Langley, Bodo Moeller*
6966 * Fix for the attack described in the paper "Recovering OpenSSL
6967 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6968 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6969 <http://eprint.iacr.org/2014/140>
6971 Thanks to Yuval Yarom and Naomi Benger for discovering this
6972 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6974 *Yuval Yarom and Naomi Benger*
6976 * TLS pad extension: draft-agl-tls-padding-03
6978 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6979 TLS client Hello record length value would otherwise be > 255 and
6980 less that 512 pad with a dummy extension containing zeroes so it
6981 is at least 512 bytes long.
6983 *Adam Langley, Steve Henson*
6985 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6987 * Fix for TLS record tampering bug. A carefully crafted invalid
6988 handshake could crash OpenSSL with a NULL pointer exception.
6989 Thanks to Anton Johansson for reporting this issues.
6992 * Keep original DTLS digest and encryption contexts in retransmission
6993 structures so we can use the previous session parameters if they need
6994 to be resent. ([CVE-2013-6450])
6998 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6999 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7000 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7001 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7002 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7003 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7005 *Rob Stradling, Adam Langley*
7007 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7009 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7010 supporting platforms or when small records were transferred.
7012 *Andy Polyakov, Steve Henson*
7014 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7016 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7018 This addresses the flaw in CBC record processing discovered by
7019 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7020 at: <http://www.isg.rhul.ac.uk/tls/>
7022 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7023 Security Group at Royal Holloway, University of London
7024 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7025 Emilia Käsper for the initial patch.
7028 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7030 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7031 ciphersuites which can be exploited in a denial of service attack.
7032 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7033 and detecting this bug and to Wolfgang Ettlinger
7034 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7039 * Return an error when checking OCSP signatures when key is NULL.
7040 This fixes a DoS attack. ([CVE-2013-0166])
7044 * Make openssl verify return errors.
7046 *Chris Palmer <palmer@google.com> and Ben Laurie*
7048 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7049 the right response is stapled. Also change SSL_get_certificate()
7050 so it returns the certificate actually sent.
7051 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7053 *Rob Stradling <rob.stradling@comodo.com>*
7055 * Fix possible deadlock when decoding public keys.
7059 * Don't use TLS 1.0 record version number in initial client hello
7064 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7066 * Sanity check record length before skipping explicit IV in TLS
7067 1.2, 1.1 and DTLS to fix DoS attack.
7069 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7070 fuzzing as a service testing platform.
7075 * Initialise tkeylen properly when encrypting CMS messages.
7076 Thanks to Solar Designer of Openwall for reporting this issue.
7080 * In FIPS mode don't try to use composite ciphers as they are not
7085 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7087 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7088 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7089 mean any application compiled against OpenSSL 1.0.0 headers setting
7090 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7091 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7092 0x10000000L Any application which was previously compiled against
7093 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7094 will need to be recompiled as a result. Letting be results in
7095 inability to disable specifically TLS 1.1 and in client context,
7096 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7100 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7101 disable just protocol X, but all protocols above X *if* there are
7102 protocols *below* X still enabled. In more practical terms it means
7103 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7104 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7105 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7110 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7112 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7113 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7114 in CRYPTO_realloc_clean.
7116 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7117 issue and to Adam Langley <agl@chromium.org> for fixing it.
7120 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7122 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7126 * Workarounds for some broken servers that "hang" if a client hello
7127 record length exceeds 255 bytes.
7129 1. Do not use record version number > TLS 1.0 in initial client
7130 hello: some (but not all) hanging servers will now work.
7131 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7132 the number of ciphers sent in the client hello. This should be
7133 set to an even number, such as 50, for example by passing:
7134 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7135 Most broken servers should now work.
7136 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7137 TLS 1.2 client support entirely.
7141 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7145 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7147 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7148 STRING form instead of a DigestInfo.
7152 * The format used for MDC2 RSA signatures is inconsistent between EVP
7153 and the RSA_sign/RSA_verify functions. This was made more apparent when
7154 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7155 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7156 the correct format in RSA_verify so both forms transparently work.
7160 * Some servers which support TLS 1.0 can choke if we initially indicate
7161 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7162 encrypted premaster secret. As a workaround use the maximum permitted
7163 client version in client hello, this should keep such servers happy
7164 and still work with previous versions of OpenSSL.
7168 * Add support for TLS/DTLS heartbeats.
7170 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7172 * Add support for SCTP.
7174 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7176 * Improved PRNG seeding for VOS.
7178 *Paul Green <Paul.Green@stratus.com>*
7180 * Extensive assembler packs updates, most notably:
7182 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7183 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7184 - x86_64: bit-sliced AES implementation;
7185 - ARM: NEON support, contemporary platforms optimizations;
7186 - s390x: z196 support;
7187 - `*`: GHASH and GF(2^m) multiplication implementations;
7191 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7192 (removal of unnecessary code)
7194 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7196 * Add TLS key material exporter from RFC 5705.
7200 * Add DTLS-SRTP negotiation from RFC 5764.
7204 * Add Next Protocol Negotiation,
7205 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7206 disabled with a no-npn flag to config or Configure. Code donated
7209 *Adam Langley <agl@google.com> and Ben Laurie*
7211 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7212 NIST-P256, NIST-P521, with constant-time single point multiplication on
7213 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7214 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7215 Code made available under Apache License version 2.0.
7217 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7218 line to include this in your build of OpenSSL, and run "make depend" (or
7219 "make update"). This enables the following EC_METHODs:
7221 EC_GFp_nistp224_method()
7222 EC_GFp_nistp256_method()
7223 EC_GFp_nistp521_method()
7225 EC_GROUP_new_by_curve_name() will automatically use these (while
7226 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7229 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7231 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7232 all platforms. Move ssize_t definition from e_os.h to the public
7233 header file e_os2.h as it now appears in public header file cms.h
7237 * New -sigopt option to the ca, req and x509 utilities. Additional
7238 signature parameters can be passed using this option and in
7243 * Add RSA PSS signing function. This will generate and set the
7244 appropriate AlgorithmIdentifiers for PSS based on those in the
7245 corresponding EVP_MD_CTX structure. No application support yet.
7249 * Support for companion algorithm specific ASN1 signing routines.
7250 New function ASN1_item_sign_ctx() signs a pre-initialised
7251 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7252 the appropriate parameters.
7256 * Add new algorithm specific ASN1 verification initialisation function
7257 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7258 handling will be the same no matter what EVP_PKEY_METHOD is used.
7259 Add a PSS handler to support verification of PSS signatures: checked
7260 against a number of sample certificates.
7264 * Add signature printing for PSS. Add PSS OIDs.
7266 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7268 * Add algorithm specific signature printing. An individual ASN1 method
7269 can now print out signatures instead of the standard hex dump.
7271 More complex signatures (e.g. PSS) can print out more meaningful
7272 information. Include DSA version that prints out the signature
7277 * Password based recipient info support for CMS library: implementing
7282 * Split password based encryption into PBES2 and PBKDF2 functions. This
7283 neatly separates the code into cipher and PBE sections and is required
7284 for some algorithms that split PBES2 into separate pieces (such as
7285 password based CMS).
7289 * Session-handling fixes:
7290 - Fix handling of connections that are resuming with a session ID,
7291 but also support Session Tickets.
7292 - Fix a bug that suppressed issuing of a new ticket if the client
7293 presented a ticket with an expired session.
7294 - Try to set the ticket lifetime hint to something reasonable.
7295 - Make tickets shorter by excluding irrelevant information.
7296 - On the client side, don't ignore renewed tickets.
7298 *Adam Langley, Bodo Moeller (Google)*
7300 * Fix PSK session representation.
7304 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7306 This work was sponsored by Intel.
7310 * Add GCM support to TLS library. Some custom code is needed to split
7311 the IV between the fixed (from PRF) and explicit (from TLS record)
7312 portions. This adds all GCM ciphersuites supported by RFC5288 and
7313 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7314 add a special AESGCM string for GCM only.
7318 * Expand range of ctrls for AES GCM. Permit setting invocation
7319 field on decrypt and retrieval of invocation field only on encrypt.
7323 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7324 As required by RFC5289 these ciphersuites cannot be used if for
7325 versions of TLS earlier than 1.2.
7329 * For FIPS capable OpenSSL interpret a NULL default public key method
7330 as unset and return the appropriate default but do *not* set the default.
7331 This means we can return the appropriate method in applications that
7332 switch between FIPS and non-FIPS modes.
7336 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7337 ENGINE is used then we cannot handle that in the FIPS module so we
7338 keep original code iff non-FIPS operations are allowed.
7342 * Add -attime option to openssl utilities.
7344 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7346 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7350 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7351 FIPS EC methods unconditionally for now.
7355 * New build option no-ec2m to disable characteristic 2 code.
7359 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7360 all cases can be covered as some introduce binary incompatibilities.
7364 * Redirect RSA operations to FIPS module including keygen,
7365 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7369 * Add similar low-level API blocking to ciphers.
7373 * low-level digest APIs are not approved in FIPS mode: any attempt
7374 to use these will cause a fatal error. Applications that *really* want
7375 to use them can use the `private_*` version instead.
7379 * Redirect cipher operations to FIPS module for FIPS builds.
7383 * Redirect digest operations to FIPS module for FIPS builds.
7387 * Update build system to add "fips" flag which will link in fipscanister.o
7388 for static and shared library builds embedding a signature if needed.
7392 * Output TLS supported curves in preference order instead of numerical
7393 order. This is currently hardcoded for the highest order curves first.
7394 This should be configurable so applications can judge speed vs strength.
7398 * Add TLS v1.2 server support for client authentication.
7402 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7407 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7408 FIPS modules versions.
7412 * Add TLS v1.2 client side support for client authentication. Keep cache
7413 of handshake records longer as we don't know the hash algorithm to use
7414 until after the certificate request message is received.
7418 * Initial TLS v1.2 client support. Add a default signature algorithms
7419 extension including all the algorithms we support. Parse new signature
7420 format in client key exchange. Relax some ECC signing restrictions for
7421 TLS v1.2 as indicated in RFC5246.
7425 * Add server support for TLS v1.2 signature algorithms extension. Switch
7426 to new signature format when needed using client digest preference.
7427 All server ciphersuites should now work correctly in TLS v1.2. No client
7428 support yet and no support for client certificates.
7432 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7433 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7434 ciphersuites. At present only RSA key exchange ciphersuites work with
7435 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7436 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7437 and version checking.
7441 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7442 with this defined it will not be affected by any changes to ssl internal
7443 structures. Add several utility functions to allow openssl application
7444 to work with OPENSSL_NO_SSL_INTERN defined.
7448 * A long standing patch to add support for SRP from EdelWeb (Peter
7449 Sylvester and Christophe Renou) was integrated.
7450 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7451 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7454 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7458 * Permit abbreviated handshakes when renegotiating using the function
7459 SSL_renegotiate_abbreviated().
7461 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7463 * Add call to ENGINE_register_all_complete() to
7464 ENGINE_load_builtin_engines(), so some implementations get used
7465 automatically instead of needing explicit application support.
7469 * Add support for TLS key exporter as described in RFC5705.
7471 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7473 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7474 a few changes are required:
7476 Add SSL_OP_NO_TLSv1_1 flag.
7477 Add TLSv1_1 methods.
7478 Update version checking logic to handle version 1.1.
7479 Add explicit IV handling (ported from DTLS code).
7480 Add command line options to s_client/s_server.
7487 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7489 * X509_ATTRIBUTE memory leak
7491 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7492 memory. This structure is used by the PKCS#7 and CMS routines so any
7493 application which reads PKCS#7 or CMS data from untrusted sources is
7494 affected. SSL/TLS is not affected.
7496 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7502 * Race condition handling PSK identify hint
7504 If PSK identity hints are received by a multi-threaded client then
7505 the values are wrongly updated in the parent SSL_CTX structure. This can
7506 result in a race condition potentially leading to a double free of the
7512 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7514 * Malformed ECParameters causes infinite loop
7516 When processing an ECParameters structure OpenSSL enters an infinite loop
7517 if the curve specified is over a specially malformed binary polynomial
7520 This can be used to perform denial of service against any
7521 system which processes public keys, certificate requests or
7522 certificates. This includes TLS clients and TLS servers with
7523 client authentication enabled.
7525 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7530 * Exploitable out-of-bounds read in X509_cmp_time
7532 X509_cmp_time does not properly check the length of the ASN1_TIME
7533 string and can read a few bytes out of bounds. In addition,
7534 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7537 An attacker can use this to craft malformed certificates and CRLs of
7538 various sizes and potentially cause a segmentation fault, resulting in
7539 a DoS on applications that verify certificates or CRLs. TLS clients
7540 that verify CRLs are affected. TLS clients and servers with client
7541 authentication enabled may be affected if they use custom verification
7544 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7545 independently by Hanno Böck.
7550 * PKCS7 crash with missing EnvelopedContent
7552 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7553 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7554 with missing content and trigger a NULL pointer dereference on parsing.
7556 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7557 structures from untrusted sources are affected. OpenSSL clients and
7558 servers are not affected.
7560 This issue was reported to OpenSSL by Michal Zalewski (Google).
7565 * CMS verify infinite loop with unknown hash function
7567 When verifying a signedData message the CMS code can enter an infinite loop
7568 if presented with an unknown hash function OID. This can be used to perform
7569 denial of service against any system which verifies signedData messages using
7571 This issue was reported to OpenSSL by Johannes Bauer.
7576 * Race condition handling NewSessionTicket
7578 If a NewSessionTicket is received by a multi-threaded client when attempting to
7579 reuse a previous ticket then a race condition can occur potentially leading to
7580 a double free of the ticket data.
7585 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7587 * Segmentation fault in ASN1_TYPE_cmp fix
7589 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7590 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7591 certificate signature algorithm consistency this can be used to crash any
7592 certificate verification operation and exploited in a DoS attack. Any
7593 application which performs certificate verification is vulnerable including
7594 OpenSSL clients and servers which enable client authentication.
7599 * ASN.1 structure reuse memory corruption fix
7601 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7602 memory corruption via an invalid write. Such reuse is and has been
7603 strongly discouraged and is believed to be rare.
7605 Applications that parse structures containing CHOICE or ANY DEFINED BY
7606 components may be affected. Certificate parsing (d2i_X509 and related
7607 functions) are however not affected. OpenSSL clients and servers are
7613 * PKCS7 NULL pointer dereferences fix
7615 The PKCS#7 parsing code does not handle missing outer ContentInfo
7616 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7617 missing content and trigger a NULL pointer dereference on parsing.
7619 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7620 otherwise parse PKCS#7 structures from untrusted sources are
7621 affected. OpenSSL clients and servers are not affected.
7623 This issue was reported to OpenSSL by Michal Zalewski (Google).
7628 * DoS via reachable assert in SSLv2 servers fix
7630 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7631 servers that both support SSLv2 and enable export cipher suites by sending
7632 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7634 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7635 (OpenSSL development team).
7640 * Use After Free following d2i_ECPrivatekey error fix
7642 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7643 could cause a use after free condition. This, in turn, could cause a double
7644 free in several private key parsing functions (such as d2i_PrivateKey
7645 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7646 for applications that receive EC private keys from untrusted
7647 sources. This scenario is considered rare.
7649 This issue was discovered by the BoringSSL project and fixed in their
7655 * X509_to_X509_REQ NULL pointer deref fix
7657 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7658 the certificate key is invalid. This function is rarely used in practice.
7660 This issue was discovered by Brian Carpenter.
7665 * Removed the export ciphers from the DEFAULT ciphers
7669 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7671 * Build fixes for the Windows and OpenVMS platforms
7673 *Matt Caswell and Richard Levitte*
7675 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7677 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7678 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7679 dereference. This could lead to a Denial Of Service attack. Thanks to
7680 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7685 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7686 dtls1_buffer_record function under certain conditions. In particular this
7687 could occur if an attacker sent repeated DTLS records with the same
7688 sequence number but for the next epoch. The memory leak could be exploited
7689 by an attacker in a Denial of Service attack through memory exhaustion.
7690 Thanks to Chris Mueller for reporting this issue.
7695 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7696 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7697 method would be set to NULL which could later result in a NULL pointer
7698 dereference. Thanks to Frank Schmirler for reporting this issue.
7703 * Abort handshake if server key exchange message is omitted for ephemeral
7706 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7707 reporting this issue.
7712 * Remove non-export ephemeral RSA code on client and server. This code
7713 violated the TLS standard by allowing the use of temporary RSA keys in
7714 non-export ciphersuites and could be used by a server to effectively
7715 downgrade the RSA key length used to a value smaller than the server
7716 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7717 INRIA or reporting this issue.
7722 * Fixed issue where DH client certificates are accepted without verification.
7723 An OpenSSL server will accept a DH certificate for client authentication
7724 without the certificate verify message. This effectively allows a client to
7725 authenticate without the use of a private key. This only affects servers
7726 which trust a client certificate authority which issues certificates
7727 containing DH keys: these are extremely rare and hardly ever encountered.
7728 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7734 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7735 results on some platforms, including x86_64. This bug occurs at random
7736 with a very low probability, and is not known to be exploitable in any
7737 way, though its exact impact is difficult to determine. Thanks to Pieter
7738 Wuille (Blockstream) who reported this issue and also suggested an initial
7739 fix. Further analysis was conducted by the OpenSSL development team and
7740 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7741 the OpenSSL core team.
7746 * Fix various certificate fingerprint issues.
7748 By using non-DER or invalid encodings outside the signed portion of a
7749 certificate the fingerprint can be changed without breaking the signature.
7750 Although no details of the signed portion of the certificate can be changed
7751 this can cause problems with some applications: e.g. those using the
7752 certificate fingerprint for blacklists.
7754 1. Reject signatures with non zero unused bits.
7756 If the BIT STRING containing the signature has non zero unused bits reject
7757 the signature. All current signature algorithms require zero unused bits.
7759 2. Check certificate algorithm consistency.
7761 Check the AlgorithmIdentifier inside TBS matches the one in the
7762 certificate signature. NB: this will result in signature failure
7763 errors for some broken certificates.
7765 Thanks to Konrad Kraszewski from Google for reporting this issue.
7767 3. Check DSA/ECDSA signatures use DER.
7769 Reencode DSA/ECDSA signatures and compare with the original received
7770 signature. Return an error if there is a mismatch.
7772 This will reject various cases including garbage after signature
7773 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7774 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7775 (negative or with leading zeroes).
7777 Further analysis was conducted and fixes were developed by Stephen Henson
7778 of the OpenSSL core team.
7784 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7786 * Session Ticket Memory Leak.
7788 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7789 integrity of that ticket is first verified. In the event of a session
7790 ticket integrity check failing, OpenSSL will fail to free memory
7791 causing a memory leak. By sending a large number of invalid session
7792 tickets an attacker could exploit this issue in a Denial Of Service
7798 * Build option no-ssl3 is incomplete.
7800 When OpenSSL is configured with "no-ssl3" as a build option, servers
7801 could accept and complete a SSL 3.0 handshake, and clients could be
7802 configured to send them.
7805 *Akamai and the OpenSSL team*
7807 * Add support for TLS_FALLBACK_SCSV.
7808 Client applications doing fallback retries should call
7809 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7812 *Adam Langley, Bodo Moeller*
7814 * Add additional DigestInfo checks.
7816 Reencode DigestInto in DER and check against the original when
7817 verifying RSA signature: this will reject any improperly encoded
7818 DigestInfo structures.
7820 Note: this is a precautionary measure and no attacks are currently known.
7824 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7826 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7827 to a denial of service attack. A malicious server can crash the client
7828 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7829 ciphersuite and sending carefully crafted handshake messages.
7831 Thanks to Felix Gröbert (Google) for discovering and researching this
7837 * By sending carefully crafted DTLS packets an attacker could cause openssl
7838 to leak memory. This can be exploited through a Denial of Service attack.
7839 Thanks to Adam Langley for discovering and researching this issue.
7844 * An attacker can force openssl to consume large amounts of memory whilst
7845 processing DTLS handshake messages. This can be exploited through a
7846 Denial of Service attack.
7847 Thanks to Adam Langley for discovering and researching this issue.
7852 * An attacker can force an error condition which causes openssl to crash
7853 whilst processing DTLS packets due to memory being freed twice. This
7854 can be exploited through a Denial of Service attack.
7855 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7861 * If a multithreaded client connects to a malicious server using a resumed
7862 session and the server sends an ec point format extension it could write
7863 up to 255 bytes to freed memory.
7865 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7871 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7872 X509_name_oneline, X509_name_print_ex et al. to leak some information
7873 from the stack. Applications may be affected if they echo pretty printing
7874 output to the attacker.
7876 Thanks to Ivan Fratric (Google) for discovering this issue.
7879 *Emilia Käsper, and Steve Henson*
7881 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7882 for corner cases. (Certain input points at infinity could lead to
7883 bogus results, with non-infinity inputs mapped to infinity too.)
7887 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7889 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7890 handshake can force the use of weak keying material in OpenSSL
7891 SSL/TLS clients and servers.
7893 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7894 researching this issue. ([CVE-2014-0224])
7896 *KIKUCHI Masashi, Steve Henson*
7898 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7899 OpenSSL DTLS client the code can be made to recurse eventually crashing
7902 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7905 *Imre Rad, Steve Henson*
7907 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7908 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7909 client or server. This is potentially exploitable to run arbitrary
7910 code on a vulnerable client or server.
7912 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7914 *Jüri Aedla, Steve Henson*
7916 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7917 are subject to a denial of service attack.
7919 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7920 this issue. ([CVE-2014-3470])
7922 *Felix Gröbert, Ivan Fratric, Steve Henson*
7924 * Harmonize version and its documentation. -f flag is used to display
7927 *mancha <mancha1@zoho.com>*
7929 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7930 in i2d_ECPrivateKey.
7932 *mancha <mancha1@zoho.com>*
7934 * Fix some double frees. These are not thought to be exploitable.
7936 *mancha <mancha1@zoho.com>*
7938 * Fix for the attack described in the paper "Recovering OpenSSL
7939 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7940 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7941 <http://eprint.iacr.org/2014/140>
7943 Thanks to Yuval Yarom and Naomi Benger for discovering this
7944 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7946 *Yuval Yarom and Naomi Benger*
7948 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7950 * Keep original DTLS digest and encryption contexts in retransmission
7951 structures so we can use the previous session parameters if they need
7952 to be resent. ([CVE-2013-6450])
7956 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7957 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7958 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7959 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7960 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7961 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7963 *Rob Stradling, Adam Langley*
7965 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7967 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7969 This addresses the flaw in CBC record processing discovered by
7970 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7971 at: <http://www.isg.rhul.ac.uk/tls/>
7973 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7974 Security Group at Royal Holloway, University of London
7975 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7976 Emilia Käsper for the initial patch.
7979 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7981 * Return an error when checking OCSP signatures when key is NULL.
7982 This fixes a DoS attack. ([CVE-2013-0166])
7986 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7987 the right response is stapled. Also change SSL_get_certificate()
7988 so it returns the certificate actually sent.
7989 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7990 (This is a backport)
7992 *Rob Stradling <rob.stradling@comodo.com>*
7994 * Fix possible deadlock when decoding public keys.
7998 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8000 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8003 * Sanity check record length before skipping explicit IV in DTLS
8006 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8007 fuzzing as a service testing platform.
8012 * Initialise tkeylen properly when encrypting CMS messages.
8013 Thanks to Solar Designer of Openwall for reporting this issue.
8017 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8019 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8020 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8021 in CRYPTO_realloc_clean.
8023 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8024 issue and to Adam Langley <agl@chromium.org> for fixing it.
8027 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8029 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8031 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8032 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8033 content decryption and always return the same error. Note: this attack
8034 needs on average 2^20 messages so it only affects automated senders. The
8035 old behaviour can be re-enabled in the CMS code by setting the
8036 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8037 an MMA defence is not necessary.
8038 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8039 this issue. ([CVE-2012-0884])
8043 * Fix CVE-2011-4619: make sure we really are receiving a
8044 client hello before rejecting multiple SGC restarts. Thanks to
8045 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8049 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8051 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8052 Thanks to Antonio Martin, Enterprise Secure Access Research and
8053 Development, Cisco Systems, Inc. for discovering this bug and
8054 preparing a fix. ([CVE-2012-0050])
8058 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8060 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8061 of the Vaudenay padding oracle attack on CBC mode encryption
8062 which enables an efficient plaintext recovery attack against
8063 the OpenSSL implementation of DTLS. Their attack exploits timing
8064 differences arising during decryption processing. A research
8065 paper describing this attack can be found at:
8066 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8067 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8068 Security Group at Royal Holloway, University of London
8069 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8070 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8071 for preparing the fix. ([CVE-2011-4108])
8073 *Robin Seggelmann, Michael Tuexen*
8075 * Clear bytes used for block padding of SSL 3.0 records.
8078 *Adam Langley (Google)*
8080 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8081 Kadianakis <desnacked@gmail.com> for discovering this issue and
8082 Adam Langley for preparing the fix. ([CVE-2011-4619])
8084 *Adam Langley (Google)*
8086 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8088 *Andrey Kulikov <amdeich@gmail.com>*
8090 * Prevent malformed RFC3779 data triggering an assertion failure.
8091 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8092 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8094 *Rob Austein <sra@hactrn.net>*
8096 * Improved PRNG seeding for VOS.
8098 *Paul Green <Paul.Green@stratus.com>*
8100 * Fix ssl_ciph.c set-up race.
8102 *Adam Langley (Google)*
8104 * Fix spurious failures in ecdsatest.c.
8106 *Emilia Käsper (Google)*
8108 * Fix the BIO_f_buffer() implementation (which was mixing different
8109 interpretations of the `..._len` fields).
8111 *Adam Langley (Google)*
8113 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8114 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8115 threads won't reuse the same blinding coefficients.
8117 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8118 lock to call BN_BLINDING_invert_ex, and avoids one use of
8119 BN_BLINDING_update for each BN_BLINDING structure (previously,
8120 the last update always remained unused).
8122 *Emilia Käsper (Google)*
8124 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8126 *Bob Buckholz (Google)*
8128 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8130 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8131 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8133 *Kaspar Brand <ossl@velox.ch>*
8135 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8136 for multi-threaded use of ECDH. ([CVE-2011-3210])
8138 *Adam Langley (Google)*
8140 * Fix x509_name_ex_d2i memory leak on bad inputs.
8144 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8145 signature public key algorithm by using OID xref utilities instead.
8146 Before this you could only use some ECC ciphersuites with SHA1 only.
8150 * Add protection against ECDSA timing attacks as mentioned in the paper
8151 by Billy Bob Brumley and Nicola Tuveri, see:
8152 <http://eprint.iacr.org/2011/232.pdf>
8154 *Billy Bob Brumley and Nicola Tuveri*
8156 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8158 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8160 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8162 * Fix bug in string printing code: if *any* escaping is enabled we must
8163 escape the escape character (backslash) or the resulting string is
8168 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8170 * Disable code workaround for ancient and obsolete Netscape browsers
8171 and servers: an attacker can use it in a ciphersuite downgrade attack.
8172 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8176 * Fixed J-PAKE implementation error, originally discovered by
8177 Sebastien Martini, further info and confirmation from Stefan
8178 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8182 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8184 * Fix extension code to avoid race conditions which can result in a buffer
8185 overrun vulnerability: resumed sessions must not be modified as they can
8186 be shared by multiple threads. CVE-2010-3864
8190 * Fix WIN32 build system to correctly link an ENGINE directory into
8195 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8197 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8200 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8202 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8204 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8205 context. The operation can be customised via the ctrl mechanism in
8206 case ENGINEs want to include additional functionality.
8210 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8214 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8215 output hashes compatible with older versions of OpenSSL.
8217 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8219 * Fix compression algorithm handling: if resuming a session use the
8220 compression algorithm of the resumed session instead of determining
8221 it from client hello again. Don't allow server to change algorithm.
8225 * Add load_crls() function to commands tidying load_certs() too. Add option
8226 to verify utility to allow additional CRLs to be included.
8230 * Update OCSP request code to permit adding custom headers to the request:
8231 some responders need this.
8235 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8238 *Julia Lawall <julia@diku.dk>*
8240 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8241 needlessly dereferenced structures, used obsolete functions and
8242 didn't handle all updated verify codes correctly.
8246 * Disable MD2 in the default configuration.
8250 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8251 indicate the initial BIO being pushed or popped. This makes it possible
8252 to determine whether the BIO is the one explicitly called or as a result
8253 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8254 it handles reference counts correctly and doesn't zero out the I/O bio
8255 when it is not being explicitly popped. WARNING: applications which
8256 included workarounds for the old buggy behaviour will need to be modified
8257 or they could free up already freed BIOs.
8261 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8262 renaming to all platforms (within the 0.9.8 branch, this was
8263 done conditionally on Netware platforms to avoid a name clash).
8265 *Guenter <lists@gknw.net>*
8267 * Add ECDHE and PSK support to DTLS.
8269 *Michael Tuexen <tuexen@fh-muenster.de>*
8271 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8276 * Add "missing" function EVP_MD_flags() (without this the only way to
8277 retrieve a digest flags is by accessing the structure directly. Update
8278 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8279 or cipher is registered as in the "from" argument. Print out all
8280 registered digests in the dgst usage message instead of manually
8281 attempting to work them out.
8285 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8286 this allows the use of compression and extensions. Change default cipher
8287 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8288 by default unless an application cipher string requests it.
8292 * Alter match criteria in PKCS12_parse(). It used to try to use local
8293 key ids to find matching certificates and keys but some PKCS#12 files
8294 don't follow the (somewhat unwritten) rules and this strategy fails.
8295 Now just gather all certificates together and the first private key
8296 then look for the first certificate that matches the key.
8300 * Support use of registered digest and cipher names for dgst and cipher
8301 commands instead of having to add each one as a special case. So now
8308 openssl dgst -sha256 foo
8310 and this works for ENGINE based algorithms too.
8314 * Update Gost ENGINE to support parameter files.
8316 *Victor B. Wagner <vitus@cryptocom.ru>*
8318 * Support GeneralizedTime in ca utility.
8320 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8322 * Enhance the hash format used for certificate directory links. The new
8323 form uses the canonical encoding (meaning equivalent names will work
8324 even if they aren't identical) and uses SHA1 instead of MD5. This form
8325 is incompatible with the older format and as a result c_rehash should
8326 be used to rebuild symbolic links.
8330 * Make PKCS#8 the default write format for private keys, replacing the
8331 traditional format. This form is standardised, more secure and doesn't
8332 include an implicit MD5 dependency.
8336 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8337 committed to OpenSSL should pass this lot as a minimum.
8341 * Add session ticket override functionality for use by EAP-FAST.
8343 *Jouni Malinen <j@w1.fi>*
8345 * Modify HMAC functions to return a value. Since these can be implemented
8346 in an ENGINE errors can occur.
8350 * Type-checked OBJ_bsearch_ex.
8354 * Type-checked OBJ_bsearch. Also some constification necessitated
8355 by type-checking. Still to come: TXT_DB, bsearch(?),
8356 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8361 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8362 seconds to a tm structure directly, instead of going through OS
8363 specific date routines. This avoids any issues with OS routines such
8364 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8365 and X509_time_adj_ex() to cover the extended range. The existing
8366 X509_time_adj() is still usable and will no longer have any date issues.
8370 * Delta CRL support. New use deltas option which will attempt to locate
8371 and search any appropriate delta CRLs available.
8373 This work was sponsored by Google.
8377 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8378 code and add additional score elements. Validate alternate CRL paths
8379 as part of the CRL checking and indicate a new error "CRL path validation
8380 error" in this case. Applications wanting additional details can use
8381 the verify callback and check the new "parent" field. If this is not
8382 NULL CRL path validation is taking place. Existing applications won't
8383 see this because it requires extended CRL support which is off by
8386 This work was sponsored by Google.
8390 * Support for freshest CRL extension.
8392 This work was sponsored by Google.
8396 * Initial indirect CRL support. Currently only supported in the CRLs
8397 passed directly and not via lookup. Process certificate issuer
8398 CRL entry extension and lookup CRL entries by bother issuer name
8399 and serial number. Check and process CRL issuer entry in IDP extension.
8401 This work was sponsored by Google.
8405 * Add support for distinct certificate and CRL paths. The CRL issuer
8406 certificate is validated separately in this case. Only enabled if
8407 an extended CRL support flag is set: this flag will enable additional
8408 CRL functionality in future.
8410 This work was sponsored by Google.
8414 * Add support for policy mappings extension.
8416 This work was sponsored by Google.
8420 * Fixes to pathlength constraint, self issued certificate handling,
8421 policy processing to align with RFC3280 and PKITS tests.
8423 This work was sponsored by Google.
8427 * Support for name constraints certificate extension. DN, email, DNS
8428 and URI types are currently supported.
8430 This work was sponsored by Google.
8434 * To cater for systems that provide a pointer-based thread ID rather
8435 than numeric, deprecate the current numeric thread ID mechanism and
8436 replace it with a structure and associated callback type. This
8437 mechanism allows a numeric "hash" to be extracted from a thread ID in
8438 either case, and on platforms where pointers are larger than 'long',
8439 mixing is done to help ensure the numeric 'hash' is usable even if it
8440 can't be guaranteed unique. The default mechanism is to use "&errno"
8441 as a pointer-based thread ID to distinguish between threads.
8443 Applications that want to provide their own thread IDs should now use
8444 CRYPTO_THREADID_set_callback() to register a callback that will call
8445 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8447 Note that ERR_remove_state() is now deprecated, because it is tied
8448 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8449 to free the current thread's error state should be replaced by
8450 ERR_remove_thread_state(NULL).
8452 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8453 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8454 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8455 application was previously providing a numeric thread callback that
8456 was inappropriate for distinguishing threads, then uniqueness might
8457 have been obtained with &errno that happened immediately in the
8458 intermediate development versions of OpenSSL; this is no longer the
8459 case, the numeric thread callback will now override the automatic use
8462 *Geoff Thorpe, with help from Bodo Moeller*
8464 * Initial support for different CRL issuing certificates. This covers a
8465 simple case where the self issued certificates in the chain exist and
8466 the real CRL issuer is higher in the existing chain.
8468 This work was sponsored by Google.
8472 * Removed effectively defunct crypto/store from the build.
8476 * Revamp of STACK to provide stronger type-checking. Still to come:
8477 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8478 ASN1_STRING, CONF_VALUE.
8482 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8483 RAM on SSL connections. This option can save about 34k per idle SSL.
8487 * Revamp of LHASH to provide stronger type-checking. Still to come:
8488 STACK, TXT_DB, bsearch, qsort.
8492 * Initial support for Cryptographic Message Syntax (aka CMS) based
8493 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8494 support for data, signedData, compressedData, digestedData and
8495 encryptedData, envelopedData types included. Scripts to check against
8496 RFC4134 examples draft and interop and consistency checks of many
8497 content types and variants.
8501 * Add options to enc utility to support use of zlib compression BIO.
8505 * Extend mk1mf to support importing of options and assembly language
8506 files from Configure script, currently only included in VC-WIN32.
8507 The assembly language rules can now optionally generate the source
8508 files from the associated perl scripts.
8512 * Implement remaining functionality needed to support GOST ciphersuites.
8513 Interop testing has been performed using CryptoPro implementations.
8515 *Victor B. Wagner <vitus@cryptocom.ru>*
8517 * s390x assembler pack.
8521 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8526 * Implement Opaque PRF Input TLS extension as specified in
8527 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8528 official specification yet and no extension type assignment by
8529 IANA exists, this extension (for now) will have to be explicitly
8530 enabled when building OpenSSL by providing the extension number
8531 to use. For example, specify an option
8533 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8535 to the "config" or "Configure" script to enable the extension,
8536 assuming extension number 0x9527 (which is a completely arbitrary
8537 and unofficial assignment based on the MD5 hash of the Internet
8538 Draft). Note that by doing so, you potentially lose
8539 interoperability with other TLS implementations since these might
8540 be using the same extension number for other purposes.
8542 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8543 opaque PRF input value to use in the handshake. This will create
8544 an internal copy of the length-'len' string at 'src', and will
8545 return non-zero for success.
8547 To get more control and flexibility, provide a callback function
8550 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8551 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8555 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8558 Callback function 'cb' will be called in handshakes, and is
8559 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8560 Argument 'arg' is for application purposes (the value as given to
8561 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8562 be provided to the callback function). The callback function
8563 has to return non-zero to report success: usually 1 to use opaque
8564 PRF input just if possible, or 2 to enforce use of the opaque PRF
8565 input. In the latter case, the library will abort the handshake
8566 if opaque PRF input is not successfully negotiated.
8568 Arguments 'peerinput' and 'len' given to the callback function
8569 will always be NULL and 0 in the case of a client. A server will
8570 see the client's opaque PRF input through these variables if
8571 available (NULL and 0 otherwise). Note that if the server
8572 provides an opaque PRF input, the length must be the same as the
8573 length of the client's opaque PRF input.
8575 Note that the callback function will only be called when creating
8576 a new session (session resumption can resume whatever was
8577 previously negotiated), and will not be called in SSL 2.0
8578 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8579 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8580 for applications that need to enforce opaque PRF input.
