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 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
44 * Fixed PVK encoder to properly query for the passphrase.
48 * Multiple fixes in the OSSL_HTTP API functions.
52 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
53 OSSL_PARAM_INTEGER data type and return error on negative numbers
54 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
55 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
59 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
63 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
67 * Multiple threading fixes.
71 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
75 * Allow fetching an operation from the provider that owns an unexportable key
76 as a fallback if that is still allowed by the property query.
80 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
82 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
87 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
88 S390X capability vector to zero. This simplifies testing of different code
89 paths on S390X architecture.
93 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
94 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
95 SP 800-38D". The communication will fail at this point.
99 * The EC_GROUP_clear_free() function is deprecated as there is nothing
100 confidential in EC_GROUP data.
104 * The byte order mark (BOM) character is ignored if encountered at the
105 beginning of a PEM-formatted file.
109 * Added CMS support for the Russian GOST algorithms.
113 * Due to move of the implementation of cryptographic operations
114 to the providers, validation of various operation parameters can
115 be postponed until the actual operation is executed where previously
116 it happened immediately when an operation parameter was set.
118 For example when setting an unsupported curve with
119 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
120 fail but later keygen operations with the EVP_PKEY_CTX will fail.
122 *OpenSSL team members and many third party contributors*
124 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
125 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
126 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
127 instead to retrieve these algorithms from a provider.
131 * On build targets where the multilib postfix is set in the build
132 configuration the libdir directory was changing based on whether
133 the lib directory with the multilib postfix exists on the system
134 or not. This unpredictable behavior was removed and eventual
135 multilib postfix is now always added to the default libdir. Use
136 `--libdir=lib` to override the libdir if adding the postfix is
141 * The triple DES key wrap functionality now conforms to RFC 3217 but is
142 no longer interoperable with OpenSSL 1.1.1.
146 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
147 function codes, this function can only cause problems for calling
152 * Add a configurable flag to output date formats as ISO 8601. Does not
153 change the default date format.
157 * Version of MSVC earlier than 1300 could get link warnings, which could
158 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
159 Support for this flag has been removed.
163 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
164 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
165 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
166 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
167 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
171 * The signatures of the functions to get and set options on SSL and
172 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
173 Some source code changes may be required.
177 * The public definitions of conf_method_st and conf_st have been
178 deprecated. They will be made opaque in a future release.
180 *Rich Salz and Tomáš Mráz*
182 * Client-initiated renegotiation is disabled by default. To allow it, use
183 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
184 flag, or the "ClientRenegotiation" config parameter as appropriate.
188 * Add "abspath" and "includedir" pragma's to config files, to prevent,
189 or modify relative pathname inclusion.
193 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
194 validated. Please consult the README-FIPS and
195 README-PROVIDERS files, as well as the migration guide.
197 *OpenSSL team members and many third party contributors*
199 * For the key types DH and DHX the allowed settable parameters are now different.
203 * The openssl commands that read keys, certificates, and CRLs now
204 automatically detect the PEM or DER format of the input files.
206 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
208 * Added enhanced PKCS#12 APIs which accept a library context.
212 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
216 * Added support for Kernel TLS (KTLS).
218 *Boris Pismenny, John Baldwin and Andrew Gallatin*
220 * Support for RFC 5746 secure renegotiation is now required by default for
221 SSL or TLS connections to succeed.
225 * The signature of the `copy` functional parameter of the
226 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
227 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
228 the signature of the `pub_decode` functional parameter of the
229 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
230 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
234 * The error return values from some control calls (ctrl) have changed.
238 * A public key check is now performed during EVP_PKEY_derive_set_peer().
242 * Many functions in the EVP_ namespace that are getters of values from
243 implementations or contexts were renamed to include get or get0 in their
244 names. Old names are provided as macro aliases for compatibility and
249 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
250 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
251 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
256 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
259 * The output from the command line applications may have minor
264 * The output from numerous "printing" may have minor changes.
268 * Windows thread synchronization uses read/write primitives (SRWLock) when
269 supported by the OS, otherwise CriticalSection continues to be used.
273 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
274 work on read only BIO source/sinks that do not support these functions.
275 This allows piping or redirection of a file BIO using stdin to be buffered
276 into memory. This is used internally in OSSL_DECODER_from_bio().
280 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
281 this function would return one of the values OSSL_STORE_INFO_NAME,
282 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
283 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
284 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
285 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
286 using this function should be amended to handle the changed return value.
290 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
291 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
292 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
293 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
294 contains more than one certificate identifier: This means that all
295 certificates referenced there MUST be part of the validation chain.
299 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
300 RC5, DESX and DES have been moved to the legacy provider.
304 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
305 RIPEMD-160 have been moved to the legacy provider.
309 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
314 * The deprecated functions EVP_PKEY_get0_RSA(),
315 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
316 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
317 well as the similarly named "get1" functions behave differently in
322 * A number of functions handling low-level keys or engines were deprecated
323 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
324 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
325 EVP_PKEY_get0_siphash().
329 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
330 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
331 will need to load the legacy crypto provider. This includes these PBE
332 algorithms which use this KDF:
333 - NID_pbeWithMD2AndDES_CBC
334 - NID_pbeWithMD5AndDES_CBC
335 - NID_pbeWithSHA1AndRC2_CBC
336 - NID_pbeWithMD2AndRC2_CBC
337 - NID_pbeWithMD5AndRC2_CBC
338 - NID_pbeWithSHA1AndDES_CBC
342 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
343 BIO_debug_callback() functions.
347 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
348 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
352 * The RAND_METHOD APIs have been deprecated.
356 * The SRP APIs have been deprecated.
360 * Add a compile time option to prevent the caching of provider fetched
361 algorithms. This is enabled by including the no-cached-fetch option
362 at configuration time.
366 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
367 count of PKCS12_DEFAULT_ITER.
369 *Tomáš Mráz and Sahana Prasad*
371 * The openssl speed command does not use low-level API calls anymore.
375 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
378 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
380 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
384 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
385 providers may supply their own group implementations (using either the "key
386 exchange" or the "key encapsulation" methods) which will automatically be
387 detected and used by libssl.
389 *Matt Caswell, Nicola Tuveri*
391 * The undocumented function X509_certificate_type() has been deprecated;
395 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
399 * Removed RSA padding mode for SSLv23 (which was only used for
400 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
401 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
406 * Deprecated the obsolete X9.31 RSA key generation related functions.
408 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
409 is not allowed to return a value > 1, this is no more taken as failure.
411 *Viktor Dukhovni and David von Oheimb*
413 * Deprecated the obsolete X9.31 RSA key generation related functions
414 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
415 BN_X931_generate_prime_ex().
419 * The default key generation method for the regular 2-prime RSA keys was
420 changed to the FIPS 186-4 B.3.6 method.
424 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
428 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
432 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
433 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
435 *Rich Salz, Richard Levitte, and David von Oheimb*
437 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
441 * Deprecated `OCSP_parse_url()`.
445 * Validation of SM2 keys has been separated from the validation of regular EC
450 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
451 switches: a validation failure triggers an early exit, returning a failure
452 exit status to the parent process.
456 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
457 to ignore unknown ciphers.
461 * The `-cipher-commands` and `-digest-commands` options
462 of the command line utility `list` have been deprecated.
463 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
467 * Added convenience functions for generating asymmetric key pairs:
468 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
469 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
473 * All of the low level EC_KEY functions have been deprecated.
475 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
477 * Deprecated all the libcrypto and libssl error string loading
482 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
483 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
488 * The `-crypt` option to the `passwd` command line tool has been removed.
492 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
497 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
501 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
502 EVP_PKEY_get1_tls_encodedpoint().
506 * The security callback, which can be customised by application code, supports
507 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
508 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
512 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
513 interface. Their functionality remains unchanged.
517 * Added new option for 'openssl list', '-providers', which will display the
518 list of loaded providers, their names, version and status. It optionally
519 displays their gettable parameters.
523 * Removed EVP_PKEY_set_alias_type().
527 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
528 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
532 * Changed all "STACK" functions to be macros instead of inline functions. Macro
533 parameters are still checked for type safety at compile time via helper
538 * Remove the RAND_DRBG API
540 *Paul Dale and Matthias St. Pierre*
542 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
543 as well as actual hostnames.
547 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
548 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
549 conversely, silently ignore DTLS protocol version bounds when configuring
550 TLS-based contexts. The commands can be repeated to set bounds of both
551 types. The same applies with the corresponding "min_protocol" and
552 "max_protocol" command-line switches, in case some application uses both TLS
555 SSL_CTX instances that are created for a fixed protocol version (e.g.
556 `TLSv1_server_method()`) also silently ignore version bounds. Previously
557 attempts to apply bounds to these protocol versions would result in an
558 error. Now only the "version-flexible" SSL_CTX instances are subject to
559 limits in configuration files in command-line options.
563 * Deprecated the `ENGINE` API. Engines should be replaced with providers
568 * Reworked the recorded ERR codes to make better space for system errors.
569 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
570 given code is a system error (true) or an OpenSSL error (false).
574 * Reworked the test perl framework to better allow parallel testing.
576 *Nicola Tuveri and David von Oheimb*
578 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
579 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
583 * 'Configure' has been changed to figure out the configuration target if
584 none is given on the command line. Consequently, the 'config' script is
585 now only a mere wrapper. All documentation is changed to only mention
588 *Rich Salz and Richard Levitte*
590 * Added a library context `OSSL_LIB_CTX` that applications as well as
591 other libraries can use to form a separate context within which
592 libcrypto operations are performed.
596 * Added various `_ex` functions to the OpenSSL API that support using
597 a non-default `OSSL_LIB_CTX`.
601 * Handshake now fails if Extended Master Secret extension is dropped
606 * Dropped interactive mode from the `openssl` program.
610 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
612 *David von Oheimb and Shane Lontis*
614 * Deprecated `EC_METHOD_get_field_type()`.
618 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
619 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
620 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
624 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
628 * Add CAdES-BES signature verification support, mostly derived
629 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
631 *Filipe Raimundo da Silva*
633 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
637 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
638 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
642 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
646 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
647 EC_KEY_precompute_mult().
651 * Deprecated EC_POINTs_mul().
655 * Removed FIPS_mode() and FIPS_mode_set().
659 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
663 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
664 EC_POINT_get_Jprojective_coordinates_GFp().
668 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
669 arrays to be more easily constructed via a series of utility functions.
670 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
671 the various push functions and finally convert to a passable OSSL_PARAM
672 array using OSSL_PARAM_BLD_to_param().
676 * The security strength of SHA1 and MD5 based signatures in TLS has been
681 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
682 contain a provider side internal key.
686 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
690 * Project text documents not yet having a proper file name extension
691 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
692 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
696 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
697 have been converted to Markdown with the goal to produce documents
698 which not only look pretty when viewed online in the browser, but
699 remain well readable inside a plain text editor.
701 To achieve this goal, a 'minimalistic' Markdown style has been applied
702 which avoids formatting elements that interfere too much with the
703 reading flow in the text file. For example, it
705 * avoids [ATX headings][] and uses [setext headings][] instead
706 (which works for `<h1>` and `<h2>` headings only).
707 * avoids [inline links][] and uses [reference links][] instead.
708 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
710 [ATX headings]: https://github.github.com/gfm/#atx-headings
711 [setext headings]: https://github.github.com/gfm/#setext-headings
712 [inline links]: https://github.github.com/gfm/#inline-link
713 [reference links]: https://github.github.com/gfm/#reference-link
714 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
715 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
717 *Matthias St. Pierre*
719 * The test suite is changed to preserve results of each test recipe.
720 A new directory test-runs/ with subdirectories named like the
721 test recipes are created in the build tree for this purpose.
725 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
726 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
727 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
729 *David von Oheimb, Martin Peylo*
731 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
732 It supports arbitrary request and response content types, GET redirection,
733 TLS, connections via HTTP(S) proxies, connections and exchange via
734 user-defined BIOs (allowing implicit connections), persistent connections,
735 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
736 The legacy OCSP-focused (and only partly documented) API
737 is retained for backward compatibility, while most of it is deprecated.
741 * Added `util/check-format.pl`, a tool for checking adherence to the
742 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
743 The checks performed are incomplete and yield some false positives.
744 Still the tool should be useful for detecting most typical glitches.
748 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
749 If domain name resolution yields multiple IP addresses all of them are tried
750 after `connect()` failures.
754 * All of the low level RSA functions have been deprecated.
758 * X509 certificates signed using SHA1 are no longer allowed at security
763 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
764 modified to use PKEY APIs. These commands are now in maintenance mode
765 and no new features will be added to them.
769 * The command line utility rsautl has been deprecated.
773 * The command line utilities genrsa and rsa have been modified to use PKEY
774 APIs. They now write PKCS#8 keys by default. These commands are now in
775 maintenance mode and no new features will be added to them.
779 * All of the low level DH functions have been deprecated.
781 *Paul Dale and Matt Caswell*
783 * All of the low level DSA functions have been deprecated.
787 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
788 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
792 * Deprecated low level ECDH and ECDSA functions.
796 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
800 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
801 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
802 a new formulation to include all the things it can be used for,
803 as well as words of caution.
807 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
811 * All of the low level HMAC functions have been deprecated.
813 *Paul Dale and David von Oheimb*
815 * Over two thousand fixes were made to the documentation, including:
816 - Common options (such as -rand/-writerand, TLS version control, etc)
817 were refactored and point to newly-enhanced descriptions in openssl.pod.
818 - Added style conformance for all options (with help from Richard Levitte),
819 documented all reported missing options, added a CI build to check
820 that all options are documented and that no unimplemented options
822 - Documented some internals, such as all use of environment variables.
823 - Addressed all internal broken L<> references.
827 * All of the low level CMAC functions have been deprecated.
831 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
832 functions have been deprecated.
834 *Paul Dale and David von Oheimb*
836 * Corrected the documentation of the return values from the `EVP_DigestSign*`
837 set of functions. The documentation mentioned negative values for some
838 errors, but this was never the case, so the mention of negative values
841 Code that followed the documentation and thereby check with something
842 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
846 * All of the low level cipher functions have been deprecated.
848 *Matt Caswell and Paul Dale*
850 * Removed include/openssl/opensslconf.h.in and replaced it with
851 include/openssl/configuration.h.in, which differs in not including
852 <openssl/macros.h>. A short header include/openssl/opensslconf.h
853 was added to include both.
855 This allows internal hacks where one might need to modify the set
856 of configured macros, for example this if deprecated symbols are
857 still supposed to be available internally:
859 #include <openssl/configuration.h>
861 #undef OPENSSL_NO_DEPRECATED
862 #define OPENSSL_SUPPRESS_DEPRECATED
864 #include <openssl/macros.h>
866 This should not be used by applications that use the exported
867 symbols, as that will lead to linking errors.
871 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
872 used in exponentiation with 512-bit moduli. No EC algorithms are
873 affected. Analysis suggests that attacks against 2-prime RSA1024,
874 3-prime RSA1536, and DSA1024 as a result of this defect would be very
875 difficult to perform and are not believed likely. Attacks against DH512
876 are considered just feasible. However, for an attack the target would
877 have to re-use the DH512 private key, which is not recommended anyway.
878 Also applications directly using the low-level API BN_mod_exp may be
879 affected if they use BN_FLG_CONSTTIME.
884 * Most memory-debug features have been deprecated, and the functionality
885 replaced with no-ops.
889 * Added documentation for the STACK API.
893 * Introduced a new method type and API, OSSL_ENCODER, to represent
894 generic encoders. These do the same sort of job that PEM writers
895 and d2i functions do, but with support for methods supplied by
896 providers, and the possibility for providers to support other
901 * Introduced a new method type and API, OSSL_DECODER, to represent
902 generic decoders. These do the same sort of job that PEM readers
903 and i2d functions do, but with support for methods supplied by
904 providers, and the possibility for providers to support other
909 * Added a .pragma directive to the syntax of configuration files, to
910 allow varying behavior in a supported and predictable manner.
911 Currently added pragma:
915 This allows dollar signs to be a keyword character unless it's
916 followed by a opening brace or parenthesis. This is useful for
917 platforms where dollar signs are commonly used in names, such as
918 volume names and system directory names on VMS.
922 * Added functionality to create an EVP_PKEY from user data.
926 * Change the interpretation of the '--api' configuration option to
927 mean that this is a desired API compatibility level with no
928 further meaning. The previous interpretation, that this would
929 also mean to remove all deprecated symbols up to and including
930 the given version, no requires that 'no-deprecated' is also used
931 in the configuration.
933 When building applications, the desired API compatibility level
934 can be set with the OPENSSL_API_COMPAT macro like before. For
935 API compatibility version below 3.0, the old style numerical
936 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
937 For version 3.0 and on, the value is expected to be the decimal
938 value calculated from the major and minor version like this:
940 MAJOR * 10000 + MINOR * 100
944 -DOPENSSL_API_COMPAT=30000 For 3.0
945 -DOPENSSL_API_COMPAT=30200 For 3.2
947 To hide declarations that are deprecated up to and including the
948 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
949 given when building the application as well.
953 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
954 access to certificate and CRL stores via URIs and OSSL_STORE
957 This adds the following functions:
959 - X509_LOOKUP_store()
960 - X509_STORE_load_file()
961 - X509_STORE_load_path()
962 - X509_STORE_load_store()
963 - SSL_add_store_cert_subjects_to_stack()
964 - SSL_CTX_set_default_verify_store()
965 - SSL_CTX_load_verify_file()
966 - SSL_CTX_load_verify_dir()
967 - SSL_CTX_load_verify_store()
971 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
972 The presence of this system service is determined at run-time.
976 * Added functionality to create an EVP_PKEY context based on data
977 for methods from providers. This takes an algorithm name and a
978 property query string and simply stores them, with the intent
979 that any operation that uses this context will use those strings
980 to fetch the needed methods implicitly, thereby making the port
981 of application written for pre-3.0 OpenSSL easier.
985 * The undocumented function NCONF_WIN32() has been deprecated; for
986 conversion details see the HISTORY section of doc/man5/config.pod
990 * Introduced the new functions EVP_DigestSignInit_ex() and
991 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
992 EVP_DigestVerifyUpdate() have been converted to functions. See the man
993 pages for further details.
997 * Over two thousand fixes were made to the documentation, including:
998 adding missing command flags, better style conformance, documentation
1001 *Rich Salz, Richard Levitte*
1003 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1004 X25519, X448, Ed25519 and Ed448.
1008 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1013 * Deprecated the public definition of `ERR_STATE` as well as the function
1014 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1019 * Added ERR functionality to give callers access to the stored function
1020 names that have replaced the older function code based functions.
1022 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1023 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1024 ERR_peek_error_all() and ERR_peek_last_error_all().
1026 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1027 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1028 ERR_func_error_string().
1032 * Extended testing to be verbose for failing tests only. The make variables
1033 VERBOSE_FAILURE or VF can be used to enable this:
1035 $ make VF=1 test # Unix
1036 $ mms /macro=(VF=1) test ! OpenVMS
1037 $ nmake VF=1 test # Windows
1041 * Added the `-copy_extensions` option to the `x509` command for use with
1042 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1043 all extensions in the request are copied to the certificate or vice versa.
1045 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1047 * Added the `-copy_extensions` option to the `req` command for use with
1048 `-x509`. When given with the `copy` or `copyall` argument,
1049 all extensions in the certification request are copied to the certificate.
1053 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1054 they generate are by default RFC 5280 compliant in the following sense:
1055 There is a subjectKeyIdentifier extension with a hash value of the public key
1056 and for not self-signed certs there is an authorityKeyIdentifier extension
1057 with a keyIdentifier field or issuer information identifying the signing key.
1058 This is done unless some configuration overrides the new default behavior,
1059 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1063 * Added several checks to `X509_verify_cert()` according to requirements in
1064 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1065 (which may be done by using the CLI option `-x509_strict`):
1066 * The basicConstraints of CA certificates must be marked critical.
1067 * CA certificates must explicitly include the keyUsage extension.
1068 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1069 * The issuer name of any certificate must not be empty.
1070 * The subject name of CA certs, certs with keyUsage crlSign,
1071 and certs without subjectAlternativeName must not be empty.
1072 * If a subjectAlternativeName extension is given it must not be empty.
1073 * The signatureAlgorithm field and the cert signature must be consistent.
1074 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1075 must not be marked critical.
1076 * The authorityKeyIdentifier must be given for X.509v3 certs
1077 unless they are self-signed.
1078 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1082 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1083 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1087 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1088 used even when parsing explicit parameters, when loading a encoded key
1089 or calling `EC_GROUP_new_from_ecpkparameters()`/
1090 `EC_GROUP_new_from_ecparameters()`.
1091 This prevents bypass of security hardening and performance gains,
1092 especially for curves with specialized EC_METHODs.
1093 By default, if a key encoded with explicit parameters is loaded and later
1094 encoded, the output is still encoded with explicit parameters, even if
1095 internally a "named" EC_GROUP is used for computation.
1099 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1100 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1101 NULL. After this change, only the cofactor parameter can be NULL. It also
1102 does some minimal sanity checks on the passed order.
1107 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1108 An attack is simple, if the first CMS_recipientInfo is valid but the
1109 second CMS_recipientInfo is chosen ciphertext. If the second
1110 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1111 encryption key will be replaced by garbage, and the message cannot be
1112 decoded, but if the RSA decryption fails, the correct encryption key is
1113 used and the recipient will not notice the attack.
1114 As a work around for this potential attack the length of the decrypted
1115 key must be equal to the cipher default key length, in case the
1116 certifiate is not given and all recipientInfo are tried out.
1117 The old behaviour can be re-enabled in the CMS code by setting the
1118 CMS_DEBUG_DECRYPT flag.
1122 * Early start up entropy quality from the DEVRANDOM seed source has been
1123 improved for older Linux systems. The RAND subsystem will wait for
1124 /dev/random to be producing output before seeding from /dev/urandom.
1125 The seeded state is stored for future library initialisations using
1126 a system global shared memory segment. The shared memory identifier
1127 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1128 the desired value. The default identifier is 114.
1132 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1133 when primes for RSA keys are computed.
1134 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1135 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1136 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1137 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1138 This avoids possible fingerprinting of newly generated RSA modules.
1142 * Correct the extended master secret constant on EBCDIC systems. Without this
1143 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1144 negotiate EMS will fail. Unfortunately this also means that TLS connections
1145 between EBCDIC systems with this fix, and EBCDIC systems without this
1146 fix will fail if they negotiate EMS.
1150 * Changed the library initialisation so that the config file is now loaded
1151 by default. This was already the case for libssl. It now occurs for both
1152 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1153 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1157 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1158 where the former acts as a replacement for `ERR_put_error()`, and the
1159 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1160 `ERR_raise_data()` adds more flexibility by taking a format string and
1161 an arbitrary number of arguments following it, to be processed with
1166 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1167 to check if a named provider is loaded and available. When called, it
1168 will also activate all fallback providers if such are still present.
1172 * Enforce a minimum DH modulus size of 512 bits.
1176 * Changed DH parameters to generate the order q subgroup instead of 2q.
1177 Previously generated DH parameters are still accepted by DH_check
1178 but DH_generate_key works around that by clearing bit 0 of the
1179 private key for those. This avoids leaking bit 0 of the private key.
1183 * Significantly reduce secure memory usage by the randomness pools.
1187 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1192 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1193 algorithms. An implementation of a key exchange algorithm can be obtained
1194 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1195 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1196 the older EVP_PKEY_derive_init() function. See the man pages for the new
1197 functions for further details.
1201 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1205 * Removed the function names from error messages and deprecated the
1210 * Removed NextStep support and the macro OPENSSL_UNISTD
1214 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1215 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1216 Also removed "export var as function" capability; we do not export
1217 variables, only functions.
1221 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1222 an error and 1 indicating success. In previous versions of OpenSSL this
1223 was a void type. If a key was set longer than the maximum possible this
1228 * Support SM2 signing and verification schemes with X509 certificate.
1232 * Use SHA256 as the default digest for TS query in the `ts` app.
1236 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1240 * Default cipher lists/suites are now available via a function, the
1241 #defines are deprecated.
1245 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1246 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1247 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1251 * Join the directories crypto/x509 and crypto/x509v3
1255 * Added command 'openssl kdf' that uses the EVP_KDF API.
1259 * Added command 'openssl mac' that uses the EVP_MAC API.
1263 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1264 as default directories. Also added the command 'openssl info'
1265 for scripting purposes.
1269 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1274 * Add prediction resistance to the DRBG reseeding process.
1278 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1279 mandated by IEEE Std 1619-2018.
1283 * Added newline escaping functionality to a filename when using openssl dgst.
1284 This output format is to replicate the output format found in the `*sum`
1285 checksum programs. This aims to preserve backward compatibility.
1287 *Matt Eaton, Richard Levitte, and Paul Dale*
1289 * Removed the heartbeat message in DTLS feature, as it has very
1290 little usage and doesn't seem to fulfill a valuable purpose.
1291 The configuration option is now deprecated.
1295 * Changed the output of 'openssl {digestname} < file' to display the
1296 digest name in its output.
1300 * Added a new generic trace API which provides support for enabling
1301 instrumentation through trace output.
1303 *Richard Levitte & Matthias St. Pierre*
1305 * Added build tests for C++. These are generated files that only do one
1306 thing, to include one public OpenSSL head file each. This tests that
1307 the public header files can be usefully included in a C++ application.
1309 This test isn't enabled by default. It can be enabled with the option
1310 'enable-buildtest-c++'.
1314 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1318 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1322 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1326 * Added KMAC to EVP_MAC.
1330 * Added property based algorithm implementation selection framework to
1335 * Added SCA hardening for modular field inversion in EC_GROUP through
1336 a new dedicated field_inv() pointer in EC_METHOD.
1337 This also addresses a leakage affecting conversions from projective
1338 to affine coordinates.
1340 *Billy Bob Brumley, Nicola Tuveri*
1342 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1343 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1344 those algorithms that were already supported through the EVP_PKEY API
1345 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1346 and scrypt are now wrappers that call EVP_KDF.
1350 * Build devcrypto engine as a dynamic engine.
1352 *Eneas U de Queiroz*
1354 * Add keyed BLAKE2 to EVP_MAC.
1358 * Fix a bug in the computation of the endpoint-pair shared secret used
1359 by DTLS over SCTP. This breaks interoperability with older versions
1360 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1361 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1362 interoperability with such broken implementations. However, enabling
1363 this switch breaks interoperability with correct implementations.
1365 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1366 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1370 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1374 * Changed the license to the Apache License v2.0.
1378 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1380 - Major releases (indicated by incrementing the MAJOR release number)
1381 may introduce incompatible API/ABI changes.
1382 - Minor releases (indicated by incrementing the MINOR release number)
1383 may introduce new features but retain API/ABI compatibility.
1384 - Patch releases (indicated by incrementing the PATCH number)
1385 are intended for bug fixes and other improvements of existing
1386 features only (like improving performance or adding documentation)
1387 and retain API/ABI compatibility.
1391 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1395 * Remove the 'dist' target and add a tarball building script. The
1396 'dist' target has fallen out of use, and it shouldn't be
1397 necessary to configure just to create a source distribution.
1401 * Recreate the OS390-Unix config target. It no longer relies on a
1402 special script like it did for OpenSSL pre-1.1.0.
1406 * Instead of having the source directories listed in Configure, add
1407 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1412 * Add GMAC to EVP_MAC.
1416 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1420 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1421 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1422 to facilitate the continued use of MACs through raw private keys in
1423 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1427 * Deprecate ECDH_KDF_X9_62().
1431 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1432 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1433 are retained for backwards compatibility.
1437 * AES-XTS mode now enforces that its two keys are different to mitigate
1438 the attacked described in "Efficient Instantiations of Tweakable
1439 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1440 Details of this attack can be obtained from:
1441 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1445 * Rename the object files, i.e. give them other names than in previous
1446 versions. Their names now include the name of the final product, as
1447 well as its type mnemonic (bin, lib, shlib).
1451 * Added new option for 'openssl list', '-objects', which will display the
1452 list of built in objects, i.e. OIDs with names.
1456 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1457 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1462 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1463 improves application performance by removing data copies and providing
1464 applications with zero-copy system calls such as sendfile and splice.
1468 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1472 * `PKCS12_parse` now maintains the order of the parsed certificates
1473 when outputting them via `*ca` (rather than reversing it).
1477 * Deprecated pthread fork support methods.
1481 * Added support for FFDHE key exchange in TLS 1.3.
1485 * Added a new concept for OpenSSL plugability: providers. This
1486 functionality is designed to replace the ENGINE API and ENGINE
1487 implementations, and to be much more dynamic, allowing provider
1488 authors to introduce new algorithms among other things, as long as
1489 there's an API that supports the algorithm type.
1491 With this concept comes a new core API for interaction between
1492 libcrypto and provider implementations. Public libcrypto functions
1493 that want to use providers do so through this core API.
1495 The main documentation for this core API is found in
1496 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1497 refer to other manuals describing the API specific for supported
1498 algorithm types (also called operations).
1505 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
1507 * Avoid loading of a dynamic engine twice.
1511 * Prioritise DANE TLSA issuer certs over peer certs
1515 * Fixed random API for MacOS prior to 10.12
1517 These MacOS versions don't support the CommonCrypto APIs
1521 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1523 * Fixed an SM2 Decryption Buffer Overflow.
1525 In order to decrypt SM2 encrypted data an application is expected to
1526 call the API function EVP_PKEY_decrypt(). Typically an application will
1527 call this function twice. The first time, on entry, the "out" parameter
1528 can be NULL and, on exit, the "outlen" parameter is populated with the
1529 buffer size required to hold the decrypted plaintext. The application
1530 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1531 again, but this time passing a non-NULL value for the "out" parameter.
1533 A bug in the implementation of the SM2 decryption code means that the
1534 calculation of the buffer size required to hold the plaintext returned
1535 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1536 size required by the second call. This can lead to a buffer overflow
1537 when EVP_PKEY_decrypt() is called by the application a second time with
1538 a buffer that is too small.
1540 A malicious attacker who is able present SM2 content for decryption to
1541 an application could cause attacker chosen data to overflow the buffer
1542 by up to a maximum of 62 bytes altering the contents of other data held
1543 after the buffer, possibly changing application behaviour or causing
1544 the application to crash. The location of the buffer is application
1545 dependent but is typically heap allocated.
1550 * Fixed various read buffer overruns processing ASN.1 strings
1552 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1553 structure which contains a buffer holding the string data and a field
1554 holding the buffer length. This contrasts with normal C strings which
1555 are repesented as a buffer for the string data which is terminated
1556 with a NUL (0) byte.
1558 Although not a strict requirement, ASN.1 strings that are parsed using
1559 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1560 well as any string whose value has been set with the ASN1_STRING_set()
1561 function will additionally NUL terminate the byte array in the
1562 ASN1_STRING structure.
1564 However, it is possible for applications to directly construct valid
1565 ASN1_STRING structures which do not NUL terminate the byte array by
1566 directly setting the "data" and "length" fields in the ASN1_STRING
1567 array. This can also happen by using the ASN1_STRING_set0() function.
1569 Numerous OpenSSL functions that print ASN.1 data have been found to
1570 assume that the ASN1_STRING byte array will be NUL terminated, even
1571 though this is not guaranteed for strings that have been directly
1572 constructed. Where an application requests an ASN.1 structure to be
1573 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1574 been directly constructed by the application without NUL terminating
1575 the "data" field, then a read buffer overrun can occur.
1577 The same thing can also occur during name constraints processing
1578 of certificates (for example if a certificate has been directly
1579 constructed by the application instead of loading it via the OpenSSL
1580 parsing functions, and the certificate contains non NUL terminated
1581 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1582 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1584 If a malicious actor can cause an application to directly construct an
1585 ASN1_STRING and then process it through one of the affected OpenSSL
1586 functions then this issue could be hit. This might result in a crash
1587 (causing a Denial of Service attack). It could also result in the
1588 disclosure of private memory contents (such as private keys, or
1589 sensitive plaintext).
1594 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1596 * Fixed a problem with verifying a certificate chain when using the
1597 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1598 the certificates present in a certificate chain. It is not set by default.
1600 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1601 the chain that have explicitly encoded elliptic curve parameters was added
1602 as an additional strict check.
1604 An error in the implementation of this check meant that the result of a
1605 previous check to confirm that certificates in the chain are valid CA
1606 certificates was overwritten. This effectively bypasses the check
1607 that non-CA certificates must not be able to issue other certificates.
1609 If a "purpose" has been configured then there is a subsequent opportunity
1610 for checks that the certificate is a valid CA. All of the named "purpose"
1611 values implemented in libcrypto perform this check. Therefore, where
1612 a purpose is set the certificate chain will still be rejected even when the
1613 strict flag has been used. A purpose is set by default in libssl client and
1614 server certificate verification routines, but it can be overridden or
1615 removed by an application.
1617 In order to be affected, an application must explicitly set the
1618 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1619 for the certificate verification or, in the case of TLS client or server
1620 applications, override the default purpose.
1625 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1626 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1627 renegotiation ClientHello omits the signature_algorithms extension (where it
1628 was present in the initial ClientHello), but includes a
1629 signature_algorithms_cert extension then a NULL pointer dereference will
1630 result, leading to a crash and a denial of service attack.
1632 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1633 (which is the default configuration). OpenSSL TLS clients are not impacted by
1637 *Peter Kästle and Samuel Sapalski*
1639 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1641 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1642 create a unique hash value based on the issuer and serial number data
1643 contained within an X509 certificate. However it was failing to correctly
1644 handle any errors that may occur while parsing the issuer field (which might
1645 occur if the issuer field is maliciously constructed). This may subsequently
1646 result in a NULL pointer deref and a crash leading to a potential denial of
1652 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1653 padding mode to correctly check for rollback attacks. This is considered a
1654 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1659 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1660 functions. Previously they could overflow the output length argument in some
1661 cases where the input length is close to the maximum permissable length for
1662 an integer on the platform. In such cases the return value from the function
1663 call would be 1 (indicating success), but the output length value would be
1664 negative. This could cause applications to behave incorrectly or crash.
1669 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1670 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1671 could be exploited in a side channel attack to recover the password. Since
1672 the attack is local host only this is outside of the current OpenSSL
1673 threat model and therefore no CVE is assigned.
1675 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1680 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1682 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1683 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1684 If an attacker can control both items being compared then this could lead
1685 to a possible denial of service attack. OpenSSL itself uses the
1686 GENERAL_NAME_cmp function for two purposes:
1687 1) Comparing CRL distribution point names between an available CRL and a
1688 CRL distribution point embedded in an X509 certificate
1689 2) When verifying that a timestamp response token signer matches the
1690 timestamp authority name (exposed via the API functions
1691 TS_RESP_verify_response and TS_RESP_verify_token)
1696 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1698 * Certificates with explicit curve parameters are now disallowed in
1699 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1703 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1704 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1705 conversely, silently ignore DTLS protocol version bounds when configuring
1706 TLS-based contexts. The commands can be repeated to set bounds of both
1707 types. The same applies with the corresponding "min_protocol" and
1708 "max_protocol" command-line switches, in case some application uses both TLS
1711 SSL_CTX instances that are created for a fixed protocol version (e.g.
1712 TLSv1_server_method()) also silently ignore version bounds. Previously
1713 attempts to apply bounds to these protocol versions would result in an
1714 error. Now only the "version-flexible" SSL_CTX instances are subject to
1715 limits in configuration files in command-line options.
1719 * Handshake now fails if Extended Master Secret extension is dropped
1724 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1726 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1728 * Fixed segmentation fault in SSL_check_chain()
1729 Server or client applications that call the SSL_check_chain() function
1730 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1731 dereference as a result of incorrect handling of the
1732 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1733 or unrecognised signature algorithm is received from the peer. This could
1734 be exploited by a malicious peer in a Denial of Service attack.
1739 * Added AES consttime code for no-asm configurations
1740 an optional constant time support for AES was added
1741 when building openssl for no-asm.
1742 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1743 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1744 At this time this feature is by default disabled.
1745 It will be enabled by default in 3.0.
1749 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1751 * Revert the change of EOF detection while reading in libssl to avoid
1752 regressions in applications depending on the current way of reporting
1753 the EOF. As the existing method is not fully accurate the change to
1754 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1755 branch and will be present in the 3.0 release.
1759 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1760 when primes for RSA keys are computed.
1761 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1762 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1763 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1764 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1765 This avoids possible fingerprinting of newly generated RSA modules.
1769 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1771 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1772 while reading in libssl then we would report an error back to the
1773 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1774 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1775 therefore give a hint as to what went wrong.
1779 * Check that ed25519 and ed448 are allowed by the security level. Previously
1780 signature algorithms not using an MD were not being checked that they were
1781 allowed by the security level.
1785 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1786 was not quite right. The behaviour was not consistent between resumption
1787 and normal handshakes, and also not quite consistent with historical
1788 behaviour. The behaviour in various scenarios has been clarified and
1789 it has been updated to make it match historical behaviour as closely as
1794 * *[VMS only]* The header files that the VMS compilers include automatically,
1795 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1796 that the C++ compiler doesn't understand. This is a shortcoming in the
1797 compiler, but can be worked around with `__cplusplus` guards.
1799 C++ applications that use OpenSSL libraries must be compiled using the
1800 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1801 functions. Otherwise, only functions with symbols of less than 31
1802 characters can be used, as the linker will not be able to successfully
1803 resolve symbols with longer names.
1807 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1808 The presence of this system service is determined at run-time.
1812 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1817 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1819 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1820 number generator (RNG). This was intended to include protection in the
1821 event of a fork() system call in order to ensure that the parent and child
1822 processes did not share the same RNG state. However this protection was not
1823 being used in the default case.
1825 A partial mitigation for this issue is that the output from a high
1826 precision timer is mixed into the RNG state so the likelihood of a parent
1827 and child process sharing state is significantly reduced.
1829 If an application already calls OPENSSL_init_crypto() explicitly using
1830 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1833 *Matthias St. Pierre*
1835 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1836 used even when parsing explicit parameters, when loading a encoded key
1837 or calling `EC_GROUP_new_from_ecpkparameters()`/
1838 `EC_GROUP_new_from_ecparameters()`.
1839 This prevents bypass of security hardening and performance gains,
1840 especially for curves with specialized EC_METHODs.
1841 By default, if a key encoded with explicit parameters is loaded and later
1842 encoded, the output is still encoded with explicit parameters, even if
1843 internally a "named" EC_GROUP is used for computation.
1847 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1848 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1849 NULL. After this change, only the cofactor parameter can be NULL. It also
1850 does some minimal sanity checks on the passed order.
1855 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1856 An attack is simple, if the first CMS_recipientInfo is valid but the
1857 second CMS_recipientInfo is chosen ciphertext. If the second
1858 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1859 encryption key will be replaced by garbage, and the message cannot be
1860 decoded, but if the RSA decryption fails, the correct encryption key is
1861 used and the recipient will not notice the attack.
1862 As a work around for this potential attack the length of the decrypted
1863 key must be equal to the cipher default key length, in case the
1864 certifiate is not given and all recipientInfo are tried out.
1865 The old behaviour can be re-enabled in the CMS code by setting the
1866 CMS_DEBUG_DECRYPT flag.
1871 * Early start up entropy quality from the DEVRANDOM seed source has been
1872 improved for older Linux systems. The RAND subsystem will wait for
1873 /dev/random to be producing output before seeding from /dev/urandom.
1874 The seeded state is stored for future library initialisations using
1875 a system global shared memory segment. The shared memory identifier
1876 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1877 the desired value. The default identifier is 114.
1881 * Correct the extended master secret constant on EBCDIC systems. Without this
1882 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1883 negotiate EMS will fail. Unfortunately this also means that TLS connections
1884 between EBCDIC systems with this fix, and EBCDIC systems without this
1885 fix will fail if they negotiate EMS.
1889 * Use Windows installation paths in the mingw builds
1891 Mingw isn't a POSIX environment per se, which means that Windows
1892 paths should be used for installation.
1897 * Changed DH_check to accept parameters with order q and 2q subgroups.
1898 With order 2q subgroups the bit 0 of the private key is not secret
1899 but DH_generate_key works around that by clearing bit 0 of the
1900 private key for those. This avoids leaking bit 0 of the private key.
1904 * Significantly reduce secure memory usage by the randomness pools.
1908 * Revert the DEVRANDOM_WAIT feature for Linux systems
1910 The DEVRANDOM_WAIT feature added a select() call to wait for the
1911 /dev/random device to become readable before reading from the
1912 /dev/urandom device.
1914 It turned out that this change had negative side effects on
1915 performance which were not acceptable. After some discussion it
1916 was decided to revert this feature and leave it up to the OS
1917 resp. the platform maintainer to ensure a proper initialization
1918 during early boot time.
1920 *Matthias St. Pierre*
1922 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1924 * Add build tests for C++. These are generated files that only do one
1925 thing, to include one public OpenSSL head file each. This tests that
1926 the public header files can be usefully included in a C++ application.
1928 This test isn't enabled by default. It can be enabled with the option
1929 'enable-buildtest-c++'.
1933 * Enable SHA3 pre-hashing for ECDSA and DSA.
1937 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1938 This changes the size when using the `genpkey` command when no size is given.
1939 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1940 generation commands to use 2048 bits by default.
1944 * Reorganize the manual pages to consistently have RETURN VALUES,
1945 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1946 util/fix-doc-nits accordingly.
1948 *Paul Yang, Joshua Lock*
1950 * Add the missing accessor EVP_PKEY_get0_engine()
1954 * Have commands like `s_client` and `s_server` output the signature scheme
1955 along with other cipher suite parameters when debugging.
1959 * Make OPENSSL_config() error agnostic again.
1963 * Do the error handling in RSA decryption constant time.
1967 * Prevent over long nonces in ChaCha20-Poly1305.
1969 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1970 for every encryption operation. RFC 7539 specifies that the nonce value
1971 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1972 and front pads the nonce with 0 bytes if it is less than 12
1973 bytes. However it also incorrectly allows a nonce to be set of up to 16
1974 bytes. In this case only the last 12 bytes are significant and any
1975 additional leading bytes are ignored.
1977 It is a requirement of using this cipher that nonce values are
1978 unique. Messages encrypted using a reused nonce value are susceptible to
1979 serious confidentiality and integrity attacks. If an application changes
1980 the default nonce length to be longer than 12 bytes and then makes a
1981 change to the leading bytes of the nonce expecting the new value to be a
1982 new unique nonce then such an application could inadvertently encrypt
1983 messages with a reused nonce.
1985 Additionally the ignored bytes in a long nonce are not covered by the
1986 integrity guarantee of this cipher. Any application that relies on the
1987 integrity of these ignored leading bytes of a long nonce may be further
1988 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1989 is safe because no such use sets such a long nonce value. However user
1990 applications that use this cipher directly and set a non-default nonce
1991 length to be longer than 12 bytes may be vulnerable.
1993 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1999 * Add DEVRANDOM_WAIT feature for Linux systems
2001 On older Linux systems where the getrandom() system call is not available,
2002 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2003 Contrary to getrandom(), the /dev/urandom device will not block during
2004 early boot when the kernel CSPRNG has not been seeded yet.
2006 To mitigate this known weakness, use select() to wait for /dev/random to
2007 become readable before reading from /dev/urandom.
2009 * Ensure that SM2 only uses SM3 as digest algorithm
2013 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2015 * Change the info callback signals for the start and end of a post-handshake
2016 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2017 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2018 confused by this and assume that a TLSv1.2 renegotiation has started. This
2019 can break KeyUpdate handling. Instead we no longer signal the start and end
2020 of a post handshake message exchange (although the messages themselves are
2021 still signalled). This could break some applications that were expecting
2022 the old signals. However without this KeyUpdate is not usable for many
2027 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2029 * Timing vulnerability in DSA signature generation
2031 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2032 timing side channel attack. An attacker could use variations in the signing
2033 algorithm to recover the private key.
2035 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2040 * Timing vulnerability in ECDSA signature generation
2042 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2043 timing side channel attack. An attacker could use variations in the signing
2044 algorithm to recover the private key.
2046 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2051 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2052 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2053 of two gigabytes and the error handling improved.
2055 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2056 categorized as a normal bug, not a security issue, because the DRBG reseeds
2057 automatically and is fully functional even without additional randomness
2058 provided by the application.
2060 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2062 * Add a new ClientHello callback. Provides a callback interface that gives
2063 the application the ability to adjust the nascent SSL object at the
2064 earliest stage of ClientHello processing, immediately after extensions have
2065 been collected but before they have been processed. In particular, this
2066 callback can adjust the supported TLS versions in response to the contents
2071 * Add SM2 base algorithm support.
2075 * s390x assembly pack: add (improved) hardware-support for the following
2076 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2077 aes-cfb/cfb8, aes-ecb.
2081 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2082 parameter is no longer accepted, as it leads to a corrupt table. NULL
2083 pem_str is reserved for alias entries only.
2087 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2088 step for prime curves. The new implementation is based on formulae from
2089 differential addition-and-doubling in homogeneous projective coordinates
2090 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2091 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2092 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2093 to work in projective coordinates.
2095 *Billy Bob Brumley, Nicola Tuveri*
2097 * Change generating and checking of primes so that the error rate of not
2098 being prime depends on the intended use based on the size of the input.
2099 For larger primes this will result in more rounds of Miller-Rabin.
2100 The maximal error rate for primes with more than 1080 bits is lowered
2103 *Kurt Roeckx, Annie Yousar*
2105 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2109 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2110 moving between systems, and to avoid confusion when a Windows build is
2111 done with mingw vs with MSVC. For POSIX installs, there's still a
2112 symlink or copy named 'tsget' to avoid that confusion as well.
2116 * Revert blinding in ECDSA sign and instead make problematic addition
2117 length-invariant. Switch even to fixed-length Montgomery multiplication.
2121 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2122 step for binary curves. The new implementation is based on formulae from
2123 differential addition-and-doubling in mixed Lopez-Dahab projective
2124 coordinates, modified to independently blind the operands.
2126 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2128 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2129 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2130 EC_METHODs to implement their own specialized "ladder step", to take
2131 advantage of more favorable coordinate systems or more efficient
2132 differential addition-and-doubling algorithms.
2134 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2136 * Modified the random device based seed sources to keep the relevant
2137 file descriptors open rather than reopening them on each access.
2138 This allows such sources to operate in a chroot() jail without
2139 the associated device nodes being available. This behaviour can be
2140 controlled using RAND_keep_random_devices_open().
2144 * Numerous side-channel attack mitigations have been applied. This may have
2145 performance impacts for some algorithms for the benefit of improved
2146 security. Specific changes are noted in this change log by their respective
2151 * AIX shared library support overhaul. Switch to AIX "natural" way of
2152 handling shared libraries, which means collecting shared objects of
2153 different versions and bitnesses in one common archive. This allows to
2154 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2155 doesn't affect the way 3rd party applications are linked, only how
2156 multi-version installation is managed.
2160 * Make ec_group_do_inverse_ord() more robust and available to other
2161 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2162 mitigations are applied to the fallback BN_mod_inverse().
2163 When using this function rather than BN_mod_inverse() directly, new
2164 EC cryptosystem implementations are then safer-by-default.
2168 * Add coordinate blinding for EC_POINT and implement projective
2169 coordinate blinding for generic prime curves as a countermeasure to
2170 chosen point SCA attacks.
2172 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2174 * Add blinding to ECDSA and DSA signatures to protect against side channel
2175 attacks discovered by Keegan Ryan (NCC Group).
2179 * Enforce checking in the `pkeyutl` command to ensure that the input
2180 length does not exceed the maximum supported digest length when performing
2181 a sign, verify or verifyrecover operation.
2185 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2186 I/O in combination with something like select() or poll() will hang. This
2187 can be turned off again using SSL_CTX_clear_mode().
2188 Many applications do not properly handle non-application data records, and
2189 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2190 around the problems in those applications, but can also break some.
2191 It's recommended to read the manpages about SSL_read(), SSL_write(),
2192 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2193 SSL_CTX_set_read_ahead() again.
2197 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2198 now allow empty (zero character) pass phrases.
2202 * Apply blinding to binary field modular inversion and remove patent
2203 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2207 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2208 binary and prime elliptic curves.
2212 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2213 constant time fixed point multiplication.
2217 * Revise elliptic curve scalar multiplication with timing attack
2218 defenses: ec_wNAF_mul redirects to a constant time implementation
2219 when computing fixed point and variable point multiplication (which
2220 in OpenSSL are mostly used with secret scalars in keygen, sign,
2221 ECDH derive operations).
2222 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2225 * Updated CONTRIBUTING
2229 * Updated DRBG / RAND to request nonce and additional low entropy
2230 randomness from the system.
2232 *Matthias St. Pierre*
2234 * Updated 'openssl rehash' to use OpenSSL consistent default.
2238 * Moved the load of the ssl_conf module to libcrypto, which helps
2239 loading engines that libssl uses before libssl is initialised.
2243 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2247 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2249 *Ingo Schwarze, Rich Salz*
2251 * Added output of accepting IP address and port for 'openssl s_server'
2255 * Added a new API for TLSv1.3 ciphersuites:
2256 SSL_CTX_set_ciphersuites()
2257 SSL_set_ciphersuites()
2261 * Memory allocation failures consistently add an error to the error
2266 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2267 in libcrypto when run as setuid/setgid.
2271 * Load any config file by default when libssl is used.
2275 * Added new public header file <openssl/rand_drbg.h> and documentation
2276 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2278 *Matthias St. Pierre*
2280 * QNX support removed (cannot find contributors to get their approval
2281 for the license change).
2285 * TLSv1.3 replay protection for early data has been implemented. See the
2286 SSL_read_early_data() man page for further details.
2290 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2291 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2292 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2293 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2294 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2295 configuration has been separated out. See the ciphers man page or the
2296 SSL_CTX_set_ciphersuites() man page for more information.
2300 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2301 in responder mode now supports the new "-multi" option, which
2302 spawns the specified number of child processes to handle OCSP
2303 requests. The "-timeout" option now also limits the OCSP
2304 responder's patience to wait to receive the full client request
2305 on a newly accepted connection. Child processes are respawned
2306 as needed, and the CA index file is automatically reloaded
2307 when changed. This makes it possible to run the "ocsp" responder
2308 as a long-running service, making the OpenSSL CA somewhat more
2309 feature-complete. In this mode, most diagnostic messages logged
2310 after entering the event loop are logged via syslog(3) rather than
2315 * Added support for X448 and Ed448. Heavily based on original work by
2320 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2321 objects loaded. This adds the functions OSSL_STORE_expect() and
2322 OSSL_STORE_find() as well as needed tools to construct searches and
2323 get the search data out of them.
2327 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2328 version of OpenSSL should review their configuration settings to ensure
2329 that they are still appropriate for TLSv1.3. For further information see:
2330 <https://wiki.openssl.org/index.php/TLS1.3>
2334 * Grand redesign of the OpenSSL random generator
2336 The default RAND method now utilizes an AES-CTR DRBG according to
2337 NIST standard SP 800-90Ar1. The new random generator is essentially
2338 a port of the default random generator from the OpenSSL FIPS 2.0
2339 object module. It is a hybrid deterministic random bit generator
2340 using an AES-CTR bit stream and which seeds and reseeds itself
2341 automatically using trusted system entropy sources.
2343 Some of its new features are:
2344 - Support for multiple DRBG instances with seed chaining.
2345 - The default RAND method makes use of a DRBG.
2346 - There is a public and private DRBG instance.
2347 - The DRBG instances are fork-safe.
2348 - Keep all global DRBG instances on the secure heap if it is enabled.
2349 - The public and private DRBG instance are per thread for lock free
2352 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2354 * Changed Configure so it only says what it does and doesn't dump
2355 so much data. Instead, ./configdata.pm should be used as a script
2356 to display all sorts of configuration data.
2360 * Added processing of "make variables" to Configure.
2364 * Added SHA512/224 and SHA512/256 algorithm support.
2368 * The last traces of Netware support, first removed in 1.1.0, have
2373 * Get rid of Makefile.shared, and in the process, make the processing
2374 of certain files (rc.obj, or the .def/.map/.opt files produced from
2375 the ordinal files) more visible and hopefully easier to trace and
2376 debug (or make silent).
2380 * Make it possible to have environment variable assignments as
2381 arguments to config / Configure.
2385 * Add multi-prime RSA (RFC 8017) support.
2389 * Add SM3 implemented according to GB/T 32905-2016
2390 *Jack Lloyd <jack.lloyd@ribose.com>,*
2391 *Ronald Tse <ronald.tse@ribose.com>,*
2392 *Erick Borsboom <erick.borsboom@ribose.com>*
2394 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2395 as documented in RFC6066.
2396 Based on a patch from Tomasz Moń
2398 *Filipe Raimundo da Silva*
2400 * Add SM4 implemented according to GB/T 32907-2016.
2401 *Jack Lloyd <jack.lloyd@ribose.com>,*
2402 *Ronald Tse <ronald.tse@ribose.com>,*
2403 *Erick Borsboom <erick.borsboom@ribose.com>*
2405 * Reimplement -newreq-nodes and ERR_error_string_n; the
2406 original author does not agree with the license change.
2410 * Add ARIA AEAD TLS support.
2414 * Some macro definitions to support VS6 have been removed. Visual
2415 Studio 6 has not worked since 1.1.0
2419 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2420 without clearing the errors.
2424 * Add "atfork" functions. If building on a system that without
2425 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2426 requirements. The RAND facility now uses/requires this.
2434 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2435 not possible to disable entirely. However, it's still possible to
2436 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2439 To disable, configure with 'no-ui-console'. 'no-ui' is still
2440 possible to use as an alias. Check at compile time with the
2441 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2442 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2446 * Add a STORE module, which implements a uniform and URI based reader of
2447 stores that can contain keys, certificates, CRLs and numerous other
2448 objects. The main API is loosely based on a few stdio functions,
2449 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2450 OSSL_STORE_error and OSSL_STORE_close.
2451 The implementation uses backends called "loaders" to implement arbitrary
2452 URI schemes. There is one built in "loader" for the 'file' scheme.
2456 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2457 then adjusted to work on FreeBSD 8.4 as well.
2458 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2459 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2463 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2464 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2465 error code calls like this:
2467 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2469 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2470 that can be encoded in C. For the foreseeable future, this will only
2473 *Richard Levitte and Tim Hudson*
2475 * Removed BSD cryptodev engine.
2479 * Add a build target 'build_all_generated', to build all generated files
2480 and only that. This can be used to prepare everything that requires
2481 things like perl for a system that lacks perl and then move everything
2482 to that system and do the rest of the build there.
2486 * In the UI interface, make it possible to duplicate the user data. This
2487 can be used by engines that need to retain the data for a longer time
2488 than just the call where this user data is passed.
2492 * Ignore the '-named_curve auto' value for compatibility of applications
2495 *Tomáš Mráz <tmraz@fedoraproject.org>*
2497 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2498 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2499 alerts across multiple records (some of which could be empty). In practice
2500 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2501 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2502 support this at all. Supporting it adds significant complexity to the
2503 record layer, and its removal is unlikely to cause interoperability
2508 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2509 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2510 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2515 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2516 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2518 *Richard Levitte, Andy Polyakov*
2520 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2525 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2526 platform rather than 'mingw'.
2530 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2531 success if they are asked to add an object which already exists
2532 in the store. This change cascades to other functions which load
2533 certificates and CRLs.
2537 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2538 facilitate stack unwinding even from assembly subroutines.
2542 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2543 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2547 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2548 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2549 which is the minimum version we support.
2553 * Certificate time validation (X509_cmp_time) enforces stricter
2554 compliance with RFC 5280. Fractional seconds and timezone offsets
2555 are no longer allowed.
2559 * Add support for ARIA
2563 * s_client will now send the Server Name Indication (SNI) extension by
2564 default unless the new "-noservername" option is used. The server name is
2565 based on the host provided to the "-connect" option unless overridden by
2566 using "-servername".
2570 * Add support for SipHash
2574 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2575 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2576 prevent issues where no progress is being made and the peer continually
2577 sends unrecognised record types, using up resources processing them.
2581 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2582 using the algorithm defined in
2583 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2587 * Heartbeat support has been removed; the ABI is changed for now.
2589 *Richard Levitte, Rich Salz*
2591 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2595 * The RSA "null" method, which was partially supported to avoid patent
2596 issues, has been replaced to always returns NULL.
2603 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2605 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2606 used even when parsing explicit parameters, when loading a encoded key
2607 or calling `EC_GROUP_new_from_ecpkparameters()`/
2608 `EC_GROUP_new_from_ecparameters()`.
2609 This prevents bypass of security hardening and performance gains,
2610 especially for curves with specialized EC_METHODs.
2611 By default, if a key encoded with explicit parameters is loaded and later
2612 encoded, the output is still encoded with explicit parameters, even if
2613 internally a "named" EC_GROUP is used for computation.
2617 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2618 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2619 NULL. After this change, only the cofactor parameter can be NULL. It also
2620 does some minimal sanity checks on the passed order.
2625 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2626 An attack is simple, if the first CMS_recipientInfo is valid but the
2627 second CMS_recipientInfo is chosen ciphertext. If the second
2628 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2629 encryption key will be replaced by garbage, and the message cannot be
2630 decoded, but if the RSA decryption fails, the correct encryption key is
2631 used and the recipient will not notice the attack.
2632 As a work around for this potential attack the length of the decrypted
2633 key must be equal to the cipher default key length, in case the
2634 certifiate is not given and all recipientInfo are tried out.
2635 The old behaviour can be re-enabled in the CMS code by setting the
2636 CMS_DEBUG_DECRYPT flag.
2641 * Use Windows installation paths in the mingw builds
2643 Mingw isn't a POSIX environment per se, which means that Windows
2644 paths should be used for installation.
2649 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2651 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2652 This changes the size when using the `genpkey` command when no size is given.
2653 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2654 generation commands to use 2048 bits by default.
2658 * Prevent over long nonces in ChaCha20-Poly1305.
2660 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2661 for every encryption operation. RFC 7539 specifies that the nonce value
2662 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2663 and front pads the nonce with 0 bytes if it is less than 12
2664 bytes. However it also incorrectly allows a nonce to be set of up to 16
2665 bytes. In this case only the last 12 bytes are significant and any
2666 additional leading bytes are ignored.
2668 It is a requirement of using this cipher that nonce values are
2669 unique. Messages encrypted using a reused nonce value are susceptible to
2670 serious confidentiality and integrity attacks. If an application changes
2671 the default nonce length to be longer than 12 bytes and then makes a
2672 change to the leading bytes of the nonce expecting the new value to be a
2673 new unique nonce then such an application could inadvertently encrypt
2674 messages with a reused nonce.
2676 Additionally the ignored bytes in a long nonce are not covered by the
2677 integrity guarantee of this cipher. Any application that relies on the
2678 integrity of these ignored leading bytes of a long nonce may be further
2679 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2680 is safe because no such use sets such a long nonce value. However user
2681 applications that use this cipher directly and set a non-default nonce
2682 length to be longer than 12 bytes may be vulnerable.
2684 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2690 * Added SCA hardening for modular field inversion in EC_GROUP through
2691 a new dedicated field_inv() pointer in EC_METHOD.
2692 This also addresses a leakage affecting conversions from projective
2693 to affine coordinates.
2695 *Billy Bob Brumley, Nicola Tuveri*
2697 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2698 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2702 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2706 * Remove the 'dist' target and add a tarball building script. The
2707 'dist' target has fallen out of use, and it shouldn't be
2708 necessary to configure just to create a source distribution.
2712 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2714 * Timing vulnerability in DSA signature generation
2716 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2717 timing side channel attack. An attacker could use variations in the signing
2718 algorithm to recover the private key.
2720 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2725 * Timing vulnerability in ECDSA signature generation
2727 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2728 timing side channel attack. An attacker could use variations in the signing
2729 algorithm to recover the private key.
2731 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2736 * Add coordinate blinding for EC_POINT and implement projective
2737 coordinate blinding for generic prime curves as a countermeasure to
2738 chosen point SCA attacks.
2740 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2742 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2744 * Client DoS due to large DH parameter
2746 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2747 malicious server can send a very large prime value to the client. This will
2748 cause the client to spend an unreasonably long period of time generating a
2749 key for this prime resulting in a hang until the client has finished. This
2750 could be exploited in a Denial Of Service attack.
2752 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2757 * Cache timing vulnerability in RSA Key Generation
2759 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2760 a cache timing side channel attack. An attacker with sufficient access to
2761 mount cache timing attacks during the RSA key generation process could
2762 recover the private key.
2764 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2765 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2770 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2771 parameter is no longer accepted, as it leads to a corrupt table. NULL
2772 pem_str is reserved for alias entries only.
2776 * Revert blinding in ECDSA sign and instead make problematic addition
2777 length-invariant. Switch even to fixed-length Montgomery multiplication.
2781 * Change generating and checking of primes so that the error rate of not
2782 being prime depends on the intended use based on the size of the input.
2783 For larger primes this will result in more rounds of Miller-Rabin.
2784 The maximal error rate for primes with more than 1080 bits is lowered
2787 *Kurt Roeckx, Annie Yousar*
2789 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2793 * Add blinding to ECDSA and DSA signatures to protect against side channel
2794 attacks discovered by Keegan Ryan (NCC Group).
2798 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2799 now allow empty (zero character) pass phrases.
2803 * Certificate time validation (X509_cmp_time) enforces stricter
2804 compliance with RFC 5280. Fractional seconds and timezone offsets
2805 are no longer allowed.
2809 * Fixed a text canonicalisation bug in CMS
2811 Where a CMS detached signature is used with text content the text goes
2812 through a canonicalisation process first prior to signing or verifying a
2813 signature. This process strips trailing space at the end of lines, converts
2814 line terminators to CRLF and removes additional trailing line terminators
2815 at the end of a file. A bug in the canonicalisation process meant that
2816 some characters, such as form-feed, were incorrectly treated as whitespace
2817 and removed. This is contrary to the specification (RFC5485). This fix
2818 could mean that detached text data signed with an earlier version of
2819 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2820 signed with a fixed OpenSSL may fail to verify with an earlier version of
2821 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2822 and use the "-binary" flag (for the "cms" command line application) or set
2823 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2827 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2829 * Constructed ASN.1 types with a recursive definition could exceed the stack
2831 Constructed ASN.1 types with a recursive definition (such as can be found
2832 in PKCS7) could eventually exceed the stack given malicious input with
2833 excessive recursion. This could result in a Denial Of Service attack. There
2834 are no such structures used within SSL/TLS that come from untrusted sources
2835 so this is considered safe.
2837 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2843 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2845 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2846 effectively reduced to only comparing the least significant bit of each
2847 byte. This allows an attacker to forge messages that would be considered as
2848 authenticated in an amount of tries lower than that guaranteed by the
2849 security claims of the scheme. The module can only be compiled by the
2850 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2852 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2858 * Add a build target 'build_all_generated', to build all generated files
2859 and only that. This can be used to prepare everything that requires
2860 things like perl for a system that lacks perl and then move everything
2861 to that system and do the rest of the build there.
2865 * Backport SSL_OP_NO_RENGOTIATION
2867 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2868 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2869 changes this is no longer possible in 1.1.0. Therefore the new
2870 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2871 1.1.0 to provide equivalent functionality.
2873 Note that if an application built against 1.1.0h headers (or above) is run
2874 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2875 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2879 * Removed the OS390-Unix config target. It relied on a script that doesn't
2884 * rsaz_1024_mul_avx2 overflow bug on x86_64
2886 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2887 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2888 Analysis suggests that attacks against RSA and DSA as a result of this
2889 defect would be very difficult to perform and are not believed likely.
2890 Attacks against DH1024 are considered just feasible, because most of the
2891 work necessary to deduce information about a private key may be performed
2892 offline. The amount of resources required for such an attack would be
2893 significant. However, for an attack on TLS to be meaningful, the server
2894 would have to share the DH1024 private key among multiple clients, which is
2895 no longer an option since CVE-2016-0701.
2897 This only affects processors that support the AVX2 but not ADX extensions
2898 like Intel Haswell (4th generation).
2900 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2901 was originally found via the OSS-Fuzz project.
2906 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2908 * bn_sqrx8x_internal carry bug on x86_64
2910 There is a carry propagating bug in the x86_64 Montgomery squaring
2911 procedure. No EC algorithms are affected. Analysis suggests that attacks
2912 against RSA and DSA as a result of this defect would be very difficult to
2913 perform and are not believed likely. Attacks against DH are considered just
2914 feasible (although very difficult) because most of the work necessary to
2915 deduce information about a private key may be performed offline. The amount
2916 of resources required for such an attack would be very significant and
2917 likely only accessible to a limited number of attackers. An attacker would
2918 additionally need online access to an unpatched system using the target
2919 private key in a scenario with persistent DH parameters and a private
2920 key that is shared between multiple clients.
2922 This only affects processors that support the BMI1, BMI2 and ADX extensions
2923 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2925 This issue was reported to OpenSSL by the OSS-Fuzz project.
2930 * Malformed X.509 IPAddressFamily could cause OOB read
2932 If an X.509 certificate has a malformed IPAddressFamily extension,
2933 OpenSSL could do a one-byte buffer overread. The most likely result
2934 would be an erroneous display of the certificate in text format.
2936 This issue was reported to OpenSSL by the OSS-Fuzz project.
2941 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2943 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2944 platform rather than 'mingw'.
2948 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2949 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2950 which is the minimum version we support.
2954 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2956 * Encrypt-Then-Mac renegotiation crash
2958 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2959 negotiated where it was not in the original handshake (or vice-versa) then
2960 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2961 and servers are affected.
2963 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2968 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2970 * Truncated packet could crash via OOB read
2972 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2973 cipher is being used, then a truncated packet can cause that host to
2974 perform an out-of-bounds read, usually resulting in a crash.
2976 This issue was reported to OpenSSL by Robert Święcki of Google.
2981 * Bad (EC)DHE parameters cause a client crash
2983 If a malicious server supplies bad parameters for a DHE or ECDHE key
2984 exchange then this can result in the client attempting to dereference a
2985 NULL pointer leading to a client crash. This could be exploited in a Denial
2988 This issue was reported to OpenSSL by Guido Vranken.
2993 * BN_mod_exp may produce incorrect results on x86_64
2995 There is a carry propagating bug in the x86_64 Montgomery squaring
2996 procedure. No EC algorithms are affected. Analysis suggests that attacks
2997 against RSA and DSA as a result of this defect would be very difficult to
2998 perform and are not believed likely. Attacks against DH are considered just
2999 feasible (although very difficult) because most of the work necessary to
3000 deduce information about a private key may be performed offline. The amount
3001 of resources required for such an attack would be very significant and
3002 likely only accessible to a limited number of attackers. An attacker would
3003 additionally need online access to an unpatched system using the target
3004 private key in a scenario with persistent DH parameters and a private
3005 key that is shared between multiple clients. For example this can occur by
3006 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3007 similar to CVE-2015-3193 but must be treated as a separate problem.
3009 This issue was reported to OpenSSL by the OSS-Fuzz project.
3014 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3016 * ChaCha20/Poly1305 heap-buffer-overflow
3018 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3019 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3020 crash. This issue is not considered to be exploitable beyond a DoS.
3022 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3027 * CMS Null dereference
3029 Applications parsing invalid CMS structures can crash with a NULL pointer
3030 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3031 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3032 structure callback if an attempt is made to free certain invalid encodings.
3033 Only CHOICE structures using a callback which do not handle NULL value are
3036 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3041 * Montgomery multiplication may produce incorrect results
3043 There is a carry propagating bug in the Broadwell-specific Montgomery
3044 multiplication procedure that handles input lengths divisible by, but
3045 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3046 and DH private keys are impossible. This is because the subroutine in
3047 question is not used in operations with the private key itself and an input
3048 of the attacker's direct choice. Otherwise the bug can manifest itself as
3049 transient authentication and key negotiation failures or reproducible
3050 erroneous outcome of public-key operations with specially crafted input.
3051 Among EC algorithms only Brainpool P-512 curves are affected and one
3052 presumably can attack ECDH key negotiation. Impact was not analyzed in
3053 detail, because pre-requisites for attack are considered unlikely. Namely
3054 multiple clients have to choose the curve in question and the server has to
3055 share the private key among them, neither of which is default behaviour.
3056 Even then only clients that chose the curve will be affected.
3058 This issue was publicly reported as transient failures and was not
3059 initially recognized as a security issue. Thanks to Richard Morgan for
3060 providing reproducible case.
3065 * Removed automatic addition of RPATH in shared libraries and executables,
3066 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3070 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3072 * Fix Use After Free for large message sizes
3074 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3075 message larger than approx 16k is received then the underlying buffer to
3076 store the incoming message is reallocated and moved. Unfortunately a
3077 dangling pointer to the old location is left which results in an attempt to
3078 write to the previously freed location. This is likely to result in a
3079 crash, however it could potentially lead to execution of arbitrary code.
3081 This issue only affects OpenSSL 1.1.0a.
3083 This issue was reported to OpenSSL by Robert Święcki.
3088 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3090 * OCSP Status Request extension unbounded memory growth
3092 A malicious client can send an excessively large OCSP Status Request
3093 extension. If that client continually requests renegotiation, sending a
3094 large OCSP Status Request extension each time, then there will be unbounded
3095 memory growth on the server. This will eventually lead to a Denial Of
3096 Service attack through memory exhaustion. Servers with a default
3097 configuration are vulnerable even if they do not support OCSP. Builds using
3098 the "no-ocsp" build time option are not affected.
3100 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3105 * SSL_peek() hang on empty record
3107 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3108 sends an empty record. This could be exploited by a malicious peer in a
3109 Denial Of Service attack.
3111 This issue was reported to OpenSSL by Alex Gaynor.
3116 * Excessive allocation of memory in tls_get_message_header() and
3117 dtls1_preprocess_fragment()
3119 A (D)TLS message includes 3 bytes for its length in the header for the
3120 message. This would allow for messages up to 16Mb in length. Messages of
3121 this length are excessive and OpenSSL includes a check to ensure that a
3122 peer is sending reasonably sized messages in order to avoid too much memory
3123 being consumed to service a connection. A flaw in the logic of version
3124 1.1.0 means that memory for the message is allocated too early, prior to
3125 the excessive message length check. Due to way memory is allocated in
3126 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3127 to service a connection. This could lead to a Denial of Service through
3128 memory exhaustion. However, the excessive message length check still takes
3129 place, and this would cause the connection to immediately fail. Assuming
3130 that the application calls SSL_free() on the failed connection in a timely
3131 manner then the 21Mb of allocated memory will then be immediately freed
3132 again. Therefore the excessive memory allocation will be transitory in
3133 nature. This then means that there is only a security impact if:
3135 1) The application does not call SSL_free() in a timely manner in the event
3136 that the connection fails
3138 2) The application is working in a constrained environment where there is
3139 very little free memory
3141 3) The attacker initiates multiple connection attempts such that there are
3142 multiple connections in a state where memory has been allocated for the
3143 connection; SSL_free() has not yet been called; and there is insufficient
3144 memory to service the multiple requests.
3146 Except in the instance of (1) above any Denial Of Service is likely to be
3147 transitory because as soon as the connection fails the memory is
3148 subsequently freed again in the SSL_free() call. However there is an
3149 increased risk during this period of application crashes due to the lack of
3150 memory - which would then mean a more serious Denial of Service.
3152 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3153 (CVE-2016-6307 and CVE-2016-6308)
3157 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3158 had to be removed. Primary reason is that vendor assembler can't
3159 assemble our modules with -KPIC flag. As result it, assembly
3160 support, was not even available as option. But its lack means
3161 lack of side-channel resistant code, which is incompatible with
3162 security by todays standards. Fortunately gcc is readily available
3163 prepackaged option, which we firmly point at...
3167 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3169 * Windows command-line tool supports UTF-8 opt-in option for arguments
3170 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3171 (to any value) allows Windows user to access PKCS#12 file generated
3172 with Windows CryptoAPI and protected with non-ASCII password, as well
3173 as files generated under UTF-8 locale on Linux also protected with
3178 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3179 have been disabled by default and removed from DEFAULT, just like RC4.
3180 See the RC4 item below to re-enable both.
3184 * The method for finding the storage location for the Windows RAND seed file
3185 has changed. First we check %RANDFILE%. If that is not set then we check
3186 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3187 all else fails we fall back to C:\.
3191 * The EVP_EncryptUpdate() function has had its return type changed from void
3192 to int. A return of 0 indicates and error while a return of 1 indicates
3197 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3198 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3199 off the constant time implementation for RSA, DSA and DH have been made
3200 no-ops and deprecated.
3204 * Windows RAND implementation was simplified to only get entropy by
3205 calling CryptGenRandom(). Various other RAND-related tickets
3208 *Joseph Wylie Yandle, Rich Salz*
3210 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3211 and `OPENSSL_LH_`, respectively. The old names are available
3212 with API compatibility. They new names are now completely documented.
3216 * Unify TYPE_up_ref(obj) methods signature.
3217 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3218 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3219 int (instead of void) like all others TYPE_up_ref() methods.
3220 So now these methods also check the return value of CRYPTO_atomic_add(),
3221 and the validity of object reference counter.
3223 *fdasilvayy@gmail.com*
3225 * With Windows Visual Studio builds, the .pdb files are installed
3226 alongside the installed libraries and executables. For a static
3227 library installation, ossl_static.pdb is the associate compiler
3228 generated .pdb file to be used when linking programs.
3232 * Remove openssl.spec. Packaging files belong with the packagers.
3236 * Automatic Darwin/OSX configuration has had a refresh, it will now
3237 recognise x86_64 architectures automatically. You can still decide
3238 to build for a different bitness with the environment variable
3239 KERNEL_BITS (can be 32 or 64), for example:
3241 KERNEL_BITS=32 ./config
3245 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3246 256 bit AES and HMAC with SHA256.
3250 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3254 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3258 * To enable users to have their own config files and build file templates,
3259 Configure looks in the directory indicated by the environment variable
3260 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3261 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3262 name and is used as is.
3266 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3267 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3268 X509_CERT_FILE_CTX was removed.
3272 * "shared" builds are now the default. To create only static libraries use
3273 the "no-shared" Configure option.
3277 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3278 All of these option have not worked for some while and are fundamental
3283 * Make various cleanup routines no-ops and mark them as deprecated. Most
3284 global cleanup functions are no longer required because they are handled
3285 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3286 Explicitly de-initing can cause problems (e.g. where a library that uses
3287 OpenSSL de-inits, but an application is still using it). The affected
3288 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3289 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3290 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3291 COMP_zlib_cleanup().
3295 * --strict-warnings no longer enables runtime debugging options
3296 such as REF_DEBUG. Instead, debug options are automatically
3297 enabled with '--debug' builds.
3299 *Andy Polyakov, Emilia Käsper*
3301 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3302 have been moved out of the public header files. New functions for managing
3303 these have been added.
3307 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3308 objects have been moved out of the public header files. New
3309 functions for managing these have been added.
3313 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3314 have been moved out of the public header files. New functions for managing
3315 these have been added.
3319 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3320 moved out of the public header files. New functions for managing these
3325 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3329 * Removed the mk1mf build scripts.
3333 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3334 it is always safe to #include a header now.
3338 * Removed the aged BC-32 config and all its supporting scripts
3342 * Removed support for Ultrix, Netware, and OS/2.
3346 * Add support for HKDF.
3348 *Alessandro Ghedini*
3350 * Add support for blake2b and blake2s
3354 * Added support for "pipelining". Ciphers that have the
3355 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3356 encryptions/decryptions simultaneously. There are currently no built-in
3357 ciphers with this property but the expectation is that engines will be able
3358 to offer it to significantly improve throughput. Support has been extended
3359 into libssl so that multiple records for a single connection can be
3360 processed in one go (for >=TLS 1.1).
3364 * Added the AFALG engine. This is an async capable engine which is able to
3365 offload work to the Linux kernel. In this initial version it only supports
3366 AES128-CBC. The kernel must be version 4.1.0 or greater.
3370 * OpenSSL now uses a new threading API. It is no longer necessary to
3371 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3372 are two supported threading models: pthreads and windows threads. It is
3373 also possible to configure OpenSSL at compile time for "no-threads". The
3374 old threading API should no longer be used. The functions have been
3375 replaced with "no-op" compatibility macros.
3377 *Alessandro Ghedini, Matt Caswell*
3379 * Modify behavior of ALPN to invoke callback after SNI/servername
3380 callback, such that updates to the SSL_CTX affect ALPN.
3384 * Add SSL_CIPHER queries for authentication and key-exchange.
3388 * Changes to the DEFAULT cipherlist:
3389 - Prefer (EC)DHE handshakes over plain RSA.
3390 - Prefer AEAD ciphers over legacy ciphers.
3391 - Prefer ECDSA over RSA when both certificates are available.
3392 - Prefer TLSv1.2 ciphers/PRF.
3393 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3398 * Change the ECC default curve list to be this, in order: x25519,
3399 secp256r1, secp521r1, secp384r1.
3403 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3404 disabled by default. They can be re-enabled using the
3405 enable-weak-ssl-ciphers option to Configure.
3409 * If the server has ALPN configured, but supports no protocols that the
3410 client advertises, send a fatal "no_application_protocol" alert.
3411 This behaviour is SHALL in RFC 7301, though it isn't universally
3412 implemented by other servers.
3416 * Add X25519 support.
3417 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3418 for public and private key encoding using the format documented in
3419 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3420 key generation and key derivation.
3422 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3427 * Deprecate SRP_VBASE_get_by_user.
3428 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3429 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3430 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3431 seed, even if the seed is configured.
3433 Users should use SRP_VBASE_get1_by_user instead. Note that in
3434 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3435 also that even though configuring the SRP seed attempts to hide
3436 invalid usernames by continuing the handshake with fake
3437 credentials, this behaviour is not constant time and no strong
3438 guarantees are made that the handshake is indistinguishable from
3439 that of a valid user.
3443 * Configuration change; it's now possible to build dynamic engines
3444 without having to build shared libraries and vice versa. This
3445 only applies to the engines in `engines/`, those in `crypto/engine/`
3446 will always be built into libcrypto (i.e. "static").
3448 Building dynamic engines is enabled by default; to disable, use
3449 the configuration option "disable-dynamic-engine".
3451 The only requirements for building dynamic engines are the
3452 presence of the DSO module and building with position independent
3453 code, so they will also automatically be disabled if configuring
3454 with "disable-dso" or "disable-pic".
3456 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3457 are also taken away from openssl/opensslconf.h, as they are
3462 * Configuration change; if there is a known flag to compile
3463 position independent code, it will always be applied on the
3464 libcrypto and libssl object files, and never on the application
3465 object files. This means other libraries that use routines from
3466 libcrypto / libssl can be made into shared libraries regardless
3467 of how OpenSSL was configured.
3469 If this isn't desirable, the configuration options "disable-pic"
3470 or "no-pic" can be used to disable the use of PIC. This will
3471 also disable building shared libraries and dynamic engines.
3475 * Removed JPAKE code. It was experimental and has no wide use.
3479 * The INSTALL_PREFIX Makefile variable has been renamed to
3480 DESTDIR. That makes for less confusion on what this variable
3481 is for. Also, the configuration option --install_prefix is
3486 * Heartbeat for TLS has been removed and is disabled by default
3487 for DTLS; configure with enable-heartbeats. Code that uses the
3488 old #define's might need to be updated.
3490 *Emilia Käsper, Rich Salz*
3492 * Rename REF_CHECK to REF_DEBUG.
3496 * New "unified" build system
3498 The "unified" build system is aimed to be a common system for all
3499 platforms we support. With it comes new support for VMS.
3501 This system builds supports building in a different directory tree
3502 than the source tree. It produces one Makefile (for unix family
3503 or lookalikes), or one descrip.mms (for VMS).
3505 The source of information to make the Makefile / descrip.mms is
3506 small files called 'build.info', holding the necessary
3507 information for each directory with source to compile, and a
3508 template in Configurations, like unix-Makefile.tmpl or
3511 With this change, the library names were also renamed on Windows
3512 and on VMS. They now have names that are closer to the standard
3513 on Unix, and include the major version number, and in certain
3514 cases, the architecture they are built for. See "Notes on shared
3515 libraries" in INSTALL.
3517 We rely heavily on the perl module Text::Template.
3521 * Added support for auto-initialisation and de-initialisation of the library.
3522 OpenSSL no longer requires explicit init or deinit routines to be called,
3523 except in certain circumstances. See the OPENSSL_init_crypto() and
3524 OPENSSL_init_ssl() man pages for further information.
3528 * The arguments to the DTLSv1_listen function have changed. Specifically the
3529 "peer" argument is now expected to be a BIO_ADDR object.
3531 * Rewrite of BIO networking library. The BIO library lacked consistent
3532 support of IPv6, and adding it required some more extensive
3533 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3534 which hold all types of addresses and chains of address information.
3535 It also introduces a new API, with functions like BIO_socket,
3536 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3537 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3538 have been adapted accordingly.
3542 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3547 * CRIME protection: disable compression by default, even if OpenSSL is
3548 compiled with zlib enabled. Applications can still enable compression
3549 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3550 using the SSL_CONF library to configure compression.