8584 * Update ssl code to support digests other than SHA1+MD5 for handshake
8587 *Victor B. Wagner <vitus@cryptocom.ru>*
8589 * Add RFC4507 support to OpenSSL. This includes the corrections in
8590 RFC4507bis. The encrypted ticket format is an encrypted encoded
8591 SSL_SESSION structure, that way new session features are automatically
8594 If a client application caches session in an SSL_SESSION structure
8595 support is transparent because tickets are now stored in the encoded
8598 The SSL_CTX structure automatically generates keys for ticket
8599 protection in servers so again support should be possible
8600 with no application modification.
8602 If a client or server wishes to disable RFC4507 support then the option
8603 SSL_OP_NO_TICKET can be set.
8605 Add a TLS extension debugging callback to allow the contents of any client
8606 or server extensions to be examined.
8608 This work was sponsored by Google.
8612 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8613 OpenSSL should now compile cleanly on gcc 4.2
8615 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8617 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8618 support including streaming MAC support: this is required for GOST
8619 ciphersuite support.
8621 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8623 * Add option -stream to use PKCS#7 streaming in smime utility. New
8624 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8625 to output in BER and PEM format.
8629 * Experimental support for use of HMAC via EVP_PKEY interface. This
8630 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8631 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8632 ENGINE support for HMAC keys which are unextractable. New -mac and
8633 -macopt options to dgst utility.
8637 * New option -sigopt to dgst utility. Update dgst to use
8638 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8639 alternative signing parameters such as X9.31 or PSS in the dgst
8644 * Change ssl_cipher_apply_rule(), the internal function that does
8645 the work each time a ciphersuite string requests enabling
8646 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8647 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8648 the order of disabled ciphersuites such that those ciphersuites
8649 that most recently went from enabled to disabled not only stay
8650 in order with respect to each other, but also have higher priority
8651 than other disabled ciphersuites the next time ciphersuites are
8654 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8655 the same ciphersuites as with "HIGH" alone, but in a specific
8656 order where the PSK ciphersuites come first (since they are the
8657 most recently disabled ciphersuites when "HIGH" is parsed).
8659 Also, change ssl_create_cipher_list() (using this new
8660 functionality) such that between otherwise identical
8661 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8666 * Change ssl_create_cipher_list() so that it automatically
8667 arranges the ciphersuites in reasonable order before starting
8668 to process the rule string. Thus, the definition for "DEFAULT"
8669 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8670 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8671 This makes it much easier to arrive at a reasonable default order
8672 in applications for which anonymous ciphers are OK (meaning
8673 that you can't actually use DEFAULT).
8675 *Bodo Moeller; suggested by Victor Duchovni*
8677 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8678 processing) into multiple integers instead of setting
8679 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8680 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8681 (These masks as well as the individual bit definitions are hidden
8682 away into the non-exported interface ssl/ssl_locl.h, so this
8683 change to the definition of the SSL_CIPHER structure shouldn't
8684 affect applications.) This give us more bits for each of these
8685 categories, so there is no longer a need to coagulate AES128 and
8686 AES256 into a single algorithm bit, and to coagulate Camellia128
8687 and Camellia256 into a single algorithm bit, which has led to all
8690 Thus, among other things, the kludge introduced in 0.9.7m and
8691 0.9.8e for masking out AES256 independently of AES128 or masking
8692 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8694 With the change, we also introduce new ciphersuite aliases that
8695 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8700 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8701 Use the leftmost N bytes of the signature input if the input is
8702 larger than the prime q (with N being the size in bytes of q).
8706 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8707 it yet and it is largely untested.
8711 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8715 * Initial incomplete changes to avoid need for function casts in OpenSSL
8716 some compilers (gcc 4.2 and later) reject their use. Safestack is
8717 reimplemented. Update ASN1 to avoid use of legacy functions.
8721 * Win32/64 targets are linked with Winsock2.
8725 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8726 to external functions. This can be used to increase CRL handling
8727 efficiency especially when CRLs are very large by (for example) storing
8728 the CRL revoked certificates in a database.
8732 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8733 new CRLs added to a directory can be used. New command line option
8734 -verify_return_error to s_client and s_server. This causes real errors
8735 to be returned by the verify callback instead of carrying on no matter
8736 what. This reflects the way a "real world" verify callback would behave.
8740 * GOST engine, supporting several GOST algorithms and public key formats.
8741 Kindly donated by Cryptocom.
8745 * Partial support for Issuing Distribution Point CRL extension. CRLs
8746 partitioned by DP are handled but no indirect CRL or reason partitioning
8747 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8748 selected via a scoring technique which handles IDP and AKID in CRLs.
8752 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8753 will ultimately be used for all verify operations: this will remove the
8754 X509_STORE dependency on certificate verification and allow alternative
8755 lookup methods. X509_STORE based implementations of these two callbacks.
8759 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8760 Modify get_crl() to find a valid (unexpired) CRL if possible.
8764 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8765 this would be called X509_CRL_cmp() but that name is already used by
8766 a function that just compares CRL issuer names. Cache several CRL
8767 extensions in X509_CRL structure and cache CRLDP in X509.
8771 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8772 this maps equivalent X509_NAME structures into a consistent structure.
8773 Name comparison can then be performed rapidly using memcmp().
8777 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8782 * Allow digests to supply their own micalg string for S/MIME type using
8783 the ctrl EVP_MD_CTRL_MICALG.
8787 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8788 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8789 ctrl. It can then customise the structure before and/or after signing
8794 * New function OBJ_add_sigid() to allow application defined signature OIDs
8795 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8796 to free up any added signature OIDs.
8800 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8801 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8802 digest and cipher tables. New options added to openssl utility:
8803 list-message-digest-algorithms and list-cipher-algorithms.
8807 * Change the array representation of binary polynomials: the list
8808 of degrees of non-zero coefficients is now terminated with -1.
8809 Previously it was terminated with 0, which was also part of the
8810 value; thus, the array representation was not applicable to
8811 polynomials where t^0 has coefficient zero. This change makes
8812 the array representation useful in a more general context.
8816 * Various modifications and fixes to SSL/TLS cipher string
8817 handling. For ECC, the code now distinguishes between fixed ECDH
8818 with RSA certificates on the one hand and with ECDSA certificates
8819 on the other hand, since these are separate ciphersuites. The
8820 unused code for Fortezza ciphersuites has been removed.
8822 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8823 (not "ECDHE"). For consistency with the code for DH
8824 certificates, use of ECDH certificates is now considered ECDH
8825 authentication, not RSA or ECDSA authentication (the latter is
8826 merely the CA's signing algorithm and not actively used in the
8829 The temporary ciphersuite alias "ECCdraft" is no longer
8830 available, and ECC ciphersuites are no longer excluded from "ALL"
8831 and "DEFAULT". The following aliases now exist for RFC 4492
8832 ciphersuites, most of these by analogy with the DH case:
8834 kECDHr - ECDH cert, signed with RSA
8835 kECDHe - ECDH cert, signed with ECDSA
8836 kECDH - ECDH cert (signed with either RSA or ECDSA)
8837 kEECDH - ephemeral ECDH
8838 ECDH - ECDH cert or ephemeral ECDH
8844 AECDH - anonymous ECDH
8845 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8849 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8850 Use correct micalg parameters depending on digest(s) in signed message.
8854 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8855 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8859 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8860 an engine to register a method. Add ENGINE lookups for methods and
8861 functional reference processing.
8865 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8866 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8871 * New -resign option to smime utility. This adds one or more signers
8872 to an existing PKCS#7 signedData structure. Also -md option to use an
8873 alternative message digest algorithm for signing.
8877 * Tidy up PKCS#7 routines and add new functions to make it easier to
8878 create PKCS7 structures containing multiple signers. Update smime
8879 application to support multiple signers.
8883 * New -macalg option to pkcs12 utility to allow setting of an alternative
8888 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8889 Reorganize PBE internals to lookup from a static table using NIDs,
8890 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8891 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8892 PRF which will be automatically used with PBES2.
8896 * Replace the algorithm specific calls to generate keys in "req" with the
8901 * Update PKCS#7 enveloped data routines to use new API. This is now
8902 supported by any public key method supporting the encrypt operation. A
8903 ctrl is added to allow the public key algorithm to examine or modify
8904 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8909 * Add a ctrl to asn1 method to allow a public key algorithm to express
8910 a default digest type to use. In most cases this will be SHA1 but some
8911 algorithms (such as GOST) need to specify an alternative digest. The
8912 return value indicates how strong the preference is 1 means optional and
8913 2 is mandatory (that is it is the only supported type). Modify
8914 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8915 use the default md. Update openssl utilities to use the default digest
8916 type for signing if it is not explicitly indicated.
8920 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8921 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8922 signing method from the key type. This effectively removes the link
8923 between digests and public key types.
8927 * Add an OID cross reference table and utility functions. Its purpose is to
8928 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8929 rsaEncryption. This will allow some of the algorithm specific hackery
8930 needed to use the correct OID to be removed.
8934 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8935 structures for PKCS7_sign(). They are now set up by the relevant public
8940 * Add provisional EC pkey method with support for ECDSA and ECDH.
8944 * Add support for key derivation (agreement) in the API, DH method and
8949 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8950 public and private key formats. As a side effect these add additional
8951 command line functionality not previously available: DSA signatures can be
8952 generated and verified using pkeyutl and DH key support and generation in
8959 *Oliver Tappe <zooey@hirschkaefer.de>*
8961 * New make target "install_html_docs" installs HTML renditions of the
8964 *Oliver Tappe <zooey@hirschkaefer.de>*
8966 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8967 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8968 support key and parameter generation and add initial key generation
8969 functionality for RSA.
8973 * Add functions for main EVP_PKEY_method operations. The undocumented
8974 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8975 `EVP_PKEY_{encrypt,decrypt}_old`.
8979 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8980 key API, doesn't do much yet.
8984 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8985 public key algorithms. New option to openssl utility:
8986 "list-public-key-algorithms" to print out info.
8990 * Implement the Supported Elliptic Curves Extension for
8991 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8995 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8996 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9000 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9001 utilities such as rsa, dsa, dsaparam etc except they process any key
9006 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9007 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9008 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9013 * Initial support for pluggable public key ASN1.
9014 De-spaghettify the public key ASN1 handling. Move public and private
9015 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9016 algorithm specific handling to a single module within the relevant
9017 algorithm directory. Add functions to allow (near) opaque processing
9018 of public and private key structures.
9022 * Implement the Supported Point Formats Extension for
9023 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9027 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9028 for the psk identity [hint] and the psk callback functions to the
9029 SSL_SESSION, SSL and SSL_CTX structure.
9032 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9036 SSL_CTX_use_psk_identity_hint
9037 SSL_get_psk_identity_hint
9038 SSL_get_psk_identity
9039 SSL_use_psk_identity_hint
9041 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9043 * Add RFC 3161 compliant time stamp request creation, response generation
9044 and response verification functionality.
9046 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9048 * Add initial support for TLS extensions, specifically for the server_name
9049 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9050 have new members for a host name. The SSL data structure has an
9051 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9052 stored in that context to allow for session resumption, even after the
9053 SSL has been switched to a new SSL_CTX in reaction to a client's
9054 server_name extension.
9056 New functions (subject to change):
9058 SSL_get_servername()
9059 SSL_get_servername_type()
9062 New CTRL codes and macros (subject to change):
9064 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9065 - SSL_CTX_set_tlsext_servername_callback()
9066 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9067 - SSL_CTX_set_tlsext_servername_arg()
9068 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9070 openssl s_client has a new '-servername ...' option.
9072 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9073 '-key2 ...', '-servername_fatal' (subject to change). This allows
9074 testing the HostName extension for a specific single host name ('-cert'
9075 and '-key' remain fallbacks for handshakes without HostName
9076 negotiation). If the unrecognized_name alert has to be sent, this by
9077 default is a warning; it becomes fatal with the '-servername_fatal'
9080 *Peter Sylvester, Remy Allais, Christophe Renou*
9082 * Whirlpool hash implementation is added.
9086 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9087 bn(64,32). Because of instruction set limitations it doesn't have
9088 any negative impact on performance. This was done mostly in order
9089 to make it possible to share assembler modules, such as bn_mul_mont
9090 implementations, between 32- and 64-bit builds without hassle.
9094 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9095 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9100 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9101 dedicated Montgomery multiplication procedure, is introduced.
9102 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9103 "64-bit" performance on certain 32-bit targets.
9107 * New option SSL_OP_NO_COMP to disable use of compression selectively
9108 in SSL structures. New SSL ctrl to set maximum send fragment size.
9109 Save memory by setting the I/O buffer sizes dynamically instead of
9110 using the maximum available value.
9114 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9115 in addition to the text details.
9119 * Very, very preliminary EXPERIMENTAL support for printing of general
9120 ASN1 structures. This currently produces rather ugly output and doesn't
9121 handle several customised structures at all.
9125 * Integrated support for PVK file format and some related formats such
9126 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9127 these in the 'rsa' and 'dsa' utilities.
9131 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9135 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9136 place for the (very old) "NETSCAPE" format certificates which are now
9137 handled using new ASN1 code equivalents.
9141 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9142 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9143 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9147 * Modify CRL distribution points extension code to print out previously
9148 unsupported fields. Enhance extension setting code to allow setting of
9153 * Add print and set support for Issuing Distribution Point CRL extension.
9157 * Change 'Configure' script to enable Camellia by default.
9164 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9166 * When rejecting SSL/TLS records due to an incorrect version number, never
9167 update s->server with a new major version number. As of
9168 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9169 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9170 the previous behavior could result in a read attempt at NULL when
9171 receiving specific incorrect SSL/TLS records once record payload
9172 protection is active. ([CVE-2010-0740])
9174 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9176 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9177 could be crashed if the relevant tables were not present (e.g. chrooted).
9179 *Tomas Hoger <thoger@redhat.com>*
9181 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9183 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9185 *Martin Olsson, Neel Mehta*
9187 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9188 accommodate for stack sorting, always a write lock!).
9192 * On some versions of WIN32 Heap32Next is very slow. This can cause
9193 excessive delays in the RAND_poll(): over a minute. As a workaround
9194 include a time check in the inner Heap32Next loop too.
9198 * The code that handled flushing of data in SSL/TLS originally used the
9199 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9200 the problem outlined in PR#1949. The fix suggested there however can
9201 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9202 of Apache). So instead simplify the code to flush unconditionally.
9203 This should be fine since flushing with no data to flush is a no op.
9207 * Handle TLS versions 2.0 and later properly and correctly use the
9208 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9209 off ancient servers have a habit of sticking around for a while...
9213 * Modify compression code so it frees up structures without using the
9214 ex_data callbacks. This works around a problem where some applications
9215 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9216 restarting) then use compression (e.g. SSL with compression) later.
9217 This results in significant per-connection memory leaks and
9218 has caused some security issues including CVE-2008-1678 and
9223 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9224 change when encrypting or decrypting.
9228 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9229 connect and renegotiate with servers which do not support RI.
9230 Until RI is more widely deployed this option is enabled by default.
9234 * Add "missing" ssl ctrls to clear options and mode.
9238 * If client attempts to renegotiate and doesn't support RI respond with
9239 a no_renegotiation alert as required by RFC5746. Some renegotiating
9240 TLS clients will continue a connection gracefully when they receive
9241 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9242 waiting for a server hello which it will never receive. Now we treat a
9243 received no_renegotiation alert as a fatal error. This is because
9244 applications requesting a renegotiation might well expect it to succeed
9245 and would have no code in place to handle the server denying it so the
9246 only safe thing to do is to terminate the connection.
9250 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9251 peer supports secure renegotiation and 0 otherwise. Print out peer
9252 renegotiation support in s_client/s_server.
9256 * Replace the highly broken and deprecated SPKAC certification method with
9257 the updated NID creation version. This should correctly handle UTF8.
9261 * Implement RFC5746. Re-enable renegotiation but require the extension
9262 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9263 turns out to be a bad idea. It has been replaced by
9264 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9265 SSL_CTX_set_options(). This is really not recommended unless you
9266 know what you are doing.
9268 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9270 * Fixes to stateless session resumption handling. Use initial_ctx when
9271 issuing and attempting to decrypt tickets in case it has changed during
9272 servername handling. Use a non-zero length session ID when attempting
9273 stateless session resumption: this makes it possible to determine if
9274 a resumption has occurred immediately after receiving server hello
9275 (several places in OpenSSL subtly assume this) instead of later in
9280 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9281 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9282 fixes for a few places where the return code is not checked
9285 *Julia Lawall <julia@diku.dk>*
9287 * Add --strict-warnings option to Configure script to include devteam
9288 warnings in other configurations.
9292 * Add support for --libdir option and LIBDIR variable in makefiles. This
9293 makes it possible to install openssl libraries in locations which
9294 have names other than "lib", for example "/usr/lib64" which some
9297 *Steve Henson, based on patch from Jeremy Utley*
9299 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9300 X690 8.9.12 and can produce some misleading textual output of OIDs.
9302 *Steve Henson, reported by Dan Kaminsky*
9304 * Delete MD2 from algorithm tables. This follows the recommendation in
9305 several standards that it is not used in new applications due to
9306 several cryptographic weaknesses. For binary compatibility reasons
9307 the MD2 API is still compiled in by default.
9311 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9316 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9317 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9320 *Guenter <lists@gknw.net>*
9322 * Fix the server certificate chain building code to use X509_verify_cert(),
9323 it used to have an ad-hoc builder which was unable to cope with anything
9324 other than a simple chain.
9326 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9328 * Don't check self signed certificate signatures in X509_verify_cert()
9329 by default (a flag can override this): it just wastes time without
9330 adding any security. As a useful side effect self signed root CAs
9331 with non-FIPS digests are now usable in FIPS mode.
9335 * In dtls1_process_out_of_seq_message() the check if the current message
9336 is already buffered was missing. For every new message was memory
9337 allocated, allowing an attacker to perform an denial of service attack
9338 with sending out of seq handshake messages until there is no memory
9339 left. Additionally every future message was buffered, even if the
9340 sequence number made no sense and would be part of another handshake.
9341 So only messages with sequence numbers less than 10 in advance will be
9342 buffered. ([CVE-2009-1378])
9344 *Robin Seggelmann, discovered by Daniel Mentz*
9346 * Records are buffered if they arrive with a future epoch to be
9347 processed after finishing the corresponding handshake. There is
9348 currently no limitation to this buffer allowing an attacker to perform
9349 a DOS attack with sending records with future epochs until there is no
9350 memory left. This patch adds the pqueue_size() function to determine
9351 the size of a buffer and limits the record buffer to 100 entries.
9354 *Robin Seggelmann, discovered by Daniel Mentz*
9356 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9357 parent structure is freed. ([CVE-2009-1379])
9361 * Handle non-blocking I/O properly in SSL_shutdown() call.
9363 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9365 * Add `2.5.4.*` OIDs
9367 *Ilya O. <vrghost@gmail.com>*
9369 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9371 * Disable renegotiation completely - this fixes a severe security
9372 problem ([CVE-2009-3555]) at the cost of breaking all
9373 renegotiation. Renegotiation can be re-enabled by setting
9374 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9375 run-time. This is really not recommended unless you know what
9380 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9382 * Don't set val to NULL when freeing up structures, it is freed up by
9383 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9384 zeroing past the valid field. ([CVE-2009-0789])
9386 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9388 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9389 checked correctly. This would allow some invalid signed attributes to
9390 appear to verify correctly. ([CVE-2009-0591])
9392 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9394 * Reject UniversalString and BMPString types with invalid lengths. This
9395 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9396 a legal length. ([CVE-2009-0590])
9400 * Set S/MIME signing as the default purpose rather than setting it
9401 unconditionally. This allows applications to override it at the store
9406 * Permit restricted recursion of ASN1 strings. This is needed in practice
9407 to handle some structures.
9411 * Improve efficiency of mem_gets: don't search whole buffer each time
9414 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9416 * New -hex option for openssl rand.
9420 * Print out UTF8String and NumericString when parsing ASN1.
9424 * Support NumericString type for name components.
9428 * Allow CC in the environment to override the automatically chosen
9429 compiler. Note that nothing is done to ensure flags work with the
9434 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9436 * Properly check EVP_VerifyFinal() and similar return values
9439 *Ben Laurie, Bodo Moeller, Google Security Team*
9441 * Enable TLS extensions by default.
9445 * Allow the CHIL engine to be loaded, whether the application is
9446 multithreaded or not. (This does not release the developer from the
9447 obligation to set up the dynamic locking callbacks.)
9449 *Sander Temme <sander@temme.net>*
9451 * Use correct exit code if there is an error in dgst command.
9453 *Steve Henson; problem pointed out by Roland Dirlewanger*
9455 * Tweak Configure so that you need to say "experimental-jpake" to enable
9456 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9460 * Add experimental JPAKE support, including demo authentication in
9461 s_client and s_server.
9465 * Set the comparison function in v3_addr_canonize().
9467 *Rob Austein <sra@hactrn.net>*
9469 * Add support for XMPP STARTTLS in s_client.
9471 *Philip Paeps <philip@freebsd.org>*
9473 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9474 to ensure that even with this option, only ciphersuites in the
9475 server's preference list will be accepted. (Note that the option
9476 applies only when resuming a session, so the earlier behavior was
9477 just about the algorithm choice for symmetric cryptography.)
9481 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9483 * Fix NULL pointer dereference if a DTLS server received
9484 ChangeCipherSpec as first record ([CVE-2009-1386]).
9488 * Fix a state transition in s3_srvr.c and d1_srvr.c
9489 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9493 * The fix in 0.9.8c that supposedly got rid of unsafe
9494 double-checked locking was incomplete for RSA blinding,
9495 addressing just one layer of what turns out to have been
9496 doubly unsafe triple-checked locking.
9498 So now fix this for real by retiring the MONT_HELPER macro
9499 in crypto/rsa/rsa_eay.c.
9501 *Bodo Moeller; problem pointed out by Marius Schilder*
9503 * Various precautionary measures:
9505 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9507 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9508 (NB: This would require knowledge of the secret session ticket key
9509 to exploit, in which case you'd be SOL either way.)
9511 - Change bn_nist.c so that it will properly handle input BIGNUMs
9512 outside the expected range.
9514 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9517 *Neel Mehta, Bodo Moeller*
9519 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9520 the load fails. Useful for distros.
9522 *Ben Laurie and the FreeBSD team*
9524 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9528 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9532 * Expand ENGINE to support engine supplied SSL client certificate functions.
9534 This work was sponsored by Logica.
9538 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9539 keystores. Support for SSL/TLS client authentication too.
9540 Not compiled unless enable-capieng specified to Configure.
9542 This work was sponsored by Logica.
9546 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9547 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9548 attribute creation routines such as certificate requests and PKCS#12
9553 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9555 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9556 handshake which could lead to a client crash as found using the
9557 Codenomicon TLS test suite ([CVE-2008-1672])
9559 *Steve Henson, Mark Cox*
9561 * Fix double free in TLS server name extensions which could lead to
9562 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9566 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9568 Clear the error queue to ensure that error entries left from
9569 older function calls do not interfere with the correct operation.
9571 *Lutz Jaenicke, Erik de Castro Lopo*
9573 * Remove root CA certificates of commercial CAs:
9575 The OpenSSL project does not recommend any specific CA and does not
9576 have any policy with respect to including or excluding any CA.
9577 Therefore it does not make any sense to ship an arbitrary selection
9578 of root CA certificates with the OpenSSL software.
9582 * RSA OAEP patches to fix two separate invalid memory reads.
9583 The first one involves inputs when 'lzero' is greater than
9584 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9585 before the beginning of from). The second one involves inputs where
9586 the 'db' section contains nothing but zeroes (there is a one-byte
9587 invalid read after the end of 'db').
9589 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9591 * Partial backport from 0.9.9-dev:
9593 Introduce bn_mul_mont (dedicated Montgomery multiplication
9594 procedure) as a candidate for BIGNUM assembler implementation.
9595 While 0.9.9-dev uses assembler for various architectures, only
9596 x86_64 is available by default here in the 0.9.8 branch, and
9597 32-bit x86 is available through a compile-time setting.
9599 To try the 32-bit x86 assembler implementation, use Configure
9600 option "enable-montasm" (which exists only for this backport).
9602 As "enable-montasm" for 32-bit x86 disclaims code stability
9603 anyway, in this constellation we activate additional code
9604 backported from 0.9.9-dev for further performance improvements,
9605 namely BN_from_montgomery_word. (To enable this otherwise,
9606 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9608 *Andy Polyakov (backport partially by Bodo Moeller)*
9610 * Add TLS session ticket callback. This allows an application to set
9611 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9612 values. This is useful for key rollover for example where several key
9613 sets may exist with different names.
9617 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9618 This was broken until now in 0.9.8 releases, such that the only way
9619 a registered ENGINE could be used (assuming it initialises
9620 successfully on the host) was to explicitly set it as the default
9621 for the relevant algorithms. This is in contradiction with 0.9.7
9622 behaviour and the documentation. With this fix, when an ENGINE is
9623 registered into a given algorithm's table of implementations, the
9624 'uptodate' flag is reset so that auto-discovery will be used next
9625 time a new context for that algorithm attempts to select an
9628 *Ian Lister (tweaked by Geoff Thorpe)*
9630 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9631 implementation in the following ways:
9633 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9636 Lack of BER streaming support means one pass streaming processing is
9637 only supported if data is detached: setting the streaming flag is
9638 ignored for embedded content.
9640 CMS support is disabled by default and must be explicitly enabled
9641 with the enable-cms configuration option.
9645 * Update the GMP engine glue to do direct copies between BIGNUM and
9646 mpz_t when openssl and GMP use the same limb size. Otherwise the
9647 existing "conversion via a text string export" trick is still used.
9649 *Paul Sheer <paulsheer@gmail.com>*
9651 * Zlib compression BIO. This is a filter BIO which compressed and
9652 uncompresses any data passed through it.
9656 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9657 RFC3394 compatible AES key wrapping.
9661 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9662 sets string data without copying. X509_ALGOR_set0() and
9663 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9664 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9665 from an X509_ATTRIBUTE structure optionally checking it occurs only
9666 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9671 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9672 to get the expected BN_FLG_CONSTTIME behavior.
9674 *Bodo Moeller (Google)*
9678 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9679 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9680 - added some more tests to do_tests.pl
9681 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9682 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9683 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9684 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9685 - various changes to netware.pl to enable gcc-cross builds on Win32
9687 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9688 - various changes to fix missing prototype warnings
9689 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9690 - added AES, WHIRLPOOL and CPUID assembler code to build files
9691 - added missing AES assembler make rules to mk1mf.pl
9692 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9694 *Guenter Knauf <eflash@gmx.net>*
9696 * Implement certificate status request TLS extension defined in RFC3546.
9697 A client can set the appropriate parameters and receive the encoded
9698 OCSP response via a callback. A server can query the supplied parameters
9699 and set the encoded OCSP response in the callback. Add simplified examples
9700 to s_client and s_server.
9704 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9707 + Binary incompatibility of ssl_ctx_st structure
9708 + DTLS interoperation with non-compliant servers
9709 + Don't call get_session_cb() without proposed session
9710 + Fix ia64 assembler code
9712 *Andy Polyakov, Steve Henson*
9714 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9716 * DTLS Handshake overhaul. There were longstanding issues with
9717 OpenSSL DTLS implementation, which were making it impossible for
9718 RFC 4347 compliant client to communicate with OpenSSL server.
9719 Unfortunately just fixing these incompatibilities would "cut off"
9720 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9721 server keeps tolerating non RFC compliant syntax. The opposite is
9722 not true, 0.9.8f client can not communicate with earlier server.
9723 This update even addresses CVE-2007-4995.
9727 * Changes to avoid need for function casts in OpenSSL: some compilers
9728 (gcc 4.2 and later) reject their use.
9729 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9732 * Add RFC4507 support to OpenSSL. This includes the corrections in
9733 RFC4507bis. The encrypted ticket format is an encrypted encoded
9734 SSL_SESSION structure, that way new session features are automatically
9737 If a client application caches session in an SSL_SESSION structure
9738 support is transparent because tickets are now stored in the encoded
9741 The SSL_CTX structure automatically generates keys for ticket
9742 protection in servers so again support should be possible
9743 with no application modification.
9745 If a client or server wishes to disable RFC4507 support then the option
9746 SSL_OP_NO_TICKET can be set.
9748 Add a TLS extension debugging callback to allow the contents of any client
9749 or server extensions to be examined.
9751 This work was sponsored by Google.
9755 * Add initial support for TLS extensions, specifically for the server_name
9756 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9757 have new members for a host name. The SSL data structure has an
9758 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9759 stored in that context to allow for session resumption, even after the
9760 SSL has been switched to a new SSL_CTX in reaction to a client's
9761 server_name extension.
9763 New functions (subject to change):
9765 SSL_get_servername()
9766 SSL_get_servername_type()
9769 New CTRL codes and macros (subject to change):
9771 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9772 - SSL_CTX_set_tlsext_servername_callback()
9773 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9774 - SSL_CTX_set_tlsext_servername_arg()
9775 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9777 openssl s_client has a new '-servername ...' option.
9779 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9780 '-key2 ...', '-servername_fatal' (subject to change). This allows
9781 testing the HostName extension for a specific single host name ('-cert'
9782 and '-key' remain fallbacks for handshakes without HostName
9783 negotiation). If the unrecognized_name alert has to be sent, this by
9784 default is a warning; it becomes fatal with the '-servername_fatal'
9787 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9789 * Add AES and SSE2 assembly language support to VC++ build.
9793 * Mitigate attack on final subtraction in Montgomery reduction.
9797 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9798 (which previously caused an internal error).
9802 * Squeeze another 10% out of IGE mode when in != out.
9806 * AES IGE mode speedup.
9808 *Dean Gaudet (Google)*
9810 * Add the Korean symmetric 128-bit cipher SEED (see
9811 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9812 add SEED ciphersuites from RFC 4162:
9814 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9815 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9816 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9817 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9819 To minimize changes between patchlevels in the OpenSSL 0.9.8
9820 series, SEED remains excluded from compilation unless OpenSSL
9821 is configured with 'enable-seed'.
9823 *KISA, Bodo Moeller*
9825 * Mitigate branch prediction attacks, which can be practical if a
9826 single processor is shared, allowing a spy process to extract
9827 information. For detailed background information, see
9828 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9829 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9830 and Necessary Software Countermeasures"). The core of the change
9831 are new versions BN_div_no_branch() and
9832 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9833 respectively, which are slower, but avoid the security-relevant
9834 conditional branches. These are automatically called by BN_div()
9835 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9836 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9837 remove a conditional branch.
9839 BN_FLG_CONSTTIME is the new name for the previous
9840 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9841 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9842 in the exponent causes BN_mod_exp_mont() to use the alternative
9843 implementation in BN_mod_exp_mont_consttime().) The old name
9844 remains as a deprecated alias.
9846 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9847 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9848 constant-time implementations for more than just exponentiation.
9849 Here too the old name is kept as a deprecated alias.
9851 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9852 the BN_BLINDING structure gets an independent copy of the
9853 modulus. This means that the previous `BIGNUM *m` argument to
9854 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9855 essentially becomes `const BIGNUM *m`, although we can't actually
9856 change this in the header file before 0.9.9. It allows
9857 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9858 enable BN_FLG_CONSTTIME.
9860 *Matthew D Wood (Intel Corp)*
9862 * In the SSL/TLS server implementation, be strict about session ID
9863 context matching (which matters if an application uses a single
9864 external cache for different purposes). Previously,
9865 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9866 set. This did ensure strict client verification, but meant that,
9867 with applications using a single external cache for quite
9868 different requirements, clients could circumvent ciphersuite
9869 restrictions for a given session ID context by starting a session
9870 in a different context.
9874 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9875 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9876 authentication-only ciphersuites.
9880 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9881 not complete and could lead to a possible single byte overflow
9882 ([CVE-2007-5135]) [Ben Laurie]
9884 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9886 * Since AES128 and AES256 (and similarly Camellia128 and
9887 Camellia256) share a single mask bit in the logic of
9888 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9889 kludge to work properly if AES128 is available and AES256 isn't
9890 (or if Camellia128 is available and Camellia256 isn't).
9894 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9895 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9896 When a point or a seed is encoded in a BIT STRING, we need to
9897 prevent the removal of trailing zero bits to get the proper DER
9898 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9899 of a NamedBitList, for which trailing 0 bits need to be removed.)
9903 * Have SSL/TLS server implementation tolerate "mismatched" record
9904 protocol version while receiving ClientHello even if the
9905 ClientHello is fragmented. (The server can't insist on the
9906 particular protocol version it has chosen before the ServerHello
9907 message has informed the client about his choice.)
9911 * Add RFC 3779 support.
9913 *Rob Austein for ARIN, Ben Laurie*
9915 * Load error codes if they are not already present instead of using a
9916 static variable. This allows them to be cleanly unloaded and reloaded.
9917 Improve header file function name parsing.
9921 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9922 or CAPABILITY handshake as required by RFCs.
9926 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9928 * Introduce limits to prevent malicious keys being able to
9929 cause a denial of service. ([CVE-2006-2940])
9931 *Steve Henson, Bodo Moeller*
9933 * Fix ASN.1 parsing of certain invalid structures that can result
9934 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9936 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9937 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9939 * Fix SSL client code which could crash if connecting to a
9940 malicious SSLv2 server. ([CVE-2006-4343])
9942 *Tavis Ormandy and Will Drewry, Google Security Team*
9944 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9945 match only those. Before that, "AES256-SHA" would be interpreted
9946 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9947 the same strength classification in 0.9.7h) as we currently only
9948 have a single AES bit in the ciphersuite description bitmap.
9949 That change, however, also applied to ciphersuite strings such as
9950 "RC4-MD5" that intentionally matched multiple ciphersuites --
9951 namely, SSL 2.0 ciphersuites in addition to the more common ones
9952 from SSL 3.0/TLS 1.0.
9954 So we change the selection algorithm again: Naming an explicit
9955 ciphersuite selects this one ciphersuite, and any other similar
9956 ciphersuite (same bitmap) from *other* protocol versions.
9957 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9958 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9960 Since SSL 2.0 does not have any ciphersuites for which the
9961 128/256 bit distinction would be relevant, this works for now.
9962 The proper fix will be to use different bits for AES128 and
9963 AES256, which would have avoided the problems from the beginning;
9964 however, bits are scarce, so we can only do this in a new release
9965 (not just a patchlevel) when we can change the SSL_CIPHER
9966 definition to split the single 'unsigned long mask' bitmap into
9967 multiple values to extend the available space.
9971 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9973 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9974 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9976 * Add AES IGE and biIGE modes.
9980 * Change the Unix randomness entropy gathering to use poll() when
9981 possible instead of select(), since the latter has some
9982 undesirable limitations.
9984 *Darryl Miles via Richard Levitte and Bodo Moeller*
9986 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9987 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9988 cannot be implicitly activated as part of, e.g., the "AES" alias.
9989 However, please upgrade to OpenSSL 0.9.9[-dev] for
9990 non-experimental use of the ECC ciphersuites to get TLS extension
9991 support, which is required for curve and point format negotiation
9992 to avoid potential handshake problems.
9996 * Disable rogue ciphersuites:
9998 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9999 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10000 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10002 The latter two were purportedly from
10003 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10006 Also deactivate the remaining ciphersuites from
10007 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10008 unofficial, and the ID has long expired.
10012 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10013 dual-core machines) and other potential thread-safety issues.
10017 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10018 versions), which is now available for royalty-free use
10019 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10020 Also, add Camellia TLS ciphersuites from RFC 4132.
10022 To minimize changes between patchlevels in the OpenSSL 0.9.8
10023 series, Camellia remains excluded from compilation unless OpenSSL
10024 is configured with 'enable-camellia'.
10028 * Disable the padding bug check when compression is in use. The padding
10029 bug check assumes the first packet is of even length, this is not
10030 necessarily true if compression is enabled and can result in false
10031 positives causing handshake failure. The actual bug test is ancient
10032 code so it is hoped that implementations will either have fixed it by
10033 now or any which still have the bug do not support compression.
10037 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10039 * When applying a cipher rule check to see if string match is an explicit
10040 cipher suite and only match that one cipher suite if it is.
10044 * Link in manifests for VC++ if needed.
10046 *Austin Ziegler <halostatue@gmail.com>*
10048 * Update support for ECC-based TLS ciphersuites according to
10049 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10050 TLS extensions, which are supported starting with the 0.9.9
10051 branch, not in the OpenSSL 0.9.8 branch).
10055 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10056 opaque EVP_CIPHER_CTX handling.
10060 * Fixes and enhancements to zlib compression code. We now only use
10061 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10062 to conform with the standards mentioned here:
10063 <http://www.zlib.net/DLL_FAQ.txt>
10064 Static zlib linking now works on Windows and the new --with-zlib-include
10065 --with-zlib-lib options to Configure can be used to supply the location
10066 of the headers and library. Gracefully handle case where zlib library
10071 * Several fixes and enhancements to the OID generation code. The old code
10072 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10073 handle numbers larger than ULONG_MAX, truncated printing and had a
10074 non standard OBJ_obj2txt() behaviour.