3554 * The signature of the session callback configured with
3555 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3556 was explicitly marked as `const unsigned char*` instead of
3561 * Always DPURIFY. Remove the use of uninitialized memory in the
3562 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3566 * Removed many obsolete configuration items, including
3567 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3568 MD2_CHAR, MD2_INT, MD2_LONG
3570 IDEA_SHORT, IDEA_LONG
3571 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3573 *Rich Salz, with advice from Andy Polyakov*
3575 * Many BN internals have been moved to an internal header file.
3577 *Rich Salz with help from Andy Polyakov*
3579 * Configuration and writing out the results from it has changed.
3580 Files such as Makefile include/openssl/opensslconf.h and are now
3581 produced through general templates, such as Makefile.in and
3582 crypto/opensslconf.h.in and some help from the perl module
3585 Also, the center of configuration information is no longer
3586 Makefile. Instead, Configure produces a perl module in
3587 configdata.pm which holds most of the config data (in the hash
3588 table %config), the target data that comes from the target
3589 configuration in one of the `Configurations/*.conf` files (in
3594 * To clarify their intended purposes, the Configure options
3595 --prefix and --openssldir change their semantics, and become more
3596 straightforward and less interdependent.
3598 --prefix shall be used exclusively to give the location INSTALLTOP
3599 where programs, scripts, libraries, include files and manuals are
3600 going to be installed. The default is now /usr/local.
3602 --openssldir shall be used exclusively to give the default
3603 location OPENSSLDIR where certificates, private keys, CRLs are
3604 managed. This is also where the default openssl.cnf gets
3606 If the directory given with this option is a relative path, the
3607 values of both the --prefix value and the --openssldir value will
3608 be combined to become OPENSSLDIR.
3609 The default for --openssldir is INSTALLTOP/ssl.
3611 Anyone who uses --openssldir to specify where OpenSSL is to be
3612 installed MUST change to use --prefix instead.
3616 * The GOST engine was out of date and therefore it has been removed. An up
3617 to date GOST engine is now being maintained in an external repository.
3618 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3619 support for GOST ciphersuites (these are only activated if a GOST engine
3624 * EGD is no longer supported by default; use enable-egd when
3627 *Ben Kaduk and Rich Salz*
3629 * The distribution now has Makefile.in files, which are used to
3630 create Makefile's when Configure is run. *Configure must be run
3631 before trying to build now.*
3635 * The return value for SSL_CIPHER_description() for error conditions
3640 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3642 Obtaining and performing DNSSEC validation of TLSA records is
3643 the application's responsibility. The application provides
3644 the TLSA records of its choice to OpenSSL, and these are then
3645 used to authenticate the peer.
3647 The TLSA records need not even come from DNS. They can, for
3648 example, be used to implement local end-entity certificate or
3649 trust-anchor "pinning", where the "pin" data takes the form
3650 of TLSA records, which can augment or replace verification
3651 based on the usual WebPKI public certification authorities.
3655 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3656 continues to support deprecated interfaces in default builds.
3657 However, applications are strongly advised to compile their
3658 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3659 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3660 or the 1.1.0 releases.
3662 In environments in which all applications have been ported to
3663 not use any deprecated interfaces OpenSSL's Configure script
3664 should be used with the --api=1.1.0 option to entirely remove
3665 support for the deprecated features from the library and
3666 unconditionally disable them in the installed headers.
3667 Essentially the same effect can be achieved with the "no-deprecated"
3668 argument to Configure, except that this will always restrict
3669 the build to just the latest API, rather than a fixed API
3672 As applications are ported to future revisions of the API,
3673 they should update their compile-time OPENSSL_API_COMPAT define
3674 accordingly, but in most cases should be able to continue to
3675 compile with later releases.
3677 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3678 0x10000000L and 0x00908000L, respectively. However those
3679 versions did not support the OPENSSL_API_COMPAT feature, and
3680 so applications are not typically tested for explicit support
3681 of just the undeprecated features of either release.
3685 * Add support for setting the minimum and maximum supported protocol.
3686 It can bet set via the SSL_set_min_proto_version() and
3687 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3688 MaxProtocol. It's recommended to use the new APIs to disable
3689 protocols instead of disabling individual protocols using
3690 SSL_set_options() or SSL_CONF's Protocol. This change also
3691 removes support for disabling TLS 1.2 in the OpenSSL TLS
3692 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3696 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3700 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3701 and integrates ECDSA and ECDH functionality into EC. Implementations can
3702 now redirect key generation and no longer need to convert to or from
3705 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3706 include the ec.h header file instead.
3710 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3711 ciphers who are no longer supported and drops support the ephemeral RSA key
3712 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3716 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3717 opaque. For HMAC_CTX, the following constructors and destructors
3720 HMAC_CTX *HMAC_CTX_new(void);
3721 void HMAC_CTX_free(HMAC_CTX *ctx);
3723 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3724 destroy such methods has been added. See EVP_MD_meth_new(3) and
3725 EVP_CIPHER_meth_new(3) for documentation.
3728 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3729 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3730 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3731 an already created structure.
3732 2) For consistency with the majority of our object creators and
3733 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3734 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3735 for deprecated builds.
3739 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3740 cryptographic operations to be performed asynchronously as long as an
3741 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3742 further details. Libssl has also had this capability integrated with the
3743 introduction of the new mode SSL_MODE_ASYNC and associated error
3744 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3745 pages. This work was developed in partnership with Intel Corp.
3749 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3750 always enabled now. If you want to disable the support you should
3751 exclude it using the list of supported ciphers. This also means that the
3752 "-no_ecdhe" option has been removed from s_server.
3756 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3757 SSL_{CTX_}set1_curves() which can set a list.
3761 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3762 curve you want to support using SSL_{CTX_}set1_curves().
3766 * State machine rewrite. The state machine code has been significantly
3767 refactored in order to remove much duplication of code and solve issues
3768 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3769 further details). This change does have some associated API changes.
3770 Notably the SSL_state() function has been removed and replaced by
3771 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3772 SSL_set_state() has been removed altogether. The previous handshake states
3773 defined in ssl.h and ssl3.h have also been removed.
3777 * All instances of the string "ssleay" in the public API were replaced
3778 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3779 Some error codes related to internal RSA_eay API's were renamed.
3783 * The demo files in crypto/threads were moved to demo/threads.
3787 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3790 *Matt Caswell, Rich Salz*
3792 * New ASN.1 embed macro.
3794 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3795 structure is not allocated: it is part of the parent. That is instead of
3803 This reduces memory fragmentation and make it impossible to accidentally
3804 set a mandatory field to NULL.
3806 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3807 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3808 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3813 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3817 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3818 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3819 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3820 DES and RC4 ciphersuites.
3824 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3825 This changes the decoding behaviour for some invalid messages,
3826 though the change is mostly in the more lenient direction, and
3827 legacy behaviour is preserved as much as possible.
3831 * Fix no-stdio build.
3832 *David Woodhouse <David.Woodhouse@intel.com> and also*
3833 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3835 * New testing framework
3836 The testing framework has been largely rewritten and is now using
3837 perl and the perl modules Test::Harness and an extended variant of
3838 Test::More called OpenSSL::Test to do its work. All test scripts in
3839 test/ have been rewritten into test recipes, and all direct calls to
3840 executables in test/Makefile have become individual recipes using the
3841 simplified testing OpenSSL::Test::Simple.
3843 For documentation on our testing modules, do:
3845 perldoc test/testlib/OpenSSL/Test/Simple.pm
3846 perldoc test/testlib/OpenSSL/Test.pm
3850 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3851 are used; the latter aborts on memory leaks (usually checked on exit).
3852 Some undocumented "set malloc, etc., hooks" functions were removed
3853 and others were changed. All are now documented.
3857 * In DSA_generate_parameters_ex, if the provided seed is too short,
3860 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3862 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3863 from RFC4279, RFC4785, RFC5487, RFC5489.
3865 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3866 original RSA_PSK patch.
3870 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3871 era flag was never set throughout the codebase (only read). Also removed
3872 SSL3_FLAGS_POP_BUFFER which was only used if
3873 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3877 * Changed the default name options in the "ca", "crl", "req" and "x509"
3878 to be "oneline" instead of "compat".
3882 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3883 not aware of clients that still exhibit this bug, and the workaround
3884 hasn't been working properly for a while.
3888 * The return type of BIO_number_read() and BIO_number_written() as well as
3889 the corresponding num_read and num_write members in the BIO structure has
3890 changed from unsigned long to uint64_t. On platforms where an unsigned
3891 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3896 * Given the pervasive nature of TLS extensions it is inadvisable to run
3897 OpenSSL without support for them. It also means that maintaining
3898 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3899 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3903 * Removed support for the two export grade static DH ciphersuites
3904 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3905 were newly added (along with a number of other static DH ciphersuites) to
3906 1.0.2. However the two export ones have *never* worked since they were
3907 introduced. It seems strange in any case to be adding new export
3908 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3912 * Version negotiation has been rewritten. In particular SSLv23_method(),
3913 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3914 and turned into macros which simply call the new preferred function names
3915 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3916 should use the new names instead. Also as part of this change the ssl23.h
3917 header file has been removed.
3921 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3922 code and the associated standard is no longer considered fit-for-purpose.
3926 * RT2547 was closed. When generating a private key, try to make the
3927 output file readable only by the owner. This behavior change might
3928 be noticeable when interacting with other software.
3930 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3935 * Added HTTP GET support to the ocsp command.
3939 * Changed default digest for the dgst and enc commands from MD5 to
3944 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3948 * Added support for TLS extended master secret from
3949 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3950 initial patch which was a great help during development.
3954 * All libssl internal structures have been removed from the public header
3955 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3956 now redundant). Users should not attempt to access internal structures
3957 directly. Instead they should use the provided API functions.
3961 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3962 Access to deprecated functions can be re-enabled by running config with
3963 "enable-deprecated". In addition applications wishing to use deprecated
3964 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3965 will, by default, disable some transitive includes that previously existed
3966 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3970 * Added support for OCB mode. OpenSSL has been granted a patent license
3971 compatible with the OpenSSL license for use of OCB. Details are available
3972 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3973 for OCB can be removed by calling config with no-ocb.
3977 * SSLv2 support has been removed. It still supports receiving a SSLv2
3978 compatible client hello.
3982 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3983 done while fixing the error code for the key-too-small case.
3985 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3987 * CA.sh has been removed; use CA.pl instead.
3991 * Removed old DES API.
3995 * Remove various unsupported platforms:
4001 Sinix/ReliantUNIX RM400
4006 16-bit platforms such as WIN16
4010 * Clean up OPENSSL_NO_xxx #define's
4011 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4012 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4013 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4014 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4015 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4016 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4017 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4018 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4019 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4020 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4024 * Cleaned up dead code
4025 Remove all but one '#ifdef undef' which is to be looked at.
4029 * Clean up calling of xxx_free routines.
4030 Just like free(), fix most of the xxx_free routines to accept
4031 NULL. Remove the non-null checks from callers. Save much code.
4035 * Add secure heap for storage of private keys (when possible).
4036 Add BIO_s_secmem(), CBIGNUM, etc.
4037 Contributed by Akamai Technologies under our Corporate CLA.
4041 * Experimental support for a new, fast, unbiased prime candidate generator,
4042 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4044 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4046 * New output format NSS in the sess_id command line tool. This allows
4047 exporting the session id and the master key in NSS keylog format.
4049 *Martin Kaiser <martin@kaiser.cx>*
4051 * Harmonize version and its documentation. -f flag is used to display
4054 *mancha <mancha1@zoho.com>*
4056 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4057 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4059 *mancha <mancha1@zoho.com>*
4061 * Fix some double frees. These are not thought to be exploitable.
4063 *mancha <mancha1@zoho.com>*
4065 * A missing bounds check in the handling of the TLS heartbeat extension
4066 can be used to reveal up to 64k of memory to a connected client or
4069 Thanks for Neel Mehta of Google Security for discovering this bug and to
4070 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4071 preparing the fix ([CVE-2014-0160])
4073 *Adam Langley, Bodo Moeller*
4075 * Fix for the attack described in the paper "Recovering OpenSSL
4076 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4077 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4078 <http://eprint.iacr.org/2014/140>
4080 Thanks to Yuval Yarom and Naomi Benger for discovering this
4081 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4083 *Yuval Yarom and Naomi Benger*
4085 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4086 this fixes a limitation in previous versions of OpenSSL.
4090 * Experimental encrypt-then-mac support.
4092 Experimental support for encrypt then mac from
4093 draft-gutmann-tls-encrypt-then-mac-02.txt
4095 To enable it set the appropriate extension number (0x42 for the test
4096 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4098 For non-compliant peers (i.e. just about everything) this should have no
4101 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4105 * Add EVP support for key wrapping algorithms, to avoid problems with
4106 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4107 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4108 algorithms and include tests cases.
4112 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4117 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4118 MGF1 digest and OAEP label.
4122 * Make openssl verify return errors.
4124 *Chris Palmer <palmer@google.com> and Ben Laurie*
4126 * New function ASN1_TIME_diff to calculate the difference between two
4127 ASN1_TIME structures or one structure and the current time.
4131 * Update fips_test_suite to support multiple command line options. New
4132 test to induce all self test errors in sequence and check expected
4137 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4138 sign or verify all in one operation.
4142 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4143 test programs and fips_test_suite. Includes functionality to parse
4144 the minimal script output of fipsalgest.pl directly.
4148 * Add authorisation parameter to FIPS_module_mode_set().
4152 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4156 * Use separate DRBG fields for internal and external flags. New function
4157 FIPS_drbg_health_check() to perform on demand health checking. Add
4158 generation tests to fips_test_suite with reduced health check interval to
4159 demonstrate periodic health checking. Add "nodh" option to
4160 fips_test_suite to skip very slow DH test.
4164 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4169 * More extensive health check for DRBG checking many more failure modes.
4170 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4171 combination: call this in fips_test_suite.
4175 * Add support for canonical generation of DSA parameter 'g'. See
4178 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4179 POST to handle HMAC cases.
4183 * Add functions FIPS_module_version() and FIPS_module_version_text()
4184 to return numerical and string versions of the FIPS module number.
4188 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4189 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4190 outside the validated module in the FIPS capable OpenSSL.
4194 * Minor change to DRBG entropy callback semantics. In some cases
4195 there is no multiple of the block length between min_len and
4196 max_len. Allow the callback to return more than max_len bytes
4197 of entropy but discard any extra: it is the callback's responsibility
4198 to ensure that the extra data discarded does not impact the
4199 requested amount of entropy.
4203 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4204 information in FIPS186-3, SP800-57 and SP800-131A.
4208 * CCM support via EVP. Interface is very similar to GCM case except we
4209 must supply all data in one chunk (i.e. no update, final) and the
4210 message length must be supplied if AAD is used. Add algorithm test
4215 * Initial version of POST overhaul. Add POST callback to allow the status
4216 of POST to be monitored and/or failures induced. Modify fips_test_suite
4217 to use callback. Always run all selftests even if one fails.
4221 * XTS support including algorithm test driver in the fips_gcmtest program.
4222 Note: this does increase the maximum key length from 32 to 64 bytes but
4223 there should be no binary compatibility issues as existing applications
4224 will never use XTS mode.
4228 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4229 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4230 performs algorithm blocking for unapproved PRNG types. Also do not
4231 set PRNG type in FIPS_mode_set(): leave this to the application.
4232 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4233 the standard OpenSSL PRNG: set additional data to a date time vector.
4237 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4238 This shouldn't present any incompatibility problems because applications
4239 shouldn't be using these directly and any that are will need to rethink
4240 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4244 * Extensive self tests and health checking required by SP800-90 DRBG.
4245 Remove strength parameter from FIPS_drbg_instantiate and always
4246 instantiate at maximum supported strength.
4250 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4254 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4258 * New function DH_compute_key_padded() to compute a DH key and pad with
4259 leading zeroes if needed: this complies with SP800-56A et al.
4263 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4264 anything, incomplete, subject to change and largely untested at present.
4268 * Modify fipscanisteronly build option to only build the necessary object
4269 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4273 * Add experimental option FIPSSYMS to give all symbols in
4274 fipscanister.o and FIPS or fips prefix. This will avoid
4275 conflicts with future versions of OpenSSL. Add perl script
4276 util/fipsas.pl to preprocess assembly language source files
4277 and rename any affected symbols.
4281 * Add selftest checks and algorithm block of non-fips algorithms in
4282 FIPS mode. Remove DES2 from selftests.
4286 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4287 return internal method without any ENGINE dependencies. Add new
4288 tiny fips sign and verify functions.
4292 * New build option no-ec2m to disable characteristic 2 code.
4296 * New build option "fipscanisteronly". This only builds fipscanister.o
4297 and (currently) associated fips utilities. Uses the file Makefile.fips
4298 instead of Makefile.org as the prototype.
4302 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4303 Update fips_gcmtest to use IV generator.
4307 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4308 setting output buffer to NULL. The `*Final` function must be
4309 called although it will not retrieve any additional data. The tag
4310 can be set or retrieved with a ctrl. The IV length is by default 12
4311 bytes (96 bits) but can be set to an alternative value. If the IV
4312 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4317 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4318 underlying do_cipher function handles all cipher semantics itself
4319 including padding and finalisation. This is useful if (for example)
4320 an ENGINE cipher handles block padding itself. The behaviour of
4321 do_cipher is subtly changed if this flag is set: the return value
4322 is the number of characters written to the output buffer (zero is
4323 no longer an error code) or a negative error code. Also if the
4324 input buffer is NULL and length 0 finalisation should be performed.
4328 * If a candidate issuer certificate is already part of the constructed
4329 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4333 * Improve forward-security support: add functions
4335 void SSL_CTX_set_not_resumable_session_callback(
4336 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4337 void SSL_set_not_resumable_session_callback(
4338 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4340 for use by SSL/TLS servers; the callback function will be called whenever a
4341 new session is created, and gets to decide whether the session may be
4342 cached to make it resumable (return 0) or not (return 1). (As by the
4343 SSL/TLS protocol specifications, the session_id sent by the server will be
4344 empty to indicate that the session is not resumable; also, the server will
4345 not generate RFC 4507 (RFC 5077) session tickets.)
4347 A simple reasonable callback implementation is to return is_forward_secure.
4348 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4349 by the SSL/TLS server library, indicating whether it can provide forward
4352 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4354 * New -verify_name option in command line utilities to set verification
4359 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4360 Add CMAC pkey methods.
4364 * Experimental renegotiation in s_server -www mode. If the client
4365 browses /reneg connection is renegotiated. If /renegcert it is
4366 renegotiated requesting a certificate.
4370 * Add an "external" session cache for debugging purposes to s_server. This
4371 should help trace issues which normally are only apparent in deployed
4372 multi-process servers.
4376 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4377 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4378 BIO_set_cipher() and some obscure PEM functions were changed so they
4379 can now return an error. The RAND changes required a change to the
4380 RAND_METHOD structure.
4384 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4385 a gcc attribute to warn if the result of a function is ignored. This
4386 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4387 whose return value is often ignored.
4391 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4392 These allow SCTs (signed certificate timestamps) to be requested and
4393 validated when establishing a connection.
4395 *Rob Percival <robpercival@google.com>*
4400 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4402 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4403 used even when parsing explicit parameters, when loading a encoded key
4404 or calling `EC_GROUP_new_from_ecpkparameters()`/
4405 `EC_GROUP_new_from_ecparameters()`.
4406 This prevents bypass of security hardening and performance gains,
4407 especially for curves with specialized EC_METHODs.
4408 By default, if a key encoded with explicit parameters is loaded and later
4409 encoded, the output is still encoded with explicit parameters, even if
4410 internally a "named" EC_GROUP is used for computation.
4414 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4415 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4416 NULL. After this change, only the cofactor parameter can be NULL. It also
4417 does some minimal sanity checks on the passed order.
4422 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4423 An attack is simple, if the first CMS_recipientInfo is valid but the
4424 second CMS_recipientInfo is chosen ciphertext. If the second
4425 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4426 encryption key will be replaced by garbage, and the message cannot be
4427 decoded, but if the RSA decryption fails, the correct encryption key is
4428 used and the recipient will not notice the attack.
4429 As a work around for this potential attack the length of the decrypted
4430 key must be equal to the cipher default key length, in case the
4431 certifiate is not given and all recipientInfo are tried out.
4432 The old behaviour can be re-enabled in the CMS code by setting the
4433 CMS_DEBUG_DECRYPT flag.
4438 * Document issue with installation paths in diverse Windows builds
4440 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4441 binaries and run-time config file.
4446 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4448 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4449 This changes the size when using the `genpkey` command when no size is given.
4450 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4451 generation commands to use 2048 bits by default.
4455 * Add FIPS support for Android Arm 64-bit
4457 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4458 Module in Version 2.0.10. For some reason, the corresponding target
4459 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4460 built with FIPS support on Android Arm 64-bit. This omission has been
4463 *Matthias St. Pierre*
4465 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4467 * 0-byte record padding oracle
4469 If an application encounters a fatal protocol error and then calls
4470 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4471 then OpenSSL can respond differently to the calling application if a 0 byte
4472 record is received with invalid padding compared to if a 0 byte record is
4473 received with an invalid MAC. If the application then behaves differently
4474 based on that in a way that is detectable to the remote peer, then this
4475 amounts to a padding oracle that could be used to decrypt data.
4477 In order for this to be exploitable "non-stitched" ciphersuites must be in
4478 use. Stitched ciphersuites are optimised implementations of certain
4479 commonly used ciphersuites. Also the application must call SSL_shutdown()
4480 twice even if a protocol error has occurred (applications should not do
4481 this but some do anyway).
4483 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4484 Aviram, with additional investigation by Steven Collison and Andrew
4485 Hourselt. It was reported to OpenSSL on 10th December 2018.
4490 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4494 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4496 * Microarchitecture timing vulnerability in ECC scalar multiplication
4498 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4499 shown to be vulnerable to a microarchitecture timing side channel attack.
4500 An attacker with sufficient access to mount local timing attacks during
4501 ECDSA signature generation could recover the private key.
4503 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4504 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4510 * Timing vulnerability in DSA signature generation
4512 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4513 timing side channel attack. An attacker could use variations in the signing
4514 algorithm to recover the private key.
4516 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4521 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4522 Module, accidentally introduced while backporting security fixes from the
4523 development branch and hindering the use of ECC in FIPS mode.
4527 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4529 * Client DoS due to large DH parameter
4531 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4532 malicious server can send a very large prime value to the client. This will
4533 cause the client to spend an unreasonably long period of time generating a
4534 key for this prime resulting in a hang until the client has finished. This
4535 could be exploited in a Denial Of Service attack.
4537 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4542 * Cache timing vulnerability in RSA Key Generation
4544 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4545 a cache timing side channel attack. An attacker with sufficient access to
4546 mount cache timing attacks during the RSA key generation process could
4547 recover the private key.
4549 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4550 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4555 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4556 parameter is no longer accepted, as it leads to a corrupt table. NULL
4557 pem_str is reserved for alias entries only.
4561 * Revert blinding in ECDSA sign and instead make problematic addition
4562 length-invariant. Switch even to fixed-length Montgomery multiplication.
4566 * Change generating and checking of primes so that the error rate of not
4567 being prime depends on the intended use based on the size of the input.
4568 For larger primes this will result in more rounds of Miller-Rabin.
4569 The maximal error rate for primes with more than 1080 bits is lowered
4572 *Kurt Roeckx, Annie Yousar*
4574 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4578 * Add blinding to ECDSA and DSA signatures to protect against side channel
4579 attacks discovered by Keegan Ryan (NCC Group).
4583 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4584 now allow empty (zero character) pass phrases.
4588 * Certificate time validation (X509_cmp_time) enforces stricter
4589 compliance with RFC 5280. Fractional seconds and timezone offsets
4590 are no longer allowed.
4594 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4596 * Constructed ASN.1 types with a recursive definition could exceed the stack
4598 Constructed ASN.1 types with a recursive definition (such as can be found
4599 in PKCS7) could eventually exceed the stack given malicious input with
4600 excessive recursion. This could result in a Denial Of Service attack. There
4601 are no such structures used within SSL/TLS that come from untrusted sources
4602 so this is considered safe.
4604 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4610 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4612 * Read/write after SSL object in error state
4614 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4615 mechanism. The intent was that if a fatal error occurred during a handshake
4616 then OpenSSL would move into the error state and would immediately fail if
4617 you attempted to continue the handshake. This works as designed for the
4618 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4619 SSL_connect()), however due to a bug it does not work correctly if
4620 SSL_read() or SSL_write() is called directly. In that scenario, if the
4621 handshake fails then a fatal error will be returned in the initial function
4622 call. If SSL_read()/SSL_write() is subsequently called by the application
4623 for the same SSL object then it will succeed and the data is passed without
4624 being decrypted/encrypted directly from the SSL/TLS record layer.
4626 In order to exploit this issue an application bug would have to be present
4627 that resulted in a call to SSL_read()/SSL_write() being issued after having
4628 already received a fatal error.
4630 This issue was reported to OpenSSL by David Benjamin (Google).
4635 * rsaz_1024_mul_avx2 overflow bug on x86_64
4637 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4638 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4639 Analysis suggests that attacks against RSA and DSA as a result of this
4640 defect would be very difficult to perform and are not believed likely.
4641 Attacks against DH1024 are considered just feasible, because most of the
4642 work necessary to deduce information about a private key may be performed
4643 offline. The amount of resources required for such an attack would be
4644 significant. However, for an attack on TLS to be meaningful, the server
4645 would have to share the DH1024 private key among multiple clients, which is
4646 no longer an option since CVE-2016-0701.
4648 This only affects processors that support the AVX2 but not ADX extensions
4649 like Intel Haswell (4th generation).
4651 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4652 was originally found via the OSS-Fuzz project.
4657 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4659 * bn_sqrx8x_internal carry bug on x86_64
4661 There is a carry propagating bug in the x86_64 Montgomery squaring
4662 procedure. No EC algorithms are affected. Analysis suggests that attacks
4663 against RSA and DSA as a result of this defect would be very difficult to
4664 perform and are not believed likely. Attacks against DH are considered just
4665 feasible (although very difficult) because most of the work necessary to
4666 deduce information about a private key may be performed offline. The amount
4667 of resources required for such an attack would be very significant and
4668 likely only accessible to a limited number of attackers. An attacker would
4669 additionally need online access to an unpatched system using the target
4670 private key in a scenario with persistent DH parameters and a private
4671 key that is shared between multiple clients.
4673 This only affects processors that support the BMI1, BMI2 and ADX extensions
4674 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4676 This issue was reported to OpenSSL by the OSS-Fuzz project.
4681 * Malformed X.509 IPAddressFamily could cause OOB read
4683 If an X.509 certificate has a malformed IPAddressFamily extension,
4684 OpenSSL could do a one-byte buffer overread. The most likely result
4685 would be an erroneous display of the certificate in text format.
4687 This issue was reported to OpenSSL by the OSS-Fuzz project.
4691 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4693 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4694 platform rather than 'mingw'.
4698 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4700 * Truncated packet could crash via OOB read
4702 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4703 cipher is being used, then a truncated packet can cause that host to
4704 perform an out-of-bounds read, usually resulting in a crash.
4706 This issue was reported to OpenSSL by Robert Święcki of Google.
4711 * BN_mod_exp may produce incorrect results on x86_64
4713 There is a carry propagating bug in the x86_64 Montgomery squaring
4714 procedure. No EC algorithms are affected. Analysis suggests that attacks
4715 against RSA and DSA as a result of this defect would be very difficult to
4716 perform and are not believed likely. Attacks against DH are considered just
4717 feasible (although very difficult) because most of the work necessary to
4718 deduce information about a private key may be performed offline. The amount
4719 of resources required for such an attack would be very significant and
4720 likely only accessible to a limited number of attackers. An attacker would
4721 additionally need online access to an unpatched system using the target
4722 private key in a scenario with persistent DH parameters and a private
4723 key that is shared between multiple clients. For example this can occur by
4724 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4725 similar to CVE-2015-3193 but must be treated as a separate problem.
4727 This issue was reported to OpenSSL by the OSS-Fuzz project.
4732 * Montgomery multiplication may produce incorrect results
4734 There is a carry propagating bug in the Broadwell-specific Montgomery
4735 multiplication procedure that handles input lengths divisible by, but
4736 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4737 and DH private keys are impossible. This is because the subroutine in
4738 question is not used in operations with the private key itself and an input
4739 of the attacker's direct choice. Otherwise the bug can manifest itself as
4740 transient authentication and key negotiation failures or reproducible
4741 erroneous outcome of public-key operations with specially crafted input.
4742 Among EC algorithms only Brainpool P-512 curves are affected and one
4743 presumably can attack ECDH key negotiation. Impact was not analyzed in
4744 detail, because pre-requisites for attack are considered unlikely. Namely
4745 multiple clients have to choose the curve in question and the server has to
4746 share the private key among them, neither of which is default behaviour.
4747 Even then only clients that chose the curve will be affected.
4749 This issue was publicly reported as transient failures and was not
4750 initially recognized as a security issue. Thanks to Richard Morgan for
4751 providing reproducible case.
4756 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4757 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4758 prevent issues where no progress is being made and the peer continually
4759 sends unrecognised record types, using up resources processing them.
4763 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4765 * Missing CRL sanity check
4767 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4768 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4769 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4771 This issue only affects the OpenSSL 1.0.2i
4776 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4778 * OCSP Status Request extension unbounded memory growth
4780 A malicious client can send an excessively large OCSP Status Request
4781 extension. If that client continually requests renegotiation, sending a
4782 large OCSP Status Request extension each time, then there will be unbounded
4783 memory growth on the server. This will eventually lead to a Denial Of
4784 Service attack through memory exhaustion. Servers with a default
4785 configuration are vulnerable even if they do not support OCSP. Builds using
4786 the "no-ocsp" build time option are not affected.
4788 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4793 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4796 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4802 * OOB write in MDC2_Update()
4804 An overflow can occur in MDC2_Update() either if called directly or
4805 through the EVP_DigestUpdate() function using MDC2. If an attacker
4806 is able to supply very large amounts of input data after a previous
4807 call to EVP_EncryptUpdate() with a partial block then a length check
4808 can overflow resulting in a heap corruption.
4810 The amount of data needed is comparable to SIZE_MAX which is impractical
4813 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4818 * Malformed SHA512 ticket DoS
4820 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4821 DoS attack where a malformed ticket will result in an OOB read which will
4824 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4825 a custom server callback and ticket lookup mechanism.
4827 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4832 * OOB write in BN_bn2dec()
4834 The function BN_bn2dec() does not check the return value of BN_div_word().
4835 This can cause an OOB write if an application uses this function with an
4836 overly large BIGNUM. This could be a problem if an overly large certificate
4837 or CRL is printed out from an untrusted source. TLS is not affected because
4838 record limits will reject an oversized certificate before it is parsed.
4840 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4845 * OOB read in TS_OBJ_print_bio()
4847 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4848 the total length the OID text representation would use and not the amount
4849 of data written. This will result in OOB reads when large OIDs are
4852 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4857 * Pointer arithmetic undefined behaviour
4859 Avoid some undefined pointer arithmetic
4861 A common idiom in the codebase is to check limits in the following manner:
4864 Where "p" points to some malloc'd data of SIZE bytes and
4867 "len" here could be from some externally supplied data (e.g. from a TLS
4870 The rules of C pointer arithmetic are such that "p + len" is only well
4871 defined where len <= SIZE. Therefore the above idiom is actually
4872 undefined behaviour.
4874 For example this could cause problems if some malloc implementation
4875 provides an address for "p" such that "p + len" actually overflows for
4876 values of len that are too big and therefore p + len < limit.
4878 This issue was reported to OpenSSL by Guido Vranken
4883 * Constant time flag not preserved in DSA signing
4885 Operations in the DSA signing algorithm should run in constant time in
4886 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4887 implementation means that a non-constant time codepath is followed for
4888 certain operations. This has been demonstrated through a cache-timing
4889 attack to be sufficient for an attacker to recover the private DSA key.
4891 This issue was reported by César Pereida (Aalto University), Billy Brumley
4892 (Tampere University of Technology), and Yuval Yarom (The University of
4893 Adelaide and NICTA).
4898 * DTLS buffered message DoS
4900 In a DTLS connection where handshake messages are delivered out-of-order
4901 those messages that OpenSSL is not yet ready to process will be buffered
4902 for later use. Under certain circumstances, a flaw in the logic means that
4903 those messages do not get removed from the buffer even though the handshake
4904 has been completed. An attacker could force up to approx. 15 messages to
4905 remain in the buffer when they are no longer required. These messages will
4906 be cleared when the DTLS connection is closed. The default maximum size for
4907 a message is 100k. Therefore the attacker could force an additional 1500k
4908 to be consumed per connection. By opening many simulataneous connections an
4909 attacker could cause a DoS attack through memory exhaustion.
4911 This issue was reported to OpenSSL by Quan Luo.
4916 * DTLS replay protection DoS
4918 A flaw in the DTLS replay attack protection mechanism means that records
4919 that arrive for future epochs update the replay protection "window" before
4920 the MAC for the record has been validated. This could be exploited by an
4921 attacker by sending a record for the next epoch (which does not have to
4922 decrypt or have a valid MAC), with a very large sequence number. This means
4923 that all subsequent legitimate packets are dropped causing a denial of
4924 service for a specific DTLS connection.
4926 This issue was reported to OpenSSL by the OCAP audit team.
4931 * Certificate message OOB reads
4933 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4934 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4935 theoretical DoS risk but this has not been observed in practice on common
4938 The messages affected are client certificate, client certificate request
4939 and server certificate. As a result the attack can only be performed
4940 against a client or a server which enables client authentication.
4942 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4947 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4949 * Prevent padding oracle in AES-NI CBC MAC check
4951 A MITM attacker can use a padding oracle attack to decrypt traffic
4952 when the connection uses an AES CBC cipher and the server support
4955 This issue was introduced as part of the fix for Lucky 13 padding
4956 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4957 constant time by making sure that always the same bytes are read and
4958 compared against either the MAC or padding bytes. But it no longer
4959 checked that there was enough data to have both the MAC and padding
4962 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4966 * Fix EVP_EncodeUpdate overflow
4968 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4969 Base64 encoding of binary data. If an attacker is able to supply very large
4970 amounts of input data then a length check can overflow resulting in a heap
4973 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4974 the `PEM_write_bio*` family of functions. These are mainly used within the
4975 OpenSSL command line applications, so any application which processes data
4976 from an untrusted source and outputs it as a PEM file should be considered
4977 vulnerable to this issue. User applications that call these APIs directly
4978 with large amounts of untrusted data may also be vulnerable.
4980 This issue was reported by Guido Vranken.
4985 * Fix EVP_EncryptUpdate overflow
4987 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4988 is able to supply very large amounts of input data after a previous call to
4989 EVP_EncryptUpdate() with a partial block then a length check can overflow
4990 resulting in a heap corruption. Following an analysis of all OpenSSL
4991 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4992 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4993 the first called function after an EVP_EncryptInit(), and therefore that
4994 specific call must be safe. The second form is where the length passed to
4995 EVP_EncryptUpdate() can be seen from the code to be some small value and
4996 therefore there is no possibility of an overflow. Since all instances are
4997 one of these two forms, it is believed that there can be no overflows in
4998 internal code due to this problem. It should be noted that
4999 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5000 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5001 of these calls have also been analysed too and it is believed there are no
5002 instances in internal usage where an overflow could occur.
5004 This issue was reported by Guido Vranken.
5009 * Prevent ASN.1 BIO excessive memory allocation
5011 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5012 a short invalid encoding can cause allocation of large amounts of memory
5013 potentially consuming excessive resources or exhausting memory.
5015 Any application parsing untrusted data through d2i BIO functions is
5016 affected. The memory based functions such as d2i_X509() are *not* affected.
5017 Since the memory based functions are used by the TLS library, TLS
5018 applications are not affected.
5020 This issue was reported by Brian Carpenter.
5027 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5028 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5029 in arbitrary stack data being returned in the buffer.
5031 This issue was reported by Guido Vranken.
5036 * Modify behavior of ALPN to invoke callback after SNI/servername
5037 callback, such that updates to the SSL_CTX affect ALPN.
5041 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5046 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5047 methods are enabled and ssl2 is disabled the methods return NULL.
5051 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5053 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5054 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5055 provide any "EXPORT" or "LOW" strength ciphers.
5059 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5060 is by default disabled at build-time. Builds that are not configured with
5061 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5062 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5063 will need to explicitly call either of:
5065 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5067 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5069 as appropriate. Even if either of those is used, or the application
5070 explicitly uses the version-specific SSLv2_method() or its client and
5071 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5072 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5073 ciphers, and SSLv2 56-bit DES are no longer available.
5078 * Fix a double-free in DSA code
5080 A double free bug was discovered when OpenSSL parses malformed DSA private
5081 keys and could lead to a DoS attack or memory corruption for applications
5082 that receive DSA private keys from untrusted sources. This scenario is
5085 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5091 * Disable SRP fake user seed to address a server memory leak.
5093 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5095 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5096 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5097 was changed to ignore the "fake user" SRP seed, even if the seed
5100 Users should use SRP_VBASE_get1_by_user instead. Note that in
5101 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5102 also that even though configuring the SRP seed attempts to hide
5103 invalid usernames by continuing the handshake with fake
5104 credentials, this behaviour is not constant time and no strong
5105 guarantees are made that the handshake is indistinguishable from
5106 that of a valid user.
5111 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5113 In the BN_hex2bn function the number of hex digits is calculated using an
5114 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5115 large values of `i` this can result in `bn_expand` not allocating any
5116 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5117 field as NULL leading to a subsequent NULL ptr deref. For very large values
5118 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5119 In this case memory is allocated to the internal BIGNUM data field, but it
5120 is insufficiently sized leading to heap corruption. A similar issue exists
5121 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5122 is ever called by user applications with very large untrusted hex/dec data.
5123 This is anticipated to be a rare occurrence.
5125 All OpenSSL internal usage of these functions use data that is not expected
5126 to be untrusted, e.g. config file data or application command line
5127 arguments. If user developed applications generate config file data based
5128 on untrusted data then it is possible that this could also lead to security
5129 consequences. This is also anticipated to be rare.
5131 This issue was reported to OpenSSL by Guido Vranken.
5136 * Fix memory issues in `BIO_*printf` functions
5138 The internal `fmtstr` function used in processing a "%s" format string in
5139 the `BIO_*printf` functions could overflow while calculating the length of a
5140 string and cause an OOB read when printing very long strings.
5142 Additionally the internal `doapr_outch` function can attempt to write to an
5143 OOB memory location (at an offset from the NULL pointer) in the event of a
5144 memory allocation failure. In 1.0.2 and below this could be caused where
5145 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5146 could be in processing a very long "%s" format string. Memory leaks can
5149 The first issue may mask the second issue dependent on compiler behaviour.
5150 These problems could enable attacks where large amounts of untrusted data
5151 is passed to the `BIO_*printf` functions. If applications use these functions
5152 in this way then they could be vulnerable. OpenSSL itself uses these
5153 functions when printing out human-readable dumps of ASN.1 data. Therefore
5154 applications that print this data could be vulnerable if the data is from
5155 untrusted sources. OpenSSL command line applications could also be
5156 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5157 as command line arguments.
5159 Libssl is not considered directly vulnerable. Additionally certificates etc
5160 received via remote connections via libssl are also unlikely to be able to
5161 trigger these issues because of message size limits enforced within libssl.
5163 This issue was reported to OpenSSL Guido Vranken.
5168 * Side channel attack on modular exponentiation
5170 A side-channel attack was found which makes use of cache-bank conflicts on
5171 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5172 of RSA keys. The ability to exploit this issue is limited as it relies on
5173 an attacker who has control of code in a thread running on the same
5174 hyper-threaded core as the victim thread which is performing decryptions.
5176 This issue was reported to OpenSSL by Yuval Yarom, The University of
5177 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5178 Nadia Heninger, University of Pennsylvania with more information at
5179 <http://cachebleed.info>.
5184 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5185 if no keysize is specified with default_bits. This fixes an
5186 omission in an earlier change that changed all RSA/DSA key generation
5187 commands to use 2048 bits by default.
5191 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5193 * DH small subgroups
5195 Historically OpenSSL only ever generated DH parameters based on "safe"
5196 primes. More recently (in version 1.0.2) support was provided for
5197 generating X9.42 style parameter files such as those required for RFC 5114
5198 support. The primes used in such files may not be "safe". Where an
5199 application is using DH configured with parameters based on primes that are
5200 not "safe" then an attacker could use this fact to find a peer's private
5201 DH exponent. This attack requires that the attacker complete multiple
5202 handshakes in which the peer uses the same private DH exponent. For example
5203 this could be used to discover a TLS server's private DH exponent if it's
5204 reusing the private DH exponent or it's using a static DH ciphersuite.
5206 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5207 TLS. It is not on by default. If the option is not set then the server
5208 reuses the same private DH exponent for the life of the server process and
5209 would be vulnerable to this attack. It is believed that many popular
5210 applications do set this option and would therefore not be at risk.
5212 The fix for this issue adds an additional check where a "q" parameter is
5213 available (as is the case in X9.42 based parameters). This detects the
5214 only known attack, and is the only possible defense for static DH
5215 ciphersuites. This could have some performance impact.
5217 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5218 default and cannot be disabled. This could have some performance impact.
5220 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5225 * SSLv2 doesn't block disabled ciphers
5227 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5228 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5229 been disabled, provided that the SSLv2 protocol was not also disabled via
5232 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5233 and Sebastian Schinzel.
5238 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5240 * BN_mod_exp may produce incorrect results on x86_64
5242 There is a carry propagating bug in the x86_64 Montgomery squaring
5243 procedure. No EC algorithms are affected. Analysis suggests that attacks
5244 against RSA and DSA as a result of this defect would be very difficult to
5245 perform and are not believed likely. Attacks against DH are considered just
5246 feasible (although very difficult) because most of the work necessary to
5247 deduce information about a private key may be performed offline. The amount
5248 of resources required for such an attack would be very significant and
5249 likely only accessible to a limited number of attackers. An attacker would
5250 additionally need online access to an unpatched system using the target
5251 private key in a scenario with persistent DH parameters and a private
5252 key that is shared between multiple clients. For example this can occur by
5253 default in OpenSSL DHE based SSL/TLS ciphersuites.
5255 This issue was reported to OpenSSL by Hanno Böck.
5260 * Certificate verify crash with missing PSS parameter
5262 The signature verification routines will crash with a NULL pointer
5263 dereference if presented with an ASN.1 signature using the RSA PSS
5264 algorithm and absent mask generation function parameter. Since these
5265 routines are used to verify certificate signature algorithms this can be
5266 used to crash any certificate verification operation and exploited in a
5267 DoS attack. Any application which performs certificate verification is
5268 vulnerable including OpenSSL clients and servers which enable client
5271 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5276 * X509_ATTRIBUTE memory leak
5278 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5279 memory. This structure is used by the PKCS#7 and CMS routines so any
5280 application which reads PKCS#7 or CMS data from untrusted sources is
5281 affected. SSL/TLS is not affected.
5283 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5289 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5290 This changes the decoding behaviour for some invalid messages,
5291 though the change is mostly in the more lenient direction, and
5292 legacy behaviour is preserved as much as possible.
5296 * In DSA_generate_parameters_ex, if the provided seed is too short,
5299 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5301 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5303 * Alternate chains certificate forgery
5305 During certificate verification, OpenSSL will attempt to find an
5306 alternative certificate chain if the first attempt to build such a chain
5307 fails. An error in the implementation of this logic can mean that an
5308 attacker could cause certain checks on untrusted certificates to be
5309 bypassed, such as the CA flag, enabling them to use a valid leaf
5310 certificate to act as a CA and "issue" an invalid certificate.
5312 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5317 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5319 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5320 incompatibility in the handling of HMAC. The previous ABI has now been
5325 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5327 * Malformed ECParameters causes infinite loop
5329 When processing an ECParameters structure OpenSSL enters an infinite loop
5330 if the curve specified is over a specially malformed binary polynomial
5333 This can be used to perform denial of service against any
5334 system which processes public keys, certificate requests or
5335 certificates. This includes TLS clients and TLS servers with
5336 client authentication enabled.
5338 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5343 * Exploitable out-of-bounds read in X509_cmp_time
5345 X509_cmp_time does not properly check the length of the ASN1_TIME
5346 string and can read a few bytes out of bounds. In addition,
5347 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5350 An attacker can use this to craft malformed certificates and CRLs of
5351 various sizes and potentially cause a segmentation fault, resulting in
5352 a DoS on applications that verify certificates or CRLs. TLS clients
5353 that verify CRLs are affected. TLS clients and servers with client
5354 authentication enabled may be affected if they use custom verification
5357 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5358 independently by Hanno Böck.
5363 * PKCS7 crash with missing EnvelopedContent
5365 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5366 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5367 with missing content and trigger a NULL pointer dereference on parsing.
5369 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5370 structures from untrusted sources are affected. OpenSSL clients and
5371 servers are not affected.
5373 This issue was reported to OpenSSL by Michal Zalewski (Google).
5378 * CMS verify infinite loop with unknown hash function
5380 When verifying a signedData message the CMS code can enter an infinite loop
5381 if presented with an unknown hash function OID. This can be used to perform
5382 denial of service against any system which verifies signedData messages using
5384 This issue was reported to OpenSSL by Johannes Bauer.
5389 * Race condition handling NewSessionTicket
5391 If a NewSessionTicket is received by a multi-threaded client when attempting to
5392 reuse a previous ticket then a race condition can occur potentially leading to
5393 a double free of the ticket data.
5398 * Only support 256-bit or stronger elliptic curves with the
5399 'ecdh_auto' setting (server) or by default (client). Of supported
5400 curves, prefer P-256 (both).
5404 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5406 * ClientHello sigalgs DoS fix
5408 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5409 invalid signature algorithms extension a NULL pointer dereference will
5410 occur. This can be exploited in a DoS attack against the server.
5412 This issue was was reported to OpenSSL by David Ramos of Stanford
5416 *Stephen Henson and Matt Caswell*
5418 * Multiblock corrupted pointer fix
5420 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5421 feature only applies on 64 bit x86 architecture platforms that support AES
5422 NI instructions. A defect in the implementation of "multiblock" can cause
5423 OpenSSL's internal write buffer to become incorrectly set to NULL when
5424 using non-blocking IO. Typically, when the user application is using a
5425 socket BIO for writing, this will only result in a failed connection.
5426 However if some other BIO is used then it is likely that a segmentation
5427 fault will be triggered, thus enabling a potential DoS attack.
5429 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5434 * Segmentation fault in DTLSv1_listen fix
5436 The DTLSv1_listen function is intended to be stateless and processes the
5437 initial ClientHello from many peers. It is common for user code to loop
5438 over the call to DTLSv1_listen until a valid ClientHello is received with
5439 an associated cookie. A defect in the implementation of DTLSv1_listen means
5440 that state is preserved in the SSL object from one invocation to the next
5441 that can lead to a segmentation fault. Errors processing the initial
5442 ClientHello can trigger this scenario. An example of such an error could be
5443 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5446 This issue was reported to OpenSSL by Per Allansson.
5451 * Segmentation fault in ASN1_TYPE_cmp fix
5453 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5454 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5455 certificate signature algorithm consistency this can be used to crash any
5456 certificate verification operation and exploited in a DoS attack. Any
5457 application which performs certificate verification is vulnerable including
5458 OpenSSL clients and servers which enable client authentication.
5463 * Segmentation fault for invalid PSS parameters fix
5465 The signature verification routines will crash with a NULL pointer
5466 dereference if presented with an ASN.1 signature using the RSA PSS
5467 algorithm and invalid parameters. Since these routines are used to verify
5468 certificate signature algorithms this can be used to crash any
5469 certificate verification operation and exploited in a DoS attack. Any
5470 application which performs certificate verification is vulnerable including
5471 OpenSSL clients and servers which enable client authentication.
5473 This issue was was reported to OpenSSL by Brian Carpenter.
5478 * ASN.1 structure reuse memory corruption fix
5480 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5481 memory corruption via an invalid write. Such reuse is and has been
5482 strongly discouraged and is believed to be rare.
5484 Applications that parse structures containing CHOICE or ANY DEFINED BY
5485 components may be affected. Certificate parsing (d2i_X509 and related
5486 functions) are however not affected. OpenSSL clients and servers are
5492 * PKCS7 NULL pointer dereferences fix
5494 The PKCS#7 parsing code does not handle missing outer ContentInfo
5495 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5496 missing content and trigger a NULL pointer dereference on parsing.
5498 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5499 otherwise parse PKCS#7 structures from untrusted sources are
5500 affected. OpenSSL clients and servers are not affected.
5502 This issue was reported to OpenSSL by Michal Zalewski (Google).
5507 * DoS via reachable assert in SSLv2 servers fix
5509 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5510 servers that both support SSLv2 and enable export cipher suites by sending
5511 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5513 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5514 (OpenSSL development team).
5519 * Empty CKE with client auth and DHE fix
5521 If client auth is used then a server can seg fault in the event of a DHE
5522 ciphersuite being selected and a zero length ClientKeyExchange message
5523 being sent by the client. This could be exploited in a DoS attack.
5528 * Handshake with unseeded PRNG fix
5530 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5531 with an unseeded PRNG. The conditions are:
5532 - The client is on a platform where the PRNG has not been seeded
5533 automatically, and the user has not seeded manually
5534 - A protocol specific client method version has been used (i.e. not
5535 SSL_client_methodv23)
5536 - A ciphersuite is used that does not require additional random data from
5537 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5539 If the handshake succeeds then the client random that has been used will
5540 have been generated from a PRNG with insufficient entropy and therefore the
5541 output may be predictable.
5543 For example using the following command with an unseeded openssl will
5544 succeed on an unpatched platform:
5546 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5551 * Use After Free following d2i_ECPrivatekey error fix
5553 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5554 could cause a use after free condition. This, in turn, could cause a double
5555 free in several private key parsing functions (such as d2i_PrivateKey
5556 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5557 for applications that receive EC private keys from untrusted
5558 sources. This scenario is considered rare.
5560 This issue was discovered by the BoringSSL project and fixed in their
5566 * X509_to_X509_REQ NULL pointer deref fix
5568 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5569 the certificate key is invalid. This function is rarely used in practice.
5571 This issue was discovered by Brian Carpenter.
5576 * Removed the export ciphers from the DEFAULT ciphers
5580 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5582 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5583 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5584 So far those who have to target multiple platforms would compromise
5585 and argue that binary targeting say ARMv5 would still execute on
5586 ARMv8. "Universal" build resolves this compromise by providing
5587 near-optimal performance even on newer platforms.
5591 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5592 (other platforms pending).
5594 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5596 * Add support for the SignedCertificateTimestampList certificate and
5597 OCSP response extensions from RFC6962.
5601 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5602 for corner cases. (Certain input points at infinity could lead to
5603 bogus results, with non-infinity inputs mapped to infinity too.)
5607 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5608 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5609 common cases are optimized and there still is room for further
5610 improvements. Vector Permutation AES for Altivec is also added.
5614 * Add support for little-endian ppc64 Linux target.
5616 *Marcelo Cerri (IBM)*
5618 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5619 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5620 are optimized and there still is room for further improvements.
5621 Both 32- and 64-bit modes are supported.
5623 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5625 * Improved ARMv7 NEON support.
5629 * Support for SPARC Architecture 2011 crypto extensions, first
5630 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5631 SHA256/512, MD5, GHASH and modular exponentiation.
5633 *Andy Polyakov, David Miller*
5635 * Accelerated modular exponentiation for Intel processors, a.k.a.
5638 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5640 * Support for new and upcoming Intel processors, including AVX2,
5641 BMI and SHA ISA extensions. This includes additional "stitched"
5642 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5645 This work was sponsored by Intel Corp.
5649 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5650 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5651 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5655 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5656 this fixes a limitation in previous versions of OpenSSL.
5660 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5661 MGF1 digest and OAEP label.
5665 * Add EVP support for key wrapping algorithms, to avoid problems with
5666 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5667 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5668 algorithms and include tests cases.
5672 * Add functions to allocate and set the fields of an ECDSA_METHOD
5675 *Douglas E. Engert, Steve Henson*
5677 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5678 difference in days and seconds between two tm or ASN1_TIME structures.
5682 * Add -rev test option to s_server to just reverse order of characters
5683 received by client and send back to server. Also prints an abbreviated
5684 summary of the connection parameters.
5688 * New option -brief for s_client and s_server to print out a brief summary
5689 of connection parameters.
5693 * Add callbacks for arbitrary TLS extensions.
5695 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5697 * New option -crl_download in several openssl utilities to download CRLs
5698 from CRLDP extension in certificates.
5702 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5706 * New function X509_CRL_diff to generate a delta CRL from the difference
5707 of two full CRLs. Add support to "crl" utility.
5711 * New functions to set lookup_crls function and to retrieve
5712 X509_STORE from X509_STORE_CTX.
5716 * Print out deprecated issuer and subject unique ID fields in
5721 * Extend OCSP I/O functions so they can be used for simple general purpose
5722 HTTP as well as OCSP. New wrapper function which can be used to download
5723 CRLs using the OCSP API.
5727 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5731 * `SSL_CONF*` functions. These provide a common framework for application
5732 configuration using configuration files or command lines.
5736 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5737 message callback and prints the results. Needs compile time option
5738 "enable-ssl-trace". New options to s_client and s_server to enable
5743 * New ctrl and macro to retrieve supported points extensions.
5744 Print out extension in s_server and s_client.
5748 * New functions to retrieve certificate signature and signature
5753 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5758 * New Suite B modes for TLS code. These use and enforce the requirements
5759 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5760 only use Suite B curves. The Suite B modes can be set by using the
5761 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5765 * New chain verification flags for Suite B levels of security. Check
5766 algorithms are acceptable when flags are set in X509_verify_cert.
5770 * Make tls1_check_chain return a set of flags indicating checks passed
5771 by a certificate chain. Add additional tests to handle client
5772 certificates: checks for matching certificate type and issuer name
5777 * If an attempt is made to use a signature algorithm not in the peer
5778 preference list abort the handshake. If client has no suitable
5779 signature algorithms in response to a certificate request do not
5780 use the certificate.
5784 * If server EC tmp key is not in client preference list abort handshake.
5788 * Add support for certificate stores in CERT structure. This makes it
5789 possible to have different stores per SSL structure or one store in
5790 the parent SSL_CTX. Include distinct stores for certificate chain
5791 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5792 to build and store a certificate chain in CERT structure: returning
5793 an error if the chain cannot be built: this will allow applications
5794 to test if a chain is correctly configured.
5796 Note: if the CERT based stores are not set then the parent SSL_CTX
5797 store is used to retain compatibility with existing behaviour.
5801 * New function ssl_set_client_disabled to set a ciphersuite disabled
5802 mask based on the current session, check mask when sending client
5803 hello and checking the requested ciphersuite.
5807 * New ctrls to retrieve and set certificate types in a certificate
5808 request message. Print out received values in s_client. If certificate
5809 types is not set with custom values set sensible values based on
5810 supported signature algorithms.
5814 * Support for distinct client and server supported signature algorithms.
5818 * Add certificate callback. If set this is called whenever a certificate
5819 is required by client or server. An application can decide which
5820 certificate chain to present based on arbitrary criteria: for example
5821 supported signature algorithms. Add very simple example to s_server.
5822 This fixes many of the problems and restrictions of the existing client
5823 certificate callback: for example you can now clear an existing
5824 certificate and specify the whole chain.
5828 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5829 the certificate can be used for (if anything). Set valid_flags field
5830 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5831 to have similar checks in it.
5833 Add new "cert_flags" field to CERT structure and include a "strict mode".
5834 This enforces some TLS certificate requirements (such as only permitting
5835 certificate signature algorithms contained in the supported algorithms
5836 extension) which some implementations ignore: this option should be used
5837 with caution as it could cause interoperability issues.
5841 * Update and tidy signature algorithm extension processing. Work out
5842 shared signature algorithms based on preferences and peer algorithms
5843 and print them out in s_client and s_server. Abort handshake if no
5844 shared signature algorithms.
5848 * Add new functions to allow customised supported signature algorithms
5849 for SSL and SSL_CTX structures. Add options to s_client and s_server
5854 * New function SSL_certs_clear() to delete all references to certificates
5855 from an SSL structure. Before this once a certificate had been added
5856 it couldn't be removed.
5860 * Integrate hostname, email address and IP address checking with certificate
5861 verification. New verify options supporting checking in openssl utility.
5865 * Fixes and wildcard matching support to hostname and email checking
5866 functions. Add manual page.
5868 *Florian Weimer (Red Hat Product Security Team)*
5870 * New functions to check a hostname email or IP address against a
5871 certificate. Add options x509 utility to print results of checks against
5876 * Fix OCSP checking.
5878 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5880 * Initial experimental support for explicitly trusted non-root CAs.
5881 OpenSSL still tries to build a complete chain to a root but if an
5882 intermediate CA has a trust setting included that is used. The first
5883 setting is used: whether to trust (e.g., -addtrust option to the x509
5888 * Add -trusted_first option which attempts to find certificates in the
5889 trusted store even if an untrusted chain is also supplied.
5893 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5894 platform support for Linux and Android.
5898 * Support for linux-x32, ILP32 environment in x86_64 framework.
5902 * Experimental multi-implementation support for FIPS capable OpenSSL.
5903 When in FIPS mode the approved implementations are used as normal,
5904 when not in FIPS mode the internal unapproved versions are used instead.
5905 This means that the FIPS capable OpenSSL isn't forced to use the
5906 (often lower performance) FIPS implementations outside FIPS mode.
5910 * Transparently support X9.42 DH parameters when calling
5911 PEM_read_bio_DHparameters. This means existing applications can handle
5912 the new parameter format automatically.
5916 * Initial experimental support for X9.42 DH parameter format: mainly
5917 to support use of 'q' parameter for RFC5114 parameters.
5921 * Add DH parameters from RFC5114 including test data to dhtest.
5925 * Support for automatic EC temporary key parameter selection. If enabled
5926 the most preferred EC parameters are automatically used instead of
5927 hardcoded fixed parameters. Now a server just has to call:
5928 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5929 support ECDH and use the most appropriate parameters.
5933 * Enhance and tidy EC curve and point format TLS extension code. Use
5934 static structures instead of allocation if default values are used.
5935 New ctrls to set curves we wish to support and to retrieve shared curves.
5936 Print out shared curves in s_server. New options to s_server and s_client
5937 to set list of supported curves.
5941 * New ctrls to retrieve supported signature algorithms and
5942 supported curve values as an array of NIDs. Extend openssl utility
5943 to print out received values.
5947 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5948 between NIDs and the more common NIST names such as "P-256". Enhance
5949 ecparam utility and ECC method to recognise the NIST names for curves.
5953 * Enhance SSL/TLS certificate chain handling to support different
5954 chains for each certificate instead of one chain in the parent SSL_CTX.
5958 * Support for fixed DH ciphersuite client authentication: where both
5959 server and client use DH certificates with common parameters.
5963 * Support for fixed DH ciphersuites: those requiring DH server
5968 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5970 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5971 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5972 X509_CINF_get_signature were reverted post internal team review.
5977 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5979 * OCSP Status Request extension unbounded memory growth
5981 A malicious client can send an excessively large OCSP Status Request
5982 extension. If that client continually requests renegotiation, sending a
5983 large OCSP Status Request extension each time, then there will be unbounded
5984 memory growth on the server. This will eventually lead to a Denial Of
5985 Service attack through memory exhaustion. Servers with a default
5986 configuration are vulnerable even if they do not support OCSP. Builds using
5987 the "no-ocsp" build time option are not affected.
5989 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5994 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5997 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6003 * OOB write in MDC2_Update()
6005 An overflow can occur in MDC2_Update() either if called directly or
6006 through the EVP_DigestUpdate() function using MDC2. If an attacker
6007 is able to supply very large amounts of input data after a previous
6008 call to EVP_EncryptUpdate() with a partial block then a length check
6009 can overflow resulting in a heap corruption.
6011 The amount of data needed is comparable to SIZE_MAX which is impractical
6014 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6019 * Malformed SHA512 ticket DoS
6021 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6022 DoS attack where a malformed ticket will result in an OOB read which will
6025 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6026 a custom server callback and ticket lookup mechanism.
6028 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6033 * OOB write in BN_bn2dec()
6035 The function BN_bn2dec() does not check the return value of BN_div_word().
6036 This can cause an OOB write if an application uses this function with an
6037 overly large BIGNUM. This could be a problem if an overly large certificate
6038 or CRL is printed out from an untrusted source. TLS is not affected because
6039 record limits will reject an oversized certificate before it is parsed.
6041 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6046 * OOB read in TS_OBJ_print_bio()
6048 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6049 the total length the OID text representation would use and not the amount
6050 of data written. This will result in OOB reads when large OIDs are
6053 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6058 * Pointer arithmetic undefined behaviour
6060 Avoid some undefined pointer arithmetic
6062 A common idiom in the codebase is to check limits in the following manner:
6065 Where "p" points to some malloc'd data of SIZE bytes and
6068 "len" here could be from some externally supplied data (e.g. from a TLS
6071 The rules of C pointer arithmetic are such that "p + len" is only well
6072 defined where len <= SIZE. Therefore the above idiom is actually
6073 undefined behaviour.
6075 For example this could cause problems if some malloc implementation
6076 provides an address for "p" such that "p + len" actually overflows for
6077 values of len that are too big and therefore p + len < limit.
6079 This issue was reported to OpenSSL by Guido Vranken
6084 * Constant time flag not preserved in DSA signing
6086 Operations in the DSA signing algorithm should run in constant time in
6087 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6088 implementation means that a non-constant time codepath is followed for
6089 certain operations. This has been demonstrated through a cache-timing
6090 attack to be sufficient for an attacker to recover the private DSA key.
6092 This issue was reported by César Pereida (Aalto University), Billy Brumley
6093 (Tampere University of Technology), and Yuval Yarom (The University of
6094 Adelaide and NICTA).
6099 * DTLS buffered message DoS
6101 In a DTLS connection where handshake messages are delivered out-of-order
6102 those messages that OpenSSL is not yet ready to process will be buffered
6103 for later use. Under certain circumstances, a flaw in the logic means that
6104 those messages do not get removed from the buffer even though the handshake
6105 has been completed. An attacker could force up to approx. 15 messages to
6106 remain in the buffer when they are no longer required. These messages will
6107 be cleared when the DTLS connection is closed. The default maximum size for
6108 a message is 100k. Therefore the attacker could force an additional 1500k
6109 to be consumed per connection. By opening many simulataneous connections an
6110 attacker could cause a DoS attack through memory exhaustion.
6112 This issue was reported to OpenSSL by Quan Luo.
6117 * DTLS replay protection DoS
6119 A flaw in the DTLS replay attack protection mechanism means that records
6120 that arrive for future epochs update the replay protection "window" before
6121 the MAC for the record has been validated. This could be exploited by an
6122 attacker by sending a record for the next epoch (which does not have to
6123 decrypt or have a valid MAC), with a very large sequence number. This means
6124 that all subsequent legitimate packets are dropped causing a denial of
6125 service for a specific DTLS connection.
6127 This issue was reported to OpenSSL by the OCAP audit team.
6132 * Certificate message OOB reads
6134 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6135 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6136 theoretical DoS risk but this has not been observed in practice on common
6139 The messages affected are client certificate, client certificate request
6140 and server certificate. As a result the attack can only be performed
6141 against a client or a server which enables client authentication.
6143 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6148 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6150 * Prevent padding oracle in AES-NI CBC MAC check
6152 A MITM attacker can use a padding oracle attack to decrypt traffic
6153 when the connection uses an AES CBC cipher and the server support
6156 This issue was introduced as part of the fix for Lucky 13 padding
6157 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6158 constant time by making sure that always the same bytes are read and
6159 compared against either the MAC or padding bytes. But it no longer
6160 checked that there was enough data to have both the MAC and padding
6163 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6168 * Fix EVP_EncodeUpdate overflow
6170 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6171 Base64 encoding of binary data. If an attacker is able to supply very large
6172 amounts of input data then a length check can overflow resulting in a heap
6175 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6176 the `PEM_write_bio*` family of functions. These are mainly used within the
6177 OpenSSL command line applications, so any application which processes data
6178 from an untrusted source and outputs it as a PEM file should be considered
6179 vulnerable to this issue. User applications that call these APIs directly
6180 with large amounts of untrusted data may also be vulnerable.
6182 This issue was reported by Guido Vranken.
6187 * Fix EVP_EncryptUpdate overflow
6189 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6190 is able to supply very large amounts of input data after a previous call to
6191 EVP_EncryptUpdate() with a partial block then a length check can overflow
6192 resulting in a heap corruption. Following an analysis of all OpenSSL
6193 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6194 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6195 the first called function after an EVP_EncryptInit(), and therefore that
6196 specific call must be safe. The second form is where the length passed to
6197 EVP_EncryptUpdate() can be seen from the code to be some small value and
6198 therefore there is no possibility of an overflow. Since all instances are
6199 one of these two forms, it is believed that there can be no overflows in
6200 internal code due to this problem. It should be noted that
6201 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6202 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6203 of these calls have also been analysed too and it is believed there are no
6204 instances in internal usage where an overflow could occur.
6206 This issue was reported by Guido Vranken.
6211 * Prevent ASN.1 BIO excessive memory allocation
6213 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6214 a short invalid encoding can casuse allocation of large amounts of memory
6215 potentially consuming excessive resources or exhausting memory.
6217 Any application parsing untrusted data through d2i BIO functions is
6218 affected. The memory based functions such as d2i_X509() are *not* affected.
6219 Since the memory based functions are used by the TLS library, TLS
6220 applications are not affected.
6222 This issue was reported by Brian Carpenter.
6229 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6230 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6231 in arbitrary stack data being returned in the buffer.
6233 This issue was reported by Guido Vranken.
6238 * Modify behavior of ALPN to invoke callback after SNI/servername
6239 callback, such that updates to the SSL_CTX affect ALPN.
6243 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6248 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6249 methods are enabled and ssl2 is disabled the methods return NULL.
6253 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6255 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6256 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6257 provide any "EXPORT" or "LOW" strength ciphers.
6261 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6262 is by default disabled at build-time. Builds that are not configured with
6263 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6264 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6265 will need to explicitly call either of:
6267 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6269 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6271 as appropriate. Even if either of those is used, or the application
6272 explicitly uses the version-specific SSLv2_method() or its client and
6273 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6274 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6275 ciphers, and SSLv2 56-bit DES are no longer available.
6280 * Fix a double-free in DSA code
6282 A double free bug was discovered when OpenSSL parses malformed DSA private
6283 keys and could lead to a DoS attack or memory corruption for applications
6284 that receive DSA private keys from untrusted sources. This scenario is
6287 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6293 * Disable SRP fake user seed to address a server memory leak.
6295 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6297 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6298 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6299 was changed to ignore the "fake user" SRP seed, even if the seed
6302 Users should use SRP_VBASE_get1_by_user instead. Note that in
6303 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6304 also that even though configuring the SRP seed attempts to hide
6305 invalid usernames by continuing the handshake with fake
6306 credentials, this behaviour is not constant time and no strong
6307 guarantees are made that the handshake is indistinguishable from
6308 that of a valid user.
6313 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6315 In the BN_hex2bn function the number of hex digits is calculated using an
6316 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6317 large values of `i` this can result in `bn_expand` not allocating any
6318 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6319 field as NULL leading to a subsequent NULL ptr deref. For very large values
6320 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6321 In this case memory is allocated to the internal BIGNUM data field, but it
6322 is insufficiently sized leading to heap corruption. A similar issue exists
6323 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6324 is ever called by user applications with very large untrusted hex/dec data.
6325 This is anticipated to be a rare occurrence.
6327 All OpenSSL internal usage of these functions use data that is not expected
6328 to be untrusted, e.g. config file data or application command line
6329 arguments. If user developed applications generate config file data based
6330 on untrusted data then it is possible that this could also lead to security
6331 consequences. This is also anticipated to be rare.
6333 This issue was reported to OpenSSL by Guido Vranken.
6338 * Fix memory issues in `BIO_*printf` functions
6340 The internal `fmtstr` function used in processing a "%s" format string in
6341 the `BIO_*printf` functions could overflow while calculating the length of a
6342 string and cause an OOB read when printing very long strings.
6344 Additionally the internal `doapr_outch` function can attempt to write to an
6345 OOB memory location (at an offset from the NULL pointer) in the event of a
6346 memory allocation failure. In 1.0.2 and below this could be caused where
6347 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6348 could be in processing a very long "%s" format string. Memory leaks can
6351 The first issue may mask the second issue dependent on compiler behaviour.
6352 These problems could enable attacks where large amounts of untrusted data
6353 is passed to the `BIO_*printf` functions. If applications use these functions
6354 in this way then they could be vulnerable. OpenSSL itself uses these
6355 functions when printing out human-readable dumps of ASN.1 data. Therefore
6356 applications that print this data could be vulnerable if the data is from
6357 untrusted sources. OpenSSL command line applications could also be
6358 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6359 as command line arguments.
6361 Libssl is not considered directly vulnerable. Additionally certificates etc
6362 received via remote connections via libssl are also unlikely to be able to
6363 trigger these issues because of message size limits enforced within libssl.
6365 This issue was reported to OpenSSL Guido Vranken.
6370 * Side channel attack on modular exponentiation
6372 A side-channel attack was found which makes use of cache-bank conflicts on
6373 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6374 of RSA keys. The ability to exploit this issue is limited as it relies on
6375 an attacker who has control of code in a thread running on the same
6376 hyper-threaded core as the victim thread which is performing decryptions.
6378 This issue was reported to OpenSSL by Yuval Yarom, The University of
6379 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6380 Nadia Heninger, University of Pennsylvania with more information at
6381 <http://cachebleed.info>.
6386 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6387 if no keysize is specified with default_bits. This fixes an
6388 omission in an earlier change that changed all RSA/DSA key generation
6389 commands to use 2048 bits by default.
6393 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6395 * Protection for DH small subgroup attacks
6397 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6398 switched on by default and cannot be disabled. This could have some
6403 * SSLv2 doesn't block disabled ciphers
6405 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6406 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6407 been disabled, provided that the SSLv2 protocol was not also disabled via
6410 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6411 and Sebastian Schinzel.
6416 * Reject DH handshakes with parameters shorter than 1024 bits.
6420 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6422 * Certificate verify crash with missing PSS parameter
6424 The signature verification routines will crash with a NULL pointer
6425 dereference if presented with an ASN.1 signature using the RSA PSS
6426 algorithm and absent mask generation function parameter. Since these
6427 routines are used to verify certificate signature algorithms this can be
6428 used to crash any certificate verification operation and exploited in a
6429 DoS attack. Any application which performs certificate verification is
6430 vulnerable including OpenSSL clients and servers which enable client
6433 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6438 * X509_ATTRIBUTE memory leak
6440 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6441 memory. This structure is used by the PKCS#7 and CMS routines so any
6442 application which reads PKCS#7 or CMS data from untrusted sources is
6443 affected. SSL/TLS is not affected.
6445 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6451 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6452 This changes the decoding behaviour for some invalid messages,
6453 though the change is mostly in the more lenient direction, and
6454 legacy behaviour is preserved as much as possible.
6458 * In DSA_generate_parameters_ex, if the provided seed is too short,
6459 use a random seed, as already documented.
6461 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6463 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6465 * Alternate chains certificate forgery
6467 During certificate verfification, OpenSSL will attempt to find an
6468 alternative certificate chain if the first attempt to build such a chain
6469 fails. An error in the implementation of this logic can mean that an
6470 attacker could cause certain checks on untrusted certificates to be
6471 bypassed, such as the CA flag, enabling them to use a valid leaf
6472 certificate to act as a CA and "issue" an invalid certificate.
6474 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6480 * Race condition handling PSK identify hint
6482 If PSK identity hints are received by a multi-threaded client then
6483 the values are wrongly updated in the parent SSL_CTX structure. This can
6484 result in a race condition potentially leading to a double free of the
6490 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6492 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6493 incompatibility in the handling of HMAC. The previous ABI has now been
6496 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6498 * Malformed ECParameters causes infinite loop
6500 When processing an ECParameters structure OpenSSL enters an infinite loop
6501 if the curve specified is over a specially malformed binary polynomial
6504 This can be used to perform denial of service against any
6505 system which processes public keys, certificate requests or
6506 certificates. This includes TLS clients and TLS servers with
6507 client authentication enabled.
6509 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6514 * Exploitable out-of-bounds read in X509_cmp_time
6516 X509_cmp_time does not properly check the length of the ASN1_TIME
6517 string and can read a few bytes out of bounds. In addition,
6518 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6521 An attacker can use this to craft malformed certificates and CRLs of
6522 various sizes and potentially cause a segmentation fault, resulting in
6523 a DoS on applications that verify certificates or CRLs. TLS clients
6524 that verify CRLs are affected. TLS clients and servers with client
6525 authentication enabled may be affected if they use custom verification
6528 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6529 independently by Hanno Böck.
6534 * PKCS7 crash with missing EnvelopedContent
6536 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6537 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6538 with missing content and trigger a NULL pointer dereference on parsing.
6540 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6541 structures from untrusted sources are affected. OpenSSL clients and
6542 servers are not affected.
6544 This issue was reported to OpenSSL by Michal Zalewski (Google).
6549 * CMS verify infinite loop with unknown hash function
6551 When verifying a signedData message the CMS code can enter an infinite loop
6552 if presented with an unknown hash function OID. This can be used to perform
6553 denial of service against any system which verifies signedData messages using
6555 This issue was reported to OpenSSL by Johannes Bauer.
6560 * Race condition handling NewSessionTicket
6562 If a NewSessionTicket is received by a multi-threaded client when attempting to
6563 reuse a previous ticket then a race condition can occur potentially leading to
6564 a double free of the ticket data.
6569 * Reject DH handshakes with parameters shorter than 768 bits.
6571 *Kurt Roeckx and Emilia Kasper*
6573 * dhparam: generate 2048-bit parameters by default.
6575 *Kurt Roeckx and Emilia Kasper*
6577 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6579 * Segmentation fault in ASN1_TYPE_cmp fix
6581 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6582 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6583 certificate signature algorithm consistency this can be used to crash any
6584 certificate verification operation and exploited in a DoS attack. Any
6585 application which performs certificate verification is vulnerable including
6586 OpenSSL clients and servers which enable client authentication.
6591 * ASN.1 structure reuse memory corruption fix
6593 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6594 memory corruption via an invalid write. Such reuse is and has been
6595 strongly discouraged and is believed to be rare.
6597 Applications that parse structures containing CHOICE or ANY DEFINED BY
6598 components may be affected. Certificate parsing (d2i_X509 and related
6599 functions) are however not affected. OpenSSL clients and servers are
6605 * PKCS7 NULL pointer dereferences fix
6607 The PKCS#7 parsing code does not handle missing outer ContentInfo
6608 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6609 missing content and trigger a NULL pointer dereference on parsing.
6611 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6612 otherwise parse PKCS#7 structures from untrusted sources are
6613 affected. OpenSSL clients and servers are not affected.
6615 This issue was reported to OpenSSL by Michal Zalewski (Google).
6620 * DoS via reachable assert in SSLv2 servers fix
6622 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6623 servers that both support SSLv2 and enable export cipher suites by sending
6624 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6626 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6627 (OpenSSL development team).
6632 * Use After Free following d2i_ECPrivatekey error fix
6634 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6635 could cause a use after free condition. This, in turn, could cause a double
6636 free in several private key parsing functions (such as d2i_PrivateKey
6637 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6638 for applications that receive EC private keys from untrusted
6639 sources. This scenario is considered rare.
6641 This issue was discovered by the BoringSSL project and fixed in their
6647 * X509_to_X509_REQ NULL pointer deref fix
6649 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6650 the certificate key is invalid. This function is rarely used in practice.
6652 This issue was discovered by Brian Carpenter.
6657 * Removed the export ciphers from the DEFAULT ciphers
6661 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6663 * Build fixes for the Windows and OpenVMS platforms
6665 *Matt Caswell and Richard Levitte*
6667 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6669 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6670 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6671 dereference. This could lead to a Denial Of Service attack. Thanks to
6672 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6677 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6678 dtls1_buffer_record function under certain conditions. In particular this
6679 could occur if an attacker sent repeated DTLS records with the same
6680 sequence number but for the next epoch. The memory leak could be exploited
6681 by an attacker in a Denial of Service attack through memory exhaustion.
6682 Thanks to Chris Mueller for reporting this issue.
6687 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6688 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6689 method would be set to NULL which could later result in a NULL pointer
6690 dereference. Thanks to Frank Schmirler for reporting this issue.
6695 * Abort handshake if server key exchange message is omitted for ephemeral
6698 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6699 reporting this issue.
6704 * Remove non-export ephemeral RSA code on client and server. This code
6705 violated the TLS standard by allowing the use of temporary RSA keys in
6706 non-export ciphersuites and could be used by a server to effectively
6707 downgrade the RSA key length used to a value smaller than the server
6708 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6709 INRIA or reporting this issue.
6714 * Fixed issue where DH client certificates are accepted without verification.
6715 An OpenSSL server will accept a DH certificate for client authentication
6716 without the certificate verify message. This effectively allows a client to
6717 authenticate without the use of a private key. This only affects servers
6718 which trust a client certificate authority which issues certificates
6719 containing DH keys: these are extremely rare and hardly ever encountered.
6720 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6726 * Ensure that the session ID context of an SSL is updated when its
6727 SSL_CTX is updated via SSL_set_SSL_CTX.
6729 The session ID context is typically set from the parent SSL_CTX,
6730 and can vary with the CTX.
6734 * Fix various certificate fingerprint issues.
6736 By using non-DER or invalid encodings outside the signed portion of a
6737 certificate the fingerprint can be changed without breaking the signature.
6738 Although no details of the signed portion of the certificate can be changed
6739 this can cause problems with some applications: e.g. those using the
6740 certificate fingerprint for blacklists.
6742 1. Reject signatures with non zero unused bits.
6744 If the BIT STRING containing the signature has non zero unused bits reject
6745 the signature. All current signature algorithms require zero unused bits.
6747 2. Check certificate algorithm consistency.
6749 Check the AlgorithmIdentifier inside TBS matches the one in the
6750 certificate signature. NB: this will result in signature failure
6751 errors for some broken certificates.
6753 Thanks to Konrad Kraszewski from Google for reporting this issue.
6755 3. Check DSA/ECDSA signatures use DER.
6757 Re-encode DSA/ECDSA signatures and compare with the original received
6758 signature. Return an error if there is a mismatch.
6760 This will reject various cases including garbage after signature
6761 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6762 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6763 (negative or with leading zeroes).
6765 Further analysis was conducted and fixes were developed by Stephen Henson
6766 of the OpenSSL core team.
6772 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6773 results on some platforms, including x86_64. This bug occurs at random
6774 with a very low probability, and is not known to be exploitable in any
6775 way, though its exact impact is difficult to determine. Thanks to Pieter
6776 Wuille (Blockstream) who reported this issue and also suggested an initial
6777 fix. Further analysis was conducted by the OpenSSL development team and
6778 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6779 the OpenSSL core team.
6784 * Do not resume sessions on the server if the negotiated protocol
6785 version does not match the session's version. Resuming with a different
6786 version, while not strictly forbidden by the RFC, is of questionable
6787 sanity and breaks all known clients.
6789 *David Benjamin, Emilia Käsper*
6791 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6792 early CCS messages during renegotiation. (Note that because
6793 renegotiation is encrypted, this early CCS was not exploitable.)
6797 * Tighten client-side session ticket handling during renegotiation:
6798 ensure that the client only accepts a session ticket if the server sends
6799 the extension anew in the ServerHello. Previously, a TLS client would
6800 reuse the old extension state and thus accept a session ticket if one was
6801 announced in the initial ServerHello.
6803 Similarly, ensure that the client requires a session ticket if one
6804 was advertised in the ServerHello. Previously, a TLS client would
6805 ignore a missing NewSessionTicket message.
6809 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6813 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6814 sends a carefully crafted handshake message, to cause OpenSSL to fail
6815 to free up to 64k of memory causing a memory leak. This could be
6816 exploited in a Denial Of Service attack. This issue affects OpenSSL
6817 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6818 whether SRTP is used or configured. Implementations of OpenSSL that
6819 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6821 The fix was developed by the OpenSSL team.
6826 * Session Ticket Memory Leak.
6828 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6829 integrity of that ticket is first verified. In the event of a session
6830 ticket integrity check failing, OpenSSL will fail to free memory
6831 causing a memory leak. By sending a large number of invalid session
6832 tickets an attacker could exploit this issue in a Denial Of Service
6838 * Build option no-ssl3 is incomplete.
6840 When OpenSSL is configured with "no-ssl3" as a build option, servers
6841 could accept and complete a SSL 3.0 handshake, and clients could be
6842 configured to send them.
6845 *Akamai and the OpenSSL team*
6847 * Add support for TLS_FALLBACK_SCSV.
6848 Client applications doing fallback retries should call
6849 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6852 *Adam Langley, Bodo Moeller*
6854 * Add additional DigestInfo checks.