10078 * Add support for building of engines under engine/ as shared libraries
10079 under VC++ build system.
10083 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10084 Hopefully, we will not see any false combination of paths any more.
10088 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10090 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10091 (part of SSL_OP_ALL). This option used to disable the
10092 countermeasure against man-in-the-middle protocol-version
10093 rollback in the SSL 2.0 server implementation, which is a bad
10094 idea. ([CVE-2005-2969])
10096 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10097 for Information Security, National Institute of Advanced Industrial
10098 Science and Technology [AIST], Japan)*
10100 * Add two function to clear and return the verify parameter flags.
10104 * Keep cipherlists sorted in the source instead of sorting them at
10105 runtime, thus removing the need for a lock.
10109 * Avoid some small subgroup attacks in Diffie-Hellman.
10111 *Nick Mathewson and Ben Laurie*
10113 * Add functions for well-known primes.
10117 * Extended Windows CE support.
10119 *Satoshi Nakamura and Andy Polyakov*
10121 * Initialize SSL_METHOD structures at compile time instead of during
10122 runtime, thus removing the need for a lock.
10126 * Make PKCS7_decrypt() work even if no certificate is supplied by
10127 attempting to decrypt each encrypted key in turn. Add support to
10132 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10134 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10137 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10141 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10142 key into the same file any more.
10146 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10150 * Add -utf8 command line and config file option to 'ca'.
10152 *Stefan <stf@udoma.org*
10154 * Removed the macro des_crypt(), as it seems to conflict with some
10155 libraries. Use DES_crypt().
10159 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10160 involves renaming the source and generated shared-libs for
10161 both. The engines will accept the corrected or legacy ids
10162 ('ncipher' and '4758_cca' respectively) when binding. NB,
10163 this only applies when building 'shared'.
10165 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10167 * Add attribute functions to EVP_PKEY structure. Modify
10168 PKCS12_create() to recognize a CSP name attribute and
10169 use it. Make -CSP option work again in pkcs12 utility.
10173 * Add new functionality to the bn blinding code:
10174 - automatic re-creation of the BN_BLINDING parameters after
10175 a fixed number of uses (currently 32)
10176 - add new function for parameter creation
10177 - introduce flags to control the update behaviour of the
10178 BN_BLINDING parameters
10179 - hide BN_BLINDING structure
10180 Add a second BN_BLINDING slot to the RSA structure to improve
10181 performance when a single RSA object is shared among several
10186 * Add support for DTLS.
10188 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10190 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10191 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10195 * Remove buggy and incomplete DH cert support from
10196 ssl/ssl_rsa.c and ssl/s3_both.c
10200 * Use SHA-1 instead of MD5 as the default digest algorithm for
10201 the `apps/openssl` commands.
10205 * Compile clean with "-Wall -Wmissing-prototypes
10206 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10207 DEBUG_SAFESTACK must also be set.
10211 * Change ./Configure so that certain algorithms can be disabled by default.
10212 The new counterpiece to "no-xxx" is "enable-xxx".
10214 The patented RC5 and MDC2 algorithms will now be disabled unless
10215 "enable-rc5" and "enable-mdc2", respectively, are specified.
10217 (IDEA remains enabled despite being patented. This is because IDEA
10218 is frequently required for interoperability, and there is no license
10219 fee for non-commercial use. As before, "no-idea" can be used to
10220 avoid this algorithm.)
10224 * Add processing of proxy certificates (see RFC 3820). This work was
10225 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10226 EGEE (Enabling Grids for E-science in Europe).
10230 * RC4 performance overhaul on modern architectures/implementations, such
10231 as Intel P4, IA-64 and AMD64.
10235 * New utility extract-section.pl. This can be used specify an alternative
10236 section number in a pod file instead of having to treat each file as
10237 a separate case in Makefile. This can be done by adding two lines to the
10240 =for comment openssl_section:XXX
10242 The blank line is mandatory.
10246 * New arguments -certform, -keyform and -pass for s_client and s_server
10247 to allow alternative format key and certificate files and passphrase
10252 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10253 update associated structures and add various utility functions.
10255 Add new policy related verify parameters, include policy checking in
10256 standard verify code. Enhance 'smime' application with extra parameters
10257 to support policy checking and print out.
10261 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10262 Nehemiah processors. These extensions support AES encryption in hardware
10263 as well as RNG (though RNG support is currently disabled).
10265 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10267 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10271 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10273 *Andy Polyakov and a number of other people*
10275 * Improved PowerPC platform support. Most notably BIGNUM assembler
10276 implementation contributed by IBM.
10278 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10280 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10281 exponent rather than 'unsigned long'. There is a corresponding change to
10282 the new 'rsa_keygen' element of the RSA_METHOD structure.
10284 *Jelte Jansen, Geoff Thorpe*
10286 * Functionality for creating the initial serial number file is now
10287 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10289 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10290 number file to 1, which is bound to cause problems. To avoid
10291 the problems while respecting compatibility between different 0.9.7
10292 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10293 CA.pl for serial number initialization. With the new release 0.9.8,
10294 we can fix the problem directly in the 'ca' utility.)
10298 * Reduced header interdependencies by declaring more opaque objects in
10299 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10300 give fewer recursive includes, which could break lazy source code - so
10301 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10302 developers should define this symbol when building and using openssl to
10303 ensure they track the recommended behaviour, interfaces, [etc], but
10304 backwards-compatible behaviour prevails when this isn't defined.
10308 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10312 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10313 This will generate a random key of the appropriate length based on the
10314 cipher context. The EVP_CIPHER can provide its own random key generation
10315 routine to support keys of a specific form. This is used in the des and
10316 3des routines to generate a key of the correct parity. Update S/MIME
10317 code to use new functions and hence generate correct parity DES keys.
10318 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10319 valid (weak or incorrect parity).
10323 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10324 as looking them up. This is useful when the verified structure may contain
10325 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10326 present unless the new PKCS7_NO_CRL flag is asserted.
10330 * Extend ASN1 oid configuration module. It now additionally accepts the
10333 shortName = some long name, 1.2.3.4
10337 * Reimplemented the BN_CTX implementation. There is now no more static
10338 limitation on the number of variables it can handle nor the depth of the
10339 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10340 information can now expand as required, and rather than having a single
10341 static array of bignums, BN_CTX now uses a linked-list of such arrays
10342 allowing it to expand on demand whilst maintaining the usefulness of
10343 BN_CTX's "bundling".
10347 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10348 to allow all RSA operations to function using a single BN_CTX.
10352 * Preliminary support for certificate policy evaluation and checking. This
10353 is initially intended to pass the tests outlined in "Conformance Testing
10354 of Relying Party Client Certificate Path Processing Logic" v1.07.
10358 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10359 remained unused and not that useful. A variety of other little bignum
10360 tweaks and fixes have also been made continuing on from the audit (see
10365 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10366 associated ASN1, EVP and SSL functions and old ASN1 macros.
10370 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10371 and this should never fail. So the return value from the use of
10372 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10373 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10377 * BN_CTX_get() should return zero-valued bignums, providing the same
10378 initialised value as BN_new().
10380 *Geoff Thorpe, suggested by Ulf Möller*
10382 * Support for inhibitAnyPolicy certificate extension.
10386 * An audit of the BIGNUM code is underway, for which debugging code is
10387 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10388 is considered valid when processing BIGNUMs, and causes execution to
10389 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10390 further steps are taken to deliberately pollute unused data in BIGNUM
10391 structures to try and expose faulty code further on. For now, openssl will
10392 (in its default mode of operation) continue to tolerate the inconsistent
10393 forms that it has tolerated in the past, but authors and packagers should
10394 consider trying openssl and their own applications when compiled with
10395 these debugging symbols defined. It will help highlight potential bugs in
10396 their own code, and will improve the test coverage for OpenSSL itself. At
10397 some point, these tighter rules will become openssl's default to improve
10398 maintainability, though the assert()s and other overheads will remain only
10399 in debugging configurations. See bn.h for more details.
10401 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10403 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10404 that can only be obtained through BN_CTX_new() (which implicitly
10405 initialises it). The presence of this function only made it possible
10406 to overwrite an existing structure (and cause memory leaks).
10410 * Because of the callback-based approach for implementing LHASH as a
10411 template type, lh_insert() adds opaque objects to hash-tables and
10412 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10413 to clean up those corresponding objects before destroying the hash table
10414 (and losing the object pointers). So some over-zealous constifications in
10415 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10416 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10417 prototyped to have "const" restrictions on the object pointers they are
10418 given (and so aren't required to cast them away any more).
10422 * The tmdiff.h API was so ugly and minimal that our own timing utility
10423 (speed) prefers to use its own implementation. The two implementations
10424 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10425 its object type properly exposed (MS_TM) instead of casting to/from
10426 `char *`. This may still change yet if someone realises MS_TM and
10428 aren't necessarily the greatest nomenclatures - but this is what was used
10429 internally to the implementation so I've used that for now.
10433 * Ensure that deprecated functions do not get compiled when
10434 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10435 the self-tests were still using deprecated key-generation functions so
10436 these have been updated also.
10440 * Reorganise PKCS#7 code to separate the digest location functionality
10441 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10442 New function PKCS7_set_digest() to set the digest type for PKCS#7
10443 digestedData type. Add additional code to correctly generate the
10444 digestedData type and add support for this type in PKCS7 initialization
10449 * New function PKCS7_set0_type_other() this initializes a PKCS7
10450 structure of type "other".
10454 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10455 sure the loop does correctly stop and breaking ("division by zero")
10456 modulus operations are not performed. The (pre-generated) prime
10457 table crypto/bn/bn_prime.h was already correct, but it could not be
10458 re-generated on some platforms because of the "division by zero"
10459 situation in the script.
10461 *Ralf S. Engelschall*
10463 * Update support for ECC-based TLS ciphersuites according to
10464 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10465 SHA-1 now is only used for "small" curves (where the
10466 representation of a field element takes up to 24 bytes); for
10467 larger curves, the field element resulting from ECDH is directly
10468 used as premaster secret.
10470 *Douglas Stebila (Sun Microsystems Laboratories)*
10472 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10473 curve secp160r1 to the tests.
10475 *Douglas Stebila (Sun Microsystems Laboratories)*
10477 * Add the possibility to load symbols globally with DSO.
10479 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10481 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10482 control of the error stack.
10486 * Add support for STORE in ENGINE.
10490 * Add the STORE type. The intention is to provide a common interface
10491 to certificate and key stores, be they simple file-based stores, or
10492 HSM-type store, or LDAP stores, or...
10493 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10497 * Add a generic structure called OPENSSL_ITEM. This can be used to
10498 pass a list of arguments to any function as well as provide a way
10499 for a function to pass data back to the caller.
10503 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10504 works like BUF_strdup() but can be used to duplicate a portion of
10505 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10510 * Add the function sk_find_ex() which works like sk_find(), but will
10511 return an index to an element even if an exact match couldn't be
10512 found. The index is guaranteed to point at the element where the
10513 searched-for key would be inserted to preserve sorting order.
10517 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10518 takes an extra flags argument for optional functionality. Currently,
10519 the following flags are defined:
10521 OBJ_BSEARCH_VALUE_ON_NOMATCH
10522 This one gets OBJ_bsearch_ex() to return a pointer to the first
10523 element where the comparing function returns a negative or zero
10526 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10527 This one gets OBJ_bsearch_ex() to return a pointer to the first
10528 element where the comparing function returns zero. This is useful
10529 if there are more than one element where the comparing function
10534 * Make it possible to create self-signed certificates with 'openssl ca'
10535 in such a way that the self-signed certificate becomes part of the
10536 CA database and uses the same mechanisms for serial number generation
10537 as all other certificate signing. The new flag '-selfsign' enables
10538 this functionality. Adapt CA.sh and CA.pl.in.
10542 * Add functionality to check the public key of a certificate request
10543 against a given private. This is useful to check that a certificate
10544 request can be signed by that key (self-signing).
10548 * Make it possible to have multiple active certificates with the same
10549 subject in the CA index file. This is done only if the keyword
10550 'unique_subject' is set to 'no' in the main CA section (default
10551 if 'CA_default') of the configuration file. The value is saved
10552 with the database itself in a separate index attribute file,
10553 named like the index file with '.attr' appended to the name.
10557 * Generate multi-valued AVAs using '+' notation in config files for
10562 * Support for nameConstraints certificate extension.
10566 * Support for policyConstraints certificate extension.
10570 * Support for policyMappings certificate extension.
10574 * Make sure the default DSA_METHOD implementation only uses its
10575 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10576 and change its own handlers to be NULL so as to remove unnecessary
10577 indirection. This lets alternative implementations fallback to the
10578 default implementation more easily.
10582 * Support for directoryName in GeneralName related extensions
10587 * Make it possible to link applications using Makefile.shared.
10588 Make that possible even when linking against static libraries!
10592 * Support for single pass processing for S/MIME signing. This now
10593 means that S/MIME signing can be done from a pipe, in addition
10594 cleartext signing (multipart/signed type) is effectively streaming
10595 and the signed data does not need to be all held in memory.
10597 This is done with a new flag PKCS7_STREAM. When this flag is set
10598 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10599 is done after the data is output (and digests calculated) in
10600 SMIME_write_PKCS7().
10604 * Add full support for -rpath/-R, both in shared libraries and
10605 applications, at least on the platforms where it's known how
10610 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10611 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10612 will now compute a table of multiples of the generator that
10613 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10614 faster (notably in the case of a single point multiplication,
10615 scalar * generator).
10617 *Nils Larsch, Bodo Moeller*
10619 * IPv6 support for certificate extensions. The various extensions
10620 which use the IP:a.b.c.d can now take IPv6 addresses using the
10621 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10626 * Added an ENGINE that implements RSA by performing private key
10627 exponentiations with the GMP library. The conversions to and from
10628 GMP's mpz_t format aren't optimised nor are any montgomery forms
10629 cached, and on x86 it appears OpenSSL's own performance has caught up.
10630 However there are likely to be other architectures where GMP could
10631 provide a boost. This ENGINE is not built in by default, but it can be
10632 specified at Configure time and should be accompanied by the necessary
10633 linker additions, eg;
10634 ./config -DOPENSSL_USE_GMP -lgmp
10638 * "openssl engine" will not display ENGINE/DSO load failure errors when
10639 testing availability of engines with "-t" - the old behaviour is
10640 produced by increasing the feature's verbosity with "-tt".
10644 * ECDSA routines: under certain error conditions uninitialized BN objects
10645 could be freed. Solution: make sure initialization is performed early
10646 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10651 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10652 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10653 software implementations. For DSA and DH, parameter generation can
10654 also be overridden by providing the appropriate method callbacks.
10658 * Change the "progress" mechanism used in key-generation and
10659 primality testing to functions that take a new BN_GENCB pointer in
10660 place of callback/argument pairs. The new API functions have `_ex`
10661 postfixes and the older functions are reimplemented as wrappers for
10662 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10663 declarations of the old functions to help (graceful) attempts to
10664 migrate to the new functions. Also, the new key-generation API
10665 functions operate on a caller-supplied key-structure and return
10666 success/failure rather than returning a key or NULL - this is to
10667 help make "keygen" another member function of RSA_METHOD etc.
10669 Example for using the new callback interface:
10671 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10672 void *my_arg = ...;
10675 BN_GENCB_set(&my_cb, my_callback, my_arg);
10677 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10678 /* For the meaning of a, b in calls to my_callback(), see the
10679 * documentation of the function that calls the callback.
10680 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10681 * my_callback should return 1 if it wants BN_is_prime_ex()
10682 * to continue, or 0 to stop.
10687 * Change the ZLIB compression method to be stateful, and make it
10688 available to TLS with the number defined in
10689 draft-ietf-tls-compression-04.txt.
10693 * Add the ASN.1 structures and functions for CertificatePair, which
10694 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10696 CertificatePair ::= SEQUENCE {
10697 forward [0] Certificate OPTIONAL,
10698 reverse [1] Certificate OPTIONAL,
10699 -- at least one of the pair shall be present -- }
10701 Also implement the PEM functions to read and write certificate
10702 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10704 This needed to be defined, mostly for the sake of the LDAP
10705 attribute crossCertificatePair, but may prove useful elsewhere as
10710 * Make it possible to inhibit symlinking of shared libraries in
10711 Makefile.shared, for Cygwin's sake.
10715 * Extend the BIGNUM API by creating a function
10716 void BN_set_negative(BIGNUM *a, int neg);
10717 and a macro that behave like
10718 int BN_is_negative(const BIGNUM *a);
10720 to avoid the need to access 'a->neg' directly in applications.
10724 * Implement fast modular reduction for pseudo-Mersenne primes
10725 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10726 EC_GROUP_new_curve_GFp() will now automatically use this
10729 *Nils Larsch <nla@trustcenter.de>*
10731 * Add new lock type (CRYPTO_LOCK_BN).
10735 * Change the ENGINE framework to automatically load engines
10736 dynamically from specific directories unless they could be
10737 found to already be built in or loaded. Move all the
10738 current engines except for the cryptodev one to a new
10739 directory engines/.
10740 The engines in engines/ are built as shared libraries if
10741 the "shared" options was given to ./Configure or ./config.
10742 Otherwise, they are inserted in libcrypto.a.
10743 /usr/local/ssl/engines is the default directory for dynamic
10744 engines, but that can be overridden at configure time through
10745 the usual use of --prefix and/or --openssldir, and at run
10746 time with the environment variable OPENSSL_ENGINES.
10748 *Geoff Thorpe and Richard Levitte*
10750 * Add Makefile.shared, a helper makefile to build shared
10751 libraries. Adapt Makefile.org.
10755 * Add version info to Win32 DLLs.
10757 *Peter 'Luna' Runestig" <peter@runestig.com>*
10759 * Add new 'medium level' PKCS#12 API. Certificates and keys
10760 can be added using this API to created arbitrary PKCS#12
10761 files while avoiding the low-level API.
10763 New options to PKCS12_create(), key or cert can be NULL and
10764 will then be omitted from the output file. The encryption
10765 algorithm NIDs can be set to -1 for no encryption, the mac
10766 iteration count can be set to 0 to omit the mac.
10768 Enhance pkcs12 utility by making the -nokeys and -nocerts
10769 options work when creating a PKCS#12 file. New option -nomac
10770 to omit the mac, NONE can be set for an encryption algorithm.
10771 New code is modified to use the enhanced PKCS12_create()
10772 instead of the low-level API.
10776 * Extend ASN1 encoder to support indefinite length constructed
10777 encoding. This can output sequences tags and octet strings in
10778 this form. Modify pk7_asn1.c to support indefinite length
10779 encoding. This is experimental and needs additional code to
10780 be useful, such as an ASN1 bio and some enhanced streaming
10783 Extend template encode functionality so that tagging is passed
10784 down to the template encoder.
10788 * Let 'openssl req' fail if an argument to '-newkey' is not
10789 recognized instead of using RSA as a default.
10793 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10794 As these are not official, they are not included in "ALL";
10795 the "ECCdraft" ciphersuite group alias can be used to select them.
10797 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10799 * Add ECDH engine support.
10801 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10803 * Add ECDH in new directory crypto/ecdh/.
10805 *Douglas Stebila (Sun Microsystems Laboratories)*
10807 * Let BN_rand_range() abort with an error after 100 iterations
10808 without success (which indicates a broken PRNG).
10812 * Change BN_mod_sqrt() so that it verifies that the input value
10813 is really the square of the return value. (Previously,
10814 BN_mod_sqrt would show GIGO behaviour.)
10818 * Add named elliptic curves over binary fields from X9.62, SECG,
10819 and WAP/WTLS; add OIDs that were still missing.
10821 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10823 * Extend the EC library for elliptic curves over binary fields
10824 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10827 EC_GF2m_simple_method
10831 EC_GROUP_new_curve_GF2m
10832 EC_GROUP_set_curve_GF2m
10833 EC_GROUP_get_curve_GF2m
10834 EC_POINT_set_affine_coordinates_GF2m
10835 EC_POINT_get_affine_coordinates_GF2m
10836 EC_POINT_set_compressed_coordinates_GF2m
10838 Point compression for binary fields is disabled by default for
10839 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10842 As binary polynomials are represented as BIGNUMs, various members
10843 of the EC_GROUP and EC_POINT data structures can be shared
10844 between the implementations for prime fields and binary fields;
10845 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10846 are essentially identical to their `..._GFp` counterparts.
10847 (For simplicity, the `..._GFp` prefix has been dropped from
10848 various internal method names.)
10850 An internal 'field_div' method (similar to 'field_mul' and
10851 'field_sqr') has been added; this is used only for binary fields.
10853 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10855 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10856 through methods ('mul', 'precompute_mult').
10858 The generic implementations (now internally called 'ec_wNAF_mul'
10859 and 'ec_wNAF_precomputed_mult') remain the default if these
10860 methods are undefined.
10862 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10864 * New function EC_GROUP_get_degree, which is defined through
10865 EC_METHOD. For curves over prime fields, this returns the bit
10866 length of the modulus.
10868 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10870 * New functions EC_GROUP_dup, EC_POINT_dup.
10871 (These simply call ..._new and ..._copy).
10873 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10875 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10876 Polynomials are represented as BIGNUMs (where the sign bit is not
10877 used) in the following functions [macros]:
10880 BN_GF2m_sub [= BN_GF2m_add]
10881 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10882 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10883 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10885 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10886 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10887 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10888 BN_GF2m_cmp [= BN_ucmp]
10890 (Note that only the 'mod' functions are actually for fields GF(2^m).
10891 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10893 For some functions, an the irreducible polynomial defining a
10894 field can be given as an 'unsigned int[]' with strictly
10895 decreasing elements giving the indices of those bits that are set;
10896 i.e., p[] represents the polynomial
10897 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10899 p[0] > p[1] > ... > p[k] = 0.
10900 This applies to the following functions:
10903 BN_GF2m_mod_mul_arr
10904 BN_GF2m_mod_sqr_arr
10905 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10906 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10907 BN_GF2m_mod_exp_arr
10908 BN_GF2m_mod_sqrt_arr
10909 BN_GF2m_mod_solve_quad_arr
10913 Conversion can be performed by the following functions:
10918 bntest.c has additional tests for binary polynomial arithmetic.
10920 Two implementations for BN_GF2m_mod_div() are available.
10921 The default algorithm simply uses BN_GF2m_mod_inv() and
10922 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10923 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10924 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10926 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10928 * Add new error code 'ERR_R_DISABLED' that can be used when some
10929 functionality is disabled at compile-time.
10931 *Douglas Stebila <douglas.stebila@sun.com>*
10933 * Change default behaviour of 'openssl asn1parse' so that more
10934 information is visible when viewing, e.g., a certificate:
10936 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10937 mode the content of non-printable OCTET STRINGs is output in a
10938 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10939 avoid the appearance of a printable string.
10941 *Nils Larsch <nla@trustcenter.de>*
10943 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10945 EC_GROUP_set_asn1_flag()
10946 EC_GROUP_get_asn1_flag()
10947 EC_GROUP_set_point_conversion_form()
10948 EC_GROUP_get_point_conversion_form()
10949 These control ASN1 encoding details:
10950 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10951 has been set to OPENSSL_EC_NAMED_CURVE.
10952 - Points are encoded in uncompressed form by default; options for
10953 asn1_for are as for point2oct, namely
10954 POINT_CONVERSION_COMPRESSED
10955 POINT_CONVERSION_UNCOMPRESSED
10956 POINT_CONVERSION_HYBRID
10958 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10960 EC_GROUP_set_seed()
10961 EC_GROUP_get0_seed()
10962 EC_GROUP_get_seed_len()
10963 This is used only for ASN1 purposes (so far).
10965 *Nils Larsch <nla@trustcenter.de>*
10967 * Add 'field_type' member to EC_METHOD, which holds the NID
10968 of the appropriate field type OID. The new function
10969 EC_METHOD_get_field_type() returns this value.
10971 *Nils Larsch <nla@trustcenter.de>*
10974 EC_POINT_point2bn()
10975 EC_POINT_bn2point()
10976 EC_POINT_point2hex()
10977 EC_POINT_hex2point()
10978 providing useful interfaces to EC_POINT_point2oct() and
10979 EC_POINT_oct2point().
10981 *Nils Larsch <nla@trustcenter.de>*
10983 * Change internals of the EC library so that the functions
10984 EC_GROUP_set_generator()
10985 EC_GROUP_get_generator()
10986 EC_GROUP_get_order()
10987 EC_GROUP_get_cofactor()
10988 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10989 to methods, which would lead to unnecessary code duplication when
10990 adding different types of curves.
10992 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10994 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10995 arithmetic, and such that modified wNAFs are generated
10996 (which avoid length expansion in many cases).
11000 * Add a function EC_GROUP_check_discriminant() (defined via
11001 EC_METHOD) that verifies that the curve discriminant is non-zero.
11003 Add a function EC_GROUP_check() that makes some sanity tests
11004 on a EC_GROUP, its generator and order. This includes
11005 EC_GROUP_check_discriminant().
11007 *Nils Larsch <nla@trustcenter.de>*
11009 * Add ECDSA in new directory crypto/ecdsa/.
11011 Add applications 'openssl ecparam' and 'openssl ecdsa'
11012 (these are based on 'openssl dsaparam' and 'openssl dsa').
11014 ECDSA support is also included in various other files across the
11015 library. Most notably,
11016 - 'openssl req' now has a '-newkey ecdsa:file' option;
11017 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11018 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11019 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11020 them suitable for ECDSA where domain parameters must be
11021 extracted before the specific public key;
11022 - ECDSA engine support has been added.
11024 *Nils Larsch <nla@trustcenter.de>*
11026 * Include some named elliptic curves, and add OIDs from X9.62,
11027 SECG, and WAP/WTLS. Each curve can be obtained from the new
11029 EC_GROUP_new_by_curve_name(),
11030 and the list of available named curves can be obtained with
11031 EC_get_builtin_curves().
11032 Also add a 'curve_name' member to EC_GROUP objects, which can be
11034 EC_GROUP_set_curve_name()
11035 EC_GROUP_get_curve_name()
11037 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11039 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11040 was actually never needed) and in BN_mul(). The removal in BN_mul()
11041 required a small change in bn_mul_part_recursive() and the addition
11042 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11043 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11044 bn_sub_words() and bn_add_words() except they take arrays with
11049 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11051 * Cleanse PEM buffers before freeing them since they may contain
11054 *Benjamin Bennett <ben@psc.edu>*
11056 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11057 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11058 authentication-only ciphersuites.
11062 * Since AES128 and AES256 share a single mask bit in the logic of
11063 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11064 kludge to work properly if AES128 is available and AES256 isn't.
11068 * Expand security boundary to match 1.1.1 module.
11072 * Remove redundant features: hash file source, editing of test vectors
11073 modify fipsld to use external fips_premain.c signature.
11077 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11078 run algorithm test programs.
11082 * Make algorithm test programs more tolerant of whitespace.
11086 * Have SSL/TLS server implementation tolerate "mismatched" record
11087 protocol version while receiving ClientHello even if the
11088 ClientHello is fragmented. (The server can't insist on the
11089 particular protocol version it has chosen before the ServerHello
11090 message has informed the client about his choice.)
11094 * Load error codes if they are not already present instead of using a
11095 static variable. This allows them to be cleanly unloaded and reloaded.
11099 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11101 * Introduce limits to prevent malicious keys being able to
11102 cause a denial of service. ([CVE-2006-2940])
11104 *Steve Henson, Bodo Moeller*
11106 * Fix ASN.1 parsing of certain invalid structures that can result
11107 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11109 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11110 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11112 * Fix SSL client code which could crash if connecting to a
11113 malicious SSLv2 server. ([CVE-2006-4343])
11115 *Tavis Ormandy and Will Drewry, Google Security Team*
11117 * Change ciphersuite string processing so that an explicit
11118 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11119 will no longer include "AES128-SHA"), and any other similar
11120 ciphersuite (same bitmap) from *other* protocol versions (so that
11121 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11122 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11123 changes from 0.9.8b and 0.9.8d.
11127 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11129 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11130 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11132 * Change the Unix randomness entropy gathering to use poll() when
11133 possible instead of select(), since the latter has some
11134 undesirable limitations.
11136 *Darryl Miles via Richard Levitte and Bodo Moeller*
11138 * Disable rogue ciphersuites:
11140 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11141 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11142 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11144 The latter two were purportedly from
11145 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11148 Also deactivate the remaining ciphersuites from
11149 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11150 unofficial, and the ID has long expired.
11154 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11155 dual-core machines) and other potential thread-safety issues.
11159 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11161 * Adapt fipsld and the build system to link against the validated FIPS
11162 module in FIPS mode.
11166 * Fixes for VC++ 2005 build under Windows.
11170 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11171 from a Windows bash shell such as MSYS. It is autodetected from the
11172 "config" script when run from a VC++ environment. Modify standard VC++
11173 build to use fipscanister.o from the GNU make build.
11177 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11179 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11180 The value now differs depending on if you build for FIPS or not.
11181 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11182 safely run with a non-FIPSed libcrypto, as it may crash because of
11183 the difference induced by this change.
11187 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11189 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11190 (part of SSL_OP_ALL). This option used to disable the
11191 countermeasure against man-in-the-middle protocol-version
11192 rollback in the SSL 2.0 server implementation, which is a bad
11193 idea. ([CVE-2005-2969])
11195 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11196 for Information Security, National Institute of Advanced Industrial
11197 Science and Technology [AIST, Japan)]*
11199 * Minimal support for X9.31 signatures and PSS padding modes. This is
11200 mainly for FIPS compliance and not fully integrated at this stage.
11204 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11205 the exponentiation using a fixed-length exponent. (Otherwise,
11206 the information leaked through timing could expose the secret key
11207 after many signatures; cf. Bleichenbacher's attack on DSA with
11212 * Make a new fixed-window mod_exp implementation the default for
11213 RSA, DSA, and DH private-key operations so that the sequence of
11214 squares and multiplies and the memory access pattern are
11215 independent of the particular secret key. This will mitigate
11216 cache-timing and potential related attacks.
11218 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11219 and this is automatically used by BN_mod_exp_mont() if the new flag
11220 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11221 will use this BN flag for private exponents unless the flag
11222 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11223 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11225 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11227 * Change the client implementation for SSLv23_method() and
11228 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11229 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11230 (Previously, the SSL 2.0 backwards compatible Client Hello
11231 message format would be used even with SSL_OP_NO_SSLv2.)
11235 * Add support for smime-type MIME parameter in S/MIME messages which some
11240 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11241 a threadsafe manner. Modify rsa code to use new function and add calls
11242 to dsa and dh code (which had race conditions before).
11246 * Include the fixed error library code in the C error file definitions
11247 instead of fixing them up at runtime. This keeps the error code
11248 structures constant.
11252 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11254 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11257 * Fixes for newer kerberos headers. NB: the casts are needed because
11258 the 'length' field is signed on one version and unsigned on another
11259 with no (?) obvious way to tell the difference, without these VC++
11260 complains. Also the "definition" of FAR (blank) is no longer included
11261 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11262 some needed definitions.
11266 * Undo Cygwin change.
11270 * Added support for proxy certificates according to RFC 3820.
11271 Because they may be a security thread to unaware applications,
11272 they must be explicitly allowed in run-time. See
11273 docs/HOWTO/proxy_certificates.txt for further information.
11277 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11279 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11280 server and client random values. Previously
11281 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11282 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11284 This change has negligible security impact because:
11286 1. Server and client random values still have 24 bytes of pseudo random
11289 2. Server and client random values are sent in the clear in the initial
11292 3. The master secret is derived using the premaster secret (48 bytes in
11293 size for static RSA ciphersuites) as well as client server and random
11296 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11299 *Stephen Henson, reported by UK NISCC*
11301 * Use Windows randomness collection on Cygwin.
11305 * Fix hang in EGD/PRNGD query when communication socket is closed
11306 prematurely by EGD/PRNGD.
11308 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11310 * Prompt for pass phrases when appropriate for PKCS12 input format.
11314 * Back-port of selected performance improvements from development
11315 branch, as well as improved support for PowerPC platforms.
11319 * Add lots of checks for memory allocation failure, error codes to indicate
11320 failure and freeing up memory if a failure occurs.
11322 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11324 * Add new -passin argument to dgst.
11328 * Perform some character comparisons of different types in X509_NAME_cmp:
11329 this is needed for some certificates that re-encode DNs into UTF8Strings
11330 (in violation of RFC3280) and can't or won't issue name rollover
11335 * Make an explicit check during certificate validation to see that
11336 the CA setting in each certificate on the chain is correct. As a
11337 side effect always do the following basic checks on extensions,
11338 not just when there's an associated purpose to the check:
11340 - if there is an unhandled critical extension (unless the user
11341 has chosen to ignore this fault)
11342 - if the path length has been exceeded (if one is set at all)
11343 - that certain extensions fit the associated purpose (if one has
11348 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11350 * Avoid a race condition when CRLs are checked in a multi threaded
11351 environment. This would happen due to the reordering of the revoked
11352 entries during signature checking and serial number lookup. Now the
11353 encoding is cached and the serial number sort performed under a lock.
11354 Add new STACK function sk_is_sorted().
11358 * Add Delta CRL to the extension code.
11362 * Various fixes to s3_pkt.c so alerts are sent properly.
11364 *David Holmes <d.holmes@f5.com>*
11366 * Reduce the chances of duplicate issuer name and serial numbers (in
11367 violation of RFC3280) using the OpenSSL certificate creation utilities.
11368 This is done by creating a random 64 bit value for the initial serial
11369 number when a serial number file is created or when a self signed
11370 certificate is created using 'openssl req -x509'. The initial serial
11371 number file is created using 'openssl x509 -next_serial' in CA.pl
11372 rather than being initialized to 1.
11376 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11378 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11379 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11381 *Joe Orton, Steve Henson*
11383 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11386 *Joe Orton, Steve Henson*
11388 * Make it possible to have multiple active certificates with the same
11389 subject in the CA index file. This is done only if the keyword
11390 'unique_subject' is set to 'no' in the main CA section (default
11391 if 'CA_default') of the configuration file. The value is saved
11392 with the database itself in a separate index attribute file,
11393 named like the index file with '.attr' appended to the name.
11397 * X509 verify fixes. Disable broken certificate workarounds when
11398 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11399 keyUsage extension present. Don't accept CRLs with unhandled critical
11400 extensions: since verify currently doesn't process CRL extensions this
11401 rejects a CRL with *any* critical extensions. Add new verify error codes
11406 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11407 A clarification of RFC2560 will require the use of OCTET STRINGs and
11408 some implementations cannot handle the current raw format. Since OpenSSL
11409 copies and compares OCSP nonces as opaque blobs without any attempt at
11410 parsing them this should not create any compatibility issues.
11414 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11415 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11416 this HMAC (and other) operations are several times slower than OpenSSL
11421 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11423 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11425 * Use the correct content when signing type "other".
11429 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11431 * Fix various bugs revealed by running the NISCC test suite:
11433 Stop out of bounds reads in the ASN1 code when presented with
11434 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11436 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11438 If verify callback ignores invalid public key errors don't try to check
11439 certificate signature with the NULL public key.
11443 * New -ignore_err option in ocsp application to stop the server
11444 exiting on the first error in a request.
11448 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11449 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11454 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11455 extra data after the compression methods not only for TLS 1.0
11456 but also for SSL 3.0 (as required by the specification).
11458 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11460 * Change X509_certificate_type() to mark the key as exported/exportable
11461 when it's 512 *bits* long, not 512 bytes.
11465 * Change AES_cbc_encrypt() so it outputs exact multiple of
11466 blocks during encryption.
11470 * Various fixes to base64 BIO and non blocking I/O. On write
11471 flushes were not handled properly if the BIO retried. On read
11472 data was not being buffered properly and had various logic bugs.
11473 This also affects blocking I/O when the data being decoded is a
11478 * Various S/MIME bugfixes and compatibility changes:
11479 output correct application/pkcs7 MIME type if
11480 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11481 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11482 of files as .eml work). Correctly handle very long lines in MIME
11487 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11489 * Countermeasure against the Klima-Pokorny-Rosa extension of
11490 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11491 a protocol version number mismatch like a decryption error
11492 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11496 * Turn on RSA blinding by default in the default implementation
11497 to avoid a timing attack. Applications that don't want it can call
11498 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11499 They would be ill-advised to do so in most cases.
11501 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11503 * Change RSA blinding code so that it works when the PRNG is not
11504 seeded (in this case, the secret RSA exponent is abused as
11505 an unpredictable seed -- if it is not unpredictable, there
11506 is no point in blinding anyway). Make RSA blinding thread-safe
11507 by remembering the creator's thread ID in rsa->blinding and
11508 having all other threads use local one-time blinding factors
11509 (this requires more computation than sharing rsa->blinding, but
11510 avoids excessive locking; and if an RSA object is not shared
11511 between threads, blinding will still be very fast).