6856 Re-encode DigestInto in DER and check against the original when
6857 verifying RSA signature: this will reject any improperly encoded
6858 DigestInfo structures.
6860 Note: this is a precautionary measure and no attacks are currently known.
6864 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6866 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6867 SRP code can be overrun an internal buffer. Add sanity check that
6868 g, A, B < N to SRP code.
6870 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6871 Group for discovering this issue.
6876 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6877 TLS 1.0 instead of higher protocol versions when the ClientHello message
6878 is badly fragmented. This allows a man-in-the-middle attacker to force a
6879 downgrade to TLS 1.0 even if both the server and the client support a
6880 higher protocol version, by modifying the client's TLS records.
6882 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6883 researching this issue.
6888 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6889 to a denial of service attack. A malicious server can crash the client
6890 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6891 ciphersuite and sending carefully crafted handshake messages.
6893 Thanks to Felix Gröbert (Google) for discovering and researching this
6899 * By sending carefully crafted DTLS packets an attacker could cause openssl
6900 to leak memory. This can be exploited through a Denial of Service attack.
6901 Thanks to Adam Langley for discovering and researching this issue.
6906 * An attacker can force openssl to consume large amounts of memory whilst
6907 processing DTLS handshake messages. This can be exploited through a
6908 Denial of Service attack.
6909 Thanks to Adam Langley for discovering and researching this issue.
6914 * An attacker can force an error condition which causes openssl to crash
6915 whilst processing DTLS packets due to memory being freed twice. This
6916 can be exploited through a Denial of Service attack.
6917 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6923 * If a multithreaded client connects to a malicious server using a resumed
6924 session and the server sends an ec point format extension it could write
6925 up to 255 bytes to freed memory.
6927 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6933 * A malicious server can crash an OpenSSL client with a null pointer
6934 dereference (read) by specifying an SRP ciphersuite even though it was not
6935 properly negotiated with the client. This can be exploited through a
6936 Denial of Service attack.
6938 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6939 discovering and researching this issue.
6944 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6945 X509_name_oneline, X509_name_print_ex et al. to leak some information
6946 from the stack. Applications may be affected if they echo pretty printing
6947 output to the attacker.
6949 Thanks to Ivan Fratric (Google) for discovering this issue.
6952 *Emilia Käsper, and Steve Henson*
6954 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6955 for corner cases. (Certain input points at infinity could lead to
6956 bogus results, with non-infinity inputs mapped to infinity too.)
6960 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6962 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6963 handshake can force the use of weak keying material in OpenSSL
6964 SSL/TLS clients and servers.
6966 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6967 researching this issue. ([CVE-2014-0224])
6969 *KIKUCHI Masashi, Steve Henson*
6971 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6972 OpenSSL DTLS client the code can be made to recurse eventually crashing
6975 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6978 *Imre Rad, Steve Henson*
6980 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6981 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6982 client or server. This is potentially exploitable to run arbitrary
6983 code on a vulnerable client or server.
6985 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6987 *Jüri Aedla, Steve Henson*
6989 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6990 are subject to a denial of service attack.
6992 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6993 this issue. ([CVE-2014-3470])
6995 *Felix Gröbert, Ivan Fratric, Steve Henson*
6997 * Harmonize version and its documentation. -f flag is used to display
7000 *mancha <mancha1@zoho.com>*
7002 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7003 in i2d_ECPrivateKey.
7005 *mancha <mancha1@zoho.com>*
7007 * Fix some double frees. These are not thought to be exploitable.
7009 *mancha <mancha1@zoho.com>*
7011 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7013 * A missing bounds check in the handling of the TLS heartbeat extension
7014 can be used to reveal up to 64k of memory to a connected client or
7017 Thanks for Neel Mehta of Google Security for discovering this bug and to
7018 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7019 preparing the fix ([CVE-2014-0160])
7021 *Adam Langley, Bodo Moeller*
7023 * Fix for the attack described in the paper "Recovering OpenSSL
7024 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7025 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7026 <http://eprint.iacr.org/2014/140>
7028 Thanks to Yuval Yarom and Naomi Benger for discovering this
7029 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7031 *Yuval Yarom and Naomi Benger*
7033 * TLS pad extension: draft-agl-tls-padding-03
7035 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7036 TLS client Hello record length value would otherwise be > 255 and
7037 less that 512 pad with a dummy extension containing zeroes so it
7038 is at least 512 bytes long.
7040 *Adam Langley, Steve Henson*
7042 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7044 * Fix for TLS record tampering bug. A carefully crafted invalid
7045 handshake could crash OpenSSL with a NULL pointer exception.
7046 Thanks to Anton Johansson for reporting this issues.
7049 * Keep original DTLS digest and encryption contexts in retransmission
7050 structures so we can use the previous session parameters if they need
7051 to be resent. ([CVE-2013-6450])
7055 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7056 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7057 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7058 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7059 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7060 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7062 *Rob Stradling, Adam Langley*
7064 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7066 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7067 supporting platforms or when small records were transferred.
7069 *Andy Polyakov, Steve Henson*
7071 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7073 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7075 This addresses the flaw in CBC record processing discovered by
7076 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7077 at: <http://www.isg.rhul.ac.uk/tls/>
7079 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7080 Security Group at Royal Holloway, University of London
7081 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7082 Emilia Käsper for the initial patch.
7085 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7087 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7088 ciphersuites which can be exploited in a denial of service attack.
7089 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7090 and detecting this bug and to Wolfgang Ettlinger
7091 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7096 * Return an error when checking OCSP signatures when key is NULL.
7097 This fixes a DoS attack. ([CVE-2013-0166])
7101 * Make openssl verify return errors.
7103 *Chris Palmer <palmer@google.com> and Ben Laurie*
7105 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7106 the right response is stapled. Also change SSL_get_certificate()
7107 so it returns the certificate actually sent.
7108 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7110 *Rob Stradling <rob.stradling@comodo.com>*
7112 * Fix possible deadlock when decoding public keys.
7116 * Don't use TLS 1.0 record version number in initial client hello
7121 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7123 * Sanity check record length before skipping explicit IV in TLS
7124 1.2, 1.1 and DTLS to fix DoS attack.
7126 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7127 fuzzing as a service testing platform.
7132 * Initialise tkeylen properly when encrypting CMS messages.
7133 Thanks to Solar Designer of Openwall for reporting this issue.
7137 * In FIPS mode don't try to use composite ciphers as they are not
7142 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7144 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7145 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7146 mean any application compiled against OpenSSL 1.0.0 headers setting
7147 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7148 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7149 0x10000000L Any application which was previously compiled against
7150 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7151 will need to be recompiled as a result. Letting be results in
7152 inability to disable specifically TLS 1.1 and in client context,
7153 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7157 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7158 disable just protocol X, but all protocols above X *if* there are
7159 protocols *below* X still enabled. In more practical terms it means
7160 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7161 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7162 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7167 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7169 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7170 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7171 in CRYPTO_realloc_clean.
7173 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7174 issue and to Adam Langley <agl@chromium.org> for fixing it.
7177 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7179 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7183 * Workarounds for some broken servers that "hang" if a client hello
7184 record length exceeds 255 bytes.
7186 1. Do not use record version number > TLS 1.0 in initial client
7187 hello: some (but not all) hanging servers will now work.
7188 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7189 the number of ciphers sent in the client hello. This should be
7190 set to an even number, such as 50, for example by passing:
7191 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7192 Most broken servers should now work.
7193 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7194 TLS 1.2 client support entirely.
7198 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7202 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7204 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7205 STRING form instead of a DigestInfo.
7209 * The format used for MDC2 RSA signatures is inconsistent between EVP
7210 and the RSA_sign/RSA_verify functions. This was made more apparent when
7211 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7212 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7213 the correct format in RSA_verify so both forms transparently work.
7217 * Some servers which support TLS 1.0 can choke if we initially indicate
7218 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7219 encrypted premaster secret. As a workaround use the maximum permitted
7220 client version in client hello, this should keep such servers happy
7221 and still work with previous versions of OpenSSL.
7225 * Add support for TLS/DTLS heartbeats.
7227 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7229 * Add support for SCTP.
7231 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7233 * Improved PRNG seeding for VOS.
7235 *Paul Green <Paul.Green@stratus.com>*
7237 * Extensive assembler packs updates, most notably:
7239 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7240 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7241 - x86_64: bit-sliced AES implementation;
7242 - ARM: NEON support, contemporary platforms optimizations;
7243 - s390x: z196 support;
7244 - `*`: GHASH and GF(2^m) multiplication implementations;
7248 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7249 (removal of unnecessary code)
7251 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7253 * Add TLS key material exporter from RFC 5705.
7257 * Add DTLS-SRTP negotiation from RFC 5764.
7261 * Add Next Protocol Negotiation,
7262 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7263 disabled with a no-npn flag to config or Configure. Code donated
7266 *Adam Langley <agl@google.com> and Ben Laurie*
7268 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7269 NIST-P256, NIST-P521, with constant-time single point multiplication on
7270 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7271 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7272 Code made available under Apache License version 2.0.
7274 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7275 line to include this in your build of OpenSSL, and run "make depend" (or
7276 "make update"). This enables the following EC_METHODs:
7278 EC_GFp_nistp224_method()
7279 EC_GFp_nistp256_method()
7280 EC_GFp_nistp521_method()
7282 EC_GROUP_new_by_curve_name() will automatically use these (while
7283 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7286 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7288 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7289 all platforms. Move ssize_t definition from e_os.h to the public
7290 header file e_os2.h as it now appears in public header file cms.h
7294 * New -sigopt option to the ca, req and x509 utilities. Additional
7295 signature parameters can be passed using this option and in
7300 * Add RSA PSS signing function. This will generate and set the
7301 appropriate AlgorithmIdentifiers for PSS based on those in the
7302 corresponding EVP_MD_CTX structure. No application support yet.
7306 * Support for companion algorithm specific ASN1 signing routines.
7307 New function ASN1_item_sign_ctx() signs a pre-initialised
7308 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7309 the appropriate parameters.
7313 * Add new algorithm specific ASN1 verification initialisation function
7314 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7315 handling will be the same no matter what EVP_PKEY_METHOD is used.
7316 Add a PSS handler to support verification of PSS signatures: checked
7317 against a number of sample certificates.
7321 * Add signature printing for PSS. Add PSS OIDs.
7323 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7325 * Add algorithm specific signature printing. An individual ASN1 method
7326 can now print out signatures instead of the standard hex dump.
7328 More complex signatures (e.g. PSS) can print out more meaningful
7329 information. Include DSA version that prints out the signature
7334 * Password based recipient info support for CMS library: implementing
7339 * Split password based encryption into PBES2 and PBKDF2 functions. This
7340 neatly separates the code into cipher and PBE sections and is required
7341 for some algorithms that split PBES2 into separate pieces (such as
7342 password based CMS).
7346 * Session-handling fixes:
7347 - Fix handling of connections that are resuming with a session ID,
7348 but also support Session Tickets.
7349 - Fix a bug that suppressed issuing of a new ticket if the client
7350 presented a ticket with an expired session.
7351 - Try to set the ticket lifetime hint to something reasonable.
7352 - Make tickets shorter by excluding irrelevant information.
7353 - On the client side, don't ignore renewed tickets.
7355 *Adam Langley, Bodo Moeller (Google)*
7357 * Fix PSK session representation.
7361 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7363 This work was sponsored by Intel.
7367 * Add GCM support to TLS library. Some custom code is needed to split
7368 the IV between the fixed (from PRF) and explicit (from TLS record)
7369 portions. This adds all GCM ciphersuites supported by RFC5288 and
7370 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7371 add a special AESGCM string for GCM only.
7375 * Expand range of ctrls for AES GCM. Permit setting invocation
7376 field on decrypt and retrieval of invocation field only on encrypt.
7380 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7381 As required by RFC5289 these ciphersuites cannot be used if for
7382 versions of TLS earlier than 1.2.
7386 * For FIPS capable OpenSSL interpret a NULL default public key method
7387 as unset and return the appropriate default but do *not* set the default.
7388 This means we can return the appropriate method in applications that
7389 switch between FIPS and non-FIPS modes.
7393 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7394 ENGINE is used then we cannot handle that in the FIPS module so we
7395 keep original code iff non-FIPS operations are allowed.
7399 * Add -attime option to openssl utilities.
7401 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7403 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7407 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7408 FIPS EC methods unconditionally for now.
7412 * New build option no-ec2m to disable characteristic 2 code.
7416 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7417 all cases can be covered as some introduce binary incompatibilities.
7421 * Redirect RSA operations to FIPS module including keygen,
7422 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7426 * Add similar low-level API blocking to ciphers.
7430 * low-level digest APIs are not approved in FIPS mode: any attempt
7431 to use these will cause a fatal error. Applications that *really* want
7432 to use them can use the `private_*` version instead.
7436 * Redirect cipher operations to FIPS module for FIPS builds.
7440 * Redirect digest operations to FIPS module for FIPS builds.
7444 * Update build system to add "fips" flag which will link in fipscanister.o
7445 for static and shared library builds embedding a signature if needed.
7449 * Output TLS supported curves in preference order instead of numerical
7450 order. This is currently hardcoded for the highest order curves first.
7451 This should be configurable so applications can judge speed vs strength.
7455 * Add TLS v1.2 server support for client authentication.
7459 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7464 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7465 FIPS modules versions.
7469 * Add TLS v1.2 client side support for client authentication. Keep cache
7470 of handshake records longer as we don't know the hash algorithm to use
7471 until after the certificate request message is received.
7475 * Initial TLS v1.2 client support. Add a default signature algorithms
7476 extension including all the algorithms we support. Parse new signature
7477 format in client key exchange. Relax some ECC signing restrictions for
7478 TLS v1.2 as indicated in RFC5246.
7482 * Add server support for TLS v1.2 signature algorithms extension. Switch
7483 to new signature format when needed using client digest preference.
7484 All server ciphersuites should now work correctly in TLS v1.2. No client
7485 support yet and no support for client certificates.
7489 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7490 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7491 ciphersuites. At present only RSA key exchange ciphersuites work with
7492 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7493 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7494 and version checking.
7498 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7499 with this defined it will not be affected by any changes to ssl internal
7500 structures. Add several utility functions to allow openssl application
7501 to work with OPENSSL_NO_SSL_INTERN defined.
7505 * A long standing patch to add support for SRP from EdelWeb (Peter
7506 Sylvester and Christophe Renou) was integrated.
7507 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7508 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7511 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7515 * Permit abbreviated handshakes when renegotiating using the function
7516 SSL_renegotiate_abbreviated().
7518 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7520 * Add call to ENGINE_register_all_complete() to
7521 ENGINE_load_builtin_engines(), so some implementations get used
7522 automatically instead of needing explicit application support.
7526 * Add support for TLS key exporter as described in RFC5705.
7528 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7530 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7531 a few changes are required:
7533 Add SSL_OP_NO_TLSv1_1 flag.
7534 Add TLSv1_1 methods.
7535 Update version checking logic to handle version 1.1.
7536 Add explicit IV handling (ported from DTLS code).
7537 Add command line options to s_client/s_server.
7544 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7546 * X509_ATTRIBUTE memory leak
7548 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7549 memory. This structure is used by the PKCS#7 and CMS routines so any
7550 application which reads PKCS#7 or CMS data from untrusted sources is
7551 affected. SSL/TLS is not affected.
7553 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7559 * Race condition handling PSK identify hint
7561 If PSK identity hints are received by a multi-threaded client then
7562 the values are wrongly updated in the parent SSL_CTX structure. This can
7563 result in a race condition potentially leading to a double free of the
7569 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7571 * Malformed ECParameters causes infinite loop
7573 When processing an ECParameters structure OpenSSL enters an infinite loop
7574 if the curve specified is over a specially malformed binary polynomial
7577 This can be used to perform denial of service against any
7578 system which processes public keys, certificate requests or
7579 certificates. This includes TLS clients and TLS servers with
7580 client authentication enabled.
7582 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7587 * Exploitable out-of-bounds read in X509_cmp_time
7589 X509_cmp_time does not properly check the length of the ASN1_TIME
7590 string and can read a few bytes out of bounds. In addition,
7591 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7594 An attacker can use this to craft malformed certificates and CRLs of
7595 various sizes and potentially cause a segmentation fault, resulting in
7596 a DoS on applications that verify certificates or CRLs. TLS clients
7597 that verify CRLs are affected. TLS clients and servers with client
7598 authentication enabled may be affected if they use custom verification
7601 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7602 independently by Hanno Böck.
7607 * PKCS7 crash with missing EnvelopedContent
7609 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7610 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7611 with missing content and trigger a NULL pointer dereference on parsing.
7613 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7614 structures from untrusted sources are affected. OpenSSL clients and
7615 servers are not affected.
7617 This issue was reported to OpenSSL by Michal Zalewski (Google).
7622 * CMS verify infinite loop with unknown hash function
7624 When verifying a signedData message the CMS code can enter an infinite loop
7625 if presented with an unknown hash function OID. This can be used to perform
7626 denial of service against any system which verifies signedData messages using
7628 This issue was reported to OpenSSL by Johannes Bauer.
7633 * Race condition handling NewSessionTicket
7635 If a NewSessionTicket is received by a multi-threaded client when attempting to
7636 reuse a previous ticket then a race condition can occur potentially leading to
7637 a double free of the ticket data.
7642 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7644 * Segmentation fault in ASN1_TYPE_cmp fix
7646 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7647 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7648 certificate signature algorithm consistency this can be used to crash any
7649 certificate verification operation and exploited in a DoS attack. Any
7650 application which performs certificate verification is vulnerable including
7651 OpenSSL clients and servers which enable client authentication.
7656 * ASN.1 structure reuse memory corruption fix
7658 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7659 memory corruption via an invalid write. Such reuse is and has been
7660 strongly discouraged and is believed to be rare.
7662 Applications that parse structures containing CHOICE or ANY DEFINED BY
7663 components may be affected. Certificate parsing (d2i_X509 and related
7664 functions) are however not affected. OpenSSL clients and servers are
7670 * PKCS7 NULL pointer dereferences fix
7672 The PKCS#7 parsing code does not handle missing outer ContentInfo
7673 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7674 missing content and trigger a NULL pointer dereference on parsing.
7676 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7677 otherwise parse PKCS#7 structures from untrusted sources are
7678 affected. OpenSSL clients and servers are not affected.
7680 This issue was reported to OpenSSL by Michal Zalewski (Google).
7685 * DoS via reachable assert in SSLv2 servers fix
7687 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7688 servers that both support SSLv2 and enable export cipher suites by sending
7689 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7691 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7692 (OpenSSL development team).
7697 * Use After Free following d2i_ECPrivatekey error fix
7699 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7700 could cause a use after free condition. This, in turn, could cause a double
7701 free in several private key parsing functions (such as d2i_PrivateKey
7702 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7703 for applications that receive EC private keys from untrusted
7704 sources. This scenario is considered rare.
7706 This issue was discovered by the BoringSSL project and fixed in their
7712 * X509_to_X509_REQ NULL pointer deref fix
7714 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7715 the certificate key is invalid. This function is rarely used in practice.
7717 This issue was discovered by Brian Carpenter.
7722 * Removed the export ciphers from the DEFAULT ciphers
7726 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7728 * Build fixes for the Windows and OpenVMS platforms
7730 *Matt Caswell and Richard Levitte*
7732 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7734 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7735 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7736 dereference. This could lead to a Denial Of Service attack. Thanks to
7737 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7742 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7743 dtls1_buffer_record function under certain conditions. In particular this
7744 could occur if an attacker sent repeated DTLS records with the same
7745 sequence number but for the next epoch. The memory leak could be exploited
7746 by an attacker in a Denial of Service attack through memory exhaustion.
7747 Thanks to Chris Mueller for reporting this issue.
7752 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7753 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7754 method would be set to NULL which could later result in a NULL pointer
7755 dereference. Thanks to Frank Schmirler for reporting this issue.
7760 * Abort handshake if server key exchange message is omitted for ephemeral
7763 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7764 reporting this issue.
7769 * Remove non-export ephemeral RSA code on client and server. This code
7770 violated the TLS standard by allowing the use of temporary RSA keys in
7771 non-export ciphersuites and could be used by a server to effectively
7772 downgrade the RSA key length used to a value smaller than the server
7773 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7774 INRIA or reporting this issue.
7779 * Fixed issue where DH client certificates are accepted without verification.
7780 An OpenSSL server will accept a DH certificate for client authentication
7781 without the certificate verify message. This effectively allows a client to
7782 authenticate without the use of a private key. This only affects servers
7783 which trust a client certificate authority which issues certificates
7784 containing DH keys: these are extremely rare and hardly ever encountered.
7785 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7791 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7792 results on some platforms, including x86_64. This bug occurs at random
7793 with a very low probability, and is not known to be exploitable in any
7794 way, though its exact impact is difficult to determine. Thanks to Pieter
7795 Wuille (Blockstream) who reported this issue and also suggested an initial
7796 fix. Further analysis was conducted by the OpenSSL development team and
7797 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7798 the OpenSSL core team.
7803 * Fix various certificate fingerprint issues.
7805 By using non-DER or invalid encodings outside the signed portion of a
7806 certificate the fingerprint can be changed without breaking the signature.
7807 Although no details of the signed portion of the certificate can be changed
7808 this can cause problems with some applications: e.g. those using the
7809 certificate fingerprint for blacklists.
7811 1. Reject signatures with non zero unused bits.
7813 If the BIT STRING containing the signature has non zero unused bits reject
7814 the signature. All current signature algorithms require zero unused bits.
7816 2. Check certificate algorithm consistency.
7818 Check the AlgorithmIdentifier inside TBS matches the one in the
7819 certificate signature. NB: this will result in signature failure
7820 errors for some broken certificates.
7822 Thanks to Konrad Kraszewski from Google for reporting this issue.
7824 3. Check DSA/ECDSA signatures use DER.
7826 Reencode DSA/ECDSA signatures and compare with the original received
7827 signature. Return an error if there is a mismatch.
7829 This will reject various cases including garbage after signature
7830 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7831 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7832 (negative or with leading zeroes).
7834 Further analysis was conducted and fixes were developed by Stephen Henson
7835 of the OpenSSL core team.
7841 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7843 * Session Ticket Memory Leak.
7845 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7846 integrity of that ticket is first verified. In the event of a session
7847 ticket integrity check failing, OpenSSL will fail to free memory
7848 causing a memory leak. By sending a large number of invalid session
7849 tickets an attacker could exploit this issue in a Denial Of Service
7855 * Build option no-ssl3 is incomplete.
7857 When OpenSSL is configured with "no-ssl3" as a build option, servers
7858 could accept and complete a SSL 3.0 handshake, and clients could be
7859 configured to send them.
7862 *Akamai and the OpenSSL team*
7864 * Add support for TLS_FALLBACK_SCSV.
7865 Client applications doing fallback retries should call
7866 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7869 *Adam Langley, Bodo Moeller*
7871 * Add additional DigestInfo checks.
7873 Reencode DigestInto in DER and check against the original when
7874 verifying RSA signature: this will reject any improperly encoded
7875 DigestInfo structures.
7877 Note: this is a precautionary measure and no attacks are currently known.
7881 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7883 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7884 to a denial of service attack. A malicious server can crash the client
7885 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7886 ciphersuite and sending carefully crafted handshake messages.
7888 Thanks to Felix Gröbert (Google) for discovering and researching this
7894 * By sending carefully crafted DTLS packets an attacker could cause openssl
7895 to leak memory. This can be exploited through a Denial of Service attack.
7896 Thanks to Adam Langley for discovering and researching this issue.
7901 * An attacker can force openssl to consume large amounts of memory whilst
7902 processing DTLS handshake messages. This can be exploited through a
7903 Denial of Service attack.
7904 Thanks to Adam Langley for discovering and researching this issue.
7909 * An attacker can force an error condition which causes openssl to crash
7910 whilst processing DTLS packets due to memory being freed twice. This
7911 can be exploited through a Denial of Service attack.
7912 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7918 * If a multithreaded client connects to a malicious server using a resumed
7919 session and the server sends an ec point format extension it could write
7920 up to 255 bytes to freed memory.
7922 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7928 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7929 X509_name_oneline, X509_name_print_ex et al. to leak some information
7930 from the stack. Applications may be affected if they echo pretty printing
7931 output to the attacker.
7933 Thanks to Ivan Fratric (Google) for discovering this issue.
7936 *Emilia Käsper, and Steve Henson*
7938 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7939 for corner cases. (Certain input points at infinity could lead to
7940 bogus results, with non-infinity inputs mapped to infinity too.)
7944 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7946 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7947 handshake can force the use of weak keying material in OpenSSL
7948 SSL/TLS clients and servers.
7950 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7951 researching this issue. ([CVE-2014-0224])
7953 *KIKUCHI Masashi, Steve Henson*
7955 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7956 OpenSSL DTLS client the code can be made to recurse eventually crashing
7959 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7962 *Imre Rad, Steve Henson*
7964 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7965 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7966 client or server. This is potentially exploitable to run arbitrary
7967 code on a vulnerable client or server.
7969 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7971 *Jüri Aedla, Steve Henson*
7973 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7974 are subject to a denial of service attack.
7976 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7977 this issue. ([CVE-2014-3470])
7979 *Felix Gröbert, Ivan Fratric, Steve Henson*
7981 * Harmonize version and its documentation. -f flag is used to display
7984 *mancha <mancha1@zoho.com>*
7986 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7987 in i2d_ECPrivateKey.
7989 *mancha <mancha1@zoho.com>*
7991 * Fix some double frees. These are not thought to be exploitable.
7993 *mancha <mancha1@zoho.com>*
7995 * Fix for the attack described in the paper "Recovering OpenSSL
7996 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7997 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7998 <http://eprint.iacr.org/2014/140>
8000 Thanks to Yuval Yarom and Naomi Benger for discovering this
8001 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8003 *Yuval Yarom and Naomi Benger*
8005 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8007 * Keep original DTLS digest and encryption contexts in retransmission
8008 structures so we can use the previous session parameters if they need
8009 to be resent. ([CVE-2013-6450])
8013 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8014 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8015 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8016 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8017 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8018 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8020 *Rob Stradling, Adam Langley*
8022 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8024 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8026 This addresses the flaw in CBC record processing discovered by
8027 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8028 at: <http://www.isg.rhul.ac.uk/tls/>
8030 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8031 Security Group at Royal Holloway, University of London
8032 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8033 Emilia Käsper for the initial patch.
8036 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8038 * Return an error when checking OCSP signatures when key is NULL.
8039 This fixes a DoS attack. ([CVE-2013-0166])
8043 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8044 the right response is stapled. Also change SSL_get_certificate()
8045 so it returns the certificate actually sent.
8046 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8047 (This is a backport)
8049 *Rob Stradling <rob.stradling@comodo.com>*
8051 * Fix possible deadlock when decoding public keys.
8055 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8057 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8060 * Sanity check record length before skipping explicit IV in DTLS
8063 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8064 fuzzing as a service testing platform.
8069 * Initialise tkeylen properly when encrypting CMS messages.
8070 Thanks to Solar Designer of Openwall for reporting this issue.
8074 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8076 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8077 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8078 in CRYPTO_realloc_clean.
8080 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8081 issue and to Adam Langley <agl@chromium.org> for fixing it.
8084 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8086 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8088 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8089 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8090 content decryption and always return the same error. Note: this attack
8091 needs on average 2^20 messages so it only affects automated senders. The
8092 old behaviour can be re-enabled in the CMS code by setting the
8093 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8094 an MMA defence is not necessary.
8095 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8096 this issue. ([CVE-2012-0884])
8100 * Fix CVE-2011-4619: make sure we really are receiving a
8101 client hello before rejecting multiple SGC restarts. Thanks to
8102 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8106 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8108 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8109 Thanks to Antonio Martin, Enterprise Secure Access Research and
8110 Development, Cisco Systems, Inc. for discovering this bug and
8111 preparing a fix. ([CVE-2012-0050])
8115 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8117 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8118 of the Vaudenay padding oracle attack on CBC mode encryption
8119 which enables an efficient plaintext recovery attack against
8120 the OpenSSL implementation of DTLS. Their attack exploits timing
8121 differences arising during decryption processing. A research
8122 paper describing this attack can be found at:
8123 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8124 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8125 Security Group at Royal Holloway, University of London
8126 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8127 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8128 for preparing the fix. ([CVE-2011-4108])
8130 *Robin Seggelmann, Michael Tuexen*
8132 * Clear bytes used for block padding of SSL 3.0 records.
8135 *Adam Langley (Google)*
8137 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8138 Kadianakis <desnacked@gmail.com> for discovering this issue and
8139 Adam Langley for preparing the fix. ([CVE-2011-4619])
8141 *Adam Langley (Google)*
8143 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8145 *Andrey Kulikov <amdeich@gmail.com>*
8147 * Prevent malformed RFC3779 data triggering an assertion failure.
8148 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8149 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8151 *Rob Austein <sra@hactrn.net>*
8153 * Improved PRNG seeding for VOS.
8155 *Paul Green <Paul.Green@stratus.com>*
8157 * Fix ssl_ciph.c set-up race.
8159 *Adam Langley (Google)*
8161 * Fix spurious failures in ecdsatest.c.
8163 *Emilia Käsper (Google)*
8165 * Fix the BIO_f_buffer() implementation (which was mixing different
8166 interpretations of the `..._len` fields).
8168 *Adam Langley (Google)*
8170 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8171 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8172 threads won't reuse the same blinding coefficients.
8174 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8175 lock to call BN_BLINDING_invert_ex, and avoids one use of
8176 BN_BLINDING_update for each BN_BLINDING structure (previously,
8177 the last update always remained unused).
8179 *Emilia Käsper (Google)*
8181 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8183 *Bob Buckholz (Google)*
8185 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8187 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8188 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8190 *Kaspar Brand <ossl@velox.ch>*
8192 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8193 for multi-threaded use of ECDH. ([CVE-2011-3210])
8195 *Adam Langley (Google)*
8197 * Fix x509_name_ex_d2i memory leak on bad inputs.
8201 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8202 signature public key algorithm by using OID xref utilities instead.
8203 Before this you could only use some ECC ciphersuites with SHA1 only.
8207 * Add protection against ECDSA timing attacks as mentioned in the paper
8208 by Billy Bob Brumley and Nicola Tuveri, see:
8209 <http://eprint.iacr.org/2011/232.pdf>
8211 *Billy Bob Brumley and Nicola Tuveri*
8213 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8215 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8217 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8219 * Fix bug in string printing code: if *any* escaping is enabled we must
8220 escape the escape character (backslash) or the resulting string is
8225 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8227 * Disable code workaround for ancient and obsolete Netscape browsers
8228 and servers: an attacker can use it in a ciphersuite downgrade attack.
8229 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8233 * Fixed J-PAKE implementation error, originally discovered by
8234 Sebastien Martini, further info and confirmation from Stefan
8235 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8239 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8241 * Fix extension code to avoid race conditions which can result in a buffer
8242 overrun vulnerability: resumed sessions must not be modified as they can
8243 be shared by multiple threads. CVE-2010-3864
8247 * Fix WIN32 build system to correctly link an ENGINE directory into
8252 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8254 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8257 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8259 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8261 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8262 context. The operation can be customised via the ctrl mechanism in
8263 case ENGINEs want to include additional functionality.
8267 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8271 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8272 output hashes compatible with older versions of OpenSSL.
8274 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8276 * Fix compression algorithm handling: if resuming a session use the
8277 compression algorithm of the resumed session instead of determining
8278 it from client hello again. Don't allow server to change algorithm.
8282 * Add load_crls() function to commands tidying load_certs() too. Add option
8283 to verify utility to allow additional CRLs to be included.
8287 * Update OCSP request code to permit adding custom headers to the request:
8288 some responders need this.
8292 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8295 *Julia Lawall <julia@diku.dk>*
8297 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8298 needlessly dereferenced structures, used obsolete functions and
8299 didn't handle all updated verify codes correctly.
8303 * Disable MD2 in the default configuration.
8307 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8308 indicate the initial BIO being pushed or popped. This makes it possible
8309 to determine whether the BIO is the one explicitly called or as a result
8310 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8311 it handles reference counts correctly and doesn't zero out the I/O bio
8312 when it is not being explicitly popped. WARNING: applications which
8313 included workarounds for the old buggy behaviour will need to be modified
8314 or they could free up already freed BIOs.
8318 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8319 renaming to all platforms (within the 0.9.8 branch, this was
8320 done conditionally on Netware platforms to avoid a name clash).
8322 *Guenter <lists@gknw.net>*
8324 * Add ECDHE and PSK support to DTLS.
8326 *Michael Tuexen <tuexen@fh-muenster.de>*
8328 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8333 * Add "missing" function EVP_MD_flags() (without this the only way to
8334 retrieve a digest flags is by accessing the structure directly. Update
8335 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8336 or cipher is registered as in the "from" argument. Print out all
8337 registered digests in the dgst usage message instead of manually
8338 attempting to work them out.
8342 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8343 this allows the use of compression and extensions. Change default cipher
8344 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8345 by default unless an application cipher string requests it.
8349 * Alter match criteria in PKCS12_parse(). It used to try to use local
8350 key ids to find matching certificates and keys but some PKCS#12 files
8351 don't follow the (somewhat unwritten) rules and this strategy fails.
8352 Now just gather all certificates together and the first private key
8353 then look for the first certificate that matches the key.
8357 * Support use of registered digest and cipher names for dgst and cipher
8358 commands instead of having to add each one as a special case. So now
8365 openssl dgst -sha256 foo
8367 and this works for ENGINE based algorithms too.
8371 * Update Gost ENGINE to support parameter files.
8373 *Victor B. Wagner <vitus@cryptocom.ru>*
8375 * Support GeneralizedTime in ca utility.
8377 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8379 * Enhance the hash format used for certificate directory links. The new
8380 form uses the canonical encoding (meaning equivalent names will work
8381 even if they aren't identical) and uses SHA1 instead of MD5. This form
8382 is incompatible with the older format and as a result c_rehash should
8383 be used to rebuild symbolic links.
8387 * Make PKCS#8 the default write format for private keys, replacing the
8388 traditional format. This form is standardised, more secure and doesn't
8389 include an implicit MD5 dependency.
8393 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8394 committed to OpenSSL should pass this lot as a minimum.
8398 * Add session ticket override functionality for use by EAP-FAST.
8400 *Jouni Malinen <j@w1.fi>*
8402 * Modify HMAC functions to return a value. Since these can be implemented
8403 in an ENGINE errors can occur.
8407 * Type-checked OBJ_bsearch_ex.
8411 * Type-checked OBJ_bsearch. Also some constification necessitated
8412 by type-checking. Still to come: TXT_DB, bsearch(?),
8413 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8418 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8419 seconds to a tm structure directly, instead of going through OS
8420 specific date routines. This avoids any issues with OS routines such
8421 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8422 and X509_time_adj_ex() to cover the extended range. The existing
8423 X509_time_adj() is still usable and will no longer have any date issues.
8427 * Delta CRL support. New use deltas option which will attempt to locate
8428 and search any appropriate delta CRLs available.
8430 This work was sponsored by Google.
8434 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8435 code and add additional score elements. Validate alternate CRL paths
8436 as part of the CRL checking and indicate a new error "CRL path validation
8437 error" in this case. Applications wanting additional details can use
8438 the verify callback and check the new "parent" field. If this is not
8439 NULL CRL path validation is taking place. Existing applications won't
8440 see this because it requires extended CRL support which is off by
8443 This work was sponsored by Google.
8447 * Support for freshest CRL extension.
8449 This work was sponsored by Google.
8453 * Initial indirect CRL support. Currently only supported in the CRLs
8454 passed directly and not via lookup. Process certificate issuer
8455 CRL entry extension and lookup CRL entries by bother issuer name
8456 and serial number. Check and process CRL issuer entry in IDP extension.
8458 This work was sponsored by Google.
8462 * Add support for distinct certificate and CRL paths. The CRL issuer
8463 certificate is validated separately in this case. Only enabled if
8464 an extended CRL support flag is set: this flag will enable additional
8465 CRL functionality in future.
8467 This work was sponsored by Google.
8471 * Add support for policy mappings extension.
8473 This work was sponsored by Google.
8477 * Fixes to pathlength constraint, self issued certificate handling,
8478 policy processing to align with RFC3280 and PKITS tests.
8480 This work was sponsored by Google.
8484 * Support for name constraints certificate extension. DN, email, DNS
8485 and URI types are currently supported.
8487 This work was sponsored by Google.
8491 * To cater for systems that provide a pointer-based thread ID rather
8492 than numeric, deprecate the current numeric thread ID mechanism and
8493 replace it with a structure and associated callback type. This
8494 mechanism allows a numeric "hash" to be extracted from a thread ID in
8495 either case, and on platforms where pointers are larger than 'long',
8496 mixing is done to help ensure the numeric 'hash' is usable even if it
8497 can't be guaranteed unique. The default mechanism is to use "&errno"
8498 as a pointer-based thread ID to distinguish between threads.
8500 Applications that want to provide their own thread IDs should now use
8501 CRYPTO_THREADID_set_callback() to register a callback that will call
8502 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8504 Note that ERR_remove_state() is now deprecated, because it is tied
8505 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8506 to free the current thread's error state should be replaced by
8507 ERR_remove_thread_state(NULL).
8509 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8510 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8511 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8512 application was previously providing a numeric thread callback that
8513 was inappropriate for distinguishing threads, then uniqueness might
8514 have been obtained with &errno that happened immediately in the
8515 intermediate development versions of OpenSSL; this is no longer the
8516 case, the numeric thread callback will now override the automatic use
8519 *Geoff Thorpe, with help from Bodo Moeller*
8521 * Initial support for different CRL issuing certificates. This covers a
8522 simple case where the self issued certificates in the chain exist and
8523 the real CRL issuer is higher in the existing chain.
8525 This work was sponsored by Google.
8529 * Removed effectively defunct crypto/store from the build.
8533 * Revamp of STACK to provide stronger type-checking. Still to come:
8534 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8535 ASN1_STRING, CONF_VALUE.
8539 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8540 RAM on SSL connections. This option can save about 34k per idle SSL.
8544 * Revamp of LHASH to provide stronger type-checking. Still to come:
8545 STACK, TXT_DB, bsearch, qsort.
8549 * Initial support for Cryptographic Message Syntax (aka CMS) based
8550 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8551 support for data, signedData, compressedData, digestedData and
8552 encryptedData, envelopedData types included. Scripts to check against
8553 RFC4134 examples draft and interop and consistency checks of many
8554 content types and variants.
8558 * Add options to enc utility to support use of zlib compression BIO.
8562 * Extend mk1mf to support importing of options and assembly language
8563 files from Configure script, currently only included in VC-WIN32.
8564 The assembly language rules can now optionally generate the source
8565 files from the associated perl scripts.
8569 * Implement remaining functionality needed to support GOST ciphersuites.
8570 Interop testing has been performed using CryptoPro implementations.
8572 *Victor B. Wagner <vitus@cryptocom.ru>*
8574 * s390x assembler pack.
8578 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8583 * Implement Opaque PRF Input TLS extension as specified in
8584 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8585 official specification yet and no extension type assignment by
8586 IANA exists, this extension (for now) will have to be explicitly
8587 enabled when building OpenSSL by providing the extension number
8588 to use. For example, specify an option
8590 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8592 to the "config" or "Configure" script to enable the extension,
8593 assuming extension number 0x9527 (which is a completely arbitrary
8594 and unofficial assignment based on the MD5 hash of the Internet
8595 Draft). Note that by doing so, you potentially lose
8596 interoperability with other TLS implementations since these might
8597 be using the same extension number for other purposes.
8599 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8600 opaque PRF input value to use in the handshake. This will create
8601 an internal copy of the length-'len' string at 'src', and will
8602 return non-zero for success.
8604 To get more control and flexibility, provide a callback function
8607 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8608 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8612 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8615 Callback function 'cb' will be called in handshakes, and is
8616 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8617 Argument 'arg' is for application purposes (the value as given to
8618 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8619 be provided to the callback function). The callback function
8620 has to return non-zero to report success: usually 1 to use opaque
8621 PRF input just if possible, or 2 to enforce use of the opaque PRF
8622 input. In the latter case, the library will abort the handshake
8623 if opaque PRF input is not successfully negotiated.
8625 Arguments 'peerinput' and 'len' given to the callback function
8626 will always be NULL and 0 in the case of a client. A server will
8627 see the client's opaque PRF input through these variables if
8628 available (NULL and 0 otherwise). Note that if the server
8629 provides an opaque PRF input, the length must be the same as the
8630 length of the client's opaque PRF input.
8632 Note that the callback function will only be called when creating
8633 a new session (session resumption can resume whatever was
8634 previously negotiated), and will not be called in SSL 2.0
8635 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8636 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8637 for applications that need to enforce opaque PRF input.
8641 * Update ssl code to support digests other than SHA1+MD5 for handshake
8644 *Victor B. Wagner <vitus@cryptocom.ru>*
8646 * Add RFC4507 support to OpenSSL. This includes the corrections in
8647 RFC4507bis. The encrypted ticket format is an encrypted encoded
8648 SSL_SESSION structure, that way new session features are automatically
8651 If a client application caches session in an SSL_SESSION structure
8652 support is transparent because tickets are now stored in the encoded
8655 The SSL_CTX structure automatically generates keys for ticket
8656 protection in servers so again support should be possible
8657 with no application modification.
8659 If a client or server wishes to disable RFC4507 support then the option
8660 SSL_OP_NO_TICKET can be set.
8662 Add a TLS extension debugging callback to allow the contents of any client
8663 or server extensions to be examined.
8665 This work was sponsored by Google.
8669 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8670 OpenSSL should now compile cleanly on gcc 4.2
8672 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8674 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8675 support including streaming MAC support: this is required for GOST
8676 ciphersuite support.
8678 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8680 * Add option -stream to use PKCS#7 streaming in smime utility. New
8681 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8682 to output in BER and PEM format.
8686 * Experimental support for use of HMAC via EVP_PKEY interface. This
8687 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8688 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8689 ENGINE support for HMAC keys which are unextractable. New -mac and
8690 -macopt options to dgst utility.
8694 * New option -sigopt to dgst utility. Update dgst to use
8695 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8696 alternative signing parameters such as X9.31 or PSS in the dgst
8701 * Change ssl_cipher_apply_rule(), the internal function that does
8702 the work each time a ciphersuite string requests enabling
8703 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8704 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8705 the order of disabled ciphersuites such that those ciphersuites
8706 that most recently went from enabled to disabled not only stay
8707 in order with respect to each other, but also have higher priority
8708 than other disabled ciphersuites the next time ciphersuites are
8711 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8712 the same ciphersuites as with "HIGH" alone, but in a specific
8713 order where the PSK ciphersuites come first (since they are the
8714 most recently disabled ciphersuites when "HIGH" is parsed).
8716 Also, change ssl_create_cipher_list() (using this new
8717 functionality) such that between otherwise identical
8718 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8723 * Change ssl_create_cipher_list() so that it automatically
8724 arranges the ciphersuites in reasonable order before starting
8725 to process the rule string. Thus, the definition for "DEFAULT"
8726 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8727 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8728 This makes it much easier to arrive at a reasonable default order
8729 in applications for which anonymous ciphers are OK (meaning
8730 that you can't actually use DEFAULT).
8732 *Bodo Moeller; suggested by Victor Duchovni*
8734 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8735 processing) into multiple integers instead of setting
8736 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8737 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8738 (These masks as well as the individual bit definitions are hidden
8739 away into the non-exported interface ssl/ssl_locl.h, so this
8740 change to the definition of the SSL_CIPHER structure shouldn't
8741 affect applications.) This give us more bits for each of these
8742 categories, so there is no longer a need to coagulate AES128 and
8743 AES256 into a single algorithm bit, and to coagulate Camellia128
8744 and Camellia256 into a single algorithm bit, which has led to all
8747 Thus, among other things, the kludge introduced in 0.9.7m and
8748 0.9.8e for masking out AES256 independently of AES128 or masking
8749 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8751 With the change, we also introduce new ciphersuite aliases that
8752 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8757 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8758 Use the leftmost N bytes of the signature input if the input is
8759 larger than the prime q (with N being the size in bytes of q).
8763 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8764 it yet and it is largely untested.
8768 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8772 * Initial incomplete changes to avoid need for function casts in OpenSSL
8773 some compilers (gcc 4.2 and later) reject their use. Safestack is
8774 reimplemented. Update ASN1 to avoid use of legacy functions.
8778 * Win32/64 targets are linked with Winsock2.
8782 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8783 to external functions. This can be used to increase CRL handling
8784 efficiency especially when CRLs are very large by (for example) storing
8785 the CRL revoked certificates in a database.
8789 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8790 new CRLs added to a directory can be used. New command line option
8791 -verify_return_error to s_client and s_server. This causes real errors
8792 to be returned by the verify callback instead of carrying on no matter
8793 what. This reflects the way a "real world" verify callback would behave.
8797 * GOST engine, supporting several GOST algorithms and public key formats.
8798 Kindly donated by Cryptocom.
8802 * Partial support for Issuing Distribution Point CRL extension. CRLs
8803 partitioned by DP are handled but no indirect CRL or reason partitioning
8804 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8805 selected via a scoring technique which handles IDP and AKID in CRLs.
8809 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8810 will ultimately be used for all verify operations: this will remove the
8811 X509_STORE dependency on certificate verification and allow alternative
8812 lookup methods. X509_STORE based implementations of these two callbacks.
8816 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8817 Modify get_crl() to find a valid (unexpired) CRL if possible.
8821 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8822 this would be called X509_CRL_cmp() but that name is already used by
8823 a function that just compares CRL issuer names. Cache several CRL
8824 extensions in X509_CRL structure and cache CRLDP in X509.
8828 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8829 this maps equivalent X509_NAME structures into a consistent structure.
8830 Name comparison can then be performed rapidly using memcmp().
8834 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8839 * Allow digests to supply their own micalg string for S/MIME type using
8840 the ctrl EVP_MD_CTRL_MICALG.
8844 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8845 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8846 ctrl. It can then customise the structure before and/or after signing
8851 * New function OBJ_add_sigid() to allow application defined signature OIDs
8852 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8853 to free up any added signature OIDs.
8857 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8858 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8859 digest and cipher tables. New options added to openssl utility:
8860 list-message-digest-algorithms and list-cipher-algorithms.
8864 * Change the array representation of binary polynomials: the list
8865 of degrees of non-zero coefficients is now terminated with -1.
8866 Previously it was terminated with 0, which was also part of the
8867 value; thus, the array representation was not applicable to
8868 polynomials where t^0 has coefficient zero. This change makes
8869 the array representation useful in a more general context.
8873 * Various modifications and fixes to SSL/TLS cipher string
8874 handling. For ECC, the code now distinguishes between fixed ECDH
8875 with RSA certificates on the one hand and with ECDSA certificates
8876 on the other hand, since these are separate ciphersuites. The
8877 unused code for Fortezza ciphersuites has been removed.
8879 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8880 (not "ECDHE"). For consistency with the code for DH
8881 certificates, use of ECDH certificates is now considered ECDH
8882 authentication, not RSA or ECDSA authentication (the latter is
8883 merely the CA's signing algorithm and not actively used in the
8886 The temporary ciphersuite alias "ECCdraft" is no longer
8887 available, and ECC ciphersuites are no longer excluded from "ALL"
8888 and "DEFAULT". The following aliases now exist for RFC 4492
8889 ciphersuites, most of these by analogy with the DH case:
8891 kECDHr - ECDH cert, signed with RSA
8892 kECDHe - ECDH cert, signed with ECDSA
8893 kECDH - ECDH cert (signed with either RSA or ECDSA)
8894 kEECDH - ephemeral ECDH
8895 ECDH - ECDH cert or ephemeral ECDH
8901 AECDH - anonymous ECDH
8902 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8906 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8907 Use correct micalg parameters depending on digest(s) in signed message.
8911 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8912 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8916 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8917 an engine to register a method. Add ENGINE lookups for methods and
8918 functional reference processing.
8922 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8923 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8928 * New -resign option to smime utility. This adds one or more signers
8929 to an existing PKCS#7 signedData structure. Also -md option to use an
8930 alternative message digest algorithm for signing.
8934 * Tidy up PKCS#7 routines and add new functions to make it easier to
8935 create PKCS7 structures containing multiple signers. Update smime
8936 application to support multiple signers.
8940 * New -macalg option to pkcs12 utility to allow setting of an alternative
8945 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8946 Reorganize PBE internals to lookup from a static table using NIDs,
8947 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8948 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8949 PRF which will be automatically used with PBES2.
8953 * Replace the algorithm specific calls to generate keys in "req" with the
8958 * Update PKCS#7 enveloped data routines to use new API. This is now
8959 supported by any public key method supporting the encrypt operation. A
8960 ctrl is added to allow the public key algorithm to examine or modify
8961 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8966 * Add a ctrl to asn1 method to allow a public key algorithm to express
8967 a default digest type to use. In most cases this will be SHA1 but some
8968 algorithms (such as GOST) need to specify an alternative digest. The
8969 return value indicates how strong the preference is 1 means optional and
8970 2 is mandatory (that is it is the only supported type). Modify
8971 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8972 use the default md. Update openssl utilities to use the default digest
8973 type for signing if it is not explicitly indicated.
8977 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8978 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8979 signing method from the key type. This effectively removes the link
8980 between digests and public key types.
8984 * Add an OID cross reference table and utility functions. Its purpose is to
8985 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8986 rsaEncryption. This will allow some of the algorithm specific hackery
8987 needed to use the correct OID to be removed.
8991 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8992 structures for PKCS7_sign(). They are now set up by the relevant public
8997 * Add provisional EC pkey method with support for ECDSA and ECDH.
9001 * Add support for key derivation (agreement) in the API, DH method and
9006 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9007 public and private key formats. As a side effect these add additional
9008 command line functionality not previously available: DSA signatures can be
9009 generated and verified using pkeyutl and DH key support and generation in
9016 *Oliver Tappe <zooey@hirschkaefer.de>*
9018 * New make target "install_html_docs" installs HTML renditions of the
9021 *Oliver Tappe <zooey@hirschkaefer.de>*
9023 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9024 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9025 support key and parameter generation and add initial key generation
9026 functionality for RSA.
9030 * Add functions for main EVP_PKEY_method operations. The undocumented
9031 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9032 `EVP_PKEY_{encrypt,decrypt}_old`.
9036 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9037 key API, doesn't do much yet.
9041 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9042 public key algorithms. New option to openssl utility:
9043 "list-public-key-algorithms" to print out info.
9047 * Implement the Supported Elliptic Curves Extension for
9048 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9052 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9053 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9057 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9058 utilities such as rsa, dsa, dsaparam etc except they process any key
9063 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9064 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9065 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9070 * Initial support for pluggable public key ASN1.
9071 De-spaghettify the public key ASN1 handling. Move public and private
9072 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9073 algorithm specific handling to a single module within the relevant
9074 algorithm directory. Add functions to allow (near) opaque processing
9075 of public and private key structures.
9079 * Implement the Supported Point Formats Extension for
9080 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9084 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9085 for the psk identity [hint] and the psk callback functions to the
9086 SSL_SESSION, SSL and SSL_CTX structure.
9089 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9093 SSL_CTX_use_psk_identity_hint
9094 SSL_get_psk_identity_hint
9095 SSL_get_psk_identity
9096 SSL_use_psk_identity_hint
9098 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9100 * Add RFC 3161 compliant time stamp request creation, response generation
9101 and response verification functionality.
9103 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9105 * Add initial support for TLS extensions, specifically for the server_name
9106 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9107 have new members for a host name. The SSL data structure has an
9108 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9109 stored in that context to allow for session resumption, even after the
9110 SSL has been switched to a new SSL_CTX in reaction to a client's
9111 server_name extension.
9113 New functions (subject to change):
9115 SSL_get_servername()
9116 SSL_get_servername_type()
9119 New CTRL codes and macros (subject to change):
9121 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9122 - SSL_CTX_set_tlsext_servername_callback()
9123 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9124 - SSL_CTX_set_tlsext_servername_arg()
9125 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9127 openssl s_client has a new '-servername ...' option.
9129 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9130 '-key2 ...', '-servername_fatal' (subject to change). This allows
9131 testing the HostName extension for a specific single host name ('-cert'
9132 and '-key' remain fallbacks for handshakes without HostName
9133 negotiation). If the unrecognized_name alert has to be sent, this by
9134 default is a warning; it becomes fatal with the '-servername_fatal'
9137 *Peter Sylvester, Remy Allais, Christophe Renou*
9139 * Whirlpool hash implementation is added.
9143 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9144 bn(64,32). Because of instruction set limitations it doesn't have
9145 any negative impact on performance. This was done mostly in order
9146 to make it possible to share assembler modules, such as bn_mul_mont
9147 implementations, between 32- and 64-bit builds without hassle.
9151 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9152 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9157 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9158 dedicated Montgomery multiplication procedure, is introduced.
9159 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9160 "64-bit" performance on certain 32-bit targets.
9164 * New option SSL_OP_NO_COMP to disable use of compression selectively
9165 in SSL structures. New SSL ctrl to set maximum send fragment size.
9166 Save memory by setting the I/O buffer sizes dynamically instead of
9167 using the maximum available value.
9171 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9172 in addition to the text details.
9176 * Very, very preliminary EXPERIMENTAL support for printing of general
9177 ASN1 structures. This currently produces rather ugly output and doesn't
9178 handle several customised structures at all.
9182 * Integrated support for PVK file format and some related formats such
9183 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9184 these in the 'rsa' and 'dsa' utilities.
9188 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9192 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9193 place for the (very old) "NETSCAPE" format certificates which are now
9194 handled using new ASN1 code equivalents.
9198 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9199 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9200 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9204 * Modify CRL distribution points extension code to print out previously
9205 unsupported fields. Enhance extension setting code to allow setting of
9210 * Add print and set support for Issuing Distribution Point CRL extension.
9214 * Change 'Configure' script to enable Camellia by default.
9221 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9223 * When rejecting SSL/TLS records due to an incorrect version number, never
9224 update s->server with a new major version number. As of
9225 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9226 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9227 the previous behavior could result in a read attempt at NULL when
9228 receiving specific incorrect SSL/TLS records once record payload
9229 protection is active. ([CVE-2010-0740])
9231 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9233 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9234 could be crashed if the relevant tables were not present (e.g. chrooted).
9236 *Tomas Hoger <thoger@redhat.com>*
9238 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9240 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9242 *Martin Olsson, Neel Mehta*
9244 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9245 accommodate for stack sorting, always a write lock!).
9249 * On some versions of WIN32 Heap32Next is very slow. This can cause
9250 excessive delays in the RAND_poll(): over a minute. As a workaround
9251 include a time check in the inner Heap32Next loop too.
9255 * The code that handled flushing of data in SSL/TLS originally used the
9256 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9257 the problem outlined in PR#1949. The fix suggested there however can
9258 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9259 of Apache). So instead simplify the code to flush unconditionally.
9260 This should be fine since flushing with no data to flush is a no op.
9264 * Handle TLS versions 2.0 and later properly and correctly use the
9265 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9266 off ancient servers have a habit of sticking around for a while...
9270 * Modify compression code so it frees up structures without using the
9271 ex_data callbacks. This works around a problem where some applications
9272 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9273 restarting) then use compression (e.g. SSL with compression) later.
9274 This results in significant per-connection memory leaks and
9275 has caused some security issues including CVE-2008-1678 and
9280 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9281 change when encrypting or decrypting.
9285 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9286 connect and renegotiate with servers which do not support RI.
9287 Until RI is more widely deployed this option is enabled by default.
9291 * Add "missing" ssl ctrls to clear options and mode.
9295 * If client attempts to renegotiate and doesn't support RI respond with
9296 a no_renegotiation alert as required by RFC5746. Some renegotiating
9297 TLS clients will continue a connection gracefully when they receive
9298 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9299 waiting for a server hello which it will never receive. Now we treat a
9300 received no_renegotiation alert as a fatal error. This is because
9301 applications requesting a renegotiation might well expect it to succeed
9302 and would have no code in place to handle the server denying it so the
9303 only safe thing to do is to terminate the connection.
9307 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9308 peer supports secure renegotiation and 0 otherwise. Print out peer
9309 renegotiation support in s_client/s_server.
9313 * Replace the highly broken and deprecated SPKAC certification method with
9314 the updated NID creation version. This should correctly handle UTF8.
9318 * Implement RFC5746. Re-enable renegotiation but require the extension
9319 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9320 turns out to be a bad idea. It has been replaced by
9321 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9322 SSL_CTX_set_options(). This is really not recommended unless you
9323 know what you are doing.
9325 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9327 * Fixes to stateless session resumption handling. Use initial_ctx when
9328 issuing and attempting to decrypt tickets in case it has changed during
9329 servername handling. Use a non-zero length session ID when attempting
9330 stateless session resumption: this makes it possible to determine if
9331 a resumption has occurred immediately after receiving server hello
9332 (several places in OpenSSL subtly assume this) instead of later in
9337 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9338 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9339 fixes for a few places where the return code is not checked
9342 *Julia Lawall <julia@diku.dk>*
9344 * Add --strict-warnings option to Configure script to include devteam
9345 warnings in other configurations.
9349 * Add support for --libdir option and LIBDIR variable in makefiles. This
9350 makes it possible to install openssl libraries in locations which
9351 have names other than "lib", for example "/usr/lib64" which some
9354 *Steve Henson, based on patch from Jeremy Utley*
9356 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9357 X690 8.9.12 and can produce some misleading textual output of OIDs.
9359 *Steve Henson, reported by Dan Kaminsky*
9361 * Delete MD2 from algorithm tables. This follows the recommendation in
9362 several standards that it is not used in new applications due to
9363 several cryptographic weaknesses. For binary compatibility reasons
9364 the MD2 API is still compiled in by default.
9368 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9373 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9374 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9377 *Guenter <lists@gknw.net>*
9379 * Fix the server certificate chain building code to use X509_verify_cert(),
9380 it used to have an ad-hoc builder which was unable to cope with anything
9381 other than a simple chain.
9383 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9385 * Don't check self signed certificate signatures in X509_verify_cert()
9386 by default (a flag can override this): it just wastes time without
9387 adding any security. As a useful side effect self signed root CAs
9388 with non-FIPS digests are now usable in FIPS mode.
9392 * In dtls1_process_out_of_seq_message() the check if the current message
9393 is already buffered was missing. For every new message was memory
9394 allocated, allowing an attacker to perform an denial of service attack
9395 with sending out of seq handshake messages until there is no memory
9396 left. Additionally every future message was buffered, even if the
9397 sequence number made no sense and would be part of another handshake.
9398 So only messages with sequence numbers less than 10 in advance will be
9399 buffered. ([CVE-2009-1378])
9401 *Robin Seggelmann, discovered by Daniel Mentz*
9403 * Records are buffered if they arrive with a future epoch to be
9404 processed after finishing the corresponding handshake. There is
9405 currently no limitation to this buffer allowing an attacker to perform
9406 a DOS attack with sending records with future epochs until there is no
9407 memory left. This patch adds the pqueue_size() function to determine
9408 the size of a buffer and limits the record buffer to 100 entries.
9411 *Robin Seggelmann, discovered by Daniel Mentz*
9413 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9414 parent structure is freed. ([CVE-2009-1379])
9418 * Handle non-blocking I/O properly in SSL_shutdown() call.
9420 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9422 * Add `2.5.4.*` OIDs
9424 *Ilya O. <vrghost@gmail.com>*
9426 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9428 * Disable renegotiation completely - this fixes a severe security
9429 problem ([CVE-2009-3555]) at the cost of breaking all
9430 renegotiation. Renegotiation can be re-enabled by setting
9431 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9432 run-time. This is really not recommended unless you know what
9437 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9439 * Don't set val to NULL when freeing up structures, it is freed up by
9440 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9441 zeroing past the valid field. ([CVE-2009-0789])
9443 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9445 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9446 checked correctly. This would allow some invalid signed attributes to
9447 appear to verify correctly. ([CVE-2009-0591])
9449 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9451 * Reject UniversalString and BMPString types with invalid lengths. This
9452 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9453 a legal length. ([CVE-2009-0590])
9457 * Set S/MIME signing as the default purpose rather than setting it
9458 unconditionally. This allows applications to override it at the store
9463 * Permit restricted recursion of ASN1 strings. This is needed in practice
9464 to handle some structures.
9468 * Improve efficiency of mem_gets: don't search whole buffer each time
9471 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9473 * New -hex option for openssl rand.
9477 * Print out UTF8String and NumericString when parsing ASN1.
9481 * Support NumericString type for name components.
9485 * Allow CC in the environment to override the automatically chosen
9486 compiler. Note that nothing is done to ensure flags work with the
9491 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9493 * Properly check EVP_VerifyFinal() and similar return values
9496 *Ben Laurie, Bodo Moeller, Google Security Team*
9498 * Enable TLS extensions by default.
9502 * Allow the CHIL engine to be loaded, whether the application is
9503 multithreaded or not. (This does not release the developer from the
9504 obligation to set up the dynamic locking callbacks.)
9506 *Sander Temme <sander@temme.net>*
9508 * Use correct exit code if there is an error in dgst command.
9510 *Steve Henson; problem pointed out by Roland Dirlewanger*
9512 * Tweak Configure so that you need to say "experimental-jpake" to enable
9513 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9517 * Add experimental JPAKE support, including demo authentication in
9518 s_client and s_server.
9522 * Set the comparison function in v3_addr_canonize().
9524 *Rob Austein <sra@hactrn.net>*
9526 * Add support for XMPP STARTTLS in s_client.
9528 *Philip Paeps <philip@freebsd.org>*
9530 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9531 to ensure that even with this option, only ciphersuites in the
9532 server's preference list will be accepted. (Note that the option
9533 applies only when resuming a session, so the earlier behavior was
9534 just about the algorithm choice for symmetric cryptography.)
9538 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9540 * Fix NULL pointer dereference if a DTLS server received
9541 ChangeCipherSpec as first record ([CVE-2009-1386]).
9545 * Fix a state transition in s3_srvr.c and d1_srvr.c
9546 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9550 * The fix in 0.9.8c that supposedly got rid of unsafe
9551 double-checked locking was incomplete for RSA blinding,
9552 addressing just one layer of what turns out to have been
9553 doubly unsafe triple-checked locking.
9555 So now fix this for real by retiring the MONT_HELPER macro
9556 in crypto/rsa/rsa_eay.c.
9558 *Bodo Moeller; problem pointed out by Marius Schilder*
9560 * Various precautionary measures:
9562 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9564 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9565 (NB: This would require knowledge of the secret session ticket key
9566 to exploit, in which case you'd be SOL either way.)
9568 - Change bn_nist.c so that it will properly handle input BIGNUMs
9569 outside the expected range.
9571 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9574 *Neel Mehta, Bodo Moeller*
9576 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9577 the load fails. Useful for distros.
9579 *Ben Laurie and the FreeBSD team*
9581 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9585 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9589 * Expand ENGINE to support engine supplied SSL client certificate functions.
9591 This work was sponsored by Logica.
9595 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9596 keystores. Support for SSL/TLS client authentication too.
9597 Not compiled unless enable-capieng specified to Configure.
9599 This work was sponsored by Logica.
9603 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9604 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9605 attribute creation routines such as certificate requests and PKCS#12
9610 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9612 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9613 handshake which could lead to a client crash as found using the
9614 Codenomicon TLS test suite ([CVE-2008-1672])
9616 *Steve Henson, Mark Cox*
9618 * Fix double free in TLS server name extensions which could lead to
9619 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9623 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9625 Clear the error queue to ensure that error entries left from
9626 older function calls do not interfere with the correct operation.
9628 *Lutz Jaenicke, Erik de Castro Lopo*
9630 * Remove root CA certificates of commercial CAs:
9632 The OpenSSL project does not recommend any specific CA and does not
9633 have any policy with respect to including or excluding any CA.
9634 Therefore it does not make any sense to ship an arbitrary selection
9635 of root CA certificates with the OpenSSL software.
9639 * RSA OAEP patches to fix two separate invalid memory reads.
9640 The first one involves inputs when 'lzero' is greater than
9641 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9642 before the beginning of from). The second one involves inputs where
9643 the 'db' section contains nothing but zeroes (there is a one-byte
9644 invalid read after the end of 'db').
9646 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9648 * Partial backport from 0.9.9-dev:
9650 Introduce bn_mul_mont (dedicated Montgomery multiplication
9651 procedure) as a candidate for BIGNUM assembler implementation.
9652 While 0.9.9-dev uses assembler for various architectures, only
9653 x86_64 is available by default here in the 0.9.8 branch, and
9654 32-bit x86 is available through a compile-time setting.
9656 To try the 32-bit x86 assembler implementation, use Configure
9657 option "enable-montasm" (which exists only for this backport).
9659 As "enable-montasm" for 32-bit x86 disclaims code stability
9660 anyway, in this constellation we activate additional code
9661 backported from 0.9.9-dev for further performance improvements,
9662 namely BN_from_montgomery_word. (To enable this otherwise,
9663 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9665 *Andy Polyakov (backport partially by Bodo Moeller)*
9667 * Add TLS session ticket callback. This allows an application to set
9668 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9669 values. This is useful for key rollover for example where several key
9670 sets may exist with different names.
9674 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9675 This was broken until now in 0.9.8 releases, such that the only way
9676 a registered ENGINE could be used (assuming it initialises
9677 successfully on the host) was to explicitly set it as the default
9678 for the relevant algorithms. This is in contradiction with 0.9.7
9679 behaviour and the documentation. With this fix, when an ENGINE is
9680 registered into a given algorithm's table of implementations, the
9681 'uptodate' flag is reset so that auto-discovery will be used next
9682 time a new context for that algorithm attempts to select an
9685 *Ian Lister (tweaked by Geoff Thorpe)*
9687 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9688 implementation in the following ways:
9690 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9693 Lack of BER streaming support means one pass streaming processing is
9694 only supported if data is detached: setting the streaming flag is
9695 ignored for embedded content.
9697 CMS support is disabled by default and must be explicitly enabled
9698 with the enable-cms configuration option.
9702 * Update the GMP engine glue to do direct copies between BIGNUM and
9703 mpz_t when openssl and GMP use the same limb size. Otherwise the
9704 existing "conversion via a text string export" trick is still used.
9706 *Paul Sheer <paulsheer@gmail.com>*
9708 * Zlib compression BIO. This is a filter BIO which compressed and
9709 uncompresses any data passed through it.
9713 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9714 RFC3394 compatible AES key wrapping.
9718 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9719 sets string data without copying. X509_ALGOR_set0() and
9720 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9721 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9722 from an X509_ATTRIBUTE structure optionally checking it occurs only
9723 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9728 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9729 to get the expected BN_FLG_CONSTTIME behavior.
9731 *Bodo Moeller (Google)*
9735 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9736 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9737 - added some more tests to do_tests.pl
9738 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9739 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9740 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9741 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9742 - various changes to netware.pl to enable gcc-cross builds on Win32
9744 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9745 - various changes to fix missing prototype warnings
9746 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9747 - added AES, WHIRLPOOL and CPUID assembler code to build files
9748 - added missing AES assembler make rules to mk1mf.pl
9749 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9751 *Guenter Knauf <eflash@gmx.net>*
9753 * Implement certificate status request TLS extension defined in RFC3546.
9754 A client can set the appropriate parameters and receive the encoded
9755 OCSP response via a callback. A server can query the supplied parameters
9756 and set the encoded OCSP response in the callback. Add simplified examples
9757 to s_client and s_server.
9761 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9764 + Binary incompatibility of ssl_ctx_st structure
9765 + DTLS interoperation with non-compliant servers
9766 + Don't call get_session_cb() without proposed session
9767 + Fix ia64 assembler code
9769 *Andy Polyakov, Steve Henson*
9771 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9773 * DTLS Handshake overhaul. There were longstanding issues with
9774 OpenSSL DTLS implementation, which were making it impossible for
9775 RFC 4347 compliant client to communicate with OpenSSL server.
9776 Unfortunately just fixing these incompatibilities would "cut off"
9777 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9778 server keeps tolerating non RFC compliant syntax. The opposite is
9779 not true, 0.9.8f client can not communicate with earlier server.
9780 This update even addresses CVE-2007-4995.
9784 * Changes to avoid need for function casts in OpenSSL: some compilers
9785 (gcc 4.2 and later) reject their use.
9786 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9789 * Add RFC4507 support to OpenSSL. This includes the corrections in
9790 RFC4507bis. The encrypted ticket format is an encrypted encoded
9791 SSL_SESSION structure, that way new session features are automatically
9794 If a client application caches session in an SSL_SESSION structure
9795 support is transparent because tickets are now stored in the encoded
9798 The SSL_CTX structure automatically generates keys for ticket
9799 protection in servers so again support should be possible
9800 with no application modification.
9802 If a client or server wishes to disable RFC4507 support then the option
9803 SSL_OP_NO_TICKET can be set.
9805 Add a TLS extension debugging callback to allow the contents of any client
9806 or server extensions to be examined.
9808 This work was sponsored by Google.
9812 * Add initial support for TLS extensions, specifically for the server_name
9813 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9814 have new members for a host name. The SSL data structure has an
9815 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9816 stored in that context to allow for session resumption, even after the
9817 SSL has been switched to a new SSL_CTX in reaction to a client's
9818 server_name extension.
9820 New functions (subject to change):
9822 SSL_get_servername()
9823 SSL_get_servername_type()
9826 New CTRL codes and macros (subject to change):
9828 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9829 - SSL_CTX_set_tlsext_servername_callback()
9830 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9831 - SSL_CTX_set_tlsext_servername_arg()
9832 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9834 openssl s_client has a new '-servername ...' option.
9836 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9837 '-key2 ...', '-servername_fatal' (subject to change). This allows
9838 testing the HostName extension for a specific single host name ('-cert'
9839 and '-key' remain fallbacks for handshakes without HostName
9840 negotiation). If the unrecognized_name alert has to be sent, this by
9841 default is a warning; it becomes fatal with the '-servername_fatal'
9844 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9846 * Add AES and SSE2 assembly language support to VC++ build.
9850 * Mitigate attack on final subtraction in Montgomery reduction.
9854 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9855 (which previously caused an internal error).
9859 * Squeeze another 10% out of IGE mode when in != out.
9863 * AES IGE mode speedup.
9865 *Dean Gaudet (Google)*
9867 * Add the Korean symmetric 128-bit cipher SEED (see
9868 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9869 add SEED ciphersuites from RFC 4162:
9871 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9872 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9873 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9874 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9876 To minimize changes between patchlevels in the OpenSSL 0.9.8
9877 series, SEED remains excluded from compilation unless OpenSSL
9878 is configured with 'enable-seed'.
9880 *KISA, Bodo Moeller*
9882 * Mitigate branch prediction attacks, which can be practical if a
9883 single processor is shared, allowing a spy process to extract
9884 information. For detailed background information, see
9885 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9886 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9887 and Necessary Software Countermeasures"). The core of the change
9888 are new versions BN_div_no_branch() and
9889 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9890 respectively, which are slower, but avoid the security-relevant
9891 conditional branches. These are automatically called by BN_div()
9892 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9893 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9894 remove a conditional branch.
9896 BN_FLG_CONSTTIME is the new name for the previous
9897 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9898 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9899 in the exponent causes BN_mod_exp_mont() to use the alternative
9900 implementation in BN_mod_exp_mont_consttime().) The old name
9901 remains as a deprecated alias.
9903 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9904 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9905 constant-time implementations for more than just exponentiation.
9906 Here too the old name is kept as a deprecated alias.
9908 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9909 the BN_BLINDING structure gets an independent copy of the
9910 modulus. This means that the previous `BIGNUM *m` argument to
9911 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9912 essentially becomes `const BIGNUM *m`, although we can't actually
9913 change this in the header file before 0.9.9. It allows
9914 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9915 enable BN_FLG_CONSTTIME.
9917 *Matthew D Wood (Intel Corp)*
9919 * In the SSL/TLS server implementation, be strict about session ID
9920 context matching (which matters if an application uses a single
9921 external cache for different purposes). Previously,
9922 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9923 set. This did ensure strict client verification, but meant that,
9924 with applications using a single external cache for quite
9925 different requirements, clients could circumvent ciphersuite
9926 restrictions for a given session ID context by starting a session
9927 in a different context.
9931 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9932 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9933 authentication-only ciphersuites.
9937 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9938 not complete and could lead to a possible single byte overflow
9939 ([CVE-2007-5135]) [Ben Laurie]
9941 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9943 * Since AES128 and AES256 (and similarly Camellia128 and
9944 Camellia256) share a single mask bit in the logic of
9945 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9946 kludge to work properly if AES128 is available and AES256 isn't
9947 (or if Camellia128 is available and Camellia256 isn't).
9951 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9952 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9953 When a point or a seed is encoded in a BIT STRING, we need to
9954 prevent the removal of trailing zero bits to get the proper DER
9955 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9956 of a NamedBitList, for which trailing 0 bits need to be removed.)
9960 * Have SSL/TLS server implementation tolerate "mismatched" record
9961 protocol version while receiving ClientHello even if the
9962 ClientHello is fragmented. (The server can't insist on the
9963 particular protocol version it has chosen before the ServerHello
9964 message has informed the client about his choice.)
9968 * Add RFC 3779 support.
9970 *Rob Austein for ARIN, Ben Laurie*
9972 * Load error codes if they are not already present instead of using a
9973 static variable. This allows them to be cleanly unloaded and reloaded.
9974 Improve header file function name parsing.
9978 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9979 or CAPABILITY handshake as required by RFCs.
9983 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9985 * Introduce limits to prevent malicious keys being able to
9986 cause a denial of service. ([CVE-2006-2940])
9988 *Steve Henson, Bodo Moeller*
9990 * Fix ASN.1 parsing of certain invalid structures that can result
9991 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9993 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9994 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9996 * Fix SSL client code which could crash if connecting to a
9997 malicious SSLv2 server. ([CVE-2006-4343])
9999 *Tavis Ormandy and Will Drewry, Google Security Team*
10001 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10002 match only those. Before that, "AES256-SHA" would be interpreted
10003 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10004 the same strength classification in 0.9.7h) as we currently only
10005 have a single AES bit in the ciphersuite description bitmap.
10006 That change, however, also applied to ciphersuite strings such as
10007 "RC4-MD5" that intentionally matched multiple ciphersuites --
10008 namely, SSL 2.0 ciphersuites in addition to the more common ones
10009 from SSL 3.0/TLS 1.0.
10011 So we change the selection algorithm again: Naming an explicit
10012 ciphersuite selects this one ciphersuite, and any other similar
10013 ciphersuite (same bitmap) from *other* protocol versions.
10014 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10015 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10017 Since SSL 2.0 does not have any ciphersuites for which the
10018 128/256 bit distinction would be relevant, this works for now.
10019 The proper fix will be to use different bits for AES128 and
10020 AES256, which would have avoided the problems from the beginning;
10021 however, bits are scarce, so we can only do this in a new release
10022 (not just a patchlevel) when we can change the SSL_CIPHER
10023 definition to split the single 'unsigned long mask' bitmap into
10024 multiple values to extend the available space.
10028 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10030 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10031 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10033 * Add AES IGE and biIGE modes.
10037 * Change the Unix randomness entropy gathering to use poll() when
10038 possible instead of select(), since the latter has some
10039 undesirable limitations.
10041 *Darryl Miles via Richard Levitte and Bodo Moeller*
10043 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10044 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10045 cannot be implicitly activated as part of, e.g., the "AES" alias.
10046 However, please upgrade to OpenSSL 0.9.9[-dev] for
10047 non-experimental use of the ECC ciphersuites to get TLS extension
10048 support, which is required for curve and point format negotiation
10049 to avoid potential handshake problems.
10053 * Disable rogue ciphersuites:
10055 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10056 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10057 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10059 The latter two were purportedly from
10060 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10063 Also deactivate the remaining ciphersuites from
10064 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10065 unofficial, and the ID has long expired.
10069 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10070 dual-core machines) and other potential thread-safety issues.
10074 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10075 versions), which is now available for royalty-free use
10076 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10077 Also, add Camellia TLS ciphersuites from RFC 4132.
10079 To minimize changes between patchlevels in the OpenSSL 0.9.8
10080 series, Camellia remains excluded from compilation unless OpenSSL
10081 is configured with 'enable-camellia'.
10085 * Disable the padding bug check when compression is in use. The padding
10086 bug check assumes the first packet is of even length, this is not
10087 necessarily true if compression is enabled and can result in false
10088 positives causing handshake failure. The actual bug test is ancient
10089 code so it is hoped that implementations will either have fixed it by
10090 now or any which still have the bug do not support compression.
10094 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10096 * When applying a cipher rule check to see if string match is an explicit
10097 cipher suite and only match that one cipher suite if it is.
10101 * Link in manifests for VC++ if needed.
10103 *Austin Ziegler <halostatue@gmail.com>*
10105 * Update support for ECC-based TLS ciphersuites according to
10106 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10107 TLS extensions, which are supported starting with the 0.9.9
10108 branch, not in the OpenSSL 0.9.8 branch).
10112 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10113 opaque EVP_CIPHER_CTX handling.
10117 * Fixes and enhancements to zlib compression code. We now only use
10118 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10119 to conform with the standards mentioned here:
10120 <http://www.zlib.net/DLL_FAQ.txt>
10121 Static zlib linking now works on Windows and the new --with-zlib-include
10122 --with-zlib-lib options to Configure can be used to supply the location
10123 of the headers and library. Gracefully handle case where zlib library
10128 * Several fixes and enhancements to the OID generation code. The old code
10129 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10130 handle numbers larger than ULONG_MAX, truncated printing and had a
10131 non standard OBJ_obj2txt() behaviour.
10135 * Add support for building of engines under engine/ as shared libraries
10136 under VC++ build system.
10140 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10141 Hopefully, we will not see any false combination of paths any more.
10145 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10147 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10148 (part of SSL_OP_ALL). This option used to disable the
10149 countermeasure against man-in-the-middle protocol-version
10150 rollback in the SSL 2.0 server implementation, which is a bad
10151 idea. ([CVE-2005-2969])
10153 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10154 for Information Security, National Institute of Advanced Industrial
10155 Science and Technology [AIST], Japan)*
10157 * Add two function to clear and return the verify parameter flags.
10161 * Keep cipherlists sorted in the source instead of sorting them at
10162 runtime, thus removing the need for a lock.
10166 * Avoid some small subgroup attacks in Diffie-Hellman.
10168 *Nick Mathewson and Ben Laurie*
10170 * Add functions for well-known primes.
10174 * Extended Windows CE support.
10176 *Satoshi Nakamura and Andy Polyakov*
10178 * Initialize SSL_METHOD structures at compile time instead of during
10179 runtime, thus removing the need for a lock.
10183 * Make PKCS7_decrypt() work even if no certificate is supplied by
10184 attempting to decrypt each encrypted key in turn. Add support to
10189 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10191 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10194 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10198 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10199 key into the same file any more.
10203 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10207 * Add -utf8 command line and config file option to 'ca'.
10209 *Stefan <stf@udoma.org*
10211 * Removed the macro des_crypt(), as it seems to conflict with some
10212 libraries. Use DES_crypt().
10216 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10217 involves renaming the source and generated shared-libs for
10218 both. The engines will accept the corrected or legacy ids
10219 ('ncipher' and '4758_cca' respectively) when binding. NB,
10220 this only applies when building 'shared'.
10222 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10224 * Add attribute functions to EVP_PKEY structure. Modify
10225 PKCS12_create() to recognize a CSP name attribute and
10226 use it. Make -CSP option work again in pkcs12 utility.
10230 * Add new functionality to the bn blinding code:
10231 - automatic re-creation of the BN_BLINDING parameters after
10232 a fixed number of uses (currently 32)
10233 - add new function for parameter creation
10234 - introduce flags to control the update behaviour of the
10235 BN_BLINDING parameters
10236 - hide BN_BLINDING structure
10237 Add a second BN_BLINDING slot to the RSA structure to improve
10238 performance when a single RSA object is shared among several
10243 * Add support for DTLS.
10245 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10247 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10248 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10252 * Remove buggy and incomplete DH cert support from
10253 ssl/ssl_rsa.c and ssl/s3_both.c
10257 * Use SHA-1 instead of MD5 as the default digest algorithm for
10258 the `apps/openssl` commands.
10262 * Compile clean with "-Wall -Wmissing-prototypes
10263 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10264 DEBUG_SAFESTACK must also be set.
10268 * Change ./Configure so that certain algorithms can be disabled by default.
10269 The new counterpiece to "no-xxx" is "enable-xxx".
10271 The patented RC5 and MDC2 algorithms will now be disabled unless
10272 "enable-rc5" and "enable-mdc2", respectively, are specified.
10274 (IDEA remains enabled despite being patented. This is because IDEA
10275 is frequently required for interoperability, and there is no license
10276 fee for non-commercial use. As before, "no-idea" can be used to
10277 avoid this algorithm.)
10281 * Add processing of proxy certificates (see RFC 3820). This work was
10282 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10283 EGEE (Enabling Grids for E-science in Europe).
10287 * RC4 performance overhaul on modern architectures/implementations, such
10288 as Intel P4, IA-64 and AMD64.