11515 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11516 ENGINE as defaults for all supported algorithms irrespective of
11517 the 'flags' parameter. 'flags' is now honoured, so applications
11518 should make sure they are passing it correctly.
11522 * Target "mingw" now allows native Windows code to be generated in
11523 the Cygwin environment as well as with the MinGW compiler.
11527 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11529 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11530 via timing by performing a MAC computation even if incorrect
11531 block cipher padding has been found. This is a countermeasure
11532 against active attacks where the attacker has to distinguish
11533 between bad padding and a MAC verification error. ([CVE-2003-0078])
11535 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11536 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11537 Martin Vuagnoux (EPFL, Ilion)*
11539 * Make the no-err option work as intended. The intention with no-err
11540 is not to have the whole error stack handling routines removed from
11541 libcrypto, it's only intended to remove all the function name and
11542 reason texts, thereby removing some of the footprint that may not
11543 be interesting if those errors aren't displayed anyway.
11545 NOTE: it's still possible for any application or module to have its
11546 own set of error texts inserted. The routines are there, just not
11547 used by default when no-err is given.
11551 * Add support for FreeBSD on IA64.
11553 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11555 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11556 Kerberos function mit_des_cbc_cksum(). Before this change,
11557 the value returned by DES_cbc_cksum() was like the one from
11558 mit_des_cbc_cksum(), except the bytes were swapped.
11560 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11562 * Allow an application to disable the automatic SSL chain building.
11563 Before this a rather primitive chain build was always performed in
11564 ssl3_output_cert_chain(): an application had no way to send the
11565 correct chain if the automatic operation produced an incorrect result.
11567 Now the chain builder is disabled if either:
11569 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11571 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11573 The reasoning behind this is that an application would not want the
11574 auto chain building to take place if extra chain certificates are
11575 present and it might also want a means of sending no additional
11576 certificates (for example the chain has two certificates and the
11581 * Add the possibility to build without the ENGINE framework.
11583 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11585 * Under Win32 gmtime() can return NULL: check return value in
11586 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11590 * DSA routines: under certain error conditions uninitialized BN objects
11591 could be freed. Solution: make sure initialization is performed early
11592 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11593 Nils Larsch <nla@trustcenter.de> via PR#459)
11597 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11598 checked on reconnect on the client side, therefore session resumption
11599 could still fail with a "ssl session id is different" error. This
11600 behaviour is masked when SSL_OP_ALL is used due to
11601 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11602 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11603 followup to PR #377.
11607 * IA-32 assembler support enhancements: unified ELF targets, support
11608 for SCO/Caldera platforms, fix for Cygwin shared build.
11612 * Add support for FreeBSD on sparc64. As a consequence, support for
11613 FreeBSD on non-x86 processors is separate from x86 processors on
11614 the config script, much like the NetBSD support.
11616 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11618 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11620 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11623 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11624 code (06) was taken as the first octet of the session ID and the last
11625 octet was ignored consequently. As a result SSLv2 client side session
11626 caching could not have worked due to the session ID mismatch between
11628 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11633 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11634 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11639 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11640 seems that in spite of existing for more than a year, many application
11641 author have done nothing to provide the necessary callbacks, which
11642 means that this particular engine will not work properly anywhere.
11643 This is a very unfortunate situation which forces us, in the name
11644 of usability, to give the hw_ncipher.c a static lock, which is part
11646 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11647 appear in 0.9.8 or later. We EXPECT application authors to have
11648 dealt properly with this when 0.9.8 is released (unless we actually
11649 make such changes in the libcrypto locking code that changes will
11650 have to be made anyway).
11654 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11655 octets have been read, EOF or an error occurs. Without this change
11656 some truncated ASN1 structures will not produce an error.
11660 * Disable Heimdal support, since it hasn't been fully implemented.
11661 Still give the possibility to force the use of Heimdal, but with
11662 warnings and a request that patches get sent to openssl-dev.
11666 * Add the VC-CE target, introduce the WINCE sysname, and add
11667 INSTALL.WCE and appropriate conditionals to make it build.
11669 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11671 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11672 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11673 edit numbers of the version.
11675 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11677 * Introduce safe string copy and catenation functions
11678 (BUF_strlcpy() and BUF_strlcat()).
11680 *Ben Laurie (CHATS) and Richard Levitte*
11682 * Avoid using fixed-size buffers for one-line DNs.
11684 *Ben Laurie (CHATS)*
11686 * Add BUF_MEM_grow_clean() to avoid information leakage when
11687 resizing buffers containing secrets, and use where appropriate.
11689 *Ben Laurie (CHATS)*
11691 * Avoid using fixed size buffers for configuration file location.
11693 *Ben Laurie (CHATS)*
11695 * Avoid filename truncation for various CA files.
11697 *Ben Laurie (CHATS)*
11699 * Use sizeof in preference to magic numbers.
11701 *Ben Laurie (CHATS)*
11703 * Avoid filename truncation in cert requests.
11705 *Ben Laurie (CHATS)*
11707 * Add assertions to check for (supposedly impossible) buffer
11710 *Ben Laurie (CHATS)*
11712 * Don't cache truncated DNS entries in the local cache (this could
11713 potentially lead to a spoofing attack).
11715 *Ben Laurie (CHATS)*
11717 * Fix various buffers to be large enough for hex/decimal
11718 representations in a platform independent manner.
11720 *Ben Laurie (CHATS)*
11722 * Add CRYPTO_realloc_clean() to avoid information leakage when
11723 resizing buffers containing secrets, and use where appropriate.
11725 *Ben Laurie (CHATS)*
11727 * Add BIO_indent() to avoid much slightly worrying code to do
11730 *Ben Laurie (CHATS)*
11732 * Convert sprintf()/BIO_puts() to BIO_printf().
11734 *Ben Laurie (CHATS)*
11736 * buffer_gets() could terminate with the buffer only half
11739 *Ben Laurie (CHATS)*
11741 * Add assertions to prevent user-supplied crypto functions from
11742 overflowing internal buffers by having large block sizes, etc.
11744 *Ben Laurie (CHATS)*
11746 * New OPENSSL_assert() macro (similar to assert(), but enabled
11749 *Ben Laurie (CHATS)*
11751 * Eliminate unused copy of key in RC4.
11753 *Ben Laurie (CHATS)*
11755 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11757 *Ben Laurie (CHATS)*
11759 * Fix off-by-one error in EGD path.
11761 *Ben Laurie (CHATS)*
11763 * If RANDFILE path is too long, ignore instead of truncating.
11765 *Ben Laurie (CHATS)*
11767 * Eliminate unused and incorrectly sized X.509 structure
11770 *Ben Laurie (CHATS)*
11772 * Eliminate unused and dangerous function knumber().
11774 *Ben Laurie (CHATS)*
11776 * Eliminate unused and dangerous structure, KSSL_ERR.
11778 *Ben Laurie (CHATS)*
11780 * Protect against overlong session ID context length in an encoded
11781 session object. Since these are local, this does not appear to be
11784 *Ben Laurie (CHATS)*
11786 * Change from security patch (see 0.9.6e below) that did not affect
11787 the 0.9.6 release series:
11789 Remote buffer overflow in SSL3 protocol - an attacker could
11790 supply an oversized master key in Kerberos-enabled versions.
11793 *Ben Laurie (CHATS)*
11795 * Change the SSL kerb5 codes to match RFC 2712.
11799 * Make -nameopt work fully for req and add -reqopt switch.
11801 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11803 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11805 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11807 * Make sure tests can be performed even if the corresponding algorithms
11808 have been removed entirely. This was also the last step to make
11809 OpenSSL compilable with DJGPP under all reasonable conditions.
11811 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11813 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11814 to allow version independent disabling of normally unselected ciphers,
11815 which may be activated as a side-effect of selecting a single cipher.
11817 (E.g., cipher list string "RSA" enables ciphersuites that are left
11818 out of "ALL" because they do not provide symmetric encryption.
11819 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11821 *Lutz Jaenicke, Bodo Moeller*
11823 * Add appropriate support for separate platform-dependent build
11824 directories. The recommended way to make a platform-dependent
11825 build directory is the following (tested on Linux), maybe with
11828 # Place yourself outside of the OpenSSL source tree. In
11829 # this example, the environment variable OPENSSL_SOURCE
11830 # is assumed to contain the absolute OpenSSL source directory.
11831 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11832 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11833 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11834 mkdir -p `dirname $F`
11835 ln -s $OPENSSL_SOURCE/$F $F
11838 To be absolutely sure not to disturb the source tree, a "make clean"
11839 is a good thing. If it isn't successful, don't worry about it,
11840 it probably means the source directory is very clean.
11844 * Make sure any ENGINE control commands make local copies of string
11845 pointers passed to them whenever necessary. Otherwise it is possible
11846 the caller may have overwritten (or deallocated) the original string
11847 data when a later ENGINE operation tries to use the stored values.
11849 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11851 * Improve diagnostics in file reading and command-line digests.
11853 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11855 * Add AES modes CFB and OFB to the object database. Correct an
11856 error in AES-CFB decryption.
11860 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11861 allows existing EVP_CIPHER_CTX structures to be reused after
11862 calling `EVP_*Final()`. This behaviour is used by encryption
11863 BIOs and some applications. This has the side effect that
11864 applications must explicitly clean up cipher contexts with
11865 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11869 * Check the values of dna and dnb in bn_mul_recursive before calling
11870 bn_mul_comba (a non zero value means the a or b arrays do not contain
11871 n2 elements) and fallback to bn_mul_normal if either is not zero.
11875 * Fix escaping of non-ASCII characters when using the -subj option
11876 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11880 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11881 form for "surname", serialNumber has no short form.
11882 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11883 therefore remove "mail" short name for "internet 7".
11884 The OID for unique identifiers in X509 certificates is
11885 x500UniqueIdentifier, not uniqueIdentifier.
11886 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11890 * Add an "init" command to the ENGINE config module and auto initialize
11891 ENGINEs. Without any "init" command the ENGINE will be initialized
11892 after all ctrl commands have been executed on it. If init=1 the
11893 ENGINE is initialized at that point (ctrls before that point are run
11894 on the uninitialized ENGINE and after on the initialized one). If
11895 init=0 then the ENGINE will not be initialized at all.
11899 * Fix the 'app_verify_callback' interface so that the user-defined
11900 argument is actually passed to the callback: In the
11901 SSL_CTX_set_cert_verify_callback() prototype, the callback
11902 declaration has been changed from
11905 int (*cb)(X509_STORE_CTX *,void *);
11906 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11907 i=s->ctx->app_verify_callback(&ctx)
11908 has been changed into
11909 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11911 To update applications using SSL_CTX_set_cert_verify_callback(),
11912 a dummy argument can be added to their callback functions.
11914 *D. K. Smetters <smetters@parc.xerox.com>*
11916 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11918 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11920 * Add and OPENSSL_LOAD_CONF define which will cause
11921 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11922 This allows older applications to transparently support certain
11923 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11924 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11925 load the config file and OPENSSL_add_all_algorithms_conf() which will
11926 always load it have also been added.
11930 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11931 Adjust NIDs and EVP layer.
11933 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11935 * Config modules support in openssl utility.
11937 Most commands now load modules from the config file,
11938 though in a few (such as version) this isn't done
11939 because it couldn't be used for anything.
11941 In the case of ca and req the config file used is
11942 the same as the utility itself: that is the -config
11943 command line option can be used to specify an
11948 * Move default behaviour from OPENSSL_config(). If appname is NULL
11949 use "openssl_conf" if filename is NULL use default openssl config file.
11953 * Add an argument to OPENSSL_config() to allow the use of an alternative
11954 config section name. Add a new flag to tolerate a missing config file
11955 and move code to CONF_modules_load_file().
11959 * Support for crypto accelerator cards from Accelerated Encryption
11960 Processing, www.aep.ie. (Use engine 'aep')
11961 The support was copied from 0.9.6c [engine] and adapted/corrected
11962 to work with the new engine framework.
11964 *AEP Inc. and Richard Levitte*
11966 * Support for SureWare crypto accelerator cards from Baltimore
11967 Technologies. (Use engine 'sureware')
11968 The support was copied from 0.9.6c [engine] and adapted
11969 to work with the new engine framework.
11973 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11974 make the newer ENGINE framework commands for the CHIL engine work.
11976 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11978 * Make it possible to produce shared libraries on ReliantUNIX.
11980 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11982 * Add the configuration target debug-linux-ppro.
11983 Make 'openssl rsa' use the general key loading routines
11984 implemented in `apps.c`, and make those routines able to
11985 handle the key format FORMAT_NETSCAPE and the variant
11988 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11990 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11992 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11994 * Add -keyform to rsautl, and document -engine.
11996 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11998 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11999 BIO_R_NO_SUCH_FILE error code rather than the generic
12000 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12004 * Add new functions
12005 ERR_peek_last_error
12006 ERR_peek_last_error_line
12007 ERR_peek_last_error_line_data.
12008 These are similar to
12010 ERR_peek_error_line
12011 ERR_peek_error_line_data,
12012 but report on the latest error recorded rather than the first one
12013 still in the error queue.
12015 *Ben Laurie, Bodo Moeller*
12017 * default_algorithms option in ENGINE config module. This allows things
12019 default_algorithms = ALL
12020 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12024 * Preliminary ENGINE config module.
12028 * New experimental application configuration code.
12032 * Change the AES code to follow the same name structure as all other
12033 symmetric ciphers, and behave the same way. Move everything to
12034 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12036 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12038 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12040 *Ben Laurie and Theo de Raadt*
12042 * Add option to output public keys in req command.
12044 *Massimiliano Pala madwolf@openca.org*
12046 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12047 (up to about 10% better than before for P-192 and P-224).
12051 * New functions/macros
12053 SSL_CTX_set_msg_callback(ctx, cb)
12054 SSL_CTX_set_msg_callback_arg(ctx, arg)
12055 SSL_set_msg_callback(ssl, cb)
12056 SSL_set_msg_callback_arg(ssl, arg)
12058 to request calling a callback function
12060 void cb(int write_p, int version, int content_type,
12061 const void *buf, size_t len, SSL *ssl, void *arg)
12063 whenever a protocol message has been completely received
12064 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12065 protocol version according to which the SSL library interprets
12066 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12067 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12068 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12069 specification (change_cipher_spec(20), alert(21), handshake(22)).
12070 'buf' and 'len' point to the actual message, 'ssl' to the
12071 SSL object, and 'arg' is the application-defined value set by
12072 SSL[_CTX]_set_msg_callback_arg().
12074 'openssl s_client' and 'openssl s_server' have new '-msg' options
12075 to enable a callback that displays all protocol messages.
12079 * Change the shared library support so shared libraries are built as
12080 soon as the corresponding static library is finished, and thereby get
12081 openssl and the test programs linked against the shared library.
12082 This still only happens when the keyword "shard" has been given to
12083 the configuration scripts.
12085 NOTE: shared library support is still an experimental thing, and
12086 backward binary compatibility is still not guaranteed.
12088 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12090 * Add support for Subject Information Access extension.
12092 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12094 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12095 additional bytes when new memory had to be allocated, not just
12096 when reusing an existing buffer.
12100 * New command line and configuration option 'utf8' for the req command.
12101 This allows field values to be specified as UTF8 strings.
12105 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12106 runs for the former and machine-readable output for the latter.
12110 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12111 of the e-mail address in the DN (i.e., it will go into a certificate
12112 extension only). The new configuration file option 'email_in_dn = no'
12113 has the same effect.
12115 *Massimiliano Pala madwolf@openca.org*
12117 * Change all functions with names starting with `des_` to be starting
12118 with `DES_` instead. Add wrappers that are compatible with libdes,
12119 but are named `_ossl_old_des_*`. Finally, add macros that map the
12120 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12121 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12122 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12125 Since we provide two compatibility mappings, the user needs to
12126 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12127 compatibility is desired. The default (i.e., when that macro
12128 isn't defined) is OpenSSL 0.9.6c compatibility.
12130 There are also macros that enable and disable the support of old
12131 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12132 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12133 are defined, the default will apply: to support the old des routines.
12135 In either case, one must include openssl/des.h to get the correct
12136 definitions. Do not try to just include openssl/des_old.h, that
12139 NOTE: This is a major break of an old API into a new one. Software
12140 authors are encouraged to switch to the `DES_` style functions. Some
12141 time in the future, des_old.h and the libdes compatibility functions
12142 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12143 default), and then completely removed.
12147 * Test for certificates which contain unsupported critical extensions.
12148 If such a certificate is found during a verify operation it is
12149 rejected by default: this behaviour can be overridden by either
12150 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12151 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12152 X509_supported_extension() has also been added which returns 1 if a
12153 particular extension is supported.
12157 * Modify the behaviour of EVP cipher functions in similar way to digests
12158 to retain compatibility with existing code.
12162 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12163 compatibility with existing code. In particular the 'ctx' parameter does
12164 not have to be to be initialized before the call to EVP_DigestInit() and
12165 it is tidied up after a call to EVP_DigestFinal(). New function
12166 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12167 EVP_MD_CTX_copy() changed to not require the destination to be
12168 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12169 requires the destination to be valid.
12171 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12172 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12176 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12177 so that complete 'Handshake' protocol structures are kept in memory
12178 instead of overwriting 'msg_type' and 'length' with 'body' data.
12182 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12184 *Massimo Santin via Richard Levitte*
12186 * Major restructuring to the underlying ENGINE code. This includes
12187 reduction of linker bloat, separation of pure "ENGINE" manipulation
12188 (initialisation, etc) from functionality dealing with implementations
12189 of specific crypto interfaces. This change also introduces integrated
12190 support for symmetric ciphers and digest implementations - so ENGINEs
12191 can now accelerate these by providing EVP_CIPHER and EVP_MD
12192 implementations of their own. This is detailed in
12193 [crypto/engine/README.md](crypto/engine/README.md)
12194 as it couldn't be adequately described here. However, there are a few
12195 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12196 were changed in the original introduction of ENGINE code have now
12197 reverted back - the hooking from this code to ENGINE is now a good
12198 deal more passive and at run-time, operations deal directly with
12199 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12200 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12201 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12202 they were not being used by the framework as there is no concept of a
12203 BIGNUM_METHOD and they could not be generalised to the new
12204 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12205 ENGINE_cpy() has been removed as it cannot be consistently defined in
12210 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12214 * Change mkdef.pl to sort symbols that get the same entry number,
12215 and make sure the automatically generated functions `ERR_load_*`
12216 become part of libeay.num as well.
12220 * New function SSL_renegotiate_pending(). This returns true once
12221 renegotiation has been requested (either SSL_renegotiate() call
12222 or HelloRequest/ClientHello received from the peer) and becomes
12223 false once a handshake has been completed.
12224 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12225 sends a HelloRequest, but does not ensure that a handshake takes
12226 place. SSL_renegotiate_pending() is useful for checking if the
12227 client has followed the request.)
12231 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12232 By default, clients may request session resumption even during
12233 renegotiation (if session ID contexts permit); with this option,
12234 session resumption is possible only in the first handshake.
12236 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12237 more bits available for options that should not be part of
12238 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12242 * Add some demos for certificate and certificate request creation.
12246 * Make maximum certificate chain size accepted from the peer application
12247 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12248 "Douglas E. Engert" <deengert@anl.gov>.
12252 * Add support for shared libraries for Unixware-7
12253 (Boyd Lynn Gerber <gerberb@zenez.com>).
12257 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12258 be done prior to destruction. Use this to unload error strings from
12259 ENGINEs that load their own error strings. NB: This adds two new API
12260 functions to "get" and "set" this destroy handler in an ENGINE.
12264 * Alter all existing ENGINE implementations (except "openssl" and
12265 "openbsd") to dynamically instantiate their own error strings. This
12266 makes them more flexible to be built both as statically-linked ENGINEs
12267 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12268 Also, add stub code to each that makes building them as self-contained
12269 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12273 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12274 implementations into applications that are completely implemented in
12275 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12276 commands that can be used to configure what shared-library to load and
12277 to control aspects of the way it is handled. Also, made an update to
12278 the [README-Engine.md](README-Engine.md) file
12279 that brings its information up-to-date and
12280 provides some information and instructions on the "dynamic" ENGINE
12281 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12285 * Make it possible to unload ranges of ERR strings with a new
12286 "ERR_unload_strings" function.
12290 * Add a copy() function to EVP_MD.
12294 * Make EVP_MD routines take a context pointer instead of just the
12295 md_data void pointer.
12299 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12300 that the digest can only process a single chunk of data
12301 (typically because it is provided by a piece of
12302 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12303 is only going to provide a single chunk of data, and hence the
12304 framework needn't accumulate the data for oneshot drivers.
12308 * As with "ERR", make it possible to replace the underlying "ex_data"
12309 functions. This change also alters the storage and management of global
12310 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12311 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12312 index counters. The API functions that use this state have been changed
12313 to take a "class_index" rather than pointers to the class's local STACK
12314 and counter, and there is now an API function to dynamically create new
12315 classes. This centralisation allows us to (a) plug a lot of the
12316 thread-safety problems that existed, and (b) makes it possible to clean
12317 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12318 such data would previously have always leaked in application code and
12319 workarounds were in place to make the memory debugging turn a blind eye
12320 to it. Application code that doesn't use this new function will still
12321 leak as before, but their memory debugging output will announce it now
12322 rather than letting it slide.
12324 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12325 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12326 has a return value to indicate success or failure.
12330 * Make it possible to replace the underlying "ERR" functions such that the
12331 global state (2 LHASH tables and 2 locks) is only used by the "default"
12332 implementation. This change also adds two functions to "get" and "set"
12333 the implementation prior to it being automatically set the first time
12334 any other ERR function takes place. Ie. an application can call "get",
12335 pass the return value to a module it has just loaded, and that module
12336 can call its own "set" function using that value. This means the
12337 module's "ERR" operations will use (and modify) the error state in the
12338 application and not in its own statically linked copy of OpenSSL code.
12342 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12343 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12344 the operation, and provides a more encapsulated way for external code
12345 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12346 to use these functions rather than manually incrementing the counts.
12348 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12352 * Add EVP test program.
12356 * Add symmetric cipher support to ENGINE. Expect the API to change!
12360 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12361 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12362 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12363 These allow a CRL to be built without having to access X509_CRL fields
12364 directly. Modify 'ca' application to use new functions.
12368 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12369 bug workarounds. Rollback attack detection is a security feature.
12370 The problem will only arise on OpenSSL servers when TLSv1 is not
12371 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12372 Software authors not wanting to support TLSv1 will have special reasons
12373 for their choice and can explicitly enable this option.
12375 *Bodo Moeller, Lutz Jaenicke*
12377 * Rationalise EVP so it can be extended: don't include a union of
12378 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12379 (similar to those existing for EVP_CIPHER_CTX).
12384 EVP_MD_CTX_init(&md); /* new function call */
12385 EVP_DigestInit(&md, EVP_sha1());
12386 EVP_DigestUpdate(&md, in, len);
12387 EVP_DigestFinal(&md, out, NULL);
12388 EVP_MD_CTX_cleanup(&md); /* new function call */
12392 * Make DES key schedule conform to the usual scheme, as well as
12393 correcting its structure. This means that calls to DES functions
12394 now have to pass a pointer to a des_key_schedule instead of a
12395 plain des_key_schedule (which was actually always a pointer
12398 des_key_schedule ks;
12400 des_set_key_checked(..., &ks);
12401 des_ncbc_encrypt(..., &ks, ...);
12403 (Note that a later change renames 'des_...' into 'DES_...'.)
12407 * Initial reduction of linker bloat: the use of some functions, such as
12408 PEM causes large amounts of unused functions to be linked in due to
12409 poor organisation. For example pem_all.c contains every PEM function
12410 which has a knock on effect of linking in large amounts of (unused)
12411 ASN1 code. Grouping together similar functions and splitting unrelated
12412 functions prevents this.
12416 * Cleanup of EVP macros.
12420 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12421 correct `_ecb suffix`.
12425 * Add initial OCSP responder support to ocsp application. The
12426 revocation information is handled using the text based index
12427 use by the ca application. The responder can either handle
12428 requests generated internally, supplied in files (for example
12429 via a CGI script) or using an internal minimal server.
12433 * Add configuration choices to get zlib compression for TLS.
12437 * Changes to Kerberos SSL for RFC 2712 compliance:
12438 1. Implemented real KerberosWrapper, instead of just using
12439 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12440 2. Implemented optional authenticator field of KerberosWrapper.
12442 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12443 and authenticator structs; see crypto/krb5/.
12445 Generalized Kerberos calls to support multiple Kerberos libraries.
12446 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12447 via Richard Levitte*
12449 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12450 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12451 values for each of the key sizes rather than having just
12452 parameters (and 'speed' generating keys each time).
12456 * Speed up EVP routines.
12459 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12460 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12461 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12462 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12464 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12465 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12466 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12469 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12471 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12475 * Added the OS2-EMX target.
12477 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12479 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12480 New functions to support `NCONF` routines in extension code.
12481 New function `CONF_set_nconf()`
12482 to allow functions which take an `NCONF` to also handle the old `LHASH`
12483 structure: this means that the old `CONF` compatible routines can be
12484 retained (in particular w.rt. extensions) without having to duplicate the
12485 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12489 * Enhance the general user interface with mechanisms for inner control
12490 and with possibilities to have yes/no kind of prompts.
12494 * Change all calls to low-level digest routines in the library and
12495 applications to use EVP. Add missing calls to HMAC_cleanup() and
12496 don't assume HMAC_CTX can be copied using memcpy().
12498 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12500 * Add the possibility to control engines through control names but with
12501 arbitrary arguments instead of just a string.
12502 Change the key loaders to take a UI_METHOD instead of a callback
12503 function pointer. NOTE: this breaks binary compatibility with earlier
12504 versions of OpenSSL [engine].
12505 Adapt the nCipher code for these new conditions and add a card insertion
12510 * Enhance the general user interface with mechanisms to better support
12511 dialog box interfaces, application-defined prompts, the possibility
12512 to use defaults (for example default passwords from somewhere else)
12513 and interrupts/cancellations.
12517 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12518 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12522 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12523 tidy up some unnecessarily weird code in 'sk_new()').
12525 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12527 * Change the key loading routines for ENGINEs to use the same kind
12528 callback (pem_password_cb) as all other routines that need this
12533 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12534 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12535 than this minimum value is recommended.
12539 * New random seeder for OpenVMS, using the system process statistics
12540 that are easily reachable.
12544 * Windows apparently can't transparently handle global
12545 variables defined in DLLs. Initialisations such as:
12547 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12549 won't compile. This is used by the any applications that need to
12550 declare their own ASN1 modules. This was fixed by adding the option
12551 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12552 needed for static libraries under Win32.
12556 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12557 setting of purpose and trust fields. New X509_STORE trust and
12558 purpose functions and tidy up setting in other SSL functions.
12562 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12563 structure. These are inherited by X509_STORE_CTX when it is
12564 initialised. This allows various defaults to be set in the
12565 X509_STORE structure (such as flags for CRL checking and custom
12566 purpose or trust settings) for functions which only use X509_STORE_CTX
12567 internally such as S/MIME.
12569 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12570 trust settings if they are not set in X509_STORE. This allows X509_STORE
12571 purposes and trust (in S/MIME for example) to override any set by default.
12573 Add command line options for CRL checking to smime, s_client and s_server
12578 * Initial CRL based revocation checking. If the CRL checking flag(s)
12579 are set then the CRL is looked up in the X509_STORE structure and
12580 its validity and signature checked, then if the certificate is found
12581 in the CRL the verify fails with a revoked error.
12583 Various new CRL related callbacks added to X509_STORE_CTX structure.
12585 Command line options added to 'verify' application to support this.
12587 This needs some additional work, such as being able to handle multiple
12588 CRLs with different times, extension based lookup (rather than just
12589 by subject name) and ultimately more complete V2 CRL extension
12594 * Add a general user interface API (crypto/ui/). This is designed
12595 to replace things like des_read_password and friends (backward
12596 compatibility functions using this new API are provided).
12597 The purpose is to remove prompting functions from the DES code
12598 section as well as provide for prompting through dialog boxes in
12599 a window system and the like.
12603 * Add "ex_data" support to ENGINE so implementations can add state at a
12604 per-structure level rather than having to store it globally.
12608 * Make it possible for ENGINE structures to be copied when retrieved by
12609 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12610 This causes the "original" ENGINE structure to act like a template,
12611 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12612 operational state can be localised to each ENGINE structure, despite the
12613 fact they all share the same "methods". New ENGINE structures returned in
12614 this case have no functional references and the return value is the single
12615 structural reference. This matches the single structural reference returned
12616 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12621 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12622 needs to match any other type at all we need to manually clear the
12627 * Changes to the "openssl engine" utility to include;
12628 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12629 about an ENGINE's available control commands.
12630 - executing control commands from command line arguments using the
12631 '-pre' and '-post' switches. '-post' is only used if '-t' is
12632 specified and the ENGINE is successfully initialised. The syntax for
12633 the individual commands are colon-separated, for example;
12634 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12638 * New dynamic control command support for ENGINEs. ENGINEs can now
12639 declare their own commands (numbers), names (strings), descriptions,
12640 and input types for run-time discovery by calling applications. A
12641 subset of these commands are implicitly classed as "executable"
12642 depending on their input type, and only these can be invoked through
12643 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12644 can be based on user input, config files, etc). The distinction is
12645 that "executable" commands cannot return anything other than a boolean
12646 result and can only support numeric or string input, whereas some
12647 discoverable commands may only be for direct use through
12648 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12649 pointers, or other custom uses. The "executable" commands are to
12650 support parameterisations of ENGINE behaviour that can be
12651 unambiguously defined by ENGINEs and used consistently across any
12652 OpenSSL-based application. Commands have been added to all the
12653 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12654 control over shared-library paths without source code alterations.
12658 * Changed all ENGINE implementations to dynamically allocate their
12659 ENGINEs rather than declaring them statically. Apart from this being
12660 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12661 this also allows the implementations to compile without using the
12662 internal engine_int.h header.
12666 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12667 'const' value. Any code that should be able to modify a RAND_METHOD
12668 should already have non-const pointers to it (ie. they should only
12669 modify their own ones).
12673 * Made a variety of little tweaks to the ENGINE code.
12674 - "atalla" and "ubsec" string definitions were moved from header files
12675 to C code. "nuron" string definitions were placed in variables
12676 rather than hard-coded - allowing parameterisation of these values
12677 later on via ctrl() commands.
12678 - Removed unused "#if 0"'d code.
12679 - Fixed engine list iteration code so it uses ENGINE_free() to release
12680 structural references.
12681 - Constified the RAND_METHOD element of ENGINE structures.
12682 - Constified various get/set functions as appropriate and added
12683 missing functions (including a catch-all ENGINE_cpy that duplicates
12684 all ENGINE values onto a new ENGINE except reference counts/state).
12685 - Removed NULL parameter checks in get/set functions. Setting a method
12686 or function to NULL is a way of cancelling out a previously set
12687 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12688 and doesn't justify the extra error symbols and code.
12689 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12690 flags from engine_int.h to engine.h.
12691 - Changed prototypes for ENGINE handler functions (init(), finish(),
12692 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12696 * Implement binary inversion algorithm for BN_mod_inverse in addition
12697 to the algorithm using long division. The binary algorithm can be
12698 used only if the modulus is odd. On 32-bit systems, it is faster
12699 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12700 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12701 up to 450 bits. In 64-bit environments, the binary algorithm
12702 appears to be advantageous for much longer moduli; here we use it
12703 for moduli up to 2048 bits.
12707 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12708 could not support the combine flag in choice fields.
12712 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12713 extensions from a certificate request to the certificate.
12717 * Allow multiple 'certopt' and 'nameopt' options to be separated
12718 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12719 file: this allows the display of the certificate about to be
12720 signed to be customised, to allow certain fields to be included
12721 or excluded and extension details. The old system didn't display
12722 multicharacter strings properly, omitted fields not in the policy
12723 and couldn't display additional details such as extensions.
12727 * Function EC_POINTs_mul for multiple scalar multiplication
12728 of an arbitrary number of elliptic curve points
12729 \sum scalars[i]*points[i],
12730 optionally including the generator defined for the EC_GROUP:
12731 scalar*generator + \sum scalars[i]*points[i].
12733 EC_POINT_mul is a simple wrapper function for the typical case
12734 that the point list has just one item (besides the optional
12739 * First EC_METHODs for curves over GF(p):
12741 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12742 operations and provides various method functions that can also
12743 operate with faster implementations of modular arithmetic.
12745 EC_GFp_mont_method() reuses most functions that are part of
12746 EC_GFp_simple_method, but uses Montgomery arithmetic.
12748 *Bodo Moeller; point addition and point doubling
12749 implementation directly derived from source code provided by
12750 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12752 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12753 crypto/ec/ec_lib.c):
12755 Curves are EC_GROUP objects (with an optional group generator)
12756 based on EC_METHODs that are built into the library.
12758 Points are EC_POINT objects based on EC_GROUP objects.
12760 Most of the framework would be able to handle curves over arbitrary
12761 finite fields, but as there are no obvious types for fields other
12762 than GF(p), some functions are limited to that for now.
12766 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12767 that the file contains a complete HTTP response.
12771 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12772 change the def and num file printf format specifier from "%-40sXXX"
12773 to "%-39s XXX". The latter will always guarantee a space after the
12774 field while the former will cause them to run together if the field
12775 is 40 of more characters long.
12779 * Constify the cipher and digest 'method' functions and structures
12780 and modify related functions to take constant EVP_MD and EVP_CIPHER
12785 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12786 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12790 * Modify `EVP_Digest*()` routines so they now return values. Although the
12791 internal software routines can never fail additional hardware versions
12796 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12798 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12799 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12802 ERR_R_NESTED_ASN1_ERROR
12804 ERR_R_MISSING_ASN1_EOS
12805 were 4 .. 9, conflicting with
12806 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12808 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12809 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12811 Add new error code 'ERR_R_INTERNAL_ERROR'.
12815 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12820 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12821 sets the subject name for a new request or supersedes the
12822 subject name in a given request. Formats that can be parsed are
12823 'CN=Some Name, OU=myOU, C=IT'
12825 'CN=Some Name/OU=myOU/C=IT'.
12827 Add options '-batch' and '-verbose' to 'openssl req'.
12829 *Massimiliano Pala <madwolf@hackmasters.net>*
12831 * Introduce the possibility to access global variables through
12832 functions on platform were that's the best way to handle exporting
12833 global variables in shared libraries. To enable this functionality,
12834 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12835 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12836 is normally done by Configure or something similar).
12838 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12839 in the source file (foo.c) like this:
12841 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12842 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12844 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12845 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12847 OPENSSL_DECLARE_GLOBAL(int,foo);
12848 #define foo OPENSSL_GLOBAL_REF(foo)
12849 OPENSSL_DECLARE_GLOBAL(double,bar);
12850 #define bar OPENSSL_GLOBAL_REF(bar)
12852 The #defines are very important, and therefore so is including the
12853 header file everywhere where the defined globals are used.
12855 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12856 of ASN.1 items, but that structure is a bit different.
12858 The largest change is in util/mkdef.pl which has been enhanced with
12859 better and easier to understand logic to choose which symbols should
12860 go into the Windows .def files as well as a number of fixes and code
12861 cleanup (among others, algorithm keywords are now sorted
12862 lexicographically to avoid constant rewrites).
12866 * In BN_div() keep a copy of the sign of 'num' before writing the
12867 result to 'rm' because if rm==num the value will be overwritten
12868 and produce the wrong result if 'num' is negative: this caused
12869 problems with BN_mod() and BN_nnmod().
12873 * Function OCSP_request_verify(). This checks the signature on an
12874 OCSP request and verifies the signer certificate. The signer
12875 certificate is just checked for a generic purpose and OCSP request
12880 * Add OCSP_check_validity() function to check the validity of OCSP
12881 responses. OCSP responses are prepared in real time and may only
12882 be a few seconds old. Simply checking that the current time lies
12883 between thisUpdate and nextUpdate max reject otherwise valid responses
12884 caused by either OCSP responder or client clock inaccuracy. Instead
12885 we allow thisUpdate and nextUpdate to fall within a certain period of
12886 the current time. The age of the response can also optionally be
12887 checked. Two new options -validity_period and -status_age added to
12892 * If signature or public key algorithm is unrecognized print out its
12893 OID rather that just UNKNOWN.
12897 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12898 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12899 ID to be generated from the issuer certificate alone which can then be
12900 passed to OCSP_id_issuer_cmp().
12904 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12905 ASN1 modules to export functions returning ASN1_ITEM pointers
12906 instead of the ASN1_ITEM structures themselves. This adds several
12907 new macros which allow the underlying ASN1 function/structure to
12908 be accessed transparently. As a result code should not use ASN1_ITEM
12909 references directly (such as &X509_it) but instead use the relevant
12910 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12911 use of the new ASN1 code on platforms where exporting structures
12912 is problematical (for example in shared libraries) but exporting
12913 functions returning pointers to structures is not.
12917 * Add support for overriding the generation of SSL/TLS session IDs.