10292 * New utility extract-section.pl. This can be used specify an alternative
10293 section number in a pod file instead of having to treat each file as
10294 a separate case in Makefile. This can be done by adding two lines to the
10297 =for comment openssl_section:XXX
10299 The blank line is mandatory.
10303 * New arguments -certform, -keyform and -pass for s_client and s_server
10304 to allow alternative format key and certificate files and passphrase
10309 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10310 update associated structures and add various utility functions.
10312 Add new policy related verify parameters, include policy checking in
10313 standard verify code. Enhance 'smime' application with extra parameters
10314 to support policy checking and print out.
10318 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10319 Nehemiah processors. These extensions support AES encryption in hardware
10320 as well as RNG (though RNG support is currently disabled).
10322 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10324 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10328 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10330 *Andy Polyakov and a number of other people*
10332 * Improved PowerPC platform support. Most notably BIGNUM assembler
10333 implementation contributed by IBM.
10335 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10337 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10338 exponent rather than 'unsigned long'. There is a corresponding change to
10339 the new 'rsa_keygen' element of the RSA_METHOD structure.
10341 *Jelte Jansen, Geoff Thorpe*
10343 * Functionality for creating the initial serial number file is now
10344 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10346 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10347 number file to 1, which is bound to cause problems. To avoid
10348 the problems while respecting compatibility between different 0.9.7
10349 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10350 CA.pl for serial number initialization. With the new release 0.9.8,
10351 we can fix the problem directly in the 'ca' utility.)
10355 * Reduced header interdependencies by declaring more opaque objects in
10356 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10357 give fewer recursive includes, which could break lazy source code - so
10358 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10359 developers should define this symbol when building and using openssl to
10360 ensure they track the recommended behaviour, interfaces, [etc], but
10361 backwards-compatible behaviour prevails when this isn't defined.
10365 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10369 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10370 This will generate a random key of the appropriate length based on the
10371 cipher context. The EVP_CIPHER can provide its own random key generation
10372 routine to support keys of a specific form. This is used in the des and
10373 3des routines to generate a key of the correct parity. Update S/MIME
10374 code to use new functions and hence generate correct parity DES keys.
10375 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10376 valid (weak or incorrect parity).
10380 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10381 as looking them up. This is useful when the verified structure may contain
10382 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10383 present unless the new PKCS7_NO_CRL flag is asserted.
10387 * Extend ASN1 oid configuration module. It now additionally accepts the
10390 shortName = some long name, 1.2.3.4
10394 * Reimplemented the BN_CTX implementation. There is now no more static
10395 limitation on the number of variables it can handle nor the depth of the
10396 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10397 information can now expand as required, and rather than having a single
10398 static array of bignums, BN_CTX now uses a linked-list of such arrays
10399 allowing it to expand on demand whilst maintaining the usefulness of
10400 BN_CTX's "bundling".
10404 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10405 to allow all RSA operations to function using a single BN_CTX.
10409 * Preliminary support for certificate policy evaluation and checking. This
10410 is initially intended to pass the tests outlined in "Conformance Testing
10411 of Relying Party Client Certificate Path Processing Logic" v1.07.
10415 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10416 remained unused and not that useful. A variety of other little bignum
10417 tweaks and fixes have also been made continuing on from the audit (see
10422 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10423 associated ASN1, EVP and SSL functions and old ASN1 macros.
10427 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10428 and this should never fail. So the return value from the use of
10429 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10430 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10434 * BN_CTX_get() should return zero-valued bignums, providing the same
10435 initialised value as BN_new().
10437 *Geoff Thorpe, suggested by Ulf Möller*
10439 * Support for inhibitAnyPolicy certificate extension.
10443 * An audit of the BIGNUM code is underway, for which debugging code is
10444 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10445 is considered valid when processing BIGNUMs, and causes execution to
10446 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10447 further steps are taken to deliberately pollute unused data in BIGNUM
10448 structures to try and expose faulty code further on. For now, openssl will
10449 (in its default mode of operation) continue to tolerate the inconsistent
10450 forms that it has tolerated in the past, but authors and packagers should
10451 consider trying openssl and their own applications when compiled with
10452 these debugging symbols defined. It will help highlight potential bugs in
10453 their own code, and will improve the test coverage for OpenSSL itself. At
10454 some point, these tighter rules will become openssl's default to improve
10455 maintainability, though the assert()s and other overheads will remain only
10456 in debugging configurations. See bn.h for more details.
10458 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10460 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10461 that can only be obtained through BN_CTX_new() (which implicitly
10462 initialises it). The presence of this function only made it possible
10463 to overwrite an existing structure (and cause memory leaks).
10467 * Because of the callback-based approach for implementing LHASH as a
10468 template type, lh_insert() adds opaque objects to hash-tables and
10469 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10470 to clean up those corresponding objects before destroying the hash table
10471 (and losing the object pointers). So some over-zealous constifications in
10472 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10473 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10474 prototyped to have "const" restrictions on the object pointers they are
10475 given (and so aren't required to cast them away any more).
10479 * The tmdiff.h API was so ugly and minimal that our own timing utility
10480 (speed) prefers to use its own implementation. The two implementations
10481 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10482 its object type properly exposed (MS_TM) instead of casting to/from
10483 `char *`. This may still change yet if someone realises MS_TM and
10485 aren't necessarily the greatest nomenclatures - but this is what was used
10486 internally to the implementation so I've used that for now.
10490 * Ensure that deprecated functions do not get compiled when
10491 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10492 the self-tests were still using deprecated key-generation functions so
10493 these have been updated also.
10497 * Reorganise PKCS#7 code to separate the digest location functionality
10498 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10499 New function PKCS7_set_digest() to set the digest type for PKCS#7
10500 digestedData type. Add additional code to correctly generate the
10501 digestedData type and add support for this type in PKCS7 initialization
10506 * New function PKCS7_set0_type_other() this initializes a PKCS7
10507 structure of type "other".
10511 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10512 sure the loop does correctly stop and breaking ("division by zero")
10513 modulus operations are not performed. The (pre-generated) prime
10514 table crypto/bn/bn_prime.h was already correct, but it could not be
10515 re-generated on some platforms because of the "division by zero"
10516 situation in the script.
10518 *Ralf S. Engelschall*
10520 * Update support for ECC-based TLS ciphersuites according to
10521 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10522 SHA-1 now is only used for "small" curves (where the
10523 representation of a field element takes up to 24 bytes); for
10524 larger curves, the field element resulting from ECDH is directly
10525 used as premaster secret.
10527 *Douglas Stebila (Sun Microsystems Laboratories)*
10529 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10530 curve secp160r1 to the tests.
10532 *Douglas Stebila (Sun Microsystems Laboratories)*
10534 * Add the possibility to load symbols globally with DSO.
10536 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10538 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10539 control of the error stack.
10543 * Add support for STORE in ENGINE.
10547 * Add the STORE type. The intention is to provide a common interface
10548 to certificate and key stores, be they simple file-based stores, or
10549 HSM-type store, or LDAP stores, or...
10550 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10554 * Add a generic structure called OPENSSL_ITEM. This can be used to
10555 pass a list of arguments to any function as well as provide a way
10556 for a function to pass data back to the caller.
10560 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10561 works like BUF_strdup() but can be used to duplicate a portion of
10562 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10567 * Add the function sk_find_ex() which works like sk_find(), but will
10568 return an index to an element even if an exact match couldn't be
10569 found. The index is guaranteed to point at the element where the
10570 searched-for key would be inserted to preserve sorting order.
10574 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10575 takes an extra flags argument for optional functionality. Currently,
10576 the following flags are defined:
10578 OBJ_BSEARCH_VALUE_ON_NOMATCH
10579 This one gets OBJ_bsearch_ex() to return a pointer to the first
10580 element where the comparing function returns a negative or zero
10583 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10584 This one gets OBJ_bsearch_ex() to return a pointer to the first
10585 element where the comparing function returns zero. This is useful
10586 if there are more than one element where the comparing function
10591 * Make it possible to create self-signed certificates with 'openssl ca'
10592 in such a way that the self-signed certificate becomes part of the
10593 CA database and uses the same mechanisms for serial number generation
10594 as all other certificate signing. The new flag '-selfsign' enables
10595 this functionality. Adapt CA.sh and CA.pl.in.
10599 * Add functionality to check the public key of a certificate request
10600 against a given private. This is useful to check that a certificate
10601 request can be signed by that key (self-signing).
10605 * Make it possible to have multiple active certificates with the same
10606 subject in the CA index file. This is done only if the keyword
10607 'unique_subject' is set to 'no' in the main CA section (default
10608 if 'CA_default') of the configuration file. The value is saved
10609 with the database itself in a separate index attribute file,
10610 named like the index file with '.attr' appended to the name.
10614 * Generate multi-valued AVAs using '+' notation in config files for
10619 * Support for nameConstraints certificate extension.
10623 * Support for policyConstraints certificate extension.
10627 * Support for policyMappings certificate extension.
10631 * Make sure the default DSA_METHOD implementation only uses its
10632 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10633 and change its own handlers to be NULL so as to remove unnecessary
10634 indirection. This lets alternative implementations fallback to the
10635 default implementation more easily.
10639 * Support for directoryName in GeneralName related extensions
10644 * Make it possible to link applications using Makefile.shared.
10645 Make that possible even when linking against static libraries!
10649 * Support for single pass processing for S/MIME signing. This now
10650 means that S/MIME signing can be done from a pipe, in addition
10651 cleartext signing (multipart/signed type) is effectively streaming
10652 and the signed data does not need to be all held in memory.
10654 This is done with a new flag PKCS7_STREAM. When this flag is set
10655 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10656 is done after the data is output (and digests calculated) in
10657 SMIME_write_PKCS7().
10661 * Add full support for -rpath/-R, both in shared libraries and
10662 applications, at least on the platforms where it's known how
10667 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10668 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10669 will now compute a table of multiples of the generator that
10670 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10671 faster (notably in the case of a single point multiplication,
10672 scalar * generator).
10674 *Nils Larsch, Bodo Moeller*
10676 * IPv6 support for certificate extensions. The various extensions
10677 which use the IP:a.b.c.d can now take IPv6 addresses using the
10678 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10683 * Added an ENGINE that implements RSA by performing private key
10684 exponentiations with the GMP library. The conversions to and from
10685 GMP's mpz_t format aren't optimised nor are any montgomery forms
10686 cached, and on x86 it appears OpenSSL's own performance has caught up.
10687 However there are likely to be other architectures where GMP could
10688 provide a boost. This ENGINE is not built in by default, but it can be
10689 specified at Configure time and should be accompanied by the necessary
10690 linker additions, eg;
10691 ./config -DOPENSSL_USE_GMP -lgmp
10695 * "openssl engine" will not display ENGINE/DSO load failure errors when
10696 testing availability of engines with "-t" - the old behaviour is
10697 produced by increasing the feature's verbosity with "-tt".
10701 * ECDSA routines: under certain error conditions uninitialized BN objects
10702 could be freed. Solution: make sure initialization is performed early
10703 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10708 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10709 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10710 software implementations. For DSA and DH, parameter generation can
10711 also be overridden by providing the appropriate method callbacks.
10715 * Change the "progress" mechanism used in key-generation and
10716 primality testing to functions that take a new BN_GENCB pointer in
10717 place of callback/argument pairs. The new API functions have `_ex`
10718 postfixes and the older functions are reimplemented as wrappers for
10719 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10720 declarations of the old functions to help (graceful) attempts to
10721 migrate to the new functions. Also, the new key-generation API
10722 functions operate on a caller-supplied key-structure and return
10723 success/failure rather than returning a key or NULL - this is to
10724 help make "keygen" another member function of RSA_METHOD etc.
10726 Example for using the new callback interface:
10728 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10729 void *my_arg = ...;
10732 BN_GENCB_set(&my_cb, my_callback, my_arg);
10734 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10735 /* For the meaning of a, b in calls to my_callback(), see the
10736 * documentation of the function that calls the callback.
10737 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10738 * my_callback should return 1 if it wants BN_is_prime_ex()
10739 * to continue, or 0 to stop.
10744 * Change the ZLIB compression method to be stateful, and make it
10745 available to TLS with the number defined in
10746 draft-ietf-tls-compression-04.txt.
10750 * Add the ASN.1 structures and functions for CertificatePair, which
10751 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10753 CertificatePair ::= SEQUENCE {
10754 forward [0] Certificate OPTIONAL,
10755 reverse [1] Certificate OPTIONAL,
10756 -- at least one of the pair shall be present -- }
10758 Also implement the PEM functions to read and write certificate
10759 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10761 This needed to be defined, mostly for the sake of the LDAP
10762 attribute crossCertificatePair, but may prove useful elsewhere as
10767 * Make it possible to inhibit symlinking of shared libraries in
10768 Makefile.shared, for Cygwin's sake.
10772 * Extend the BIGNUM API by creating a function
10773 void BN_set_negative(BIGNUM *a, int neg);
10774 and a macro that behave like
10775 int BN_is_negative(const BIGNUM *a);
10777 to avoid the need to access 'a->neg' directly in applications.
10781 * Implement fast modular reduction for pseudo-Mersenne primes
10782 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10783 EC_GROUP_new_curve_GFp() will now automatically use this
10786 *Nils Larsch <nla@trustcenter.de>*
10788 * Add new lock type (CRYPTO_LOCK_BN).
10792 * Change the ENGINE framework to automatically load engines
10793 dynamically from specific directories unless they could be
10794 found to already be built in or loaded. Move all the
10795 current engines except for the cryptodev one to a new
10796 directory engines/.
10797 The engines in engines/ are built as shared libraries if
10798 the "shared" options was given to ./Configure or ./config.
10799 Otherwise, they are inserted in libcrypto.a.
10800 /usr/local/ssl/engines is the default directory for dynamic
10801 engines, but that can be overridden at configure time through
10802 the usual use of --prefix and/or --openssldir, and at run
10803 time with the environment variable OPENSSL_ENGINES.
10805 *Geoff Thorpe and Richard Levitte*
10807 * Add Makefile.shared, a helper makefile to build shared
10808 libraries. Adapt Makefile.org.
10812 * Add version info to Win32 DLLs.
10814 *Peter 'Luna' Runestig" <peter@runestig.com>*
10816 * Add new 'medium level' PKCS#12 API. Certificates and keys
10817 can be added using this API to created arbitrary PKCS#12
10818 files while avoiding the low-level API.
10820 New options to PKCS12_create(), key or cert can be NULL and
10821 will then be omitted from the output file. The encryption
10822 algorithm NIDs can be set to -1 for no encryption, the mac
10823 iteration count can be set to 0 to omit the mac.
10825 Enhance pkcs12 utility by making the -nokeys and -nocerts
10826 options work when creating a PKCS#12 file. New option -nomac
10827 to omit the mac, NONE can be set for an encryption algorithm.
10828 New code is modified to use the enhanced PKCS12_create()
10829 instead of the low-level API.
10833 * Extend ASN1 encoder to support indefinite length constructed
10834 encoding. This can output sequences tags and octet strings in
10835 this form. Modify pk7_asn1.c to support indefinite length
10836 encoding. This is experimental and needs additional code to
10837 be useful, such as an ASN1 bio and some enhanced streaming
10840 Extend template encode functionality so that tagging is passed
10841 down to the template encoder.
10845 * Let 'openssl req' fail if an argument to '-newkey' is not
10846 recognized instead of using RSA as a default.
10850 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10851 As these are not official, they are not included in "ALL";
10852 the "ECCdraft" ciphersuite group alias can be used to select them.
10854 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10856 * Add ECDH engine support.
10858 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10860 * Add ECDH in new directory crypto/ecdh/.
10862 *Douglas Stebila (Sun Microsystems Laboratories)*
10864 * Let BN_rand_range() abort with an error after 100 iterations
10865 without success (which indicates a broken PRNG).
10869 * Change BN_mod_sqrt() so that it verifies that the input value
10870 is really the square of the return value. (Previously,
10871 BN_mod_sqrt would show GIGO behaviour.)
10875 * Add named elliptic curves over binary fields from X9.62, SECG,
10876 and WAP/WTLS; add OIDs that were still missing.
10878 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10880 * Extend the EC library for elliptic curves over binary fields
10881 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10884 EC_GF2m_simple_method
10888 EC_GROUP_new_curve_GF2m
10889 EC_GROUP_set_curve_GF2m
10890 EC_GROUP_get_curve_GF2m
10891 EC_POINT_set_affine_coordinates_GF2m
10892 EC_POINT_get_affine_coordinates_GF2m
10893 EC_POINT_set_compressed_coordinates_GF2m
10895 Point compression for binary fields is disabled by default for
10896 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10899 As binary polynomials are represented as BIGNUMs, various members
10900 of the EC_GROUP and EC_POINT data structures can be shared
10901 between the implementations for prime fields and binary fields;
10902 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10903 are essentially identical to their `..._GFp` counterparts.
10904 (For simplicity, the `..._GFp` prefix has been dropped from
10905 various internal method names.)
10907 An internal 'field_div' method (similar to 'field_mul' and
10908 'field_sqr') has been added; this is used only for binary fields.
10910 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10912 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10913 through methods ('mul', 'precompute_mult').
10915 The generic implementations (now internally called 'ec_wNAF_mul'
10916 and 'ec_wNAF_precomputed_mult') remain the default if these
10917 methods are undefined.
10919 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10921 * New function EC_GROUP_get_degree, which is defined through
10922 EC_METHOD. For curves over prime fields, this returns the bit
10923 length of the modulus.
10925 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10927 * New functions EC_GROUP_dup, EC_POINT_dup.
10928 (These simply call ..._new and ..._copy).
10930 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10932 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10933 Polynomials are represented as BIGNUMs (where the sign bit is not
10934 used) in the following functions [macros]:
10937 BN_GF2m_sub [= BN_GF2m_add]
10938 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10939 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10940 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10942 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10943 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10944 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10945 BN_GF2m_cmp [= BN_ucmp]
10947 (Note that only the 'mod' functions are actually for fields GF(2^m).
10948 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10950 For some functions, an the irreducible polynomial defining a
10951 field can be given as an 'unsigned int[]' with strictly
10952 decreasing elements giving the indices of those bits that are set;
10953 i.e., p[] represents the polynomial
10954 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10956 p[0] > p[1] > ... > p[k] = 0.
10957 This applies to the following functions:
10960 BN_GF2m_mod_mul_arr
10961 BN_GF2m_mod_sqr_arr
10962 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10963 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10964 BN_GF2m_mod_exp_arr
10965 BN_GF2m_mod_sqrt_arr
10966 BN_GF2m_mod_solve_quad_arr
10970 Conversion can be performed by the following functions:
10975 bntest.c has additional tests for binary polynomial arithmetic.
10977 Two implementations for BN_GF2m_mod_div() are available.
10978 The default algorithm simply uses BN_GF2m_mod_inv() and
10979 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10980 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10981 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10983 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10985 * Add new error code 'ERR_R_DISABLED' that can be used when some
10986 functionality is disabled at compile-time.
10988 *Douglas Stebila <douglas.stebila@sun.com>*
10990 * Change default behaviour of 'openssl asn1parse' so that more
10991 information is visible when viewing, e.g., a certificate:
10993 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10994 mode the content of non-printable OCTET STRINGs is output in a
10995 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10996 avoid the appearance of a printable string.
10998 *Nils Larsch <nla@trustcenter.de>*
11000 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11002 EC_GROUP_set_asn1_flag()
11003 EC_GROUP_get_asn1_flag()
11004 EC_GROUP_set_point_conversion_form()
11005 EC_GROUP_get_point_conversion_form()
11006 These control ASN1 encoding details:
11007 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11008 has been set to OPENSSL_EC_NAMED_CURVE.
11009 - Points are encoded in uncompressed form by default; options for
11010 asn1_for are as for point2oct, namely
11011 POINT_CONVERSION_COMPRESSED
11012 POINT_CONVERSION_UNCOMPRESSED
11013 POINT_CONVERSION_HYBRID
11015 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11017 EC_GROUP_set_seed()
11018 EC_GROUP_get0_seed()
11019 EC_GROUP_get_seed_len()
11020 This is used only for ASN1 purposes (so far).
11022 *Nils Larsch <nla@trustcenter.de>*
11024 * Add 'field_type' member to EC_METHOD, which holds the NID
11025 of the appropriate field type OID. The new function
11026 EC_METHOD_get_field_type() returns this value.
11028 *Nils Larsch <nla@trustcenter.de>*
11031 EC_POINT_point2bn()
11032 EC_POINT_bn2point()
11033 EC_POINT_point2hex()
11034 EC_POINT_hex2point()
11035 providing useful interfaces to EC_POINT_point2oct() and
11036 EC_POINT_oct2point().
11038 *Nils Larsch <nla@trustcenter.de>*
11040 * Change internals of the EC library so that the functions
11041 EC_GROUP_set_generator()
11042 EC_GROUP_get_generator()
11043 EC_GROUP_get_order()
11044 EC_GROUP_get_cofactor()
11045 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11046 to methods, which would lead to unnecessary code duplication when
11047 adding different types of curves.
11049 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11051 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11052 arithmetic, and such that modified wNAFs are generated
11053 (which avoid length expansion in many cases).
11057 * Add a function EC_GROUP_check_discriminant() (defined via
11058 EC_METHOD) that verifies that the curve discriminant is non-zero.
11060 Add a function EC_GROUP_check() that makes some sanity tests
11061 on a EC_GROUP, its generator and order. This includes
11062 EC_GROUP_check_discriminant().
11064 *Nils Larsch <nla@trustcenter.de>*
11066 * Add ECDSA in new directory crypto/ecdsa/.
11068 Add applications 'openssl ecparam' and 'openssl ecdsa'
11069 (these are based on 'openssl dsaparam' and 'openssl dsa').
11071 ECDSA support is also included in various other files across the
11072 library. Most notably,
11073 - 'openssl req' now has a '-newkey ecdsa:file' option;
11074 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11075 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11076 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11077 them suitable for ECDSA where domain parameters must be
11078 extracted before the specific public key;
11079 - ECDSA engine support has been added.
11081 *Nils Larsch <nla@trustcenter.de>*
11083 * Include some named elliptic curves, and add OIDs from X9.62,
11084 SECG, and WAP/WTLS. Each curve can be obtained from the new
11086 EC_GROUP_new_by_curve_name(),
11087 and the list of available named curves can be obtained with
11088 EC_get_builtin_curves().
11089 Also add a 'curve_name' member to EC_GROUP objects, which can be
11091 EC_GROUP_set_curve_name()
11092 EC_GROUP_get_curve_name()
11094 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11096 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11097 was actually never needed) and in BN_mul(). The removal in BN_mul()
11098 required a small change in bn_mul_part_recursive() and the addition
11099 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11100 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11101 bn_sub_words() and bn_add_words() except they take arrays with
11106 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11108 * Cleanse PEM buffers before freeing them since they may contain
11111 *Benjamin Bennett <ben@psc.edu>*
11113 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11114 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11115 authentication-only ciphersuites.
11119 * Since AES128 and AES256 share a single mask bit in the logic of
11120 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11121 kludge to work properly if AES128 is available and AES256 isn't.
11125 * Expand security boundary to match 1.1.1 module.
11129 * Remove redundant features: hash file source, editing of test vectors
11130 modify fipsld to use external fips_premain.c signature.
11134 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11135 run algorithm test programs.
11139 * Make algorithm test programs more tolerant of whitespace.
11143 * Have SSL/TLS server implementation tolerate "mismatched" record
11144 protocol version while receiving ClientHello even if the
11145 ClientHello is fragmented. (The server can't insist on the
11146 particular protocol version it has chosen before the ServerHello
11147 message has informed the client about his choice.)
11151 * Load error codes if they are not already present instead of using a
11152 static variable. This allows them to be cleanly unloaded and reloaded.
11156 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11158 * Introduce limits to prevent malicious keys being able to
11159 cause a denial of service. ([CVE-2006-2940])
11161 *Steve Henson, Bodo Moeller*
11163 * Fix ASN.1 parsing of certain invalid structures that can result
11164 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11166 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11167 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11169 * Fix SSL client code which could crash if connecting to a
11170 malicious SSLv2 server. ([CVE-2006-4343])
11172 *Tavis Ormandy and Will Drewry, Google Security Team*
11174 * Change ciphersuite string processing so that an explicit
11175 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11176 will no longer include "AES128-SHA"), and any other similar
11177 ciphersuite (same bitmap) from *other* protocol versions (so that
11178 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11179 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11180 changes from 0.9.8b and 0.9.8d.
11184 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11186 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11187 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11189 * Change the Unix randomness entropy gathering to use poll() when
11190 possible instead of select(), since the latter has some
11191 undesirable limitations.
11193 *Darryl Miles via Richard Levitte and Bodo Moeller*
11195 * Disable rogue ciphersuites:
11197 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11198 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11199 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11201 The latter two were purportedly from
11202 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11205 Also deactivate the remaining ciphersuites from
11206 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11207 unofficial, and the ID has long expired.
11211 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11212 dual-core machines) and other potential thread-safety issues.
11216 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11218 * Adapt fipsld and the build system to link against the validated FIPS
11219 module in FIPS mode.
11223 * Fixes for VC++ 2005 build under Windows.
11227 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11228 from a Windows bash shell such as MSYS. It is autodetected from the
11229 "config" script when run from a VC++ environment. Modify standard VC++
11230 build to use fipscanister.o from the GNU make build.
11234 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11236 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11237 The value now differs depending on if you build for FIPS or not.
11238 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11239 safely run with a non-FIPSed libcrypto, as it may crash because of
11240 the difference induced by this change.
11244 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11246 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11247 (part of SSL_OP_ALL). This option used to disable the
11248 countermeasure against man-in-the-middle protocol-version
11249 rollback in the SSL 2.0 server implementation, which is a bad
11250 idea. ([CVE-2005-2969])
11252 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11253 for Information Security, National Institute of Advanced Industrial
11254 Science and Technology [AIST, Japan)]*
11256 * Minimal support for X9.31 signatures and PSS padding modes. This is
11257 mainly for FIPS compliance and not fully integrated at this stage.
11261 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11262 the exponentiation using a fixed-length exponent. (Otherwise,
11263 the information leaked through timing could expose the secret key
11264 after many signatures; cf. Bleichenbacher's attack on DSA with
11269 * Make a new fixed-window mod_exp implementation the default for
11270 RSA, DSA, and DH private-key operations so that the sequence of
11271 squares and multiplies and the memory access pattern are
11272 independent of the particular secret key. This will mitigate
11273 cache-timing and potential related attacks.
11275 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11276 and this is automatically used by BN_mod_exp_mont() if the new flag
11277 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11278 will use this BN flag for private exponents unless the flag
11279 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11280 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11282 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11284 * Change the client implementation for SSLv23_method() and
11285 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11286 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11287 (Previously, the SSL 2.0 backwards compatible Client Hello
11288 message format would be used even with SSL_OP_NO_SSLv2.)
11292 * Add support for smime-type MIME parameter in S/MIME messages which some
11297 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11298 a threadsafe manner. Modify rsa code to use new function and add calls
11299 to dsa and dh code (which had race conditions before).
11303 * Include the fixed error library code in the C error file definitions
11304 instead of fixing them up at runtime. This keeps the error code
11305 structures constant.
11309 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11311 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11314 * Fixes for newer kerberos headers. NB: the casts are needed because
11315 the 'length' field is signed on one version and unsigned on another
11316 with no (?) obvious way to tell the difference, without these VC++
11317 complains. Also the "definition" of FAR (blank) is no longer included
11318 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11319 some needed definitions.
11323 * Undo Cygwin change.
11327 * Added support for proxy certificates according to RFC 3820.
11328 Because they may be a security thread to unaware applications,
11329 they must be explicitly allowed in run-time. See
11330 docs/HOWTO/proxy_certificates.txt for further information.
11334 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11336 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11337 server and client random values. Previously
11338 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11339 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11341 This change has negligible security impact because:
11343 1. Server and client random values still have 24 bytes of pseudo random
11346 2. Server and client random values are sent in the clear in the initial
11349 3. The master secret is derived using the premaster secret (48 bytes in
11350 size for static RSA ciphersuites) as well as client server and random
11353 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11356 *Stephen Henson, reported by UK NISCC*
11358 * Use Windows randomness collection on Cygwin.
11362 * Fix hang in EGD/PRNGD query when communication socket is closed
11363 prematurely by EGD/PRNGD.
11365 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11367 * Prompt for pass phrases when appropriate for PKCS12 input format.
11371 * Back-port of selected performance improvements from development
11372 branch, as well as improved support for PowerPC platforms.
11376 * Add lots of checks for memory allocation failure, error codes to indicate
11377 failure and freeing up memory if a failure occurs.
11379 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11381 * Add new -passin argument to dgst.
11385 * Perform some character comparisons of different types in X509_NAME_cmp:
11386 this is needed for some certificates that re-encode DNs into UTF8Strings
11387 (in violation of RFC3280) and can't or won't issue name rollover
11392 * Make an explicit check during certificate validation to see that
11393 the CA setting in each certificate on the chain is correct. As a
11394 side effect always do the following basic checks on extensions,
11395 not just when there's an associated purpose to the check:
11397 - if there is an unhandled critical extension (unless the user
11398 has chosen to ignore this fault)
11399 - if the path length has been exceeded (if one is set at all)
11400 - that certain extensions fit the associated purpose (if one has
11405 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11407 * Avoid a race condition when CRLs are checked in a multi threaded
11408 environment. This would happen due to the reordering of the revoked
11409 entries during signature checking and serial number lookup. Now the
11410 encoding is cached and the serial number sort performed under a lock.
11411 Add new STACK function sk_is_sorted().
11415 * Add Delta CRL to the extension code.
11419 * Various fixes to s3_pkt.c so alerts are sent properly.
11421 *David Holmes <d.holmes@f5.com>*
11423 * Reduce the chances of duplicate issuer name and serial numbers (in
11424 violation of RFC3280) using the OpenSSL certificate creation utilities.
11425 This is done by creating a random 64 bit value for the initial serial
11426 number when a serial number file is created or when a self signed
11427 certificate is created using 'openssl req -x509'. The initial serial
11428 number file is created using 'openssl x509 -next_serial' in CA.pl
11429 rather than being initialized to 1.
11433 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11435 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11436 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11438 *Joe Orton, Steve Henson*
11440 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11443 *Joe Orton, Steve Henson*
11445 * Make it possible to have multiple active certificates with the same
11446 subject in the CA index file. This is done only if the keyword
11447 'unique_subject' is set to 'no' in the main CA section (default
11448 if 'CA_default') of the configuration file. The value is saved
11449 with the database itself in a separate index attribute file,
11450 named like the index file with '.attr' appended to the name.
11454 * X509 verify fixes. Disable broken certificate workarounds when
11455 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11456 keyUsage extension present. Don't accept CRLs with unhandled critical
11457 extensions: since verify currently doesn't process CRL extensions this
11458 rejects a CRL with *any* critical extensions. Add new verify error codes
11463 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11464 A clarification of RFC2560 will require the use of OCTET STRINGs and
11465 some implementations cannot handle the current raw format. Since OpenSSL
11466 copies and compares OCSP nonces as opaque blobs without any attempt at
11467 parsing them this should not create any compatibility issues.
11471 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11472 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11473 this HMAC (and other) operations are several times slower than OpenSSL
11478 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11480 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11482 * Use the correct content when signing type "other".
11486 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11488 * Fix various bugs revealed by running the NISCC test suite:
11490 Stop out of bounds reads in the ASN1 code when presented with
11491 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11493 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11495 If verify callback ignores invalid public key errors don't try to check
11496 certificate signature with the NULL public key.
11500 * New -ignore_err option in ocsp application to stop the server
11501 exiting on the first error in a request.
11505 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11506 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11511 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11512 extra data after the compression methods not only for TLS 1.0
11513 but also for SSL 3.0 (as required by the specification).
11515 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11517 * Change X509_certificate_type() to mark the key as exported/exportable
11518 when it's 512 *bits* long, not 512 bytes.
11522 * Change AES_cbc_encrypt() so it outputs exact multiple of
11523 blocks during encryption.
11527 * Various fixes to base64 BIO and non blocking I/O. On write
11528 flushes were not handled properly if the BIO retried. On read
11529 data was not being buffered properly and had various logic bugs.
11530 This also affects blocking I/O when the data being decoded is a
11535 * Various S/MIME bugfixes and compatibility changes:
11536 output correct application/pkcs7 MIME type if
11537 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11538 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11539 of files as .eml work). Correctly handle very long lines in MIME
11544 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11546 * Countermeasure against the Klima-Pokorny-Rosa extension of
11547 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11548 a protocol version number mismatch like a decryption error
11549 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11553 * Turn on RSA blinding by default in the default implementation
11554 to avoid a timing attack. Applications that don't want it can call
11555 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11556 They would be ill-advised to do so in most cases.
11558 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11560 * Change RSA blinding code so that it works when the PRNG is not
11561 seeded (in this case, the secret RSA exponent is abused as
11562 an unpredictable seed -- if it is not unpredictable, there
11563 is no point in blinding anyway). Make RSA blinding thread-safe
11564 by remembering the creator's thread ID in rsa->blinding and
11565 having all other threads use local one-time blinding factors
11566 (this requires more computation than sharing rsa->blinding, but
11567 avoids excessive locking; and if an RSA object is not shared
11568 between threads, blinding will still be very fast).
11572 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11573 ENGINE as defaults for all supported algorithms irrespective of
11574 the 'flags' parameter. 'flags' is now honoured, so applications
11575 should make sure they are passing it correctly.
11579 * Target "mingw" now allows native Windows code to be generated in
11580 the Cygwin environment as well as with the MinGW compiler.
11584 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11586 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11587 via timing by performing a MAC computation even if incorrect
11588 block cipher padding has been found. This is a countermeasure
11589 against active attacks where the attacker has to distinguish
11590 between bad padding and a MAC verification error. ([CVE-2003-0078])
11592 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11593 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11594 Martin Vuagnoux (EPFL, Ilion)*
11596 * Make the no-err option work as intended. The intention with no-err
11597 is not to have the whole error stack handling routines removed from
11598 libcrypto, it's only intended to remove all the function name and
11599 reason texts, thereby removing some of the footprint that may not
11600 be interesting if those errors aren't displayed anyway.
11602 NOTE: it's still possible for any application or module to have its
11603 own set of error texts inserted. The routines are there, just not
11604 used by default when no-err is given.
11608 * Add support for FreeBSD on IA64.
11610 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11612 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11613 Kerberos function mit_des_cbc_cksum(). Before this change,
11614 the value returned by DES_cbc_cksum() was like the one from
11615 mit_des_cbc_cksum(), except the bytes were swapped.
11617 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11619 * Allow an application to disable the automatic SSL chain building.
11620 Before this a rather primitive chain build was always performed in
11621 ssl3_output_cert_chain(): an application had no way to send the
11622 correct chain if the automatic operation produced an incorrect result.
11624 Now the chain builder is disabled if either:
11626 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11628 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11630 The reasoning behind this is that an application would not want the
11631 auto chain building to take place if extra chain certificates are
11632 present and it might also want a means of sending no additional
11633 certificates (for example the chain has two certificates and the
11638 * Add the possibility to build without the ENGINE framework.
11640 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11642 * Under Win32 gmtime() can return NULL: check return value in
11643 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11647 * DSA routines: under certain error conditions uninitialized BN objects
11648 could be freed. Solution: make sure initialization is performed early
11649 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11650 Nils Larsch <nla@trustcenter.de> via PR#459)
11654 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11655 checked on reconnect on the client side, therefore session resumption
11656 could still fail with a "ssl session id is different" error. This
11657 behaviour is masked when SSL_OP_ALL is used due to
11658 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11659 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11660 followup to PR #377.
11664 * IA-32 assembler support enhancements: unified ELF targets, support
11665 for SCO/Caldera platforms, fix for Cygwin shared build.
11669 * Add support for FreeBSD on sparc64. As a consequence, support for
11670 FreeBSD on non-x86 processors is separate from x86 processors on
11671 the config script, much like the NetBSD support.
11673 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11675 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11677 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11680 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11681 code (06) was taken as the first octet of the session ID and the last
11682 octet was ignored consequently. As a result SSLv2 client side session
11683 caching could not have worked due to the session ID mismatch between
11685 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11690 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11691 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11696 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11697 seems that in spite of existing for more than a year, many application
11698 author have done nothing to provide the necessary callbacks, which
11699 means that this particular engine will not work properly anywhere.
11700 This is a very unfortunate situation which forces us, in the name
11701 of usability, to give the hw_ncipher.c a static lock, which is part
11703 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11704 appear in 0.9.8 or later. We EXPECT application authors to have
11705 dealt properly with this when 0.9.8 is released (unless we actually
11706 make such changes in the libcrypto locking code that changes will
11707 have to be made anyway).
11711 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11712 octets have been read, EOF or an error occurs. Without this change
11713 some truncated ASN1 structures will not produce an error.
11717 * Disable Heimdal support, since it hasn't been fully implemented.
11718 Still give the possibility to force the use of Heimdal, but with
11719 warnings and a request that patches get sent to openssl-dev.
11723 * Add the VC-CE target, introduce the WINCE sysname, and add
11724 INSTALL.WCE and appropriate conditionals to make it build.
11726 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11728 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11729 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11730 edit numbers of the version.
11732 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11734 * Introduce safe string copy and catenation functions
11735 (BUF_strlcpy() and BUF_strlcat()).
11737 *Ben Laurie (CHATS) and Richard Levitte*
11739 * Avoid using fixed-size buffers for one-line DNs.
11741 *Ben Laurie (CHATS)*
11743 * Add BUF_MEM_grow_clean() to avoid information leakage when
11744 resizing buffers containing secrets, and use where appropriate.
11746 *Ben Laurie (CHATS)*
11748 * Avoid using fixed size buffers for configuration file location.
11750 *Ben Laurie (CHATS)*
11752 * Avoid filename truncation for various CA files.
11754 *Ben Laurie (CHATS)*
11756 * Use sizeof in preference to magic numbers.
11758 *Ben Laurie (CHATS)*
11760 * Avoid filename truncation in cert requests.
11762 *Ben Laurie (CHATS)*
11764 * Add assertions to check for (supposedly impossible) buffer
11767 *Ben Laurie (CHATS)*
11769 * Don't cache truncated DNS entries in the local cache (this could
11770 potentially lead to a spoofing attack).
11772 *Ben Laurie (CHATS)*
11774 * Fix various buffers to be large enough for hex/decimal
11775 representations in a platform independent manner.
11777 *Ben Laurie (CHATS)*
11779 * Add CRYPTO_realloc_clean() to avoid information leakage when
11780 resizing buffers containing secrets, and use where appropriate.
11782 *Ben Laurie (CHATS)*
11784 * Add BIO_indent() to avoid much slightly worrying code to do
11787 *Ben Laurie (CHATS)*
11789 * Convert sprintf()/BIO_puts() to BIO_printf().
11791 *Ben Laurie (CHATS)*
11793 * buffer_gets() could terminate with the buffer only half
11796 *Ben Laurie (CHATS)*
11798 * Add assertions to prevent user-supplied crypto functions from
11799 overflowing internal buffers by having large block sizes, etc.
11801 *Ben Laurie (CHATS)*
11803 * New OPENSSL_assert() macro (similar to assert(), but enabled
11806 *Ben Laurie (CHATS)*
11808 * Eliminate unused copy of key in RC4.
11810 *Ben Laurie (CHATS)*
11812 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11814 *Ben Laurie (CHATS)*
11816 * Fix off-by-one error in EGD path.
11818 *Ben Laurie (CHATS)*
11820 * If RANDFILE path is too long, ignore instead of truncating.
11822 *Ben Laurie (CHATS)*
11824 * Eliminate unused and incorrectly sized X.509 structure
11827 *Ben Laurie (CHATS)*
11829 * Eliminate unused and dangerous function knumber().
11831 *Ben Laurie (CHATS)*
11833 * Eliminate unused and dangerous structure, KSSL_ERR.
11835 *Ben Laurie (CHATS)*
11837 * Protect against overlong session ID context length in an encoded
11838 session object. Since these are local, this does not appear to be
11841 *Ben Laurie (CHATS)*
11843 * Change from security patch (see 0.9.6e below) that did not affect
11844 the 0.9.6 release series:
11846 Remote buffer overflow in SSL3 protocol - an attacker could
11847 supply an oversized master key in Kerberos-enabled versions.
11850 *Ben Laurie (CHATS)*
11852 * Change the SSL kerb5 codes to match RFC 2712.
11856 * Make -nameopt work fully for req and add -reqopt switch.