12918 These callbacks can be registered either in an SSL_CTX or per SSL.
12919 The purpose of this is to allow applications to control, if they wish,
12920 the arbitrary values chosen for use as session IDs, particularly as it
12921 can be useful for session caching in multiple-server environments. A
12922 command-line switch for testing this (and any client code that wishes
12923 to use such a feature) has been added to "s_server".
12925 *Geoff Thorpe, Lutz Jaenicke*
12927 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12928 of the form `#if defined(...) || defined(...) || ...` and
12929 `#if !defined(...) && !defined(...) && ...`. This also avoids
12930 the growing number of special cases it was previously handling.
12934 * Make all configuration macros available for application by making
12935 sure they are available in opensslconf.h, by giving them names starting
12936 with `OPENSSL_` to avoid conflicts with other packages and by making
12937 sure e_os2.h will cover all platform-specific cases together with
12939 Additionally, it is now possible to define configuration/platform-
12940 specific names (called "system identities"). In the C code, these
12941 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12942 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12943 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12948 * New option -set_serial to 'req' and 'x509' this allows the serial
12949 number to use to be specified on the command line. Previously self
12950 signed certificates were hard coded with serial number 0 and the
12951 CA options of 'x509' had to use a serial number in a file which was
12956 * New options to 'ca' utility to support V2 CRL entry extensions.
12957 Currently CRL reason, invalidity date and hold instruction are
12958 supported. Add new CRL extensions to V3 code and some new objects.
12962 * New function EVP_CIPHER_CTX_set_padding() this is used to
12963 disable standard block padding (aka PKCS#5 padding) in the EVP
12964 API, which was previously mandatory. This means that the data is
12965 not padded in any way and so the total length much be a multiple
12966 of the block size, otherwise an error occurs.
12970 * Initial (incomplete) OCSP SSL support.
12974 * New function OCSP_parse_url(). This splits up a URL into its host,
12975 port and path components: primarily to parse OCSP URLs. New -url
12976 option to ocsp utility.
12980 * New nonce behavior. The return value of OCSP_check_nonce() now
12981 reflects the various checks performed. Applications can decide
12982 whether to tolerate certain situations such as an absent nonce
12983 in a response when one was present in a request: the ocsp application
12984 just prints out a warning. New function OCSP_add1_basic_nonce()
12985 this is to allow responders to include a nonce in a response even if
12986 the request is nonce-less.
12990 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12991 skipped when using openssl x509 multiple times on a single input file,
12992 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12996 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12997 set string type: to handle setting ASN1_TIME structures. Fix ca
12998 utility to correctly initialize revocation date of CRLs.
13002 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13003 the clients preferred ciphersuites and rather use its own preferences.
13004 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13005 Internet Explorer by ensuring unchanged hash method during stepup.
13006 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13010 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13011 to aes and add a new 'exist' option to print out symbols that don't
13016 * Additional options to ocsp utility to allow flags to be set and
13017 additional certificates supplied.
13021 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13022 OCSP client a number of certificate to only verify the response
13027 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13028 handle the new API. Currently only ECB, CBC modes supported. Add new
13031 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13032 Encryption Standard (AES) Ciphersuites for Transport Layer
13033 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13034 not enabled by default and were not part of the "ALL" ciphersuite
13035 alias because they were not yet official; they could be
13036 explicitly requested by specifying the "AESdraft" ciphersuite
13037 group alias. In the final release of OpenSSL 0.9.7, the group
13038 alias is called "AES" and is part of "ALL".)
13040 *Ben Laurie, Steve Henson, Bodo Moeller*
13042 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13043 request to response.
13047 * Functions for OCSP responders. OCSP_request_onereq_count(),
13048 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13049 extract information from a certificate request. OCSP_response_create()
13050 creates a response and optionally adds a basic response structure.
13051 OCSP_basic_add1_status() adds a complete single response to a basic
13052 response and returns the OCSP_SINGLERESP structure just added (to allow
13053 extensions to be included for example). OCSP_basic_add1_cert() adds a
13054 certificate to a basic response and OCSP_basic_sign() signs a basic
13055 response with various flags. New helper functions ASN1_TIME_check()
13056 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13057 (converts ASN1_TIME to GeneralizedTime).
13061 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13062 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13063 structure from a certificate. X509_pubkey_digest() digests the public_key
13064 contents: this is used in various key identifiers.
13068 * Make sk_sort() tolerate a NULL argument.
13070 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13072 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13073 passed by the function are trusted implicitly. If any of them signed the
13074 response then it is assumed to be valid and is not verified.
13078 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13079 to data. This was previously part of the PKCS7 ASN1 code. This
13080 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13081 *Steve Henson, reported by Kenneth R. Robinette
13082 <support@securenetterm.com>*
13084 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13085 routines: without these tracing memory leaks is very painful.
13086 Fix leaks in PKCS12 and PKCS7 routines.
13090 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13091 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13092 effectively meant GeneralizedTime would never be used. Now it
13093 is initialised to -1 but X509_time_adj() now has to check the value
13094 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13095 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13096 *Steve Henson, reported by Kenneth R. Robinette
13097 <support@securenetterm.com>*
13099 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13100 result in a zero length in the ASN1_INTEGER structure which was
13101 not consistent with the structure when d2i_ASN1_INTEGER() was used
13102 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13103 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13104 where it did not print out a minus for negative ASN1_INTEGER.
13108 * Add summary printout to ocsp utility. The various functions which
13109 convert status values to strings have been renamed to:
13110 OCSP_response_status_str(), OCSP_cert_status_str() and
13111 OCSP_crl_reason_str() and are no longer static. New options
13112 to verify nonce values and to disable verification. OCSP response
13113 printout format cleaned up.
13117 * Add additional OCSP certificate checks. These are those specified
13118 in RFC2560. This consists of two separate checks: the CA of the
13119 certificate being checked must either be the OCSP signer certificate
13120 or the issuer of the OCSP signer certificate. In the latter case the
13121 OCSP signer certificate must contain the OCSP signing extended key
13122 usage. This check is performed by attempting to match the OCSP
13123 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13124 in the OCSP_CERTID structures of the response.
13128 * Initial OCSP certificate verification added to OCSP_basic_verify()
13129 and related routines. This uses the standard OpenSSL certificate
13130 verify routines to perform initial checks (just CA validity) and
13131 to obtain the certificate chain. Then additional checks will be
13132 performed on the chain. Currently the root CA is checked to see
13133 if it is explicitly trusted for OCSP signing. This is used to set
13134 a root CA as a global signing root: that is any certificate that
13135 chains to that CA is an acceptable OCSP signing certificate.
13139 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13140 extensions from a separate configuration file.
13141 As when reading extensions from the main configuration file,
13142 the '-extensions ...' option may be used for specifying the
13145 *Massimiliano Pala <madwolf@comune.modena.it>*
13147 * New OCSP utility. Allows OCSP requests to be generated or
13148 read. The request can be sent to a responder and the output
13149 parsed, outputted or printed in text form. Not complete yet:
13150 still needs to check the OCSP response validity.
13154 * New subcommands for 'openssl ca':
13155 `openssl ca -status <serial>` prints the status of the cert with
13156 the given serial number (according to the index file).
13157 `openssl ca -updatedb` updates the expiry status of certificates
13160 *Massimiliano Pala <madwolf@comune.modena.it>*
13162 * New '-newreq-nodes' command option to CA.pl. This is like
13163 '-newreq', but calls 'openssl req' with the '-nodes' option
13164 so that the resulting key is not encrypted.
13166 *Damien Miller <djm@mindrot.org>*
13168 * New configuration for the GNU Hurd.
13170 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13172 * Initial code to implement OCSP basic response verify. This
13173 is currently incomplete. Currently just finds the signer's
13174 certificate and verifies the signature on the response.
13178 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13179 value of OPENSSLDIR. This is available via the new '-d' option
13180 to 'openssl version', and is also included in 'openssl version -a'.
13184 * Allowing defining memory allocation callbacks that will be given
13185 file name and line number information in additional arguments
13186 (a `const char*` and an int). The basic functionality remains, as
13187 well as the original possibility to just replace malloc(),
13188 realloc() and free() by functions that do not know about these
13189 additional arguments. To register and find out the current
13190 settings for extended allocation functions, the following
13191 functions are provided:
13193 CRYPTO_set_mem_ex_functions
13194 CRYPTO_set_locked_mem_ex_functions
13195 CRYPTO_get_mem_ex_functions
13196 CRYPTO_get_locked_mem_ex_functions
13198 These work the same way as CRYPTO_set_mem_functions and friends.
13199 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13200 extended allocation function is enabled.
13201 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13202 a conventional allocation function is enabled.
13204 *Richard Levitte, Bodo Moeller*
13206 * Finish off removing the remaining LHASH function pointer casts.
13207 There should no longer be any prototype-casting required when using
13208 the LHASH abstraction, and any casts that remain are "bugs". See
13209 the callback types and macros at the head of lhash.h for details
13210 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13214 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13215 If /dev/[u]random devices are not available or do not return enough
13216 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13218 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13219 /etc/entropy will be queried once each in this sequence, querying stops
13220 when enough entropy was collected without querying more sockets.
13224 * Change the Unix RAND_poll() variant to be able to poll several
13225 random devices, as specified by DEVRANDOM, until a sufficient amount
13226 of data has been collected. We spend at most 10 ms on each file
13227 (select timeout) and read in non-blocking mode. DEVRANDOM now
13228 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13229 (previously it was just the string "/dev/urandom"), so on typical
13230 platforms the 10 ms delay will never occur.
13231 Also separate out the Unix variant to its own file, rand_unix.c.
13232 For VMS, there's a currently-empty rand_vms.c.
13236 * Move OCSP client related routines to ocsp_cl.c. These
13237 provide utility functions which an application needing
13238 to issue a request to an OCSP responder and analyse the
13239 response will typically need: as opposed to those which an
13240 OCSP responder itself would need which will be added later.
13242 OCSP_request_sign() signs an OCSP request with an API similar
13243 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13244 response. OCSP_response_get1_basic() extracts basic response
13245 from response. OCSP_resp_find_status(): finds and extracts status
13246 information from an OCSP_CERTID structure (which will be created
13247 when the request structure is built). These are built from lower
13248 level functions which work on OCSP_SINGLERESP structures but
13249 won't normally be used unless the application wishes to examine
13250 extensions in the OCSP response for example.
13252 Replace nonce routines with a pair of functions.
13253 OCSP_request_add1_nonce() adds a nonce value and optionally
13254 generates a random value. OCSP_check_nonce() checks the
13255 validity of the nonce in an OCSP response.
13259 * Change function OCSP_request_add() to OCSP_request_add0_id().
13260 This doesn't copy the supplied OCSP_CERTID and avoids the
13261 need to free up the newly created id. Change return type
13262 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13263 This can then be used to add extensions to the request.
13264 Deleted OCSP_request_new(), since most of its functionality
13265 is now in OCSP_REQUEST_new() (and the case insensitive name
13266 clash) apart from the ability to set the request name which
13267 will be added elsewhere.
13271 * Update OCSP API. Remove obsolete extensions argument from
13272 various functions. Extensions are now handled using the new
13273 OCSP extension code. New simple OCSP HTTP function which
13274 can be used to send requests and parse the response.
13278 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13279 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13280 uses the special reorder version of SET OF to sort the attributes
13281 and reorder them to match the encoded order. This resolves a long
13282 standing problem: a verify on a PKCS7 structure just after signing
13283 it used to fail because the attribute order did not match the
13284 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13285 it uses the received order. This is necessary to tolerate some broken
13286 software that does not order SET OF. This is handled by encoding
13287 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13288 to produce the required SET OF.
13292 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13293 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13294 files to get correct declarations of the ASN.1 item variables.
13298 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13299 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13300 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13301 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13302 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13303 ASN1_ITEM and no wrapper functions.
13307 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13308 replace the old function pointer based I/O routines. Change most of
13309 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13313 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13314 lines, recognize more "algorithms" that can be deselected, and make
13315 it complain about algorithm deselection that isn't recognised.
13319 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13320 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13321 to use new functions. Add NO_ASN1_OLD which can be set to remove
13322 some old style ASN1 functions: this can be used to determine if old
13323 code will still work when these eventually go away.
13327 * New extension functions for OCSP structures, these follow the
13328 same conventions as certificates and CRLs.
13332 * New function X509V3_add1_i2d(). This automatically encodes and
13333 adds an extension. Its behaviour can be customised with various
13334 flags to append, replace or delete. Various wrappers added for
13335 certificates and CRLs.
13339 * Fix to avoid calling the underlying ASN1 print routine when
13340 an extension cannot be parsed. Correct a typo in the
13341 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13345 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13346 entries for variables.
13350 * Add functionality to `apps/openssl.c` for detecting locking
13351 problems: As the program is single-threaded, all we have
13352 to do is register a locking callback using an array for
13353 storing which locks are currently held by the program.
13357 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13358 SSL_get_ex_data_X509_STORE_idx(), which is used in
13359 ssl_verify_cert_chain() and thus can be called at any time
13360 during TLS/SSL handshakes so that thread-safety is essential.
13361 Unfortunately, the ex_data design is not at all suited
13362 for multi-threaded use, so it probably should be abolished.
13366 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13368 *Broadcom, tweaked and integrated by Geoff Thorpe*
13370 * Move common extension printing code to new function
13371 X509V3_print_extensions(). Reorganise OCSP print routines and
13372 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13376 * New function X509_signature_print() to remove duplication in some
13381 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13382 set (this was treated exactly the same as SET OF previously). This
13383 is used to reorder the STACK representing the structure to match the
13384 encoding. This will be used to get round a problem where a PKCS7
13385 structure which was signed could not be verified because the STACK
13386 order did not reflect the encoded order.
13390 * Reimplement the OCSP ASN1 module using the new code.
13394 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13395 for its ASN1 operations. The old style function pointers still exist
13396 for now but they will eventually go away.
13400 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13401 completely replaces the old ASN1 functionality with a table driven
13402 encoder and decoder which interprets an ASN1_ITEM structure describing
13403 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13404 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13405 has also been converted to the new form.
13409 * Change BN_mod_exp_recp so that negative moduli are tolerated
13410 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13411 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13412 for negative moduli.
13416 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13417 of not touching the result's sign bit.
13421 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13426 * Changed the LHASH code to use prototypes for callbacks, and created
13427 macros to declare and implement thin (optionally static) functions
13428 that provide type-safety and avoid function pointer casting for the
13429 type-specific callbacks.
13433 * Added Kerberos Cipher Suites to be used with TLS, as written in
13435 *Veers Staats <staatsvr@asc.hpc.mil>,
13436 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13438 * Reformat the FAQ so the different questions and answers can be divided
13439 in sections depending on the subject.
13443 * Have the zlib compression code load ZLIB.DLL dynamically under
13448 * New function BN_mod_sqrt for computing square roots modulo a prime
13449 (using the probabilistic Tonelli-Shanks algorithm unless
13450 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13451 be handled deterministically).
13453 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13455 * Make BN_mod_inverse faster by explicitly handling small quotients
13456 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13457 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13461 * New function BN_kronecker.
13465 * Fix BN_gcd so that it works on negative inputs; the result is
13466 positive unless both parameters are zero.
13467 Previously something reasonably close to an infinite loop was
13468 possible because numbers could be growing instead of shrinking
13469 in the implementation of Euclid's algorithm.
13473 * Fix BN_is_word() and BN_is_one() macros to take into account the
13474 sign of the number in question.
13476 Fix BN_is_word(a,w) to work correctly for w == 0.
13478 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13479 because its test if the absolute value of 'a' equals 'w'.
13480 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13481 it exists mostly for use in the implementations of BN_is_zero(),
13482 BN_is_one(), and BN_is_word().
13486 * New function BN_swap.
13490 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13491 the exponentiation functions are more likely to produce reasonable
13492 results on negative inputs.
13496 * Change BN_mod_mul so that the result is always non-negative.
13497 Previously, it could be negative if one of the factors was negative;
13498 I don't think anyone really wanted that behaviour.
13502 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13503 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13504 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13505 and add new functions:
13514 BN_mod_lshift1_quick
13516 BN_mod_lshift_quick
13518 These functions always generate non-negative results.
13520 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13521 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13523 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13524 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13525 be reduced modulo `m`.
13527 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13530 The following entry accidentally appeared in the CHANGES file
13531 distributed with OpenSSL 0.9.7. The modifications described in
13532 it do *not* apply to OpenSSL 0.9.7.
13534 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13535 was actually never needed) and in BN_mul(). The removal in BN_mul()
13536 required a small change in bn_mul_part_recursive() and the addition
13537 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13538 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13539 bn_sub_words() and bn_add_words() except they take arrays with
13545 * In 'openssl passwd', verify passwords read from the terminal
13546 unless the '-salt' option is used (which usually means that
13547 verification would just waste user's time since the resulting
13548 hash is going to be compared with some given password hash)
13549 or the new '-noverify' option is used.
13551 This is an incompatible change, but it does not affect
13552 non-interactive use of 'openssl passwd' (passwords on the command
13553 line, '-stdin' option, '-in ...' option) and thus should not
13554 cause any problems.
13558 * Remove all references to RSAref, since there's no more need for it.
13562 * Make DSO load along a path given through an environment variable
13563 (SHLIB_PATH) with shl_load().
13567 * Constify the ENGINE code as a result of BIGNUM constification.
13568 Also constify the RSA code and most things related to it. In a
13569 few places, most notable in the depth of the ASN.1 code, ugly
13570 casts back to non-const were required (to be solved at a later
13575 * Make it so the openssl application has all engines loaded by default.
13579 * Constify the BIGNUM routines a little more.
13583 * Add the following functions:
13585 ENGINE_load_cswift()
13587 ENGINE_load_atalla()
13588 ENGINE_load_nuron()
13589 ENGINE_load_builtin_engines()
13591 That way, an application can itself choose if external engines that
13592 are built-in in OpenSSL shall ever be used or not. The benefit is
13593 that applications won't have to be linked with libdl or other dso
13594 libraries unless it's really needed.
13596 Changed 'openssl engine' to load all engines on demand.
13597 Changed the engine header files to avoid the duplication of some
13598 declarations (they differed!).
13602 * 'openssl engine' can now list capabilities.
13606 * Better error reporting in 'openssl engine'.
13610 * Never call load_dh_param(NULL) in s_server.
13614 * Add engine application. It can currently list engines by name and
13615 identity, and test if they are actually available.
13619 * Improve RPM specification file by forcing symbolic linking and making
13620 sure the installed documentation is also owned by root.root.
13622 *Damien Miller <djm@mindrot.org>*
13624 * Give the OpenSSL applications more possibilities to make use of
13625 keys (public as well as private) handled by engines.
13629 * Add OCSP code that comes from CertCo.
13633 * Add VMS support for the Rijndael code.
13637 * Added untested support for Nuron crypto accelerator.
13641 * Add support for external cryptographic devices. This code was
13642 previously distributed separately as the "engine" branch.
13644 *Geoff Thorpe, Richard Levitte*
13646 * Rework the filename-translation in the DSO code. It is now possible to
13647 have far greater control over how a "name" is turned into a filename
13648 depending on the operating environment and any oddities about the
13649 different shared library filenames on each system.
13653 * Support threads on FreeBSD-elf in Configure.
13657 * Fix for SHA1 assembly problem with MASM: it produces
13658 warnings about corrupt line number information when assembling
13659 with debugging information. This is caused by the overlapping
13662 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13665 NCONF_get_number() has no error checking at all. As a replacement,
13666 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13667 promoted strongly. The old NCONF_get_number is kept around for
13668 binary backward compatibility.
13669 Make it possible for methods to load from something other than a BIO,
13670 by providing a function pointer that is given a name instead of a BIO.
13671 For example, this could be used to load configuration data from an
13676 * Fix for non blocking accept BIOs. Added new I/O special reason
13677 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13678 with non blocking I/O was not possible because no retry code was
13679 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13684 * Added the beginnings of Rijndael support.
13688 * Fix for bug in DirectoryString mask setting. Add support for
13689 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13690 to allow certificate printing to more controllable, additional
13691 'certopt' option to 'x509' to allow new printing options to be
13696 * Clean old EAY MD5 hack from e_os.h.
13700 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13702 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13703 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13705 *Joe Orton, Steve Henson*
13707 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13709 * Fix additional bug revealed by the NISCC test suite:
13711 Stop bug triggering large recursion when presented with
13712 certain ASN.1 tags ([CVE-2003-0851])
13716 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13718 * Fix various bugs revealed by running the NISCC test suite:
13720 Stop out of bounds reads in the ASN1 code when presented with
13721 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13723 If verify callback ignores invalid public key errors don't try to check
13724 certificate signature with the NULL public key.
13728 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13729 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13734 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13735 extra data after the compression methods not only for TLS 1.0
13736 but also for SSL 3.0 (as required by the specification).
13738 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13740 * Change X509_certificate_type() to mark the key as exported/exportable
13741 when it's 512 *bits* long, not 512 bytes.
13745 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13747 * Countermeasure against the Klima-Pokorny-Rosa extension of
13748 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13749 a protocol version number mismatch like a decryption error
13750 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13754 * Turn on RSA blinding by default in the default implementation
13755 to avoid a timing attack. Applications that don't want it can call
13756 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13757 They would be ill-advised to do so in most cases.
13759 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13761 * Change RSA blinding code so that it works when the PRNG is not
13762 seeded (in this case, the secret RSA exponent is abused as
13763 an unpredictable seed -- if it is not unpredictable, there
13764 is no point in blinding anyway). Make RSA blinding thread-safe
13765 by remembering the creator's thread ID in rsa->blinding and
13766 having all other threads use local one-time blinding factors
13767 (this requires more computation than sharing rsa->blinding, but
13768 avoids excessive locking; and if an RSA object is not shared
13769 between threads, blinding will still be very fast).
13773 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13775 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13776 via timing by performing a MAC computation even if incorrect
13777 block cipher padding has been found. This is a countermeasure
13778 against active attacks where the attacker has to distinguish
13779 between bad padding and a MAC verification error. ([CVE-2003-0078])
13781 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13782 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13783 Martin Vuagnoux (EPFL, Ilion)*
13785 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13787 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13788 memory from its contents. This is done with a counter that will
13789 place alternating values in each byte. This can be used to solve
13790 two issues: 1) the removal of calls to memset() by highly optimizing
13791 compilers, and 2) cleansing with other values than 0, since those can
13792 be read through on certain media, for example a swap space on disk.
13796 * Bugfix: client side session caching did not work with external caching,
13797 because the session->cipher setting was not restored when reloading
13798 from the external cache. This problem was masked, when
13799 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13800 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13804 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13805 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13807 *Zeev Lieber <zeev-l@yahoo.com>*
13809 * Undo an undocumented change introduced in 0.9.6e which caused
13810 repeated calls to OpenSSL_add_all_ciphers() and
13811 OpenSSL_add_all_digests() to be ignored, even after calling
13816 * Change the default configuration reader to deal with last line not
13817 being properly terminated.
13821 * Change X509_NAME_cmp() so it applies the special rules on handling
13822 DN values that are of type PrintableString, as well as RDNs of type
13823 emailAddress where the value has the type ia5String.
13825 *stefank@valicert.com via Richard Levitte*
13827 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13828 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13829 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13830 the bitwise-OR of the two for use by the majority of applications
13831 wanting this behaviour, and update the docs. The documented
13832 behaviour and actual behaviour were inconsistent and had been
13833 changing anyway, so this is more a bug-fix than a behavioural
13836 *Geoff Thorpe, diagnosed by Nadav Har'El*
13838 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13839 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13843 * Fix initialization code race conditions in
13844 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13845 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13846 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13847 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13848 ssl2_get_cipher_by_char(),
13849 ssl3_get_cipher_by_char().
13851 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13853 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13854 the cached sessions are flushed, as the remove_cb() might use ex_data
13855 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13856 (see [openssl.org #212]).
13858 *Geoff Thorpe, Lutz Jaenicke*
13860 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13861 length, instead of the encoding length to d2i_ASN1_OBJECT.
13865 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13867 * [In 0.9.6g-engine release:]
13868 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13870 *Lynn Gazis <lgazis@rainbow.com>*
13872 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13874 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13875 and get fix the header length calculation.
13876 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13877 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13879 * Use proper error handling instead of 'assertions' in buffer
13880 overflow checks added in 0.9.6e. This prevents DoS (the
13881 assertions could call abort()).
13883 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13885 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13887 * Add various sanity checks to asn1_get_length() to reject
13888 the ASN1 length bytes if they exceed sizeof(long), will appear
13889 negative or the content length exceeds the length of the
13892 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13894 * Fix cipher selection routines: ciphers without encryption had no flags
13895 for the cipher strength set and where therefore not handled correctly
13896 by the selection routines (PR #130).
13900 * Fix EVP_dsa_sha macro.
13905 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13906 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13907 that was added in OpenSSL 0.9.6d.
13909 As the countermeasure turned out to be incompatible with some
13910 broken SSL implementations, the new option is part of SSL_OP_ALL.
13911 SSL_OP_ALL is usually employed when compatibility with weird SSL
13912 implementations is desired (e.g. '-bugs' option to 's_client' and
13913 's_server'), so the new option is automatically set in many
13918 * Changes in security patch:
13920 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13921 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13922 Air Force Materiel Command, USAF, under agreement number
13925 * Add various sanity checks to asn1_get_length() to reject
13926 the ASN1 length bytes if they exceed sizeof(long), will appear
13927 negative or the content length exceeds the length of the
13928 supplied buffer. ([CVE-2002-0659])
13930 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13932 * Assertions for various potential buffer overflows, not known to
13933 happen in practice.
13935 *Ben Laurie (CHATS)*
13937 * Various temporary buffers to hold ASCII versions of integers were
13938 too small for 64 bit platforms. ([CVE-2002-0655])
13939 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13941 * Remote buffer overflow in SSL3 protocol - an attacker could
13942 supply an oversized session ID to a client. ([CVE-2002-0656])
13944 *Ben Laurie (CHATS)*
13946 * Remote buffer overflow in SSL2 protocol - an attacker could
13947 supply an oversized client master key. ([CVE-2002-0656])
13949 *Ben Laurie (CHATS)*
13951 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13953 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13954 encoded as NULL) with id-dsa-with-sha1.
13956 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13958 * Check various `X509_...()` return values in `apps/req.c`.
13960 *Nils Larsch <nla@trustcenter.de>*
13962 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13963 an end-of-file condition would erroneously be flagged, when the CRLF
13964 was just at the end of a processed block. The bug was discovered when
13965 processing data through a buffering memory BIO handing the data to a
13966 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13967 <ptsekov@syntrex.com> and Nedelcho Stanev.
13971 * Implement a countermeasure against a vulnerability recently found
13972 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13973 before application data chunks to avoid the use of known IVs
13974 with data potentially chosen by the attacker.
13978 * Fix length checks in ssl3_get_client_hello().
13982 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13983 to prevent ssl3_read_internal() from incorrectly assuming that
13984 ssl3_read_bytes() found application data while handshake
13985 processing was enabled when in fact s->s3->in_read_app_data was
13986 merely automatically cleared during the initial handshake.
13988 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13990 * Fix object definitions for Private and Enterprise: they were not
13991 recognized in their shortname (=lowercase) representation. Extend
13992 obj_dat.pl to issue an error when using undefined keywords instead
13993 of silently ignoring the problem (Svenning Sorensen
13994 <sss@sss.dnsalias.net>).
13998 * Fix DH_generate_parameters() so that it works for 'non-standard'
13999 generators, i.e. generators other than 2 and 5. (Previously, the
14000 code did not properly initialise the 'add' and 'rem' values to
14001 BN_generate_prime().)
14003 In the new general case, we do not insist that 'generator' is
14004 actually a primitive root: This requirement is rather pointless;
14005 a generator of the order-q subgroup is just as good, if not
14010 * Map new X509 verification errors to alerts. Discovered and submitted by
14011 Tom Wu <tom@arcot.com>.
14015 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14016 returning non-zero before the data has been completely received
14017 when using non-blocking I/O.
14019 *Bodo Moeller; problem pointed out by John Hughes*
14021 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14023 *Ben Laurie, Lutz Jaenicke*
14025 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14026 Yoram Zahavi <YoramZ@gilian.com>).
14030 * Add information about CygWin 1.3 and on, and preserve proper
14031 configuration for the versions before that.
14033 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14035 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14036 check whether we deal with a copy of a session and do not delete from
14037 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14038 <izhar@checkpoint.com>.
14042 * Do not store session data into the internal session cache, if it
14043 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14044 flag is set). Proposed by Aslam <aslam@funk.com>.
14048 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14053 * [In 0.9.6d-engine release:]
14054 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14056 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14058 * Add the configuration target linux-s390x.
14060 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14062 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14063 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14064 variable as an indication that a ClientHello message has been
14065 received. As the flag value will be lost between multiple
14066 invocations of ssl3_accept when using non-blocking I/O, the
14067 function may not be aware that a handshake has actually taken
14068 place, thus preventing a new session from being added to the
14071 To avoid this problem, we now set s->new_session to 2 instead of
14072 using a local variable.
14074 *Lutz Jaenicke, Bodo Moeller*
14076 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14077 if the SSL_R_LENGTH_MISMATCH error is detected.
14079 *Geoff Thorpe, Bodo Moeller*
14081 * New 'shared_ldflag' column in Configure platform table.
14085 * Fix EVP_CIPHER_mode macro.
14087 *"Dan S. Camper" <dan@bti.net>*
14089 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14090 type, we must throw them away by setting rr->length to 0.
14092 *D P Chang <dpc@qualys.com>*
14094 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14096 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14097 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14098 worked incorrectly for those cases where range = `10..._2` and
14099 `3*range` is two bits longer than range.)
14103 * Only add signing time to PKCS7 structures if it is not already
14108 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14109 OBJ_ld_ce should be OBJ_id_ce.
14110 Also some ip-pda OIDs in crypto/objects/objects.txt were
14111 incorrect (cf. RFC 3039).
14113 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14115 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14116 returns early because it has nothing to do.
14118 *Andy Schneider <andy.schneider@bjss.co.uk>*
14120 * [In 0.9.6c-engine release:]
14121 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14123 *Andy Schneider <andy.schneider@bjss.co.uk>*
14125 * [In 0.9.6c-engine release:]
14126 Add support for Cryptographic Appliance's keyserver technology.
14127 (Use engine 'keyclient')
14129 *Cryptographic Appliances and Geoff Thorpe*
14131 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14132 is called via tools/c89.sh because arguments have to be
14133 rearranged (all '-L' options must appear before the first object
14136 *Richard Shapiro <rshapiro@abinitio.com>*
14138 * [In 0.9.6c-engine release:]
14139 Add support for Broadcom crypto accelerator cards, backported
14142 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14144 * [In 0.9.6c-engine release:]
14145 Add support for SureWare crypto accelerator cards from
14146 Baltimore Technologies. (Use engine 'sureware')
14148 *Baltimore Technologies and Mark Cox*
14150 * [In 0.9.6c-engine release:]
14151 Add support for crypto accelerator cards from Accelerated
14152 Encryption Processing, www.aep.ie. (Use engine 'aep')
14154 *AEP Inc. and Mark Cox*
14156 * Add a configuration entry for gcc on UnixWare.
14158 *Gary Benson <gbenson@redhat.com>*
14160 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14161 messages are stored in a single piece (fixed-length part and
14162 variable-length part combined) and fix various bugs found on the way.
14166 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14167 instead. BIO_gethostbyname() does not know what timeouts are
14168 appropriate, so entries would stay in cache even when they have
14170 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14172 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14173 faced with a pathologically small ClientHello fragment that does
14174 not contain client_version: Instead of aborting with an error,
14175 simply choose the highest available protocol version (i.e.,
14176 TLS 1.0 unless it is disabled). In practice, ClientHello
14177 messages are never sent like this, but this change gives us
14178 strictly correct behaviour at least for TLS.
14182 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14183 never resets s->method to s->ctx->method when called from within
14184 one of the SSL handshake functions.
14186 *Bodo Moeller; problem pointed out by Niko Baric*
14188 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14189 (sent using the client's version number) if client_version is
14190 smaller than the protocol version in use. Also change
14191 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14192 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14193 the client will at least see that alert.
14197 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14202 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14203 client receives HelloRequest while in a handshake.
14205 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14207 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14208 should end in 'break', not 'goto end' which circumvents various
14209 cleanups done in state SSL_ST_OK. But session related stuff
14210 must be disabled for SSL_ST_OK in the case that we just sent a
14213 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14214 before just sending a HelloRequest.
14216 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14218 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14219 reveal whether illegal block cipher padding was found or a MAC
14220 verification error occurred. (Neither SSLerr() codes nor alerts
14221 are directly visible to potential attackers, but the information
14222 may leak via logfiles.)
14224 Similar changes are not required for the SSL 2.0 implementation
14225 because the number of padding bytes is sent in clear for SSL 2.0,
14226 and the extra bytes are just ignored. However ssl/s2_pkt.c
14227 failed to verify that the purported number of padding bytes is in
14232 * Add OpenUNIX-8 support including shared libraries
14233 (Boyd Lynn Gerber <gerberb@zenez.com>).
14237 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14238 'wristwatch attack' using huge encoding parameters (cf.
14239 James H. Manger's CRYPTO 2001 paper). Note that the
14240 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14241 encoding parameters and hence was not vulnerable.
14245 * BN_sqr() bug fix.
14247 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14249 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14250 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14251 followed by modular reduction.
14253 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14255 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14256 equivalent based on BN_pseudo_rand() instead of BN_rand().
14260 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14261 This function was broken, as the check for a new client hello message
14262 to handle SGC did not allow these large messages.
14263 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14267 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14271 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14272 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14276 * Rework the configuration and shared library support for Tru64 Unix.
14277 The configuration part makes use of modern compiler features and
14278 still retains old compiler behavior for those that run older versions
14279 of the OS. The shared library support part includes a variant that
14280 uses the RPATH feature, and is available through the special
14281 configuration target "alpha-cc-rpath", which will never be selected
14284 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14286 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14287 with the same message size as in ssl3_get_certificate_request().
14288 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14289 messages might inadvertently be reject as too long.
14291 *Petr Lampa <lampa@fee.vutbr.cz>*
14293 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14297 * Modified SSL library such that the verify_callback that has been set
14298 specifically for an SSL object with SSL_set_verify() is actually being
14299 used. Before the change, a verify_callback set with this function was
14300 ignored and the verify_callback() set in the SSL_CTX at the time of
14301 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14302 to allow the necessary settings.
14306 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14307 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14308 done automatically (in contradiction to the requirements of the C
14309 standard). This made problems when used from OpenSSH.
14313 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14314 dh->length and always used
14316 BN_rand_range(priv_key, dh->p).
14318 BN_rand_range() is not necessary for Diffie-Hellman, and this
14319 specific range makes Diffie-Hellman unnecessarily inefficient if
14320 dh->length (recommended exponent length) is much smaller than the
14321 length of dh->p. We could use BN_rand_range() if the order of
14322 the subgroup was stored in the DH structure, but we only have
14327 BN_rand(priv_key, l, ...)
14329 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14336 RSA_eay_public_encrypt
14337 RSA_eay_private_decrypt
14338 RSA_eay_private_encrypt (signing)
14339 RSA_eay_public_decrypt (signature verification)
14341 (default implementations for RSA_public_encrypt,
14342 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14343 always reject numbers >= n.
14347 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14348 to synchronize access to 'locking_thread'. This is necessary on
14349 systems where access to 'locking_thread' (an 'unsigned long'
14350 variable) is not atomic.
14354 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14355 *before* setting the 'crypto_lock_rand' flag. The previous code had
14356 a race condition if 0 is a valid thread ID.
14358 *Travis Vitek <vitek@roguewave.com>*
14360 * Add support for shared libraries under Irix.
14362 *Albert Chin-A-Young <china@thewrittenword.com>*
14364 * Add configuration option to build on Linux on both big-endian and
14365 little-endian MIPS.
14367 *Ralf Baechle <ralf@uni-koblenz.de>*
14369 * Add the possibility to create shared libraries on HP-UX.
14373 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14375 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14376 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14377 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14378 PRNG state recovery was possible based on the output of
14379 one PRNG request appropriately sized to gain knowledge on
14380 'md' followed by enough consecutive 1-byte PRNG requests
14381 to traverse all of 'state'.
14383 1. When updating 'md_local' (the current thread's copy of 'md')
14384 during PRNG output generation, hash all of the previous
14385 'md_local' value, not just the half used for PRNG output.
14387 2. Make the number of bytes from 'state' included into the hash
14388 independent from the number of PRNG bytes requested.
14390 The first measure alone would be sufficient to avoid
14391 Markku-Juhani's attack. (Actually it had never occurred
14392 to me that the half of 'md_local' used for chaining was the
14393 half from which PRNG output bytes were taken -- I had always
14394 assumed that the secret half would be used.) The second
14395 measure makes sure that additional data from 'state' is never
14396 mixed into 'md_local' in small portions; this heuristically
14397 further strengthens the PRNG.
14401 * Fix crypto/bn/asm/mips3.s.
14405 * When only the key is given to "enc", the IV is undefined. Print out
14406 an error message in this case.
14410 * Handle special case when X509_NAME is empty in X509 printing routines.
14414 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14415 positive and less than q.
14419 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14420 used: it isn't thread safe and the add_lock_callback should handle
14423 *Paul Rose <Paul.Rose@bridge.com>*
14425 * Verify that incoming data obeys the block size in
14426 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14432 *Ulf Möller, Bodo Möller*
14434 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14435 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14436 when fixing the server behaviour for backwards-compatible 'client
14437 hello' messages. (Note that the attack is impractical against
14438 SSL 3.0 and TLS 1.0 anyway because length and version checking
14439 means that the probability of guessing a valid ciphertext is
14440 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14443 Before 0.9.5, the countermeasure (hide the error by generating a
14444 random 'decryption result') did not work properly because
14445 ERR_clear_error() was missing, meaning that SSL_get_error() would
14446 detect the supposedly ignored error.
14448 Both problems are now fixed.
14452 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14453 (previously it was 1024).
14457 * Fix for compatibility mode trust settings: ignore trust settings
14458 unless some valid trust or reject settings are present.
14462 * Fix for blowfish EVP: its a variable length cipher.
14466 * Fix various bugs related to DSA S/MIME verification. Handle missing
14467 parameters in DSA public key structures and return an error in the
14468 DSA routines if parameters are absent.
14472 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14473 in the current directory if neither $RANDFILE nor $HOME was set.
14474 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14475 caused some confusion to Windows users who haven't defined $HOME.
14476 Thus RAND_file_name() is changed again: e_os.h can define a
14477 DEFAULT_HOME, which will be used if $HOME is not set.
14478 For Windows, we use "C:"; on other platforms, we still require
14479 environment variables.
14481 * Move 'if (!initialized) RAND_poll()' into regions protected by
14482 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14483 having multiple threads call RAND_poll() concurrently.
14487 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14488 combination of a flag and a thread ID variable.
14489 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14490 flag), *other* threads can enter ssleay_add_bytes without obeying
14491 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14492 that they do not hold after the first thread unsets add_do_not_lock).
14496 * Change bctest again: '-x' expressions are not available in all
14497 versions of 'test'.
14501 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14503 * Fix a couple of memory leaks in PKCS7_dataDecode()
14505 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14507 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14508 the default extension for executables, if any. Also, make the perl
14509 scripts that use symlink() to test if it really exists and use "cp"
14510 if it doesn't. All this made OpenSSL compilable and installable in
14515 * Fix for asn1_GetSequence() for indefinite length constructed data.
14516 If SEQUENCE is length is indefinite just set c->slen to the total
14517 amount of data available.
14519 *Steve Henson, reported by shige@FreeBSD.org*
14521 *This change does not apply to 0.9.7.*
14523 * Change bctest to avoid here-documents inside command substitution
14524 (workaround for FreeBSD /bin/sh bug).
14525 For compatibility with Ultrix, avoid shell functions (introduced
14526 in the bctest version that searches along $PATH).
14530 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14531 with des_encrypt() defined on some operating systems, like Solaris
14536 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14537 On the Importance of Eliminating Errors in Cryptographic
14538 Computations, J. Cryptology 14 (2001) 2, 101-119,
14539 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14543 * MIPS assembler BIGNUM division bug fix.
14547 * Disabled incorrect Alpha assembler code.
14551 * Fix PKCS#7 decode routines so they correctly update the length
14552 after reading an EOC for the EXPLICIT tag.
14556 *This change does not apply to 0.9.7.*
14558 * Fix bug in PKCS#12 key generation routines. This was triggered
14559 if a 3DES key was generated with a 0 initial byte. Include
14560 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14561 (but broken) behaviour.
14565 * Enhance bctest to search for a working bc along $PATH and print
14568 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14570 * Fix memory leaks in err.c: free err_data string if necessary;
14571 don't write to the wrong index in ERR_set_error_data.
14575 * Implement ssl23_peek (analogous to ssl23_read), which previously
14580 * Replace rdtsc with `_emit` statements for VC++ version 5.
14582 *Jeremy Cooper <jeremy@baymoo.org>*
14584 * Make it possible to reuse SSLv2 sessions.
14588 * In copy_email() check for >= 0 as a return value for
14589 X509_NAME_get_index_by_NID() since 0 is a valid index.
14591 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14593 * Avoid coredump with unsupported or invalid public keys by checking if
14594 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14595 PKCS7_verify() fails with non detached data.
14599 * Don't use getenv in library functions when run as setuid/setgid.
14600 New function OPENSSL_issetugid().
14604 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14605 due to incorrect handling of multi-threading:
14607 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14609 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14611 3. Count how many times MemCheck_off() has been called so that
14612 nested use can be treated correctly. This also avoids
14613 inband-signalling in the previous code (which relied on the
14614 assumption that thread ID 0 is impossible).
14618 * Add "-rand" option also to s_client and s_server.
14622 * Fix CPU detection on Irix 6.x.
14623 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14624 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14626 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14631 *This change does not apply to 0.9.7.*
14633 * Use the cached encoding of an X509_NAME structure rather than
14634 copying it. This is apparently the reason for the libsafe "errors"
14635 but the code is actually correct.
14639 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14640 Bleichenbacher's DSA attack.
14641 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14642 to be set and top=0 forces the highest bit to be set; top=-1 is new
14643 and leaves the highest bit random.
14645 *Ulf Moeller, Bodo Moeller*
14647 * In the `NCONF_...`-based implementations for `CONF_...` queries
14648 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14649 a temporary CONF structure with the data component set to NULL
14650 (which gives segmentation faults in lh_retrieve).
14651 Instead, use NULL for the CONF pointer in CONF_get_string and
14652 CONF_get_number (which may use environment variables) and directly
14653 return NULL from CONF_get_section.
14657 * Fix potential buffer overrun for EBCDIC.
14661 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14662 keyUsage if basicConstraints absent for a CA.
14666 * Make SMIME_write_PKCS7() write mail header values with a format that
14667 is more generally accepted (no spaces before the semicolon), since
14668 some programs can't parse those values properly otherwise. Also make
14669 sure BIO's that break lines after each write do not create invalid
14674 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14675 macros previously used would not encode an empty SEQUENCE OF
14676 and break the signature.
14680 *This change does not apply to 0.9.7.*
14682 * Zero the premaster secret after deriving the master secret in
14687 * Add some EVP_add_digest_alias registrations (as found in
14688 OpenSSL_add_all_digests()) to SSL_library_init()
14689 aka OpenSSL_add_ssl_algorithms(). This provides improved
14690 compatibility with peers using X.509 certificates
14691 with unconventional AlgorithmIdentifier OIDs.
14695 * Fix for Irix with NO_ASM.
14697 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14699 * ./config script fixes.
14701 *Ulf Moeller, Richard Levitte*
14703 * Fix 'openssl passwd -1'.
14707 * Change PKCS12_key_gen_asc() so it can cope with non null
14708 terminated strings whose length is passed in the passlen
14709 parameter, for example from PEM callbacks. This was done
14710 by adding an extra length parameter to asc2uni().
14712 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14714 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14715 call failed, free the DSA structure.
14719 * Fix to uni2asc() to cope with zero length Unicode strings.
14720 These are present in some PKCS#12 files.
14724 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14725 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14726 when writing a 32767 byte record.
14728 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14730 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14731 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14733 (RSA objects have a reference count access to which is protected
14734 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14735 so they are meant to be shared between threads.)
14736 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14737 "Reddie, Steven" <Steven.Reddie@ca.com>*
14739 * Fix a deadlock in CRYPTO_mem_leaks().
14743 * Use better test patterns in bntest.
14747 * rand_win.c fix for Borland C.
14751 * BN_rshift bugfix for n == 0.
14755 * Add a 'bctest' script that checks for some known 'bc' bugs
14756 so that 'make test' does not abort just because 'bc' is broken.
14760 * Store verify_result within SSL_SESSION also for client side to
14761 avoid potential security hole. (Re-used sessions on the client side
14762 always resulted in verify_result==X509_V_OK, not using the original
14763 result of the server certificate verification.)
14767 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14768 SSL3_RT_APPLICATION_DATA, return 0.
14769 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14774 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14775 releases, have been re-implemented by renaming the previous
14776 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14777 and ssl3_read_internal, respectively, and adding 'peek' parameters
14778 to them. The new ssl[23]_{read,peek} functions are calls to
14779 ssl[23]_read_internal with the 'peek' flag set appropriately.
14780 A 'peek' parameter has also been added to ssl3_read_bytes, which
14781 does the actual work for ssl3_read_internal.
14785 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14786 the method-specific "init()" handler. Also clean up ex_data after
14787 calling the method-specific "finish()" handler. Previously, this was
14788 happening the other way round.
14792 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14793 The previous value, 12, was not always sufficient for BN_mod_exp().
14797 * Make sure that shared libraries get the internal name engine with
14798 the full version number and not just 0. This should mark the
14799 shared libraries as not backward compatible. Of course, this should
14800 be changed again when we can guarantee backward binary compatibility.
14804 * Fix typo in get_cert_by_subject() in by_dir.c
14806 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14808 * Rework the system to generate shared libraries:
14810 - Make note of the expected extension for the shared libraries and
14811 if there is a need for symbolic links from for example libcrypto.so.0
14812 to libcrypto.so.0.9.7. There is extended info in Configure for
14815 - Make as few rebuilds of the shared libraries as possible.
14817 - Still avoid linking the OpenSSL programs with the shared libraries.
14819 - When installing, install the shared libraries separately from the
14824 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14826 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14827 and not in SSL_clear because the latter is also used by the
14828 accept/connect functions; previously, the settings made by
14829 SSL_set_read_ahead would be lost during the handshake.
14831 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14833 * Correct util/mkdef.pl to be selective about disabled algorithms.
14834 Previously, it would create entries for disabled algorithms no
14839 * Added several new manual pages for SSL_* function.
14843 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14845 * In ssl23_get_client_hello, generate an error message when faced
14846 with an initial SSL 3.0/TLS record that is too small to contain the
14847 first two bytes of the ClientHello message, i.e. client_version.
14848 (Note that this is a pathologic case that probably has never happened
14849 in real life.) The previous approach was to use the version number
14850 from the record header as a substitute; but our protocol choice
14851 should not depend on that one because it is not authenticated
14852 by the Finished messages.
14856 * More robust randomness gathering functions for Windows.
14858 *Jeffrey Altman <jaltman@columbia.edu>*
14860 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14861 not set then we don't setup the error code for issuer check errors
14862 to avoid possibly overwriting other errors which the callback does
14863 handle. If an application does set the flag then we assume it knows
14864 what it is doing and can handle the new informational codes
14869 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14870 a general "ANY" type, as such it should be able to decode anything
14871 including tagged types. However it didn't check the class so it would
14872 wrongly interpret tagged types in the same way as their universal
14873 counterpart and unknown types were just rejected. Changed so that the
14874 tagged and unknown types are handled in the same way as a SEQUENCE:
14875 that is the encoding is stored intact. There is also a new type
14876 "V_ASN1_OTHER" which is used when the class is not universal, in this
14877 case we have no idea what the actual type is so we just lump them all
14882 * On VMS, stdout may very well lead to a file that is written to
14883 in a record-oriented fashion. That means that every write() will
14884 write a separate record, which will be read separately by the
14885 programs trying to read from it. This can be very confusing.
14887 The solution is to put a BIO filter in the way that will buffer
14888 text until a linefeed is reached, and then write everything a
14889 line at a time, so every record written will be an actual line,
14890 not chunks of lines and not (usually doesn't happen, but I've
14891 seen it once) several lines in one record. BIO_f_linebuffer() is
14894 Currently, it's a VMS-only method, because that's where it has
14895 been tested well enough.
14899 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14900 it can return incorrect results.
14901 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14902 but it was in 0.9.6-beta[12].)
14906 * Disable the check for content being present when verifying detached
14907 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14908 include zero length content when signing messages.
14912 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14913 BIO_ctrl (for BIO pairs).
14917 * Add DSO method for VMS.
14921 * Bug fix: Montgomery multiplication could produce results with the
14926 * Add RPM specification openssl.spec and modify it to build three
14927 packages. The default package contains applications, application
14928 documentation and run-time libraries. The devel package contains
14929 include files, static libraries and function documentation. The
14930 doc package contains the contents of the doc directory. The original
14931 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14935 * Add a large number of documentation files for many SSL routines.
14937 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14939 * Add a configuration entry for Sony News 4.
14941 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14943 * Don't set the two most significant bits to one when generating a
14944 random number < q in the DSA library.
14948 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14949 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14950 the underlying transport is blocking) if a handshake took place.
14951 (The default behaviour is needed by applications such as s_client
14952 and s_server that use select() to determine when to use SSL_read;
14953 but for applications that know in advance when to expect data, it
14954 just makes things more complicated.)
14958 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14963 * Add a few more EBCDIC conditionals that make `req` and `x509`
14964 work better on such systems.
14966 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14968 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14969 Update PKCS12_parse() so it copies the friendlyName and the
14970 keyid to the certificates aux info.
14974 * Fix bug in PKCS7_verify() which caused an infinite loop
14975 if there was more than one signature.
14977 *Sven Uszpelkat <su@celocom.de>*
14979 * Major change in util/mkdef.pl to include extra information
14980 about each symbol, as well as presenting variables as well
14981 as functions. This change means that there's n more need
14982 to rebuild the .num files when some algorithms are excluded.
14986 * Allow the verify time to be set by an application,
14987 rather than always using the current time.
14991 * Phase 2 verify code reorganisation. The certificate
14992 verify code now looks up an issuer certificate by a
14993 number of criteria: subject name, authority key id
14994 and key usage. It also verifies self signed certificates
14995 by the same criteria. The main comparison function is
14996 X509_check_issued() which performs these checks.
14998 Lot of changes were necessary in order to support this
14999 without completely rewriting the lookup code.
15001 Authority and subject key identifier are now cached.
15003 The LHASH 'certs' is X509_STORE has now been replaced
15004 by a STACK_OF(X509_OBJECT). This is mainly because an
15005 LHASH can't store or retrieve multiple objects with
15006 the same hash value.
15008 As a result various functions (which were all internal
15009 use only) have changed to handle the new X509_STORE
15010 structure. This will break anything that messed round
15011 with X509_STORE internally.
15013 The functions X509_STORE_add_cert() now checks for an
15014 exact match, rather than just subject name.
15016 The X509_STORE API doesn't directly support the retrieval
15017 of multiple certificates matching a given criteria, however
15018 this can be worked round by performing a lookup first
15019 (which will fill the cache with candidate certificates)
15020 and then examining the cache for matches. This is probably
15021 the best we can do without throwing out X509_LOOKUP
15022 entirely (maybe later...).
15024 The X509_VERIFY_CTX structure has been enhanced considerably.
15026 All certificate lookup operations now go via a get_issuer()
15027 callback. Although this currently uses an X509_STORE it
15028 can be replaced by custom lookups. This is a simple way
15029 to bypass the X509_STORE hackery necessary to make this
15030 work and makes it possible to use more efficient techniques
15031 in future. A very simple version which uses a simple
15032 STACK for its trusted certificate store is also provided
15033 using X509_STORE_CTX_trusted_stack().
15035 The verify_cb() and verify() callbacks now have equivalents
15036 in the X509_STORE_CTX structure.
15038 X509_STORE_CTX also has a 'flags' field which can be used
15039 to customise the verify behaviour.
15043 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15044 excludes S/MIME capabilities.
15048 * When a certificate request is read in keep a copy of the
15049 original encoding of the signed data and use it when outputting
15050 again. Signatures then use the original encoding rather than
15051 a decoded, encoded version which may cause problems if the
15052 request is improperly encoded.
15056 * For consistency with other BIO_puts implementations, call
15057 buffer_write(b, ...) directly in buffer_puts instead of calling
15060 In BIO_puts, increment b->num_write as in BIO_write.
15062 *Peter.Sylvester@EdelWeb.fr*
15064 * Fix BN_mul_word for the case where the word is 0. (We have to use
15065 BN_zero, we may not return a BIGNUM with an array consisting of
15066 words set to zero.)
15070 * Avoid calling abort() from within the library when problems are
15071 detected, except if preprocessor symbols have been defined
15072 (such as REF_CHECK, BN_DEBUG etc.).
15076 * New openssl application 'rsautl'. This utility can be
15077 used for low-level RSA operations. DER public key
15078 BIO/fp routines also added.
15082 * New Configure entry and patches for compiling on QNX 4.
15084 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15086 * A demo state-machine implementation was sponsored by
15087 Nuron (<http://www.nuron.com/>) and is now available in
15088 demos/state_machine.
15092 * New options added to the 'dgst' utility for signature
15093 generation and verification.
15097 * Unrecognized PKCS#7 content types are now handled via a
15098 catch all ASN1_TYPE structure. This allows unsupported
15099 types to be stored as a "blob" and an application can
15100 encode and decode it manually.
15104 * Fix various signed/unsigned issues to make a_strex.c
15105 compile under VC++.
15107 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15109 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15110 length if passed a buffer. ASN1_INTEGER_to_BN failed
15111 if passed a NULL BN and its argument was negative.
15113 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15115 * Modification to PKCS#7 encoding routines to output definite
15116 length encoding. Since currently the whole structures are in
15117 memory there's not real point in using indefinite length
15118 constructed encoding. However if OpenSSL is compiled with
15119 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15123 * Added BIO_vprintf() and BIO_vsnprintf().
15127 * Added more prefixes to parse for in the strings written
15128 through a logging bio, to cover all the levels that are available
15129 through syslog. The prefixes are now:
15131 PANIC, EMERG, EMR => LOG_EMERG
15132 ALERT, ALR => LOG_ALERT
15133 CRIT, CRI => LOG_CRIT
15134 ERROR, ERR => LOG_ERR
15135 WARNING, WARN, WAR => LOG_WARNING
15136 NOTICE, NOTE, NOT => LOG_NOTICE
15137 INFO, INF => LOG_INFO
15138 DEBUG, DBG => LOG_DEBUG
15140 and as before, if none of those prefixes are present at the
15141 beginning of the string, LOG_ERR is chosen.
15143 On Win32, the `LOG_*` levels are mapped according to this:
15145 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15146 LOG_WARNING => EVENTLOG_WARNING_TYPE
15147 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15151 * Made it possible to reconfigure with just the configuration
15152 argument "reconf" or "reconfigure". The command line arguments
15153 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15154 and are retrieved from there when reconfiguring.
15160 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15162 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15166 * The obj_dat.pl script was messing up the sorting of object
15167 names. The reason was that it compared the quoted version
15168 of strings as a result "OCSP" > "OCSP Signing" because
15169 " > SPACE. Changed script to store unquoted versions of
15170 names and add quotes on output. It was also omitting some
15171 names from the lookup table if they were given a default
15172 value (that is if SN is missing it is given the same
15173 value as LN and vice versa), these are now added on the
15174 grounds that if an object has a name we should be able to
15175 look it up. Finally added warning output when duplicate
15176 short or long names are found.
15180 * Changes needed for Tandem NSK.
15182 *Scott Uroff <scott@xypro.com>*
15184 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15185 RSA_padding_check_SSLv23(), special padding was never detected
15186 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15187 version rollback attacks was not effective.
15189 In s23_clnt.c, don't use special rollback-attack detection padding
15190 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15191 client; similarly, in s23_srvr.c, don't do the rollback check if
15192 SSL 2.0 is the only protocol enabled in the server.
15196 * Make it possible to get hexdumps of unprintable data with 'openssl
15197 asn1parse'. By implication, the functions ASN1_parse_dump() and
15198 BIO_dump_indent() are added.
15202 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15203 these print out strings and name structures based on various
15204 flags including RFC2253 support and proper handling of
15205 multibyte characters. Added options to the 'x509' utility
15206 to allow the various flags to be set.
15210 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15211 Also change the functions X509_cmp_current_time() and
15212 X509_gmtime_adj() work with an ASN1_TIME structure,
15213 this will enable certificates using GeneralizedTime in validity
15214 dates to be checked.
15218 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15219 negative public key encodings) on by default,
15220 NO_NEG_PUBKEY_BUG can be set to disable it.
15224 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15225 content octets. An i2c_ASN1_OBJECT is unnecessary because
15226 the encoding can be trivially obtained from the structure.
15230 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15231 not read locks (`CRYPTO_r_[un]lock`).
15235 * A first attempt at creating official support for shared
15236 libraries through configuration. I've kept it so the
15237 default is static libraries only, and the OpenSSL programs
15238 are always statically linked for now, but there are
15239 preparations for dynamic linking in place.
15240 This has been tested on Linux and Tru64.
15244 * Randomness polling function for Win9x, as described in:
15245 Peter Gutmann, Software Generation of Practically Strong
15250 * Fix so PRNG is seeded in req if using an already existing
15255 * New options to smime application. -inform and -outform
15256 allow alternative formats for the S/MIME message including
15257 PEM and DER. The -content option allows the content to be
15258 specified separately. This should allow things like Netscape
15259 form signing output easier to verify.
15263 * Fix the ASN1 encoding of tags using the 'long form'.
15267 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15268 STRING types. These convert content octets to and from the
15269 underlying type. The actual tag and length octets are
15270 already assumed to have been read in and checked. These
15271 are needed because all other string types have virtually
15272 identical handling apart from the tag. By having versions
15273 of the ASN1 functions that just operate on content octets
15274 IMPLICIT tagging can be handled properly. It also allows
15275 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15276 and ASN1_INTEGER are identical apart from the tag.
15280 * Change the handling of OID objects as follows:
15282 - New object identifiers are inserted in objects.txt, following
15283 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15284 - objects.pl is used to process obj_mac.num and create a new
15286 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15289 This is currently kind of a hack, and the perl code in objects.pl
15290 isn't very elegant, but it works as I intended. The simplest way
15291 to check that it worked correctly is to look in obj_dat.h and
15292 check the array nid_objs and make sure the objects haven't moved
15293 around (this is important!). Additions are OK, as well as
15294 consistent name changes.
15298 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15302 * Addition of the command line parameter '-rand file' to 'openssl req'.
15303 The given file adds to whatever has already been seeded into the
15304 random pool through the RANDFILE configuration file option or
15305 environment variable, or the default random state file.
15309 * mkstack.pl now sorts each macro group into lexical order.
15310 Previously the output order depended on the order the files
15311 appeared in the directory, resulting in needless rewriting
15316 * Patches to make OpenSSL compile under Win32 again. Mostly
15317 work arounds for the VC++ problem that it treats func() as
15318 func(void). Also stripped out the parts of mkdef.pl that
15319 added extra typesafe functions: these no longer exist.
15323 * Reorganisation of the stack code. The macros are now all
15324 collected in safestack.h . Each macro is defined in terms of
15325 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15326 DEBUG_SAFESTACK is now handled in terms of function casts,
15327 this has the advantage of retaining type safety without the
15328 use of additional functions. If DEBUG_SAFESTACK is not defined
15329 then the non typesafe macros are used instead. Also modified the
15330 mkstack.pl script to handle the new form. Needs testing to see
15331 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15332 the default if no major problems. Similar behaviour for ASN1_SET_OF
15333 and PKCS12_STACK_OF.
15337 * When some versions of IIS use the 'NET' form of private key the
15338 key derivation algorithm is different. Normally MD5(password) is
15339 used as a 128 bit RC4 key. In the modified case
15340 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15341 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15342 as the old Netscape_RSA functions except they have an additional
15343 'sgckey' parameter which uses the modified algorithm. Also added
15344 an -sgckey command line option to the rsa utility. Thanks to
15345 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15346 algorithm to openssl-dev.
15350 * The evp_local.h macros were using 'c.##kname' which resulted in
15351 invalid expansion on some systems (SCO 5.0.5 for example).
15352 Corrected to 'c.kname'.
15354 *Phillip Porch <root@theporch.com>*
15356 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15357 a STACK of email addresses from a certificate or request, these look
15358 in the subject name and the subject alternative name extensions and
15359 omit any duplicate addresses.
15363 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15364 This makes DSA verification about 2 % faster.
15368 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15369 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15370 plus overhead for 1024 bit moduli).
15371 This makes exponentiations about 0.5 % faster for 1024 bit
15372 exponents (as measured by "openssl speed rsa2048").
15376 * Rename memory handling macros to avoid conflicts with other
15378 Malloc => OPENSSL_malloc
15379 Malloc_locked => OPENSSL_malloc_locked
15380 Realloc => OPENSSL_realloc
15381 Free => OPENSSL_free
15385 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15386 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15390 * CygWin32 support.
15392 *John Jarvie <jjarvie@newsguy.com>*
15394 * The type-safe stack code has been rejigged. It is now only compiled
15395 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15396 by default all type-specific stack functions are "#define"d back to
15397 standard stack functions. This results in more streamlined output
15398 but retains the type-safety checking possibilities of the original
15403 * The STACK code has been cleaned up, and certain type declarations
15404 that didn't make a lot of sense have been brought in line. This has
15405 also involved a cleanup of sorts in safestack.h to more correctly
15406 map type-safe stack functions onto their plain stack counterparts.
15407 This work has also resulted in a variety of "const"ifications of
15408 lots of the code, especially `_cmp` operations which should normally
15409 be prototyped with "const" parameters anyway.
15413 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15414 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15415 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15416 where all of 'md' is used each time the PRNG is used, but 'state'
15417 is used only indexed by a cyclic counter. As entropy may not be
15418 well distributed from the beginning, 'md' is important as a
15419 chaining variable. However, the output function chains only half
15420 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15421 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15422 in all of 'state' being rewritten, with the new values depending
15423 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15427 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15428 the handshake is continued after ssl_verify_cert_chain();
15429 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15430 can lead to 'unexplainable' connection aborts later.
15432 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15434 * Major EVP API cipher revision.
15435 Add hooks for extra EVP features. This allows various cipher
15436 parameters to be set in the EVP interface. Support added for variable
15437 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15438 setting of RC2 and RC5 parameters.
15440 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15443 Remove lots of duplicated code from the EVP library. For example *every*
15444 cipher init() function handles the 'iv' in the same way according to the
15445 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15446 for CFB and OFB modes they zero ctx->num.
15448 New functionality allows removal of S/MIME code RC2 hack.
15450 Most of the routines have the same form and so can be declared in terms
15453 By shifting this to the top level EVP_CipherInit() it can be removed from
15454 all individual ciphers. If the cipher wants to handle IVs or keys
15455 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15458 Change lots of functions like EVP_EncryptUpdate() to now return a
15459 value: although software versions of the algorithms cannot fail
15460 any installed hardware versions can.
15464 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15465 this option is set, tolerate broken clients that send the negotiated
15466 protocol version number instead of the requested protocol version
15471 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15472 i.e. non-zero for export ciphersuites, zero otherwise.
15473 Previous versions had this flag inverted, inconsistent with
15474 rsa_tmp_cb (..._TMP_RSA_CB).
15476 *Bodo Moeller; problem reported by Amit Chopra*
15478 * Add missing DSA library text string. Work around for some IIS
15479 key files with invalid SEQUENCE encoding.
15483 * Add a document (doc/standards.txt) that list all kinds of standards
15484 and so on that are implemented in OpenSSL.
15488 * Enhance c_rehash script. Old version would mishandle certificates
15489 with the same subject name hash and wouldn't handle CRLs at all.
15490 Added -fingerprint option to crl utility, to support new c_rehash
15495 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15499 * Fix for SSL server purpose checking. Server checking was
15500 rejecting certificates which had extended key usage present
15501 but no ssl client purpose.
15503 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15505 * Make PKCS#12 code work with no password. The PKCS#12 spec
15506 is a little unclear about how a blank password is handled.
15507 Since the password in encoded as a BMPString with terminating
15508 double NULL a zero length password would end up as just the
15509 double NULL. However no password at all is different and is
15510 handled differently in the PKCS#12 key generation code. NS
15511 treats a blank password as zero length. MSIE treats it as no
15512 password on export: but it will try both on import. We now do
15513 the same: PKCS12_parse() tries zero length and no password if
15514 the password is set to "" or NULL (NULL is now a valid password:
15515 it wasn't before) as does the pkcs12 application.
15519 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15520 perror when PEM_read_bio_X509_REQ fails, the error message must
15521 be obtained from the error queue.
15525 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15526 it in ERR_remove_state if appropriate, and change ERR_get_state
15527 accordingly to avoid race conditions (this is necessary because
15528 thread_hash is no longer constant once set).
15532 * Bugfix for linux-elf makefile.one.
15536 * RSA_get_default_method() will now cause a default
15537 RSA_METHOD to be chosen if one doesn't exist already.
15538 Previously this was only set during a call to RSA_new()
15539 or RSA_new_method(NULL) meaning it was possible for
15540 RSA_get_default_method() to return NULL.
15544 * Added native name translation to the existing DSO code
15545 that will convert (if the flag to do so is set) filenames
15546 that are sufficiently small and have no path information
15547 into a canonical native form. Eg. "blah" converted to
15548 "libblah.so" or "blah.dll" etc.
15552 * New function ERR_error_string_n(e, buf, len) which is like
15553 ERR_error_string(e, buf), but writes at most 'len' bytes
15554 including the 0 terminator. For ERR_error_string_n, 'buf'
15557 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15559 * CONF library reworked to become more general. A new CONF
15560 configuration file reader "class" is implemented as well as a
15561 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15562 old `CONF_*` functions are still there, but are reimplemented to
15563 work in terms of the new functions. Also, a set of functions
15564 to handle the internal storage of the configuration data is
15565 provided to make it easier to write new configuration file
15566 reader "classes" (I can definitely see something reading a
15567 configuration file in XML format, for example), called `_CONF_*`,
15568 or "the configuration storage API"...
15570 The new configuration file reading functions are:
15572 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15573 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15575 NCONF_default, NCONF_WIN32
15577 NCONF_dump_fp, NCONF_dump_bio
15579 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15580 NCONF_new creates a new CONF object. This works in the same way
15581 as other interfaces in OpenSSL, like the BIO interface.
15582 `NCONF_dump_*` dump the internal storage of the configuration file,
15583 which is useful for debugging. All other functions take the same
15584 arguments as the old `CONF_*` functions with the exception of the
15585 first that must be a `CONF *` instead of a `LHASH *`.
15587 To make it easier to use the new classes with the old `CONF_*` functions,
15588 the function CONF_set_default_method is provided.
15592 * Add '-tls1' option to 'openssl ciphers', which was already
15593 mentioned in the documentation but had not been implemented.
15594 (This option is not yet really useful because even the additional
15595 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15599 * Initial DSO code added into libcrypto for letting OpenSSL (and
15600 OpenSSL-based applications) load shared libraries and bind to
15601 them in a portable way.
15603 *Geoff Thorpe, with contributions from Richard Levitte*
15605 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15607 * Make sure _lrotl and _lrotr are only used with MSVC.
15609 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15610 (the default implementation of RAND_status).
15612 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15613 to '-clrext' (= clear extensions), as intended and documented.
15614 *Bodo Moeller; inconsistency pointed out by Michael Attili
15615 <attili@amaxo.com>*
15617 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15618 was larger than the MD block size.
15620 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15622 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15623 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15624 using the passed key: if the passed key was a private key the result
15625 of X509_print(), for example, would be to print out all the private key
15630 * des_quad_cksum() byte order bug fix.
15631 *Ulf Möller, using the problem description in krb4-0.9.7, where
15632 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15634 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15637 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15639 * For easily testing in shell scripts whether some command
15640 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15641 returns with exit code 0 iff no command of the given name is available.
15642 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15643 the output goes to stdout and nothing is printed to stderr.
15644 Additional arguments are always ignored.
15646 Since for each cipher there is a command of the same name,
15647 the 'no-cipher' compilation switches can be tested this way.
15649 ('openssl no-XXX' is not able to detect pseudo-commands such
15650 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15654 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15658 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15659 is set; it will be thrown away anyway because each handshake creates
15661 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15662 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15663 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15664 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15668 * New s_client option -ign_eof: EOF at stdin is ignored, and
15669 'Q' and 'R' lose their special meanings (quit/renegotiate).
15670 This is part of what -quiet does; unlike -quiet, -ign_eof
15671 does not suppress any output.
15675 * Add compatibility options to the purpose and trust code. The
15676 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15677 accepts a certificate or CA, this was the previous behaviour,
15678 with all the associated security issues.
15680 X509_TRUST_COMPAT is the old trust behaviour: only and
15681 automatically trust self signed roots in certificate store. A
15682 new trust setting X509_TRUST_DEFAULT is used to specify that
15683 a purpose has no associated trust setting and it should instead
15684 use the value in the default purpose.
15688 * Fix the PKCS#8 DSA private key code so it decodes keys again
15689 and fix a memory leak.
15693 * In util/mkerr.pl (which implements 'make errors'), preserve
15694 reason strings from the previous version of the .c file, as
15695 the default to have only downcase letters (and digits) in
15696 automatically generated reasons codes is not always appropriate.
15700 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15701 using strerror. Previously, ERR_reason_error_string() returned
15702 library names as reason strings for SYSerr; but SYSerr is a special
15703 case where small numbers are errno values, not library numbers.
15707 * Add '-dsaparam' option to 'openssl dhparam' application. This
15708 converts DSA parameters into DH parameters. (When creating parameters,
15709 DSA_generate_parameters is used.)
15713 * Include 'length' (recommended exponent length) in C code generated
15714 by 'openssl dhparam -C'.
15718 * The second argument to set_label in perlasm was already being used
15719 so couldn't be used as a "file scope" flag. Moved to third argument
15724 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15725 instead of RAND_bytes for encryption IVs and salts.
15729 * Include RAND_status() into RAND_METHOD instead of implementing
15730 it only for md_rand.c Otherwise replacing the PRNG by calling
15731 RAND_set_rand_method would be impossible.
15735 * Don't let DSA_generate_key() enter an infinite loop if the random
15736 number generation fails.
15740 * New 'rand' application for creating pseudo-random output.
15744 * Added configuration support for Linux/IA64
15746 *Rolf Haberrecker <rolf@suse.de>*
15748 * Assembler module support for Mingw32.
15752 * Shared library support for HPUX (in shlib/).
15754 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15756 * Shared library support for Solaris gcc.
15758 *Lutz Behnke <behnke@trustcenter.de>*
15760 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15762 * PKCS7_encrypt() was adding text MIME headers twice because they
15763 were added manually and by SMIME_crlf_copy().
15767 * In bntest.c don't call BN_rand with zero bits argument.
15769 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15771 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15772 case was implemented. This caused BN_div_recp() to fail occasionally.
15776 * Add an optional second argument to the set_label() in the perl
15777 assembly language builder. If this argument exists and is set
15778 to 1 it signals that the assembler should use a symbol whose
15779 scope is the entire file, not just the current function. This
15780 is needed with MASM which uses the format label:: for this scope.
15782 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15784 * Change the ASN1 types so they are typedefs by default. Before
15785 almost all types were #define'd to ASN1_STRING which was causing
15786 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15791 * Change names of new functions to the new get1/get0 naming
15792 convention: After 'get1', the caller owns a reference count
15793 and has to call `..._free`; 'get0' returns a pointer to some
15794 data structure without incrementing reference counters.
15795 (Some of the existing 'get' functions increment a reference
15796 counter, some don't.)
15797 Similarly, 'set1' and 'add1' functions increase reference
15798 counters or duplicate objects.
15802 * Allow for the possibility of temp RSA key generation failure:
15803 the code used to assume it always worked and crashed on failure.
15807 * Fix potential buffer overrun problem in BIO_printf().
15808 *Ulf Möller, using public domain code by Patrick Powell; problem
15809 pointed out by David Sacerdote <das33@cornell.edu>*
15811 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15812 RAND_egd() and RAND_status(). In the command line application,
15813 the EGD socket can be specified like a seed file using RANDFILE
15818 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15819 Some CAs (e.g. Verisign) distribute certificates in this form.
15823 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15824 list to exclude them. This means that no special compilation option
15825 is needed to use anonymous DH: it just needs to be included in the
15830 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15831 EVP_MD_type. The old functionality is available in a new macro called
15832 EVP_MD_md(). Change code that uses it and update docs.
15836 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15837 where the `void *` argument is replaced by a function pointer argument.
15838 Previously `void *` was abused to point to functions, which works on
15839 many platforms, but is not correct. As these functions are usually
15840 called by macros defined in OpenSSL header files, most source code
15841 should work without changes.
15845 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15846 sections with information on -D... compiler switches used for
15847 compiling the library so that applications can see them. To enable
15848 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15849 must be defined. E.g.,
15850 #define OPENSSL_ALGORITHM_DEFINES
15851 #include <openssl/opensslconf.h>
15852 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15854 *Richard Levitte, Ulf and Bodo Möller*
15856 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15861 * Change the 'other' type in certificate aux info to a STACK_OF
15862 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15863 the required ASN1 format: arbitrary types determined by an OID.
15867 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15868 argument to 'req'. This is not because the function is newer or
15869 better than others it just uses the work 'NEW' in the certificate
15870 request header lines. Some software needs this.
15874 * Reorganise password command line arguments: now passwords can be
15875 obtained from various sources. Delete the PEM_cb function and make
15876 it the default behaviour: i.e. if the callback is NULL and the
15877 usrdata argument is not NULL interpret it as a null terminated pass
15878 phrase. If usrdata and the callback are NULL then the pass phrase
15879 is prompted for as usual.
15883 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15884 the support is automatically enabled. The resulting binaries will
15885 autodetect the card and use it if present.
15887 *Ben Laurie and Compaq Inc.*
15889 * Work around for Netscape hang bug. This sends certificate request
15890 and server done in one record. Since this is perfectly legal in the
15891 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15892 the bugs/SSLv3 entry for more info.
15896 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15900 * Add -rand argument to smime and pkcs12 applications and read/write
15905 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15909 * Add command line password options to the remaining applications.
15913 * Bug fix for BN_div_recp() for numerators with an even number of
15918 * More tests in bntest.c, and changed test_bn output.
15922 * ./config recognizes MacOS X now.
15926 * Bug fix for BN_div() when the first words of num and divisor are
15927 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15931 * Add support for various broken PKCS#8 formats, and command line
15932 options to produce them.
15936 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15937 get temporary BIGNUMs from a BN_CTX.
15941 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15946 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15947 include a #define from the old name to the new. The original intent
15948 was that statically linked binaries could for example just call
15949 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15950 link with digests. This never worked because SSLeay_add_all_digests()
15951 and SSLeay_add_all_ciphers() were in the same source file so calling
15952 one would link with the other. They are now in separate source files.
15956 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15960 * Use a less unusual form of the Miller-Rabin primality test (it used
15961 a binary algorithm for exponentiation integrated into the Miller-Rabin
15962 loop, our standard modexp algorithms are faster).
15966 * Support for the EBCDIC character set completed.
15968 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15970 * Source code cleanups: use const where appropriate, eliminate casts,
15971 use `void *` instead of `char *` in lhash.
15975 * Bugfix: ssl3_send_server_key_exchange was not restartable
15976 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15977 this the server could overwrite ephemeral keys that the client
15982 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15983 using 50 iterations of the Rabin-Miller test.
15985 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15986 iterations of the Rabin-Miller test as required by the appendix
15987 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15988 As BN_is_prime_fasttest includes trial division, DSA parameter
15989 generation becomes much faster.
15991 This implies a change for the callback functions in DSA_is_prime
15992 and DSA_generate_parameters: The callback function is called once
15993 for each positive witness in the Rabin-Miller test, not just
15994 occasionally in the inner loop; and the parameters to the
15995 callback function now provide an iteration count for the outer
15996 loop rather than for the current invocation of the inner loop.
15997 DSA_generate_parameters additionally can call the callback
15998 function with an 'iteration count' of -1, meaning that a
15999 candidate has passed the trial division test (when q is generated
16000 from an application-provided seed, trial division is skipped).
16004 * New function BN_is_prime_fasttest that optionally does trial
16005 division before starting the Rabin-Miller test and has
16006 an additional BN_CTX * argument (whereas BN_is_prime always
16007 has to allocate at least one BN_CTX).
16008 'callback(1, -1, cb_arg)' is called when a number has passed the
16009 trial division stage.
16013 * Fix for bug in CRL encoding. The validity dates weren't being handled
16018 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16022 * New function BN_pseudo_rand().
16026 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16027 bignum version of BN_from_montgomery() with the working code from
16028 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16033 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16034 made it impossible to use the same SSL_SESSION data structure in
16035 SSL2 clients in multiple threads.
16039 * The return value of RAND_load_file() no longer counts bytes obtained
16040 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16041 to seed the PRNG (previously an explicit byte count was required).
16043 *Ulf Möller, Bodo Möller*
16045 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16046 used `char *` instead of `void *` and had casts all over the place.
16050 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16054 * Retain source code compatibility for BN_prime_checks macro:
16055 BN_is_prime(..., BN_prime_checks, ...) now uses
16056 BN_prime_checks_for_size to determine the appropriate number of
16057 Rabin-Miller iterations.
16061 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16062 DH_CHECK_P_NOT_SAFE_PRIME.
16063 (Check if this is true? OpenPGP calls them "strong".)
16067 * Merge the functionality of "dh" and "gendh" programs into a new program
16068 "dhparam". The old programs are retained for now but will handle DH keys
16069 (instead of parameters) in future.
16073 * Make the ciphers, s_server and s_client programs check the return values
16074 when a new cipher list is set.
16078 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16079 ciphers. Before when the 56bit ciphers were enabled the sorting was
16082 The syntax for the cipher sorting has been extended to support sorting by
16083 cipher-strength (using the strength_bits hard coded in the tables).
16084 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16086 Fix a bug in the cipher-command parser: when supplying a cipher command
16087 string with an "undefined" symbol (neither command nor alphanumeric
16088 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16089 an error is flagged.
16091 Due to the strength-sorting extension, the code of the
16092 ssl_create_cipher_list() function was completely rearranged. I hope that
16093 the readability was also increased :-)
16095 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16097 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16098 for the first serial number and places 2 in the serial number file. This
16099 avoids problems when the root CA is created with serial number zero and
16100 the first user certificate has the same issuer name and serial number
16105 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16106 the new code. Add documentation for this stuff.
16110 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16111 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16112 structures and behave in an analogous way to the X509v3 functions:
16113 they shouldn't be called directly but wrapper functions should be used
16116 So we also now have some wrapper functions that call the X509at functions
16117 when passed certificate requests. (TO DO: similar things can be done with
16118 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16119 things. Some of these need some d2i or i2d and print functionality
16120 because they handle more complex structures.)
16124 * Add missing #ifndefs that caused missing symbols when building libssl
16125 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16126 NO_RSA in `ssl/s2*.c`.
16128 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16130 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16131 has a return value which indicates the quality of the random data
16132 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16133 error queue. New function RAND_pseudo_bytes() generates output that is
16134 guaranteed to be unique but not unpredictable. RAND_add is like
16135 RAND_seed, but takes an extra argument for an entropy estimate
16136 (RAND_seed always assumes full entropy).
16140 * Do more iterations of Rabin-Miller probable prime test (specifically,
16141 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16142 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16143 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16144 false-positive rate of at most 2^-80 for random input.
16148 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16152 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16153 in the 0.9.5 release), this returns the chain
16154 from an X509_CTX structure with a dup of the stack and all
16155 the X509 reference counts upped: so the stack will exist
16156 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16159 Also make SSL_SESSION_print() print out the verify return
16164 * Add manpage for the pkcs12 command. Also change the default
16165 behaviour so MAC iteration counts are used unless the new
16166 -nomaciter option is used. This improves file security and
16167 only older versions of MSIE (4.0 for example) need it.
16171 * Honor the no-xxx Configure options when creating .DEF files.
16175 * Add PKCS#10 attributes to field table: challengePassword,
16176 unstructuredName and unstructuredAddress. These are taken from
16177 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16178 international characters are used.
16180 More changes to X509_ATTRIBUTE code: allow the setting of types
16181 based on strings. Remove the 'loc' parameter when adding
16182 attributes because these will be a SET OF encoding which is sorted
16187 * Initial changes to the 'req' utility to allow request generation
16188 automation. This will allow an application to just generate a template
16189 file containing all the field values and have req construct the
16192 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16193 used all over the place including certificate requests and PKCS#7
16194 structures. They are currently handled manually where necessary with
16195 some primitive wrappers for PKCS#7. The new functions behave in a
16196 manner analogous to the X509 extension functions: they allow
16197 attributes to be looked up by NID and added.
16199 Later something similar to the X509V3 code would be desirable to
16200 automatically handle the encoding, decoding and printing of the
16201 more complex types. The string types like challengePassword can
16202 be handled by the string table functions.
16204 Also modified the multi byte string table handling. Now there is
16205 a 'global mask' which masks out certain types. The table itself
16206 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16207 is useful when for example there is only one permissible type
16208 (as in countryName) and using the mask might result in no valid
16213 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16214 SSL_get_peer_finished to allow applications to obtain the latest
16215 Finished messages sent to the peer or expected from the peer,
16216 respectively. (SSL_get_peer_finished is usually the Finished message
16217 actually received from the peer, otherwise the protocol will be aborted.)
16219 As the Finished message are message digests of the complete handshake
16220 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16221 be used for external authentication procedures when the authentication
16222 provided by SSL/TLS is not desired or is not enough.
16226 * Enhanced support for Alpha Linux is added. Now ./config checks if
16227 the host supports BWX extension and if Compaq C is present on the
16228 $PATH. Just exploiting of the BWX extension results in 20-30%
16229 performance kick for some algorithms, e.g. DES and RC4 to mention
16230 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16235 * Add support for MS "fast SGC". This is arguably a violation of the
16236 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16237 weak crypto and after checking the certificate is SGC a second one
16238 with strong crypto. MS SGC stops the first handshake after receiving
16239 the server certificate message and sends a second client hello. Since
16240 a server will typically do all the time consuming operations before
16241 expecting any further messages from the client (server key exchange
16242 is the most expensive) there is little difference between the two.
16244 To get OpenSSL to support MS SGC we have to permit a second client
16245 hello message after we have sent server done. In addition we have to
16246 reset the MAC if we do get this second client hello.
16250 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16251 if a DER encoded private key is RSA or DSA traditional format. Changed
16252 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16253 format DER encoded private key. Newer code should use PKCS#8 format which
16254 has the key type encoded in the ASN1 structure. Added DER private key
16255 support to pkcs8 application.
16259 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16260 ciphersuites has been selected (as required by the SSL 3/TLS 1
16261 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16262 is set, we interpret this as a request to violate the specification
16263 (the worst that can happen is a handshake failure, and 'correct'
16264 behaviour would result in a handshake failure anyway).
16268 * In SSL_CTX_add_session, take into account that there might be multiple
16269 SSL_SESSION structures with the same session ID (e.g. when two threads
16270 concurrently obtain them from an external cache).
16271 The internal cache can handle only one SSL_SESSION with a given ID,
16272 so if there's a conflict, we now throw out the old one to achieve
16277 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16278 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16279 some routines that use cipher OIDs: some ciphers do not have OIDs
16280 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16285 * Simplify the trust setting structure and code. Now we just have
16286 two sequences of OIDs for trusted and rejected settings. These will
16287 typically have values the same as the extended key usage extension
16288 and any application specific purposes.
16290 The trust checking code now has a default behaviour: it will just
16291 check for an object with the same NID as the passed id. Functions can
16292 be provided to override either the default behaviour or the behaviour
16293 for a given id. SSL client, server and email already have functions
16294 in place for compatibility: they check the NID and also return "trusted"
16295 if the certificate is self signed.
16299 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16300 traditional format into an EVP_PKEY structure.
16304 * Add a password callback function PEM_cb() which either prompts for
16305 a password if usr_data is NULL or otherwise assumes it is a null
16306 terminated password. Allow passwords to be passed on command line
16307 environment or config files in a few more utilities.
16311 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16312 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16313 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16314 Update documentation.
16318 * Support for ASN1 "NULL" type. This could be handled before by using
16319 ASN1_TYPE but there wasn't any function that would try to read a NULL
16320 and produce an error if it couldn't. For compatibility we also have
16321 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16322 don't allocate anything because they don't need to.
16326 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16329 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16331 * Rebuild of the memory allocation routines used by OpenSSL code and
16332 possibly others as well. The purpose is to make an interface that
16333 provide hooks so anyone can build a separate set of allocation and
16334 deallocation routines to be used by OpenSSL, for example memory
16335 pool implementations, or something else, which was previously hard
16336 since Malloc(), Realloc() and Free() were defined as macros having
16337 the values malloc, realloc and free, respectively (except for Win32
16338 compilations). The same is provided for memory debugging code.
16339 OpenSSL already comes with functionality to find memory leaks, but
16340 this gives people a chance to debug other memory problems.
16342 With these changes, a new set of functions and macros have appeared:
16344 CRYPTO_set_mem_debug_functions() [F]
16345 CRYPTO_get_mem_debug_functions() [F]
16346 CRYPTO_dbg_set_options() [F]
16347 CRYPTO_dbg_get_options() [F]
16348 CRYPTO_malloc_debug_init() [M]
16350 The memory debug functions are NULL by default, unless the library
16351 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16352 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16353 gives the standard debugging functions that come with OpenSSL) or
16354 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16355 provided by the library user) must be used. When the standard
16356 debugging functions are used, CRYPTO_dbg_set_options can be used to
16357 request additional information:
16358 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16359 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16361 Also, things like CRYPTO_set_mem_functions will always give the
16362 expected result (the new set of functions is used for allocation
16363 and deallocation) at all times, regardless of platform and compiler
16366 To finish it up, some functions that were never use in any other
16367 way than through macros have a new API and new semantic:
16369 CRYPTO_dbg_malloc()
16370 CRYPTO_dbg_realloc()
16373 All macros of value have retained their old syntax.
16375 *Richard Levitte and Bodo Moeller*
16377 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16378 ordering of SMIMECapabilities wasn't in "strength order" and there
16379 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16384 * Some ASN1 types with illegal zero length encoding (INTEGER,
16385 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16387 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16389 * Merge in my S/MIME library for OpenSSL. This provides a simple
16390 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16391 functionality to handle multipart/signed properly) and a utility
16392 called 'smime' to call all this stuff. This is based on code I
16393 originally wrote for Celo who have kindly allowed it to be
16394 included in OpenSSL.
16398 * Add variants des_set_key_checked and des_set_key_unchecked of
16399 des_set_key (aka des_key_sched). Global variable des_check_key
16400 decides which of these is called by des_set_key; this way
16401 des_check_key behaves as it always did, but applications and
16402 the library itself, which was buggy for des_check_key == 1,
16403 have a cleaner way to pick the version they need.
16407 * New function PKCS12_newpass() which changes the password of a
16412 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16413 dynamic mix. In both cases the ids can be used as an index into the
16414 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16415 functions so they accept a list of the field values and the
16416 application doesn't need to directly manipulate the X509_TRUST
16421 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16426 * Modify the way the V3 extension code looks up extensions. This now
16427 works in a similar way to the object code: we have some "standard"
16428 extensions in a static table which is searched with OBJ_bsearch()
16429 and the application can add dynamic ones if needed. The file
16430 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16431 updated whenever a new extension is added to the core code and kept
16432 in ext_nid order. There is a simple program 'tabtest.c' which checks
16433 this. New extensions are not added too often so this file can readily
16434 be maintained manually.
16436 There are two big advantages in doing things this way. The extensions
16437 can be looked up immediately and no longer need to be "added" using
16438 X509V3_add_standard_extensions(): this function now does nothing.
16439 Side note: I get *lots* of email saying the extension code doesn't
16440 work because people forget to call this function.
16441 Also no dynamic allocation is done unless new extensions are added:
16442 so if we don't add custom extensions there is no need to call
16443 X509V3_EXT_cleanup().
16447 * Modify enc utility's salting as follows: make salting the default. Add a
16448 magic header, so unsalted files fail gracefully instead of just decrypting
16449 to garbage. This is because not salting is a big security hole, so people
16450 should be discouraged from doing it.
16454 * Fixes and enhancements to the 'x509' utility. It allowed a message
16455 digest to be passed on the command line but it only used this
16456 parameter when signing a certificate. Modified so all relevant
16457 operations are affected by the digest parameter including the
16458 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16459 DSA key was used because it didn't fix the digest.
16463 * Initial certificate chain verify code. Currently tests the untrusted
16464 certificates for consistency with the verify purpose (which is set
16465 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16467 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16468 this is because it will reject chains with invalid extensions whereas
16469 every previous version of OpenSSL and SSLeay made no checks at all.
16471 Trust code: checks the root CA for the relevant trust settings. Trust
16472 settings have an initial value consistent with the verify purpose: e.g.
16473 if the verify purpose is for SSL client use it expects the CA to be
16474 trusted for SSL client use. However the default value can be changed to
16475 permit custom trust settings: one example of this would be to only trust
16476 certificates from a specific "secure" set of CAs.
16478 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16479 which should be used for version portability: especially since the
16480 verify structure is likely to change more often now.
16482 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16483 to set them. If not set then assume SSL clients will verify SSL servers
16486 Two new options to the verify program: -untrusted allows a set of
16487 untrusted certificates to be passed in and -purpose which sets the
16488 intended purpose of the certificate. If a purpose is set then the
16489 new chain verify code is used to check extension consistency.
16493 * Support for the authority information access extension.
16497 * Modify RSA and DSA PEM read routines to transparently handle
16498 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16499 public keys in a format compatible with certificate
16500 SubjectPublicKeyInfo structures. Unfortunately there were already
16501 functions called *_PublicKey_* which used various odd formats so
16502 these are retained for compatibility: however the DSA variants were
16503 never in a public release so they have been deleted. Changed dsa/rsa
16504 utilities to handle the new format: note no releases ever handled public
16505 keys so we should be OK.
16507 The primary motivation for this change is to avoid the same fiasco
16508 that dogs private keys: there are several incompatible private key
16509 formats some of which are standard and some OpenSSL specific and
16510 require various evil hacks to allow partial transparent handling and
16511 even then it doesn't work with DER formats. Given the option anything
16512 other than PKCS#8 should be dumped: but the other formats have to
16513 stay in the name of compatibility.
16515 With public keys and the benefit of hindsight one standard format
16516 is used which works with EVP_PKEY, RSA or DSA structures: though
16517 it clearly returns an error if you try to read the wrong kind of key.
16519 Added a -pubkey option to the 'x509' utility to output the public key.
16520 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16521 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16522 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16523 that do the same as the `EVP_PKEY_assign_*()` except they up the
16524 reference count of the added key (they don't "swallow" the
16529 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16530 CRLs would fail if the file contained no certificates or no CRLs:
16531 added a new function to read in both types and return the number
16532 read: this means that if none are read it will be an error. The
16533 DER versions of the certificate and CRL reader would always fail
16534 because it isn't possible to mix certificates and CRLs in DER format
16535 without choking one or the other routine. Changed this to just read
16536 a certificate: this is the best we can do. Also modified the code
16537 in `apps/verify.c` to take notice of return codes: it was previously
16538 attempting to read in certificates from NULL pointers and ignoring
16539 any errors: this is one reason why the cert and CRL reader seemed
16540 to work. It doesn't check return codes from the default certificate
16541 routines: these may well fail if the certificates aren't installed.
16545 * Code to support otherName option in GeneralName.
16549 * First update to verify code. Change the verify utility
16550 so it warns if it is passed a self signed certificate:
16551 for consistency with the normal behaviour. X509_verify
16552 has been modified to it will now verify a self signed
16553 certificate if *exactly* the same certificate appears
16554 in the store: it was previously impossible to trust a
16555 single self signed certificate. This means that:
16556 openssl verify ss.pem
16557 now gives a warning about a self signed certificate but
16558 openssl verify -CAfile ss.pem ss.pem
16563 * For servers, store verify_result in SSL_SESSION data structure
16564 (and add it to external session representation).
16565 This is needed when client certificate verifications fails,
16566 but an application-provided verification callback (set by
16567 SSL_CTX_set_cert_verify_callback) allows accepting the session
16568 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16569 but returns 1): When the session is reused, we have to set
16570 ssl->verify_result to the appropriate error code to avoid
16573 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16575 * Fix a bug in the new PKCS#7 code: it didn't consider the
16576 case in PKCS7_dataInit() where the signed PKCS7 structure
16577 didn't contain any existing data because it was being created.
16579 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16581 * Add a salt to the key derivation routines in enc.c. This
16582 forms the first 8 bytes of the encrypted file. Also add a
16583 -S option to allow a salt to be input on the command line.
16587 * New function X509_cmp(). Oddly enough there wasn't a function
16588 to compare two certificates. We do this by working out the SHA1
16589 hash and comparing that. X509_cmp() will be needed by the trust
16594 * SSL_get1_session() is like SSL_get_session(), but increments
16595 the reference count in the SSL_SESSION returned.
16597 *Geoff Thorpe <geoff@eu.c2.net>*
16599 * Fix for 'req': it was adding a null to request attributes.
16600 Also change the X509_LOOKUP and X509_INFO code to handle
16601 certificate auxiliary information.
16605 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16610 * Add the possibility to add extra information to the memory leak
16611 detecting output, to form tracebacks, showing from where each
16612 allocation was originated: CRYPTO_push_info("constant string") adds
16613 the string plus current file name and line number to a per-thread
16614 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16615 is like calling CYRPTO_pop_info() until the stack is empty.
16616 Also updated memory leak detection code to be multi-thread-safe.
16620 * Add options -text and -noout to pkcs7 utility and delete the
16621 encryption options which never did anything. Update docs.
16625 * Add options to some of the utilities to allow the pass phrase
16626 to be included on either the command line (not recommended on
16627 OSes like Unix) or read from the environment. Update the
16628 manpages and fix a few bugs.
16632 * Add a few manpages for some of the openssl commands.
16636 * Fix the -revoke option in ca. It was freeing up memory twice,
16637 leaking and not finding already revoked certificates.
16641 * Extensive changes to support certificate auxiliary information.
16642 This involves the use of X509_CERT_AUX structure and X509_AUX
16643 functions. An X509_AUX function such as PEM_read_X509_AUX()
16644 can still read in a certificate file in the usual way but it
16645 will also read in any additional "auxiliary information". By
16646 doing things this way a fair degree of compatibility can be
16647 retained: existing certificates can have this information added
16648 using the new 'x509' options.
16650 Current auxiliary information includes an "alias" and some trust
16651 settings. The trust settings will ultimately be used in enhanced
16652 certificate chain verification routines: currently a certificate
16653 can only be trusted if it is self signed and then it is trusted
16658 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16659 The problem was that one of the replacement routines had not been working
16660 since SSLeay releases. For now the offending routine has been replaced
16661 with non-optimised assembler. Even so, this now gives around 95%
16662 performance improvement for 1024 bit RSA signs.
16666 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16667 handling. Most clients have the effective key size in bits equal to
16668 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16669 A few however don't do this and instead use the size of the decrypted key
16670 to determine the RC2 key length and the AlgorithmIdentifier to determine
16671 the effective key length. In this case the effective key length can still
16672 be 40 bits but the key length can be 168 bits for example. This is fixed
16673 by manually forcing an RC2 key into the EVP_PKEY structure because the
16674 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16675 the key length and effective key length are equal.
16679 * Add a bunch of functions that should simplify the creation of
16680 X509_NAME structures. Now you should be able to do:
16681 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16682 and have it automatically work out the correct field type and fill in
16683 the structures. The more adventurous can try:
16684 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16685 and it will (hopefully) work out the correct multibyte encoding.
16689 * Change the 'req' utility to use the new field handling and multibyte
16690 copy routines. Before the DN field creation was handled in an ad hoc
16691 way in req, ca, and x509 which was rather broken and didn't support
16692 BMPStrings or UTF8Strings. Since some software doesn't implement
16693 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16694 using the dirstring_type option. See the new comment in the default
16695 openssl.cnf for more info.
16699 * Make crypto/rand/md_rand.c more robust:
16700 - Assure unique random numbers after fork().
16701 - Make sure that concurrent threads access the global counter and
16702 md serializably so that we never lose entropy in them
16703 or use exactly the same state in multiple threads.
16704 Access to the large state is not always serializable because
16705 the additional locking could be a performance killer, and
16706 md should be large enough anyway.
16710 * New file `apps/app_rand.c` with commonly needed functionality
16711 for handling the random seed file.
16713 Use the random seed file in some applications that previously did not:
16715 dsaparam -genkey (which also ignored its '-rand' option),
16718 x509 (when signing).
16719 Except on systems with /dev/urandom, it is crucial to have a random
16720 seed file at least for key creation, DSA signing, and for DH exchanges;
16721 for RSA signatures we could do without one.
16723 gendh and gendsa (unlike genrsa) used to read only the first byte
16724 of each file listed in the '-rand' option. The function as previously
16725 found in genrsa is now in app_rand.c and is used by all programs
16726 that support '-rand'.
16730 * In RAND_write_file, use mode 0600 for creating files;
16731 don't just chmod when it may be too late.
16735 * Report an error from X509_STORE_load_locations
16736 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16740 * New function ASN1_mbstring_copy() this copies a string in either
16741 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16742 into an ASN1_STRING type. A mask of permissible types is passed
16743 and it chooses the "minimal" type to use or an error if not type
16748 * Add function equivalents to the various macros in asn1.h. The old
16749 macros are retained with an `M_` prefix. Code inside the library can
16750 use the `M_` macros. External code (including the openssl utility)
16751 should *NOT* in order to be "shared library friendly".
16755 * Add various functions that can check a certificate's extensions
16756 to see if it usable for various purposes such as SSL client,
16757 server or S/MIME and CAs of these types. This is currently
16758 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16759 verification. Also added a -purpose flag to x509 utility to
16760 print out all the purposes.
16764 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16769 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16770 for, obtain and decode and extension and obtain its critical flag.
16771 This allows all the necessary extension code to be handled in a
16772 single function call.
16776 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16777 platforms. See crypto/rc4/rc4_enc.c for further details.
16781 * New -noout option to asn1parse. This causes no output to be produced
16782 its main use is when combined with -strparse and -out to extract data
16783 from a file (which may not be in ASN.1 format).
16787 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16788 when producing the local key id.
16790 *Richard Levitte <levitte@stacken.kth.se>*
16792 * New option -dhparam in s_server. This allows a DH parameter file to be
16793 stated explicitly. If it is not stated then it tries the first server
16794 certificate file. The previous behaviour hard coded the filename
16799 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16800 a public key to be input or output. For example:
16801 openssl rsa -in key.pem -pubout -out pubkey.pem
16802 Also added necessary DSA public key functions to handle this.
16806 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16807 in the message. This was handled by allowing
16808 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16810 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16812 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16813 to the end of the strings whereas this didn't. This would cause problems
16814 if strings read with d2i_ASN1_bytes() were later modified.
16816 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16818 * Fix for base64 decode bug. When a base64 bio reads only one line of
16819 data and it contains EOF it will end up returning an error. This is
16820 caused by input 46 bytes long. The cause is due to the way base64
16821 BIOs find the start of base64 encoded data. They do this by trying a
16822 trial decode on each line until they find one that works. When they
16823 do a flag is set and it starts again knowing it can pass all the
16824 data directly through the decoder. Unfortunately it doesn't reset
16825 the context it uses. This means that if EOF is reached an attempt
16826 is made to pass two EOFs through the context and this causes the
16827 resulting error. This can also cause other problems as well. As is
16828 usual with these problems it takes *ages* to find and the fix is
16829 trivial: move one line.
16831 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16833 * Ugly workaround to get s_client and s_server working under Windows. The
16834 old code wouldn't work because it needed to select() on sockets and the
16835 tty (for keypresses and to see if data could be written). Win32 only
16836 supports select() on sockets so we select() with a 1s timeout on the
16837 sockets and then see if any characters are waiting to be read, if none
16838 are present then we retry, we also assume we can always write data to
16839 the tty. This isn't nice because the code then blocks until we've
16840 received a complete line of data and it is effectively polling the
16841 keyboard at 1s intervals: however it's quite a bit better than not
16842 working at all :-) A dedicated Windows application might handle this
16843 with an event loop for example.
16847 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16848 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16849 will be called when RSA_sign() and RSA_verify() are used. This is useful
16850 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16851 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16852 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16853 This necessitated the support of an extra signature type NID_md5_sha1
16854 for SSL signatures and modifications to the SSL library to use it instead
16855 of calling RSA_public_decrypt() and RSA_private_encrypt().
16859 * Add new -verify -CAfile and -CApath options to the crl program, these
16860 will lookup a CRL issuers certificate and verify the signature in a
16861 similar way to the verify program. Tidy up the crl program so it
16862 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16863 less strict. It will now permit CRL extensions even if it is not
16864 a V2 CRL: this will allow it to tolerate some broken CRLs.
16868 * Initialize all non-automatic variables each time one of the openssl
16869 sub-programs is started (this is necessary as they may be started
16870 multiple times from the "OpenSSL>" prompt).
16872 *Lennart Bang, Bodo Moeller*
16874 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16875 removing all other RSA functionality (this is what NO_RSA does). This
16876 is so (for example) those in the US can disable those operations covered
16877 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16882 * Non-copying interface to BIO pairs.
16883 (still largely untested)
16887 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16888 ASCII string. This was handled independently in various places before.
16892 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16893 UTF8 strings a character at a time.
16897 * Use client_version from client hello to select the protocol
16898 (s23_srvr.c) and for RSA client key exchange verification
16899 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16903 * Add various utility functions to handle SPKACs, these were previously
16904 handled by poking round in the structure internals. Added new function
16905 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16906 print, verify and generate SPKACs. Based on an original idea from
16907 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16911 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16915 * Allow the config file extension section to be overwritten on the
16916 command line. Based on an original idea from Massimiliano Pala
16917 <madwolf@comune.modena.it>. The new option is called -extensions
16918 and can be applied to ca, req and x509. Also -reqexts to override
16919 the request extensions in req and -crlexts to override the crl extensions
16924 * Add new feature to the SPKAC handling in ca. Now you can include
16925 the same field multiple times by preceding it by "XXXX." for example:
16928 this is the same syntax as used in the req config file.
16932 * Allow certificate extensions to be added to certificate requests. These
16933 are specified in a 'req_extensions' option of the req section of the
16934 config file. They can be printed out with the -text option to req but
16935 are otherwise ignored at present.
16939 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16940 data read consists of only the final block it would not decrypted because
16941 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16942 A misplaced 'break' also meant the decrypted final block might not be
16943 copied until the next read.
16947 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16948 a few extra parameters to the DH structure: these will be useful if
16949 for example we want the value of 'q' or implement X9.42 DH.
16953 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16954 provides hooks that allow the default DSA functions or functions on a
16955 "per key" basis to be replaced. This allows hardware acceleration and
16956 hardware key storage to be handled without major modification to the
16957 library. Also added low-level modexp hooks and CRYPTO_EX structure and
16958 associated functions.
16962 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16963 as "read only": it can't be written to and the buffer it points to will
16964 not be freed. Reading from a read only BIO is much more efficient than
16965 a normal memory BIO. This was added because there are several times when
16966 an area of memory needs to be read from a BIO. The previous method was
16967 to create a memory BIO and write the data to it, this results in two
16968 copies of the data and an O(n^2) reading algorithm. There is a new
16969 function BIO_new_mem_buf() which creates a read only memory BIO from
16970 an area of memory. Also modified the PKCS#7 routines to use read only
16975 * Bugfix: ssl23_get_client_hello did not work properly when called in
16976 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16977 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16978 but a retry condition occurred while trying to read the rest.
16982 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16983 NID_pkcs7_encrypted by default: this was wrong since this should almost
16984 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16985 the encrypted data type: this is a more sensible place to put it and it
16986 allows the PKCS#12 code to be tidied up that duplicated this
16991 * Changed obj_dat.pl script so it takes its input and output files on
16992 the command line. This should avoid shell escape redirection problems
16997 * Initial support for certificate extension requests, these are included
16998 in things like Xenroll certificate requests. Included functions to allow
16999 extensions to be obtained and added.
17003 * -crlf option to s_client and s_server for sending newlines as
17004 CRLF (as required by many protocols).
17008 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17010 * Install libRSAglue.a when OpenSSL is built with RSAref.
17012 *Ralf S. Engelschall*
17014 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17016 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17018 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17023 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17024 DH parameters/keys (q is lost during that conversion, but the resulting
17025 DH parameters contain its length).
17027 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17028 much faster than DH_generate_parameters (which creates parameters
17029 where `p = 2*q + 1`), and also the smaller q makes DH computations
17030 much more efficient (160-bit exponentiation instead of 1024-bit
17031 exponentiation); so this provides a convenient way to support DHE
17032 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17033 utter importance to use
17034 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17036 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17037 when such DH parameters are used, because otherwise small subgroup
17038 attacks may become possible!
17042 * Avoid memory leak in i2d_DHparams.
17046 * Allow the -k option to be used more than once in the enc program:
17047 this allows the same encrypted message to be read by multiple recipients.
17051 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17052 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17053 it will always use the numerical form of the OID, even if it has a short
17058 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17059 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17060 otherwise bn_mod_exp was called. In the case of hardware keys for example
17061 no private key components need be present and it might store extra data
17062 in the RSA structure, which cannot be accessed from bn_mod_exp.
17063 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17064 private key operations.
17068 * Added support for SPARC Linux.
17072 * pem_password_cb function type incompatibly changed from
17073 typedef int pem_password_cb(char *buf, int size, int rwflag);
17075 ....(char *buf, int size, int rwflag, void *userdata);
17076 so that applications can pass data to their callbacks:
17077 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17078 additional void * argument, which is just handed through whenever
17079 the password callback is called.
17081 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17083 New function SSL_CTX_set_default_passwd_cb_userdata.
17085 Compatibility note: As many C implementations push function arguments
17086 onto the stack in reverse order, the new library version is likely to
17087 interoperate with programs that have been compiled with the old
17088 pem_password_cb definition (PEM_whatever takes some data that
17089 happens to be on the stack as its last argument, and the callback
17090 just ignores this garbage); but there is no guarantee whatsoever that
17093 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17094 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17095 problems not only on Windows, but also on some Unix platforms.
17096 To avoid problematic command lines, these definitions are now in an
17097 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17098 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17102 * MIPS III/IV assembler module is reimplemented.
17106 * More DES library cleanups: remove references to srand/rand and
17107 delete an unused file.
17111 * Add support for the free Netwide assembler (NASM) under Win32,
17112 since not many people have MASM (ml) and it can be hard to obtain.
17113 This is currently experimental but it seems to work OK and pass all
17114 the tests. Check out INSTALL.W32 for info.
17118 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17119 without temporary keys kept an extra copy of the server key,
17120 and connections with temporary keys did not free everything in case
17125 * New function RSA_check_key and new openssl rsa option -check
17126 for verifying the consistency of RSA keys.
17128 *Ulf Moeller, Bodo Moeller*
17130 * Various changes to make Win32 compile work:
17131 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17132 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17133 comparison" warnings.
17134 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17138 * Add a debugging option to PKCS#5 v2 key generation function: when
17139 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17140 derived keys are printed to stderr.
17144 * Copy the flags in ASN1_STRING_dup().
17146 *Roman E. Pavlov <pre@mo.msk.ru>*
17148 * The x509 application mishandled signing requests containing DSA
17149 keys when the signing key was also DSA and the parameters didn't match.
17151 It was supposed to omit the parameters when they matched the signing key:
17152 the verifying software was then supposed to automatically use the CA's
17153 parameters if they were absent from the end user certificate.
17155 Omitting parameters is no longer recommended. The test was also
17156 the wrong way round! This was probably due to unusual behaviour in
17157 EVP_cmp_parameters() which returns 1 if the parameters match.
17158 This meant that parameters were omitted when they *didn't* match and
17159 the certificate was useless. Certificates signed with 'ca' didn't have
17162 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17164 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17165 The interface is as follows:
17166 Applications can use
17167 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17168 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17169 "off" is now the default.
17170 The library internally uses
17171 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17172 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17173 to disable memory-checking temporarily.
17175 Some inconsistent states that previously were possible (and were
17176 even the default) are now avoided.
17178 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17179 with each memory chunk allocated; this is occasionally more helpful
17180 than just having a counter.
17182 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17184 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17189 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17190 which largely parallels "options", but is for changing API behaviour,
17191 whereas "options" are about protocol behaviour.
17192 Initial "mode" flags are:
17194 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17195 a single record has been written.
17196 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17197 retries use the same buffer location.
17198 (But all of the contents must be
17203 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17206 * Fix problems with no-hmac etc.
17208 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17210 * New functions RSA_get_default_method(), RSA_set_method() and
17211 RSA_get_method(). These allows replacement of RSA_METHODs without having
17212 to mess around with the internals of an RSA structure.
17216 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17217 Also really enable memory leak checks in openssl.c and in some
17220 *Chad C. Mulligan, Bodo Moeller*
17222 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17223 up the length of negative integers. This has now been simplified to just
17224 store the length when it is first determined and use it later, rather
17225 than trying to keep track of where data is copied and updating it to
17227 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17229 * Add a new function PKCS7_signatureVerify. This allows the verification
17230 of a PKCS#7 signature but with the signing certificate passed to the
17231 function itself. This contrasts with PKCS7_dataVerify which assumes the
17232 certificate is present in the PKCS#7 structure. This isn't always the
17233 case: certificates can be omitted from a PKCS#7 structure and be
17234 distributed by "out of band" means (such as a certificate database).
17238 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17239 function prototypes in pem.h, also change util/mkdef.pl to add the
17240 necessary function names.
17244 * mk1mf.pl (used by Windows builds) did not properly read the
17245 options set by Configure in the top level Makefile, and Configure
17246 was not even able to write more than one option correctly.
17247 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17251 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17252 file to be loaded from a BIO or FILE pointer. The BIO version will
17253 for example allow memory BIOs to contain config info.
17257 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17258 Whoever hopes to achieve shared-library compatibility across versions
17259 must use this, not the compile-time macro.
17260 (Exercise 0.9.4: Which is the minimum library version required by
17262 Note: All this applies only to multi-threaded programs, others don't
17267 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17268 through a BIO pair triggered the default case, i.e.
17269 SSLerr(...,SSL_R_UNKNOWN_STATE).
17273 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17274 can use the SSL library even if none of the specific BIOs is
17279 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17280 for the encoded length.
17282 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17284 * Add initial documentation of the X509V3 functions.
17288 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17289 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17290 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17291 secure PKCS#8 private key format with a high iteration count.
17295 * Fix determination of Perl interpreter: A perl or perl5
17296 *directory* in $PATH was also accepted as the interpreter.
17298 *Ralf S. Engelschall*
17300 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17301 wrong with it but it was very old and did things like calling
17302 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17303 unusual formatting.
17307 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17308 to use the new extension code.
17312 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17313 with macros. This should make it easier to change their form, add extra
17314 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17319 * Add to configuration table a new entry that can specify an alternative
17320 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17321 according to Mark Crispin <MRC@Panda.COM>.
17325 * DES CBC did not update the IV. Weird.
17329 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17330 Changing the behaviour of the former might break existing programs --
17331 where IV updating is needed, des_ncbc_encrypt can be used.
17334 * When bntest is run from "make test" it drives bc to check its
17335 calculations, as well as internally checking them. If an internal check
17336 fails, it needs to cause bc to give a non-zero result or make test carries
17337 on without noticing the failure. Fixed.
17341 * DES library cleanups.
17345 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17346 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17347 ciphers. NOTE: although the key derivation function has been verified
17348 against some published test vectors it has not been extensively tested
17349 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17354 * Instead of "mkdir -p", which is not fully portable, use new
17355 Perl script "util/mkdir-p.pl".
17359 * Rewrite the way password based encryption (PBE) is handled. It used to
17360 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17361 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17362 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17363 the 'parameter' field of the AlgorithmIdentifier is passed to the
17364 underlying key generation function so it must do its own ASN1 parsing.
17365 This has also changed the EVP_PBE_CipherInit() function which now has a
17366 'parameter' argument instead of literal salt and iteration count values
17367 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17371 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17372 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17373 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17374 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17375 value was just used as a "magic string" and not used directly its
17376 value doesn't matter.
17380 * Introduce some semblance of const correctness to BN. Shame C doesn't
17385 * "linux-sparc64" configuration (ultrapenguin).
17387 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17388 "linux-sparc" configuration.
17390 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17392 * config now generates no-xxx options for missing ciphers.
17396 * Support the EBCDIC character set (work in progress).
17397 File ebcdic.c not yet included because it has a different license.
17399 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17401 * Support BS2000/OSD-POSIX.
17403 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17405 * Make callbacks for key generation use `void *` instead of `char *`.
17409 * Make S/MIME samples compile (not yet tested).
17413 * Additional typesafe stacks.
17417 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17421 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17423 * New configuration variant "sco5-gcc".
17425 * Updated some demos.
17427 *Sean O Riordain, Wade Scholine*
17429 * Add missing BIO_free at exit of pkcs12 application.
17433 * Fix memory leak in conf.c.
17437 * Updates for Win32 to assembler version of MD5.
17441 * Set #! path to perl in `apps/der_chop` to where we found it
17442 instead of using a fixed path.
17446 * SHA library changes for irix64-mips4-cc.
17450 * Improvements for VMS support.
17454 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17456 * Bignum library bug fix. IRIX 6 passes "make test" now!
17457 This also avoids the problems with SC4.2 and unpatched SC5.
17459 *Andy Polyakov <appro@fy.chalmers.se>*
17461 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17462 These are required because of the typesafe stack would otherwise break
17463 existing code. If old code used a structure member which used to be STACK
17464 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17465 sk_num or sk_value it would produce an error because the num, data members
17466 are not present in STACK_OF. Now it just produces a warning. sk_set
17467 replaces the old method of assigning a value to sk_value
17468 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17469 that does this will no longer work (and should use sk_set instead) but
17470 this could be regarded as a "questionable" behaviour anyway.
17474 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17475 correctly handle encrypted S/MIME data.
17479 * Change type of various DES function arguments from des_cblock
17480 (which means, in function argument declarations, pointer to char)
17481 to des_cblock * (meaning pointer to array with 8 char elements),
17482 which allows the compiler to do more typechecking; it was like
17483 that back in SSLeay, but with lots of ugly casts.
17485 Introduce new type const_des_cblock.
17489 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17490 problems: find RecipientInfo structure that matches recipient certificate
17491 and initialise the ASN1 structures properly based on passed cipher.
17495 * Belatedly make the BN tests actually check the results.
17499 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17500 to and from BNs: it was completely broken. New compilation option
17501 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17502 key elements as negative integers.
17506 * Reorganize and speed up MD5.
17508 *Andy Polyakov <appro@fy.chalmers.se>*
17512 *Richard Levitte <richard@levitte.org>*
17514 * New option -out to asn1parse to allow the parsed structure to be
17515 output to a file. This is most useful when combined with the -strparse
17516 option to examine the output of things like OCTET STRINGS.
17520 * Make SSL library a little more fool-proof by not requiring any longer
17521 that `SSL_set_{accept,connect}_state` be called before
17522 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17523 in many applications because usually everything *appeared* to work as
17524 intended anyway -- now it really works as intended).
17528 * Move openssl.cnf out of lib/.
17532 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17533 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17534 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17536 *Ralf S. Engelschall*
17538 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17539 handle PKCS#7 enveloped data properly.
17541 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17543 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17544 copying pointers. The cert_st handling is changed by this in
17545 various ways (and thus what used to be known as ctx->default_cert
17546 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17547 any longer when s->cert does not give us what we need).
17548 ssl_cert_instantiate becomes obsolete by this change.
17549 As soon as we've got the new code right (possibly it already is?),
17550 we have solved a couple of bugs of the earlier code where s->cert
17551 was used as if it could not have been shared with other SSL structures.
17553 Note that using the SSL API in certain dirty ways now will result
17554 in different behaviour than observed with earlier library versions:
17555 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17556 does not influence s as it used to.
17558 In order to clean up things more thoroughly, inside SSL_SESSION
17559 we don't use CERT any longer, but a new structure SESS_CERT
17560 that holds per-session data (if available); currently, this is
17561 the peer's certificate chain and, for clients, the server's certificate
17562 and temporary key. CERT holds only those values that can have
17563 meaningful defaults in an SSL_CTX.
17567 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17568 from the internal representation. Various PKCS#7 fixes: remove some
17569 evil casts and set the enc_dig_alg field properly based on the signing
17574 * Allow PKCS#12 password to be set from the command line or the
17575 environment. Let 'ca' get its config file name from the environment
17576 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17581 * Allow certificate policies extension to use an IA5STRING for the
17582 organization field. This is contrary to the PKIX definition but
17583 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17588 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17589 without disallowing inline assembler and the like for non-pedantic builds.
17593 * Support Borland C++ builder.
17595 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17601 * SHA-1 cleanups and performance enhancements.
17603 *Andy Polyakov <appro@fy.chalmers.se>*
17605 * Sparc v8plus assembler for the bignum library.
17607 *Andy Polyakov <appro@fy.chalmers.se>*
17609 * Accept any -xxx and +xxx compiler options in Configure.
17613 * Update HPUX configuration.
17617 * Add missing `sk_<type>_unshift()` function to safestack.h
17619 *Ralf S. Engelschall*
17621 * New function SSL_CTX_use_certificate_chain_file that sets the
17622 "extra_cert"s in addition to the certificate. (This makes sense
17623 only for "PEM" format files, as chains as a whole are not
17628 * Support verify_depth from the SSL API.
17629 x509_vfy.c had what can be considered an off-by-one-error:
17630 Its depth (which was not part of the external interface)
17631 was actually counting the number of certificates in a chain;
17632 now it really counts the depth.
17636 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17637 instead of X509err, which often resulted in confusing error
17638 messages since the error codes are not globally unique
17639 (e.g. an alleged error in ssl3_accept when a certificate
17640 didn't match the private key).
17642 * New function SSL_CTX_set_session_id_context that allows to set a default
17643 value (so that you don't need SSL_set_session_id_context for each
17644 connection using the SSL_CTX).
17648 * OAEP decoding bug fix.
17652 * Support INSTALL_PREFIX for package builders, as proposed by
17657 * New Configure options "threads" and "no-threads". For systems
17658 where the proper compiler options are known (currently Solaris
17659 and Linux), "threads" is the default.
17663 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17667 * Install various scripts to $(OPENSSLDIR)/misc, not to
17668 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17669 such as /usr/local/bin.
17673 * "make linux-shared" to build shared libraries.
17675 *Niels Poppe <niels@netbox.org>*
17677 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17681 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17682 extension adding in x509 utility.
17686 * Remove NOPROTO sections and error code comments.
17690 * Partial rewrite of the DEF file generator to now parse the ANSI
17695 * New Configure options --prefix=DIR and --openssldir=DIR.
17699 * Complete rewrite of the error code script(s). It is all now handled
17700 by one script at the top level which handles error code gathering,
17701 header rewriting and C source file generation. It should be much better
17702 than the old method: it now uses a modified version of Ulf's parser to
17703 read the ANSI prototypes in all header files (thus the old K&R definitions
17704 aren't needed for error creation any more) and do a better job of
17705 translating function codes into names. The old 'ASN1 error code embedded
17706 in a comment' is no longer necessary and it doesn't use .err files which
17707 have now been deleted. Also the error code call doesn't have to appear all
17708 on one line (which resulted in some large lines...).
17712 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17716 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17717 0 (which usually indicates a closed connection), but continue reading.
17721 * Fix some race conditions.
17725 * Add support for CRL distribution points extension. Add Certificate
17726 Policies and CRL distribution points documentation.
17730 * Move the autogenerated header file parts to crypto/opensslconf.h.
17734 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17735 8 of keying material. Merlin has also confirmed interop with this fix
17736 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17738 *Merlin Hughes <merlin@baltimore.ie>*
17740 * Fix lots of warnings.
17742 *Richard Levitte <levitte@stacken.kth.se>*
17744 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17745 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17747 *Richard Levitte <levitte@stacken.kth.se>*
17749 * Fix problems with sizeof(long) == 8.
17751 *Andy Polyakov <appro@fy.chalmers.se>*
17753 * Change functions to ANSI C.
17757 * Fix typos in error codes.
17759 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17761 * Remove defunct assembler files from Configure.
17765 * SPARC v8 assembler BIGNUM implementation.
17767 *Andy Polyakov <appro@fy.chalmers.se>*
17769 * Support for Certificate Policies extension: both print and set.
17770 Various additions to support the r2i method this uses.
17774 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17775 return a const string when you are expecting an allocated buffer.
17779 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17780 types DirectoryString and DisplayText.
17784 * Add code to allow r2i extensions to access the configuration database,
17785 add an LHASH database driver and add several ctx helper functions.
17789 * Fix an evil bug in bn_expand2() which caused various BN functions to
17790 fail when they extended the size of a BIGNUM.
17794 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17795 support typesafe stack.
17799 * Fix typo in SSL_[gs]et_options().
17801 *Nils Frostberg <nils@medcom.se>*
17803 * Delete various functions and files that belonged to the (now obsolete)
17804 old X509V3 handling code.
17808 * New Configure option "rsaref".
17812 * Don't auto-generate pem.h.
17816 * Introduce type-safe ASN.1 SETs.
17820 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17822 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17824 * Introduce type-safe STACKs. This will almost certainly break lots of code
17825 that links with OpenSSL (well at least cause lots of warnings), but fear
17826 not: the conversion is trivial, and it eliminates loads of evil casts. A
17827 few STACKed things have been converted already. Feel free to convert more.
17828 In the fullness of time, I'll do away with the STACK type altogether.
17832 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17833 specified in `<certfile>` by updating the entry in the index.txt file.
17834 This way one no longer has to edit the index.txt file manually for
17835 revoking a certificate. The -revoke option does the gory details now.
17837 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17839 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17840 `-text` option at all and this way the `-noout -text` combination was
17841 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17843 *Ralf S. Engelschall*
17845 * Make sure a corresponding plain text error message exists for the
17846 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17847 verify callback function determined that a certificate was revoked.
17849 *Ralf S. Engelschall*
17851 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17852 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17853 all available ciphers including rc5, which was forgotten until now.
17854 In order to let the testing shell script know which algorithms
17855 are available, a new (up to now undocumented) command
17856 `openssl list-cipher-commands` is used.
17860 * Bugfix: s_client occasionally would sleep in select() when
17861 it should have checked SSL_pending() first.
17865 * New functions DSA_do_sign and DSA_do_verify to provide access to
17866 the raw DSA values prior to ASN.1 encoding.
17870 * Tweaks to Configure
17872 *Niels Poppe <niels@netbox.org>*
17874 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17879 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17883 * New config option to avoid instructions that are illegal on the 80386.
17884 The default code is faster, but requires at least a 486.
17888 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17889 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17890 same as SSL2_VERSION anyway.
17894 * New "-showcerts" option for s_client.
17898 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17899 application. Various cleanups and fixes.
17903 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17904 modify error routines to work internally. Add error codes and PBE init
17905 to library startup routines.
17909 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17910 packing functions to asn1 and evp. Changed function names and error
17911 codes along the way.
17915 * PKCS12 integration: and so it begins... First of several patches to
17916 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17917 objects to objects.h
17921 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17922 and display support for Thawte strong extranet extension.
17926 * Add LinuxPPC support.
17928 *Jeff Dubrule <igor@pobox.org>*
17930 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17931 bn_div_words in alpha.s.
17933 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17935 * Make sure the RSA OAEP test is skipped under -DRSAref because
17936 OAEP isn't supported when OpenSSL is built with RSAref.
17938 *Ulf Moeller <ulf@fitug.de>*
17940 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17941 so they no longer are missing under -DNOPROTO.
17943 *Soren S. Jorvang <soren@t.dk>*
17945 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17947 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17948 doesn't work when the session is reused. Coming soon!
17952 * Fix a security hole, that allows sessions to be reused in the wrong
17953 context thus bypassing client cert protection! All software that uses
17954 client certs and session caches in multiple contexts NEEDS PATCHING to
17955 allow session reuse! A fuller solution is in the works.
17957 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17959 * Some more source tree cleanups (removed obsolete files
17960 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17961 permission on "config" script to be executable) and a fix for the INSTALL
17964 *Ulf Moeller <ulf@fitug.de>*
17966 * Remove some legacy and erroneous uses of malloc, free instead of
17969 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17971 * Make rsa_oaep_test return non-zero on error.
17973 *Ulf Moeller <ulf@fitug.de>*
17975 * Add support for native Solaris shared libraries. Configure
17976 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17977 if someone would make that last step automatic.
17979 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17981 * ctx_size was not built with the right compiler during "make links". Fixed.
17985 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17986 except NULL ciphers". This means the default cipher list will no longer
17987 enable NULL ciphers. They need to be specifically enabled e.g. with
17988 the string "DEFAULT:eNULL".
17992 * Fix to RSA private encryption routines: if p < q then it would
17993 occasionally produce an invalid result. This will only happen with
17994 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17998 * Be less restrictive and allow also `perl util/perlpath.pl
17999 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18000 because this way one can also use an interpreter named `perl5` (which is
18001 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18002 installed as `perl`).
18004 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18006 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18008 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18010 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18011 advapi32.lib to Win32 build and change the pem test comparison
18012 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18013 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18014 and crypto/des/ede_cbcm_enc.c.
18018 * DES quad checksum was broken on big-endian architectures. Fixed.
18022 * Comment out two functions in bio.h that aren't implemented. Fix up the
18023 Win32 test batch file so it (might) work again. The Win32 test batch file
18024 is horrible: I feel ill....
18028 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18029 in e_os.h. Audit of header files to check ANSI and non ANSI
18030 sections: 10 functions were absent from non ANSI section and not exported
18031 from Windows DLLs. Fixed up libeay.num for new functions.
18035 * Make `openssl version` output lines consistent.
18037 *Ralf S. Engelschall*
18039 * Fix Win32 symbol export lists for BIO functions: Added
18040 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18041 to ms/libeay{16,32}.def.
18043 *Ralf S. Engelschall*
18045 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18046 fine under Unix and passes some trivial tests I've now added. But the
18047 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18048 added to make sure no one expects that this stuff really works in the
18049 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18050 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18053 *Ralf S. Engelschall*
18055 * Fix the generation of two part addresses in perl.
18057 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18059 * Add config entry for Linux on MIPS.
18061 *John Tobey <jtobey@channel1.com>*
18063 * Make links whenever Configure is run, unless we are on Windoze.
18067 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18068 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18073 * Add a useful kludge to allow package maintainers to specify compiler and
18074 other platforms details on the command line without having to patch the
18075 Configure script every time: One now can use
18076 `perl Configure <id>:<details>`,
18077 i.e. platform ids are allowed to have details appended
18078 to them (separated by colons). This is treated as there would be a static
18079 pre-configured entry in Configure's %table under key `<id>` with value
18080 `<details>` and `perl Configure <id>` is called. So, when you want to
18081 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18082 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18083 now, which overrides the FreeBSD-elf entry on-the-fly.
18085 *Ralf S. Engelschall*
18087 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18091 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18092 on the `perl Configure ...` command line. This way one can compile
18093 OpenSSL libraries with Position Independent Code (PIC) which is needed
18094 for linking it into DSOs.
18096 *Ralf S. Engelschall*
18098 * Remarkably, export ciphers were totally broken and no-one had noticed!
18103 * Cleaned up the LICENSE document: The official contact for any license
18104 questions now is the OpenSSL core team under openssl-core@openssl.org.
18105 And add a paragraph about the dual-license situation to make sure people
18106 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18107 to the OpenSSL toolkit.
18109 *Ralf S. Engelschall*
18111 * General source tree makefile cleanups: Made `making xxx in yyy...`
18112 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18113 Additionally cleaned up the `make links` target: Remove unnecessary
18114 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18115 to speed processing and no longer clutter the display with confusing
18116 stuff. Instead only the actually done links are displayed.
18118 *Ralf S. Engelschall*
18120 * Permit null encryption ciphersuites, used for authentication only. It used
18121 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18122 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18127 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18128 signed attributes when verifying signatures (this would break them),
18129 the detached data encoding was wrong and public keys obtained using
18130 X509_get_pubkey() weren't freed.
18134 * Add text documentation for the BUFFER functions. Also added a work around
18135 to a Win95 console bug. This was triggered by the password read stuff: the
18136 last character typed gets carried over to the next fread(). If you were
18137 generating a new cert request using 'req' for example then the last
18138 character of the passphrase would be CR which would then enter the first
18143 * Added the new 'Includes OpenSSL Cryptography Software' button as
18144 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18145 button and can be used by applications based on OpenSSL to show the
18146 relationship to the OpenSSL project.
18148 *Ralf S. Engelschall*
18150 * Remove confusing variables in function signatures in files
18151 ssl/ssl_lib.c and ssl/ssl.h.
18153 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18155 * Don't install bss_file.c under PREFIX/include/
18157 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18159 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18160 functions that return function pointers and has support for NT specific
18161 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18162 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18163 unsigned to signed types: this was killing the Win32 compile.
18167 * Add new certificate file to stack functions,
18168 SSL_add_dir_cert_subjects_to_stack() and
18169 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18170 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18171 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18172 This means that Apache-SSL and similar packages don't have to mess around
18173 to add as many CAs as they want to the preferred list.
18177 * Experiment with doxygen documentation. Currently only partially applied to
18179 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18180 openssl.doxy as the configuration file.
18184 * Get rid of remaining C++-style comments which strict C compilers hate.
18186 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18188 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18189 compiled in by default: it has problems with large keys.
18193 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18194 DH private keys and/or callback functions which directly correspond to
18195 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18196 is needed for applications which have to configure certificates on a
18197 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18199 For the RSA certificate situation is makes no difference, but
18200 for the DSA certificate situation this fixes the "no shared cipher"
18201 problem where the OpenSSL cipher selection procedure failed because the
18202 temporary keys were not overtaken from the context and the API provided
18203 no way to reconfigure them.
18204 The new functions now let applications reconfigure the stuff and they
18205 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18206 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18207 non-public-API function ssl_cert_instantiate() is used as a helper
18208 function and also to reduce code redundancy inside ssl_rsa.c.
18210 *Ralf S. Engelschall*
18212 * Move s_server -dcert and -dkey options out of the undocumented feature
18213 area because they are useful for the DSA situation and should be
18214 recognized by the users.
18216 *Ralf S. Engelschall*
18218 * Fix the cipher decision scheme for export ciphers: the export bits are
18219 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18220 SSL_EXP_MASK. So, the original variable has to be used instead of the
18221 already masked variable.
18223 *Richard Levitte <levitte@stacken.kth.se>*
18225 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18227 *Richard Levitte <levitte@stacken.kth.se>*
18229 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18230 from `int` to `unsigned int` because it is a length and initialized by
18231 EVP_DigestFinal() which expects an `unsigned int *`.
18233 *Richard Levitte <levitte@stacken.kth.se>*
18235 * Don't hard-code path to Perl interpreter on shebang line of Configure
18236 script. Instead use the usual Shell->Perl transition trick.
18238 *Ralf S. Engelschall*
18240 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18241 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18242 -noout -modulus` as it's already the case for `openssl rsa -noout
18243 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18244 currently the public key is printed (a decision which was already done by
18245 `openssl dsa -modulus` in the past) which serves a similar purpose.
18246 Additionally the NO_RSA no longer completely removes the whole -modulus
18247 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18250 *Ralf S. Engelschall*
18252 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18253 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18255 *Arne Ansper <arne@ats.cyber.ee>*
18257 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18258 to be added. Now both 'req' and 'ca' can use new objects defined in the
18263 * Add cool BIO that does syslog (or event log on NT).
18265 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18267 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18268 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18269 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18270 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18274 * Add preliminary config info for new extension code.
18278 * Make RSA_NO_PADDING really use no padding.
18280 *Ulf Moeller <ulf@fitug.de>*
18282 * Generate errors when private/public key check is done.
18286 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18287 for some CRL extensions and new objects added.
18291 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18292 key usage extension and fuller support for authority key id.
18296 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18297 padding method for RSA, which is recommended for new applications in PKCS
18298 #1 v2.0 (RFC 2437, October 1998).
18299 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18300 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18301 against Bleichbacher's attack on RSA.
18302 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18305 * Updates to the new SSL compression code
18307 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18309 * Fix so that the version number in the master secret, when passed
18310 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18311 (because the server will not accept higher), that the version number
18312 is 0x03,0x01, not 0x03,0x00
18314 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18316 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18317 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18318 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18322 * Support for RAW extensions where an arbitrary extension can be
18323 created by including its DER encoding. See `apps/openssl.cnf` for
18328 * Make sure latest Perl versions don't interpret some generated C array
18329 code as Perl array code in the crypto/err/err_genc.pl script.
18331 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18333 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18334 not many people have the assembler. Various Win32 compilation fixes and
18335 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18336 build instructions.
18340 * Modify configure script 'Configure' to automatically create crypto/date.h
18341 file under Win32 and also build pem.h from pem.org. New script
18342 util/mkfiles.pl to create the MINFO file on environments that can't do a
18343 'make files': perl util/mkfiles.pl >MINFO should work.
18347 * Major rework of DES function declarations, in the pursuit of correctness
18348 and purity. As a result, many evil casts evaporated, and some weirdness,
18349 too. You may find this causes warnings in your code. Zapping your evil
18350 casts will probably fix them. Mostly.
18354 * Fix for a typo in asn1.h. Bug fix to object creation script
18355 obj_dat.pl. It considered a zero in an object definition to mean
18356 "end of object": none of the objects in objects.h have any zeros
18357 so it wasn't spotted.
18359 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18361 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18362 Masking (CBCM). In the absence of test vectors, the best I have been able
18363 to do is check that the decrypt undoes the encrypt, so far. Send me test
18364 vectors if you have them.
18368 * Correct calculation of key length for export ciphers (too much space was
18369 allocated for null ciphers). This has not been tested!
18373 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18374 message is now correct (it understands "crypto" and "ssl" on its
18375 command line). There is also now an "update" option. This will update
18376 the util/ssleay.num and util/libeay.num files with any new functions.
18378 perl util/mkdef.pl crypto ssl update
18379 it will update them.
18383 * Overhauled the Perl interface:
18384 - ported BN stuff to OpenSSL's different BN library
18385 - made the perl/ source tree CVS-aware
18386 - renamed the package from SSLeay to OpenSSL (the files still contain
18387 their history because I've copied them in the repository)
18388 - removed obsolete files (the test scripts will be replaced
18389 by better Test::Harness variants in the future)
18391 *Ralf S. Engelschall*
18393 * First cut for a very conservative source tree cleanup:
18394 1. merge various obsolete readme texts into doc/ssleay.txt
18395 where we collect the old documents and readme texts.
18396 2. remove the first part of files where I'm already sure that we no
18397 longer need them because of three reasons: either they are just temporary
18398 files which were left by Eric or they are preserved original files where
18399 I've verified that the diff is also available in the CVS via "cvs diff
18400 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18401 the crypto/md/ stuff).
18403 *Ralf S. Engelschall*
18405 * More extension code. Incomplete support for subject and issuer alt
18406 name, issuer and authority key id. Change the i2v function parameters
18407 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18408 what that's for :-) Fix to ASN1 macro which messed up
18409 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18413 * Preliminary support for ENUMERATED type. This is largely copied from the
18418 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18420 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18422 * Make sure `make rehash` target really finds the `openssl` program.
18424 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18426 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18427 like to hear about it if this slows down other processors.
18431 * Add CygWin32 platform information to Configure script.
18433 *Alan Batie <batie@aahz.jf.intel.com>*
18435 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18437 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18439 * New program nseq to manipulate netscape certificate sequences
18443 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18448 * Fixes to BN code. Previously the default was to define BN_RECURSION
18449 but the BN code had some problems that would cause failures when
18450 doing certificate verification and some other functions.
18452 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18454 * Add ASN1 and PEM code to support netscape certificate sequences.
18458 * Add ASN1 and PEM code to support netscape certificate sequences.
18462 * Add several PKIX and private extended key usage OIDs.
18466 * Modify the 'ca' program to handle the new extension code. Modify
18467 openssl.cnf for new extension format, add comments.
18471 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18472 and add a sample to openssl.cnf so req -x509 now adds appropriate
18477 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18478 error code, add initial support to X509_print() and x509 application.
18482 * Takes a deep breath and start adding X509 V3 extension support code. Add
18483 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18484 stuff is currently isolated and isn't even compiled yet.
18488 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18489 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18490 Removed the versions check from X509 routines when loading extensions:
18491 this allows certain broken certificates that don't set the version
18492 properly to be processed.
18496 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18497 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18498 can still be regenerated with "make depend".
18502 * Spelling mistake in C version of CAST-128.
18504 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18506 * Changes to the error generation code. The perl script err-code.pl
18507 now reads in the old error codes and retains the old numbers, only
18508 adding new ones if necessary. It also only changes the .err files if new
18509 codes are added. The makefiles have been modified to only insert errors
18510 when needed (to avoid needlessly modifying header files). This is done
18511 by only inserting errors if the .err file is newer than the auto generated
18512 C file. To rebuild all the error codes from scratch (the old behaviour)
18513 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18514 or delete all the .err files.
18518 * CAST-128 was incorrectly implemented for short keys. The C version has
18519 been fixed, but is untested. The assembler versions are also fixed, but
18520 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18521 to regenerate it if needed.
18522 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18523 Hagino <itojun@kame.net>*
18525 * File was opened incorrectly in randfile.c.
18527 *Ulf Möller <ulf@fitug.de>*
18529 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18530 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18531 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18532 al: it's just almost always a UTCTime. Note this patch adds new error
18533 codes so do a "make errors" if there are problems.
18537 * Correct Linux 1 recognition in config.
18539 *Ulf Möller <ulf@fitug.de>*
18541 * Remove pointless MD5 hash when using DSA keys in ca.
18543 *Anonymous <nobody@replay.com>*
18545 * Generate an error if given an empty string as a cert directory. Also
18546 generate an error if handed NULL (previously returned 0 to indicate an
18547 error, but didn't set one).
18549 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18551 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18555 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18556 parameters. This was causing a warning which killed off the Win32 compile.
18560 * Remove C++ style comments from crypto/bn/bn_local.h.
18562 *Neil Costigan <neil.costigan@celocom.com>*
18564 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18565 based on a text string, looking up short and long names and finally
18566 "dot" format. The "dot" format stuff didn't work. Added new function
18567 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18568 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18569 OID is not part of the table.
18573 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18574 X509_LOOKUP_by_alias().
18578 * Sort openssl functions by name.
18582 * Get the `gendsa` command working and add it to the `list` command. Remove
18583 encryption from sample DSA keys (in case anyone is interested the password
18588 * Make *all* `*_free` functions accept a NULL pointer.
18590 *Frans Heymans <fheymans@isaserver.be>*
18592 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18595 *Anonymous <nobody@replay.com>*
18597 * s_server should send the CAfile as acceptable CAs, not its own cert.
18599 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18601 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18603 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18605 * Temp key "for export" tests were wrong in s3_srvr.c.
18607 *Anonymous <nobody@replay.com>*
18609 * Add prototype for temp key callback functions
18610 SSL_CTX_set_tmp_{rsa,dh}_callback().
18614 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18615 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18619 * X509_name_add_entry() freed the wrong thing after an error.
18621 *Arne Ansper <arne@ats.cyber.ee>*
18623 * rsa_eay.c would attempt to free a NULL context.
18625 *Arne Ansper <arne@ats.cyber.ee>*
18627 * BIO_s_socket() had a broken should_retry() on Windoze.
18629 *Arne Ansper <arne@ats.cyber.ee>*
18631 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18633 *Arne Ansper <arne@ats.cyber.ee>*
18635 * Make sure the already existing X509_STORE->depth variable is initialized
18636 in X509_STORE_new(), but document the fact that this variable is still
18637 unused in the certificate verification process.
18639 *Ralf S. Engelschall*
18641 * Fix the various library and `apps/` files to free up pkeys obtained from
18642 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18646 * Fix reference counting in X509_PUBKEY_get(). This makes
18647 demos/maurice/example2.c work, amongst others, probably.
18649 *Steve Henson and Ben Laurie*
18651 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18652 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18653 are no longer created. This way we have a single and consistent command
18654 line interface `openssl <command>`, similar to `cvs <command>`.
18656 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18658 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18659 BIT STRING wrapper always have zero unused bits.
18663 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18667 * Make the top-level INSTALL documentation easier to understand.
18671 * Makefiles updated to exit if an error occurs in a sub-directory
18672 make (including if user presses ^C) [Paul Sutton]
18674 * Make Montgomery context stuff explicit in RSA data structure.
18678 * Fix build order of pem and err to allow for generated pem.h.
18682 * Fix renumbering bug in X509_NAME_delete_entry().
18686 * Enhanced the err-ins.pl script so it makes the error library number
18687 global and can add a library name. This is needed for external ASN1 and
18688 other error libraries.
18692 * Fixed sk_insert which never worked properly.
18696 * Fix ASN1 macros so they can handle indefinite length constructed
18697 EXPLICIT tags. Some non standard certificates use these: they can now
18702 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18703 into a single doc/ssleay.txt bundle. This way the information is still
18704 preserved but no longer messes up this directory. Now it's new room for
18705 the new set of documentation files.
18707 *Ralf S. Engelschall*
18709 * SETs were incorrectly DER encoded. This was a major pain, because they
18710 shared code with SEQUENCEs, which aren't coded the same. This means that
18711 almost everything to do with SETs or SEQUENCEs has either changed name or
18712 number of arguments.
18714 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18716 * Fix test data to work with the above.
18720 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18721 was already fixed by Eric for 0.9.1 it seems.
18723 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18725 * Autodetect FreeBSD3.
18729 * Fix various bugs in Configure. This affects the following platforms:
18733 unixware-2.0-pentium
18738 * Eliminate generated files from CVS. Reorder tests to regenerate files
18739 before they are needed.
18743 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18747 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18749 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18750 changed SSLeay to OpenSSL in version strings.
18752 *Ralf S. Engelschall*
18754 * Some fixups to the top-level documents.
18758 * Fixed the nasty bug where rsaref.h was not found under compile-time
18759 because the symlink to include/ was missing.
18761 *Ralf S. Engelschall*
18763 * Incorporated the popular no-RSA/DSA-only patches
18764 which allow to compile a RSA-free SSLeay.
18766 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18768 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18769 when "ssleay" is still not found.
18771 *Ralf S. Engelschall*
18773 * Added more platforms to Configure: Cray T3E, HPUX 11,
18775 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18777 * Updated the README file.
18779 *Ralf S. Engelschall*
18781 * Added various .cvsignore files in the CVS repository subdirs
18782 to make a "cvs update" really silent.
18784 *Ralf S. Engelschall*
18786 * Recompiled the error-definition header files and added
18787 missing symbols to the Win32 linker tables.
18789 *Ralf S. Engelschall*
18791 * Cleaned up the top-level documents;
18792 o new files: CHANGES and LICENSE
18793 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18794 o merged COPYRIGHT into LICENSE
18795 o removed obsolete TODO file
18796 o renamed MICROSOFT to INSTALL.W32
18798 *Ralf S. Engelschall*
18800 * Removed dummy files from the 0.9.1b source tree:
18801 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18802 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18803 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18804 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18805 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18807 *Ralf S. Engelschall*
18809 * Added various platform portability fixes.
18813 * The Genesis of the OpenSSL rpject:
18814 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18815 Young and Tim J. Hudson created while they were working for C2Net until
18818 *The OpenSSL Project*
18820 ### Changes between 0.9.0b and 0.9.1b [not released]
18822 * Updated a few CA certificates under certs/
18826 * Changed some BIGNUM api stuff.
18830 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18831 DGUX x86, Linux Alpha, etc.
18835 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18836 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18841 * Add -strparse option to asn1pars program which parses nested
18844 *Dr Stephen Henson <shenson@bigfoot.com>*
18846 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18850 * DSA fix for "ca" program.
18854 * Added "-genkey" option to "dsaparam" program.
18858 * Added RIPE MD160 (rmd160) message digest.
18862 * Added -a (all) option to "ssleay version" command.
18866 * Added PLATFORM define which is the id given to Configure.
18870 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18874 * Extended the ASN.1 parser routines.
18878 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18882 * Added a BN_CTX to the BN library.
18886 * Fixed the weak key values in DES library
18890 * Changed API in EVP library for cipher aliases.
18894 * Added support for RC2/64bit cipher.
18898 * Converted the lhash library to the crypto/mem.c functions.
18902 * Added more recognized ASN.1 object ids.
18906 * Added more RSA padding checks for SSL/TLS.
18910 * Added BIO proxy/filter functionality.
18914 * Added extra_certs to SSL_CTX which can be used
18915 send extra CA certificates to the client in the CA cert chain sending
18916 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18920 * Now Fortezza is denied in the authentication phase because
18921 this is key exchange mechanism is not supported by SSLeay at all.
18925 * Additional PKCS1 checks.
18929 * Support the string "TLSv1" for all TLS v1 ciphers.
18933 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18934 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18938 * Fixed a few memory leaks.
18942 * Fixed various code and comment typos.
18946 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18947 bytes sent in the client random.
18949 *Edward Bishop <ebishop@spyglass.com>*
18953 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18954 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18955 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18956 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18957 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18958 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18959 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18960 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18961 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18962 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18963 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18964 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18965 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18966 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18967 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18968 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18969 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18970 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18971 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18972 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18973 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18974 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18975 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18976 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18977 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18978 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18979 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18980 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18981 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18982 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18983 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18984 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18985 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18986 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18987 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18988 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18989 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18990 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18991 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18992 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18993 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18994 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18995 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18996 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18997 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18998 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18999 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19000 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19001 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19002 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19003 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19004 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19005 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19006 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19007 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19008 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19009 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19010 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19011 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19012 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19013 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19014 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19015 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19016 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19017 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19018 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19019 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19020 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19021 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19022 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19023 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19024 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19025 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19026 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19027 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19028 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19029 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19030 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19031 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19032 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19033 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19034 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19035 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19036 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19037 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19038 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19039 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19040 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19041 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19042 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19043 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19044 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19045 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19046 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19047 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19048 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19049 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19050 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19051 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19052 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19053 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19054 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19055 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19056 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19057 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19058 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19059 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19060 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19061 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19062 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19063 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19064 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19065 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19066 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19067 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19068 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19069 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19070 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19071 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19072 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19073 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19074 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19075 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19076 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19077 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19078 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19079 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19080 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19081 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19082 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19083 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19084 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19085 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19086 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19087 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19088 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19089 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19090 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19091 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19092 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19093 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19094 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19095 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19096 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19097 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19098 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19099 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19100 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19101 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19102 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19103 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19104 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19105 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19106 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19107 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19108 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19109 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19110 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19111 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19112 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19113 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19114 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655