11858 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11860 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11862 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11864 * Make sure tests can be performed even if the corresponding algorithms
11865 have been removed entirely. This was also the last step to make
11866 OpenSSL compilable with DJGPP under all reasonable conditions.
11868 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11870 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11871 to allow version independent disabling of normally unselected ciphers,
11872 which may be activated as a side-effect of selecting a single cipher.
11874 (E.g., cipher list string "RSA" enables ciphersuites that are left
11875 out of "ALL" because they do not provide symmetric encryption.
11876 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11878 *Lutz Jaenicke, Bodo Moeller*
11880 * Add appropriate support for separate platform-dependent build
11881 directories. The recommended way to make a platform-dependent
11882 build directory is the following (tested on Linux), maybe with
11885 # Place yourself outside of the OpenSSL source tree. In
11886 # this example, the environment variable OPENSSL_SOURCE
11887 # is assumed to contain the absolute OpenSSL source directory.
11888 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11889 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11890 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11891 mkdir -p `dirname $F`
11892 ln -s $OPENSSL_SOURCE/$F $F
11895 To be absolutely sure not to disturb the source tree, a "make clean"
11896 is a good thing. If it isn't successful, don't worry about it,
11897 it probably means the source directory is very clean.
11901 * Make sure any ENGINE control commands make local copies of string
11902 pointers passed to them whenever necessary. Otherwise it is possible
11903 the caller may have overwritten (or deallocated) the original string
11904 data when a later ENGINE operation tries to use the stored values.
11906 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11908 * Improve diagnostics in file reading and command-line digests.
11910 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11912 * Add AES modes CFB and OFB to the object database. Correct an
11913 error in AES-CFB decryption.
11917 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11918 allows existing EVP_CIPHER_CTX structures to be reused after
11919 calling `EVP_*Final()`. This behaviour is used by encryption
11920 BIOs and some applications. This has the side effect that
11921 applications must explicitly clean up cipher contexts with
11922 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11926 * Check the values of dna and dnb in bn_mul_recursive before calling
11927 bn_mul_comba (a non zero value means the a or b arrays do not contain
11928 n2 elements) and fallback to bn_mul_normal if either is not zero.
11932 * Fix escaping of non-ASCII characters when using the -subj option
11933 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11937 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11938 form for "surname", serialNumber has no short form.
11939 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11940 therefore remove "mail" short name for "internet 7".
11941 The OID for unique identifiers in X509 certificates is
11942 x500UniqueIdentifier, not uniqueIdentifier.
11943 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11947 * Add an "init" command to the ENGINE config module and auto initialize
11948 ENGINEs. Without any "init" command the ENGINE will be initialized
11949 after all ctrl commands have been executed on it. If init=1 the
11950 ENGINE is initialized at that point (ctrls before that point are run
11951 on the uninitialized ENGINE and after on the initialized one). If
11952 init=0 then the ENGINE will not be initialized at all.
11956 * Fix the 'app_verify_callback' interface so that the user-defined
11957 argument is actually passed to the callback: In the
11958 SSL_CTX_set_cert_verify_callback() prototype, the callback
11959 declaration has been changed from
11962 int (*cb)(X509_STORE_CTX *,void *);
11963 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11964 i=s->ctx->app_verify_callback(&ctx)
11965 has been changed into
11966 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11968 To update applications using SSL_CTX_set_cert_verify_callback(),
11969 a dummy argument can be added to their callback functions.
11971 *D. K. Smetters <smetters@parc.xerox.com>*
11973 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11975 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11977 * Add and OPENSSL_LOAD_CONF define which will cause
11978 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11979 This allows older applications to transparently support certain
11980 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11981 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11982 load the config file and OPENSSL_add_all_algorithms_conf() which will
11983 always load it have also been added.
11987 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11988 Adjust NIDs and EVP layer.
11990 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11992 * Config modules support in openssl utility.
11994 Most commands now load modules from the config file,
11995 though in a few (such as version) this isn't done
11996 because it couldn't be used for anything.
11998 In the case of ca and req the config file used is
11999 the same as the utility itself: that is the -config
12000 command line option can be used to specify an
12005 * Move default behaviour from OPENSSL_config(). If appname is NULL
12006 use "openssl_conf" if filename is NULL use default openssl config file.
12010 * Add an argument to OPENSSL_config() to allow the use of an alternative
12011 config section name. Add a new flag to tolerate a missing config file
12012 and move code to CONF_modules_load_file().
12016 * Support for crypto accelerator cards from Accelerated Encryption
12017 Processing, www.aep.ie. (Use engine 'aep')
12018 The support was copied from 0.9.6c [engine] and adapted/corrected
12019 to work with the new engine framework.
12021 *AEP Inc. and Richard Levitte*
12023 * Support for SureWare crypto accelerator cards from Baltimore
12024 Technologies. (Use engine 'sureware')
12025 The support was copied from 0.9.6c [engine] and adapted
12026 to work with the new engine framework.
12030 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12031 make the newer ENGINE framework commands for the CHIL engine work.
12033 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12035 * Make it possible to produce shared libraries on ReliantUNIX.
12037 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12039 * Add the configuration target debug-linux-ppro.
12040 Make 'openssl rsa' use the general key loading routines
12041 implemented in `apps.c`, and make those routines able to
12042 handle the key format FORMAT_NETSCAPE and the variant
12045 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12047 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12049 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12051 * Add -keyform to rsautl, and document -engine.
12053 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12055 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12056 BIO_R_NO_SUCH_FILE error code rather than the generic
12057 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12061 * Add new functions
12062 ERR_peek_last_error
12063 ERR_peek_last_error_line
12064 ERR_peek_last_error_line_data.
12065 These are similar to
12067 ERR_peek_error_line
12068 ERR_peek_error_line_data,
12069 but report on the latest error recorded rather than the first one
12070 still in the error queue.
12072 *Ben Laurie, Bodo Moeller*
12074 * default_algorithms option in ENGINE config module. This allows things
12076 default_algorithms = ALL
12077 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12081 * Preliminary ENGINE config module.
12085 * New experimental application configuration code.
12089 * Change the AES code to follow the same name structure as all other
12090 symmetric ciphers, and behave the same way. Move everything to
12091 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12093 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12095 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12097 *Ben Laurie and Theo de Raadt*
12099 * Add option to output public keys in req command.
12101 *Massimiliano Pala madwolf@openca.org*
12103 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12104 (up to about 10% better than before for P-192 and P-224).
12108 * New functions/macros
12110 SSL_CTX_set_msg_callback(ctx, cb)
12111 SSL_CTX_set_msg_callback_arg(ctx, arg)
12112 SSL_set_msg_callback(ssl, cb)
12113 SSL_set_msg_callback_arg(ssl, arg)
12115 to request calling a callback function
12117 void cb(int write_p, int version, int content_type,
12118 const void *buf, size_t len, SSL *ssl, void *arg)
12120 whenever a protocol message has been completely received
12121 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12122 protocol version according to which the SSL library interprets
12123 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12124 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12125 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12126 specification (change_cipher_spec(20), alert(21), handshake(22)).
12127 'buf' and 'len' point to the actual message, 'ssl' to the
12128 SSL object, and 'arg' is the application-defined value set by
12129 SSL[_CTX]_set_msg_callback_arg().
12131 'openssl s_client' and 'openssl s_server' have new '-msg' options
12132 to enable a callback that displays all protocol messages.
12136 * Change the shared library support so shared libraries are built as
12137 soon as the corresponding static library is finished, and thereby get
12138 openssl and the test programs linked against the shared library.
12139 This still only happens when the keyword "shard" has been given to
12140 the configuration scripts.
12142 NOTE: shared library support is still an experimental thing, and
12143 backward binary compatibility is still not guaranteed.
12145 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12147 * Add support for Subject Information Access extension.
12149 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12151 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12152 additional bytes when new memory had to be allocated, not just
12153 when reusing an existing buffer.
12157 * New command line and configuration option 'utf8' for the req command.
12158 This allows field values to be specified as UTF8 strings.
12162 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12163 runs for the former and machine-readable output for the latter.
12167 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12168 of the e-mail address in the DN (i.e., it will go into a certificate
12169 extension only). The new configuration file option 'email_in_dn = no'
12170 has the same effect.
12172 *Massimiliano Pala madwolf@openca.org*
12174 * Change all functions with names starting with `des_` to be starting
12175 with `DES_` instead. Add wrappers that are compatible with libdes,
12176 but are named `_ossl_old_des_*`. Finally, add macros that map the
12177 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12178 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12179 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12182 Since we provide two compatibility mappings, the user needs to
12183 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12184 compatibility is desired. The default (i.e., when that macro
12185 isn't defined) is OpenSSL 0.9.6c compatibility.
12187 There are also macros that enable and disable the support of old
12188 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12189 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12190 are defined, the default will apply: to support the old des routines.
12192 In either case, one must include openssl/des.h to get the correct
12193 definitions. Do not try to just include openssl/des_old.h, that
12196 NOTE: This is a major break of an old API into a new one. Software
12197 authors are encouraged to switch to the `DES_` style functions. Some
12198 time in the future, des_old.h and the libdes compatibility functions
12199 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12200 default), and then completely removed.
12204 * Test for certificates which contain unsupported critical extensions.
12205 If such a certificate is found during a verify operation it is
12206 rejected by default: this behaviour can be overridden by either
12207 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12208 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12209 X509_supported_extension() has also been added which returns 1 if a
12210 particular extension is supported.
12214 * Modify the behaviour of EVP cipher functions in similar way to digests
12215 to retain compatibility with existing code.
12219 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12220 compatibility with existing code. In particular the 'ctx' parameter does
12221 not have to be to be initialized before the call to EVP_DigestInit() and
12222 it is tidied up after a call to EVP_DigestFinal(). New function
12223 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12224 EVP_MD_CTX_copy() changed to not require the destination to be
12225 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12226 requires the destination to be valid.
12228 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12229 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12233 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12234 so that complete 'Handshake' protocol structures are kept in memory
12235 instead of overwriting 'msg_type' and 'length' with 'body' data.
12239 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12241 *Massimo Santin via Richard Levitte*
12243 * Major restructuring to the underlying ENGINE code. This includes
12244 reduction of linker bloat, separation of pure "ENGINE" manipulation
12245 (initialisation, etc) from functionality dealing with implementations
12246 of specific crypto interfaces. This change also introduces integrated
12247 support for symmetric ciphers and digest implementations - so ENGINEs
12248 can now accelerate these by providing EVP_CIPHER and EVP_MD
12249 implementations of their own. This is detailed in
12250 [crypto/engine/README.md](crypto/engine/README.md)
12251 as it couldn't be adequately described here. However, there are a few
12252 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12253 were changed in the original introduction of ENGINE code have now
12254 reverted back - the hooking from this code to ENGINE is now a good
12255 deal more passive and at run-time, operations deal directly with
12256 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12257 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12258 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12259 they were not being used by the framework as there is no concept of a
12260 BIGNUM_METHOD and they could not be generalised to the new
12261 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12262 ENGINE_cpy() has been removed as it cannot be consistently defined in
12267 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12271 * Change mkdef.pl to sort symbols that get the same entry number,
12272 and make sure the automatically generated functions `ERR_load_*`
12273 become part of libeay.num as well.
12277 * New function SSL_renegotiate_pending(). This returns true once
12278 renegotiation has been requested (either SSL_renegotiate() call
12279 or HelloRequest/ClientHello received from the peer) and becomes
12280 false once a handshake has been completed.
12281 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12282 sends a HelloRequest, but does not ensure that a handshake takes
12283 place. SSL_renegotiate_pending() is useful for checking if the
12284 client has followed the request.)
12288 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12289 By default, clients may request session resumption even during
12290 renegotiation (if session ID contexts permit); with this option,
12291 session resumption is possible only in the first handshake.
12293 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12294 more bits available for options that should not be part of
12295 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12299 * Add some demos for certificate and certificate request creation.
12303 * Make maximum certificate chain size accepted from the peer application
12304 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12305 "Douglas E. Engert" <deengert@anl.gov>.
12309 * Add support for shared libraries for Unixware-7
12310 (Boyd Lynn Gerber <gerberb@zenez.com>).
12314 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12315 be done prior to destruction. Use this to unload error strings from
12316 ENGINEs that load their own error strings. NB: This adds two new API
12317 functions to "get" and "set" this destroy handler in an ENGINE.
12321 * Alter all existing ENGINE implementations (except "openssl" and
12322 "openbsd") to dynamically instantiate their own error strings. This
12323 makes them more flexible to be built both as statically-linked ENGINEs
12324 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12325 Also, add stub code to each that makes building them as self-contained
12326 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12330 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12331 implementations into applications that are completely implemented in
12332 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12333 commands that can be used to configure what shared-library to load and
12334 to control aspects of the way it is handled. Also, made an update to
12335 the [README-Engine.md](README-Engine.md) file
12336 that brings its information up-to-date and
12337 provides some information and instructions on the "dynamic" ENGINE
12338 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12342 * Make it possible to unload ranges of ERR strings with a new
12343 "ERR_unload_strings" function.
12347 * Add a copy() function to EVP_MD.
12351 * Make EVP_MD routines take a context pointer instead of just the
12352 md_data void pointer.
12356 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12357 that the digest can only process a single chunk of data
12358 (typically because it is provided by a piece of
12359 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12360 is only going to provide a single chunk of data, and hence the
12361 framework needn't accumulate the data for oneshot drivers.
12365 * As with "ERR", make it possible to replace the underlying "ex_data"
12366 functions. This change also alters the storage and management of global
12367 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12368 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12369 index counters. The API functions that use this state have been changed
12370 to take a "class_index" rather than pointers to the class's local STACK
12371 and counter, and there is now an API function to dynamically create new
12372 classes. This centralisation allows us to (a) plug a lot of the
12373 thread-safety problems that existed, and (b) makes it possible to clean
12374 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12375 such data would previously have always leaked in application code and
12376 workarounds were in place to make the memory debugging turn a blind eye
12377 to it. Application code that doesn't use this new function will still
12378 leak as before, but their memory debugging output will announce it now
12379 rather than letting it slide.
12381 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12382 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12383 has a return value to indicate success or failure.
12387 * Make it possible to replace the underlying "ERR" functions such that the
12388 global state (2 LHASH tables and 2 locks) is only used by the "default"
12389 implementation. This change also adds two functions to "get" and "set"
12390 the implementation prior to it being automatically set the first time
12391 any other ERR function takes place. Ie. an application can call "get",
12392 pass the return value to a module it has just loaded, and that module
12393 can call its own "set" function using that value. This means the
12394 module's "ERR" operations will use (and modify) the error state in the
12395 application and not in its own statically linked copy of OpenSSL code.
12399 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12400 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12401 the operation, and provides a more encapsulated way for external code
12402 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12403 to use these functions rather than manually incrementing the counts.
12405 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12409 * Add EVP test program.
12413 * Add symmetric cipher support to ENGINE. Expect the API to change!
12417 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12418 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12419 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12420 These allow a CRL to be built without having to access X509_CRL fields
12421 directly. Modify 'ca' application to use new functions.
12425 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12426 bug workarounds. Rollback attack detection is a security feature.
12427 The problem will only arise on OpenSSL servers when TLSv1 is not
12428 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12429 Software authors not wanting to support TLSv1 will have special reasons
12430 for their choice and can explicitly enable this option.
12432 *Bodo Moeller, Lutz Jaenicke*
12434 * Rationalise EVP so it can be extended: don't include a union of
12435 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12436 (similar to those existing for EVP_CIPHER_CTX).
12441 EVP_MD_CTX_init(&md); /* new function call */
12442 EVP_DigestInit(&md, EVP_sha1());
12443 EVP_DigestUpdate(&md, in, len);
12444 EVP_DigestFinal(&md, out, NULL);
12445 EVP_MD_CTX_cleanup(&md); /* new function call */
12449 * Make DES key schedule conform to the usual scheme, as well as
12450 correcting its structure. This means that calls to DES functions
12451 now have to pass a pointer to a des_key_schedule instead of a
12452 plain des_key_schedule (which was actually always a pointer
12455 des_key_schedule ks;
12457 des_set_key_checked(..., &ks);
12458 des_ncbc_encrypt(..., &ks, ...);
12460 (Note that a later change renames 'des_...' into 'DES_...'.)
12464 * Initial reduction of linker bloat: the use of some functions, such as
12465 PEM causes large amounts of unused functions to be linked in due to
12466 poor organisation. For example pem_all.c contains every PEM function
12467 which has a knock on effect of linking in large amounts of (unused)
12468 ASN1 code. Grouping together similar functions and splitting unrelated
12469 functions prevents this.
12473 * Cleanup of EVP macros.
12477 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12478 correct `_ecb suffix`.
12482 * Add initial OCSP responder support to ocsp application. The
12483 revocation information is handled using the text based index
12484 use by the ca application. The responder can either handle
12485 requests generated internally, supplied in files (for example
12486 via a CGI script) or using an internal minimal server.
12490 * Add configuration choices to get zlib compression for TLS.
12494 * Changes to Kerberos SSL for RFC 2712 compliance:
12495 1. Implemented real KerberosWrapper, instead of just using
12496 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12497 2. Implemented optional authenticator field of KerberosWrapper.
12499 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12500 and authenticator structs; see crypto/krb5/.
12502 Generalized Kerberos calls to support multiple Kerberos libraries.
12503 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12504 via Richard Levitte*
12506 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12507 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12508 values for each of the key sizes rather than having just
12509 parameters (and 'speed' generating keys each time).
12513 * Speed up EVP routines.
12516 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12517 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12518 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12519 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12521 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12522 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12523 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12526 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12528 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12532 * Added the OS2-EMX target.
12534 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12536 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12537 New functions to support `NCONF` routines in extension code.
12538 New function `CONF_set_nconf()`
12539 to allow functions which take an `NCONF` to also handle the old `LHASH`
12540 structure: this means that the old `CONF` compatible routines can be
12541 retained (in particular w.rt. extensions) without having to duplicate the
12542 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12546 * Enhance the general user interface with mechanisms for inner control
12547 and with possibilities to have yes/no kind of prompts.
12551 * Change all calls to low-level digest routines in the library and
12552 applications to use EVP. Add missing calls to HMAC_cleanup() and
12553 don't assume HMAC_CTX can be copied using memcpy().
12555 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12557 * Add the possibility to control engines through control names but with
12558 arbitrary arguments instead of just a string.
12559 Change the key loaders to take a UI_METHOD instead of a callback
12560 function pointer. NOTE: this breaks binary compatibility with earlier
12561 versions of OpenSSL [engine].
12562 Adapt the nCipher code for these new conditions and add a card insertion
12567 * Enhance the general user interface with mechanisms to better support
12568 dialog box interfaces, application-defined prompts, the possibility
12569 to use defaults (for example default passwords from somewhere else)
12570 and interrupts/cancellations.
12574 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12575 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12579 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12580 tidy up some unnecessarily weird code in 'sk_new()').
12582 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12584 * Change the key loading routines for ENGINEs to use the same kind
12585 callback (pem_password_cb) as all other routines that need this
12590 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12591 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12592 than this minimum value is recommended.
12596 * New random seeder for OpenVMS, using the system process statistics
12597 that are easily reachable.
12601 * Windows apparently can't transparently handle global
12602 variables defined in DLLs. Initialisations such as:
12604 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12606 won't compile. This is used by the any applications that need to
12607 declare their own ASN1 modules. This was fixed by adding the option
12608 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12609 needed for static libraries under Win32.
12613 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12614 setting of purpose and trust fields. New X509_STORE trust and
12615 purpose functions and tidy up setting in other SSL functions.
12619 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12620 structure. These are inherited by X509_STORE_CTX when it is
12621 initialised. This allows various defaults to be set in the
12622 X509_STORE structure (such as flags for CRL checking and custom
12623 purpose or trust settings) for functions which only use X509_STORE_CTX
12624 internally such as S/MIME.
12626 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12627 trust settings if they are not set in X509_STORE. This allows X509_STORE
12628 purposes and trust (in S/MIME for example) to override any set by default.
12630 Add command line options for CRL checking to smime, s_client and s_server
12635 * Initial CRL based revocation checking. If the CRL checking flag(s)
12636 are set then the CRL is looked up in the X509_STORE structure and
12637 its validity and signature checked, then if the certificate is found
12638 in the CRL the verify fails with a revoked error.
12640 Various new CRL related callbacks added to X509_STORE_CTX structure.
12642 Command line options added to 'verify' application to support this.
12644 This needs some additional work, such as being able to handle multiple
12645 CRLs with different times, extension based lookup (rather than just
12646 by subject name) and ultimately more complete V2 CRL extension
12651 * Add a general user interface API (crypto/ui/). This is designed
12652 to replace things like des_read_password and friends (backward
12653 compatibility functions using this new API are provided).
12654 The purpose is to remove prompting functions from the DES code
12655 section as well as provide for prompting through dialog boxes in
12656 a window system and the like.
12660 * Add "ex_data" support to ENGINE so implementations can add state at a
12661 per-structure level rather than having to store it globally.
12665 * Make it possible for ENGINE structures to be copied when retrieved by
12666 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12667 This causes the "original" ENGINE structure to act like a template,
12668 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12669 operational state can be localised to each ENGINE structure, despite the
12670 fact they all share the same "methods". New ENGINE structures returned in
12671 this case have no functional references and the return value is the single
12672 structural reference. This matches the single structural reference returned
12673 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12678 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12679 needs to match any other type at all we need to manually clear the
12684 * Changes to the "openssl engine" utility to include;
12685 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12686 about an ENGINE's available control commands.
12687 - executing control commands from command line arguments using the
12688 '-pre' and '-post' switches. '-post' is only used if '-t' is
12689 specified and the ENGINE is successfully initialised. The syntax for
12690 the individual commands are colon-separated, for example;
12691 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12695 * New dynamic control command support for ENGINEs. ENGINEs can now
12696 declare their own commands (numbers), names (strings), descriptions,
12697 and input types for run-time discovery by calling applications. A
12698 subset of these commands are implicitly classed as "executable"
12699 depending on their input type, and only these can be invoked through
12700 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12701 can be based on user input, config files, etc). The distinction is
12702 that "executable" commands cannot return anything other than a boolean
12703 result and can only support numeric or string input, whereas some
12704 discoverable commands may only be for direct use through
12705 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12706 pointers, or other custom uses. The "executable" commands are to
12707 support parameterisations of ENGINE behaviour that can be
12708 unambiguously defined by ENGINEs and used consistently across any
12709 OpenSSL-based application. Commands have been added to all the
12710 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12711 control over shared-library paths without source code alterations.
12715 * Changed all ENGINE implementations to dynamically allocate their
12716 ENGINEs rather than declaring them statically. Apart from this being
12717 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12718 this also allows the implementations to compile without using the
12719 internal engine_int.h header.
12723 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12724 'const' value. Any code that should be able to modify a RAND_METHOD
12725 should already have non-const pointers to it (ie. they should only
12726 modify their own ones).
12730 * Made a variety of little tweaks to the ENGINE code.
12731 - "atalla" and "ubsec" string definitions were moved from header files
12732 to C code. "nuron" string definitions were placed in variables
12733 rather than hard-coded - allowing parameterisation of these values
12734 later on via ctrl() commands.
12735 - Removed unused "#if 0"'d code.
12736 - Fixed engine list iteration code so it uses ENGINE_free() to release
12737 structural references.
12738 - Constified the RAND_METHOD element of ENGINE structures.
12739 - Constified various get/set functions as appropriate and added
12740 missing functions (including a catch-all ENGINE_cpy that duplicates
12741 all ENGINE values onto a new ENGINE except reference counts/state).
12742 - Removed NULL parameter checks in get/set functions. Setting a method
12743 or function to NULL is a way of cancelling out a previously set
12744 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12745 and doesn't justify the extra error symbols and code.
12746 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12747 flags from engine_int.h to engine.h.
12748 - Changed prototypes for ENGINE handler functions (init(), finish(),
12749 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12753 * Implement binary inversion algorithm for BN_mod_inverse in addition
12754 to the algorithm using long division. The binary algorithm can be
12755 used only if the modulus is odd. On 32-bit systems, it is faster
12756 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12757 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12758 up to 450 bits. In 64-bit environments, the binary algorithm
12759 appears to be advantageous for much longer moduli; here we use it
12760 for moduli up to 2048 bits.
12764 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12765 could not support the combine flag in choice fields.
12769 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12770 extensions from a certificate request to the certificate.
12774 * Allow multiple 'certopt' and 'nameopt' options to be separated
12775 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12776 file: this allows the display of the certificate about to be
12777 signed to be customised, to allow certain fields to be included
12778 or excluded and extension details. The old system didn't display
12779 multicharacter strings properly, omitted fields not in the policy
12780 and couldn't display additional details such as extensions.
12784 * Function EC_POINTs_mul for multiple scalar multiplication
12785 of an arbitrary number of elliptic curve points
12786 \sum scalars[i]*points[i],
12787 optionally including the generator defined for the EC_GROUP:
12788 scalar*generator + \sum scalars[i]*points[i].
12790 EC_POINT_mul is a simple wrapper function for the typical case
12791 that the point list has just one item (besides the optional
12796 * First EC_METHODs for curves over GF(p):
12798 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12799 operations and provides various method functions that can also
12800 operate with faster implementations of modular arithmetic.
12802 EC_GFp_mont_method() reuses most functions that are part of
12803 EC_GFp_simple_method, but uses Montgomery arithmetic.
12805 *Bodo Moeller; point addition and point doubling
12806 implementation directly derived from source code provided by
12807 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12809 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12810 crypto/ec/ec_lib.c):
12812 Curves are EC_GROUP objects (with an optional group generator)
12813 based on EC_METHODs that are built into the library.
12815 Points are EC_POINT objects based on EC_GROUP objects.
12817 Most of the framework would be able to handle curves over arbitrary
12818 finite fields, but as there are no obvious types for fields other
12819 than GF(p), some functions are limited to that for now.
12823 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12824 that the file contains a complete HTTP response.
12828 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12829 change the def and num file printf format specifier from "%-40sXXX"
12830 to "%-39s XXX". The latter will always guarantee a space after the
12831 field while the former will cause them to run together if the field
12832 is 40 of more characters long.
12836 * Constify the cipher and digest 'method' functions and structures
12837 and modify related functions to take constant EVP_MD and EVP_CIPHER
12842 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12843 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12847 * Modify `EVP_Digest*()` routines so they now return values. Although the
12848 internal software routines can never fail additional hardware versions
12853 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12855 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12856 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12859 ERR_R_NESTED_ASN1_ERROR
12861 ERR_R_MISSING_ASN1_EOS
12862 were 4 .. 9, conflicting with
12863 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12865 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12866 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12868 Add new error code 'ERR_R_INTERNAL_ERROR'.
12872 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12877 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12878 sets the subject name for a new request or supersedes the
12879 subject name in a given request. Formats that can be parsed are
12880 'CN=Some Name, OU=myOU, C=IT'
12882 'CN=Some Name/OU=myOU/C=IT'.
12884 Add options '-batch' and '-verbose' to 'openssl req'.
12886 *Massimiliano Pala <madwolf@hackmasters.net>*
12888 * Introduce the possibility to access global variables through
12889 functions on platform were that's the best way to handle exporting
12890 global variables in shared libraries. To enable this functionality,
12891 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12892 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12893 is normally done by Configure or something similar).
12895 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12896 in the source file (foo.c) like this:
12898 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12899 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12901 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12902 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12904 OPENSSL_DECLARE_GLOBAL(int,foo);
12905 #define foo OPENSSL_GLOBAL_REF(foo)
12906 OPENSSL_DECLARE_GLOBAL(double,bar);
12907 #define bar OPENSSL_GLOBAL_REF(bar)
12909 The #defines are very important, and therefore so is including the
12910 header file everywhere where the defined globals are used.
12912 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12913 of ASN.1 items, but that structure is a bit different.
12915 The largest change is in util/mkdef.pl which has been enhanced with
12916 better and easier to understand logic to choose which symbols should
12917 go into the Windows .def files as well as a number of fixes and code
12918 cleanup (among others, algorithm keywords are now sorted
12919 lexicographically to avoid constant rewrites).
12923 * In BN_div() keep a copy of the sign of 'num' before writing the
12924 result to 'rm' because if rm==num the value will be overwritten
12925 and produce the wrong result if 'num' is negative: this caused
12926 problems with BN_mod() and BN_nnmod().
12930 * Function OCSP_request_verify(). This checks the signature on an
12931 OCSP request and verifies the signer certificate. The signer
12932 certificate is just checked for a generic purpose and OCSP request
12937 * Add OCSP_check_validity() function to check the validity of OCSP
12938 responses. OCSP responses are prepared in real time and may only
12939 be a few seconds old. Simply checking that the current time lies
12940 between thisUpdate and nextUpdate max reject otherwise valid responses
12941 caused by either OCSP responder or client clock inaccuracy. Instead
12942 we allow thisUpdate and nextUpdate to fall within a certain period of
12943 the current time. The age of the response can also optionally be
12944 checked. Two new options -validity_period and -status_age added to
12949 * If signature or public key algorithm is unrecognized print out its
12950 OID rather that just UNKNOWN.
12954 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12955 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12956 ID to be generated from the issuer certificate alone which can then be
12957 passed to OCSP_id_issuer_cmp().
12961 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12962 ASN1 modules to export functions returning ASN1_ITEM pointers
12963 instead of the ASN1_ITEM structures themselves. This adds several
12964 new macros which allow the underlying ASN1 function/structure to
12965 be accessed transparently. As a result code should not use ASN1_ITEM
12966 references directly (such as &X509_it) but instead use the relevant
12967 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12968 use of the new ASN1 code on platforms where exporting structures
12969 is problematical (for example in shared libraries) but exporting
12970 functions returning pointers to structures is not.
12974 * Add support for overriding the generation of SSL/TLS session IDs.
12975 These callbacks can be registered either in an SSL_CTX or per SSL.
12976 The purpose of this is to allow applications to control, if they wish,
12977 the arbitrary values chosen for use as session IDs, particularly as it
12978 can be useful for session caching in multiple-server environments. A
12979 command-line switch for testing this (and any client code that wishes
12980 to use such a feature) has been added to "s_server".
12982 *Geoff Thorpe, Lutz Jaenicke*
12984 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12985 of the form `#if defined(...) || defined(...) || ...` and
12986 `#if !defined(...) && !defined(...) && ...`. This also avoids
12987 the growing number of special cases it was previously handling.
12991 * Make all configuration macros available for application by making
12992 sure they are available in opensslconf.h, by giving them names starting
12993 with `OPENSSL_` to avoid conflicts with other packages and by making
12994 sure e_os2.h will cover all platform-specific cases together with
12996 Additionally, it is now possible to define configuration/platform-
12997 specific names (called "system identities"). In the C code, these
12998 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12999 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13000 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13005 * New option -set_serial to 'req' and 'x509' this allows the serial
13006 number to use to be specified on the command line. Previously self
13007 signed certificates were hard coded with serial number 0 and the
13008 CA options of 'x509' had to use a serial number in a file which was
13013 * New options to 'ca' utility to support V2 CRL entry extensions.
13014 Currently CRL reason, invalidity date and hold instruction are
13015 supported. Add new CRL extensions to V3 code and some new objects.
13019 * New function EVP_CIPHER_CTX_set_padding() this is used to
13020 disable standard block padding (aka PKCS#5 padding) in the EVP
13021 API, which was previously mandatory. This means that the data is
13022 not padded in any way and so the total length much be a multiple
13023 of the block size, otherwise an error occurs.
13027 * Initial (incomplete) OCSP SSL support.
13031 * New function OCSP_parse_url(). This splits up a URL into its host,
13032 port and path components: primarily to parse OCSP URLs. New -url
13033 option to ocsp utility.
13037 * New nonce behavior. The return value of OCSP_check_nonce() now
13038 reflects the various checks performed. Applications can decide
13039 whether to tolerate certain situations such as an absent nonce
13040 in a response when one was present in a request: the ocsp application
13041 just prints out a warning. New function OCSP_add1_basic_nonce()
13042 this is to allow responders to include a nonce in a response even if
13043 the request is nonce-less.
13047 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13048 skipped when using openssl x509 multiple times on a single input file,
13049 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13053 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13054 set string type: to handle setting ASN1_TIME structures. Fix ca
13055 utility to correctly initialize revocation date of CRLs.
13059 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13060 the clients preferred ciphersuites and rather use its own preferences.
13061 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13062 Internet Explorer by ensuring unchanged hash method during stepup.
13063 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13067 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13068 to aes and add a new 'exist' option to print out symbols that don't
13073 * Additional options to ocsp utility to allow flags to be set and
13074 additional certificates supplied.
13078 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13079 OCSP client a number of certificate to only verify the response
13084 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13085 handle the new API. Currently only ECB, CBC modes supported. Add new
13088 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13089 Encryption Standard (AES) Ciphersuites for Transport Layer
13090 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13091 not enabled by default and were not part of the "ALL" ciphersuite
13092 alias because they were not yet official; they could be
13093 explicitly requested by specifying the "AESdraft" ciphersuite
13094 group alias. In the final release of OpenSSL 0.9.7, the group
13095 alias is called "AES" and is part of "ALL".)
13097 *Ben Laurie, Steve Henson, Bodo Moeller*
13099 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13100 request to response.
13104 * Functions for OCSP responders. OCSP_request_onereq_count(),
13105 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13106 extract information from a certificate request. OCSP_response_create()
13107 creates a response and optionally adds a basic response structure.
13108 OCSP_basic_add1_status() adds a complete single response to a basic
13109 response and returns the OCSP_SINGLERESP structure just added (to allow
13110 extensions to be included for example). OCSP_basic_add1_cert() adds a
13111 certificate to a basic response and OCSP_basic_sign() signs a basic
13112 response with various flags. New helper functions ASN1_TIME_check()
13113 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13114 (converts ASN1_TIME to GeneralizedTime).
13118 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13119 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13120 structure from a certificate. X509_pubkey_digest() digests the public_key
13121 contents: this is used in various key identifiers.
13125 * Make sk_sort() tolerate a NULL argument.
13127 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13129 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13130 passed by the function are trusted implicitly. If any of them signed the
13131 response then it is assumed to be valid and is not verified.
13135 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13136 to data. This was previously part of the PKCS7 ASN1 code. This
13137 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13138 *Steve Henson, reported by Kenneth R. Robinette
13139 <support@securenetterm.com>*
13141 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13142 routines: without these tracing memory leaks is very painful.
13143 Fix leaks in PKCS12 and PKCS7 routines.
13147 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13148 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13149 effectively meant GeneralizedTime would never be used. Now it
13150 is initialised to -1 but X509_time_adj() now has to check the value
13151 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13152 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13153 *Steve Henson, reported by Kenneth R. Robinette
13154 <support@securenetterm.com>*
13156 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13157 result in a zero length in the ASN1_INTEGER structure which was
13158 not consistent with the structure when d2i_ASN1_INTEGER() was used
13159 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13160 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13161 where it did not print out a minus for negative ASN1_INTEGER.
13165 * Add summary printout to ocsp utility. The various functions which
13166 convert status values to strings have been renamed to:
13167 OCSP_response_status_str(), OCSP_cert_status_str() and
13168 OCSP_crl_reason_str() and are no longer static. New options
13169 to verify nonce values and to disable verification. OCSP response
13170 printout format cleaned up.
13174 * Add additional OCSP certificate checks. These are those specified
13175 in RFC2560. This consists of two separate checks: the CA of the
13176 certificate being checked must either be the OCSP signer certificate
13177 or the issuer of the OCSP signer certificate. In the latter case the
13178 OCSP signer certificate must contain the OCSP signing extended key
13179 usage. This check is performed by attempting to match the OCSP
13180 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13181 in the OCSP_CERTID structures of the response.
13185 * Initial OCSP certificate verification added to OCSP_basic_verify()
13186 and related routines. This uses the standard OpenSSL certificate
13187 verify routines to perform initial checks (just CA validity) and
13188 to obtain the certificate chain. Then additional checks will be
13189 performed on the chain. Currently the root CA is checked to see
13190 if it is explicitly trusted for OCSP signing. This is used to set
13191 a root CA as a global signing root: that is any certificate that
13192 chains to that CA is an acceptable OCSP signing certificate.
13196 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13197 extensions from a separate configuration file.
13198 As when reading extensions from the main configuration file,
13199 the '-extensions ...' option may be used for specifying the
13202 *Massimiliano Pala <madwolf@comune.modena.it>*
13204 * New OCSP utility. Allows OCSP requests to be generated or
13205 read. The request can be sent to a responder and the output
13206 parsed, outputted or printed in text form. Not complete yet:
13207 still needs to check the OCSP response validity.
13211 * New subcommands for 'openssl ca':
13212 `openssl ca -status <serial>` prints the status of the cert with
13213 the given serial number (according to the index file).
13214 `openssl ca -updatedb` updates the expiry status of certificates
13217 *Massimiliano Pala <madwolf@comune.modena.it>*
13219 * New '-newreq-nodes' command option to CA.pl. This is like
13220 '-newreq', but calls 'openssl req' with the '-nodes' option
13221 so that the resulting key is not encrypted.
13223 *Damien Miller <djm@mindrot.org>*
13225 * New configuration for the GNU Hurd.
13227 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13229 * Initial code to implement OCSP basic response verify. This
13230 is currently incomplete. Currently just finds the signer's
13231 certificate and verifies the signature on the response.
13235 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13236 value of OPENSSLDIR. This is available via the new '-d' option
13237 to 'openssl version', and is also included in 'openssl version -a'.
13241 * Allowing defining memory allocation callbacks that will be given
13242 file name and line number information in additional arguments
13243 (a `const char*` and an int). The basic functionality remains, as
13244 well as the original possibility to just replace malloc(),
13245 realloc() and free() by functions that do not know about these
13246 additional arguments. To register and find out the current
13247 settings for extended allocation functions, the following
13248 functions are provided:
13250 CRYPTO_set_mem_ex_functions
13251 CRYPTO_set_locked_mem_ex_functions
13252 CRYPTO_get_mem_ex_functions
13253 CRYPTO_get_locked_mem_ex_functions
13255 These work the same way as CRYPTO_set_mem_functions and friends.
13256 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13257 extended allocation function is enabled.
13258 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13259 a conventional allocation function is enabled.
13261 *Richard Levitte, Bodo Moeller*
13263 * Finish off removing the remaining LHASH function pointer casts.
13264 There should no longer be any prototype-casting required when using
13265 the LHASH abstraction, and any casts that remain are "bugs". See
13266 the callback types and macros at the head of lhash.h for details
13267 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13271 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13272 If /dev/[u]random devices are not available or do not return enough
13273 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13275 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13276 /etc/entropy will be queried once each in this sequence, querying stops
13277 when enough entropy was collected without querying more sockets.
13281 * Change the Unix RAND_poll() variant to be able to poll several
13282 random devices, as specified by DEVRANDOM, until a sufficient amount
13283 of data has been collected. We spend at most 10 ms on each file
13284 (select timeout) and read in non-blocking mode. DEVRANDOM now
13285 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13286 (previously it was just the string "/dev/urandom"), so on typical
13287 platforms the 10 ms delay will never occur.
13288 Also separate out the Unix variant to its own file, rand_unix.c.
13289 For VMS, there's a currently-empty rand_vms.c.
13293 * Move OCSP client related routines to ocsp_cl.c. These
13294 provide utility functions which an application needing
13295 to issue a request to an OCSP responder and analyse the
13296 response will typically need: as opposed to those which an
13297 OCSP responder itself would need which will be added later.
13299 OCSP_request_sign() signs an OCSP request with an API similar
13300 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13301 response. OCSP_response_get1_basic() extracts basic response
13302 from response. OCSP_resp_find_status(): finds and extracts status
13303 information from an OCSP_CERTID structure (which will be created
13304 when the request structure is built). These are built from lower
13305 level functions which work on OCSP_SINGLERESP structures but
13306 won't normally be used unless the application wishes to examine
13307 extensions in the OCSP response for example.
13309 Replace nonce routines with a pair of functions.
13310 OCSP_request_add1_nonce() adds a nonce value and optionally
13311 generates a random value. OCSP_check_nonce() checks the
13312 validity of the nonce in an OCSP response.
13316 * Change function OCSP_request_add() to OCSP_request_add0_id().
13317 This doesn't copy the supplied OCSP_CERTID and avoids the
13318 need to free up the newly created id. Change return type
13319 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13320 This can then be used to add extensions to the request.
13321 Deleted OCSP_request_new(), since most of its functionality
13322 is now in OCSP_REQUEST_new() (and the case insensitive name
13323 clash) apart from the ability to set the request name which
13324 will be added elsewhere.
13328 * Update OCSP API. Remove obsolete extensions argument from
13329 various functions. Extensions are now handled using the new
13330 OCSP extension code. New simple OCSP HTTP function which
13331 can be used to send requests and parse the response.
13335 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13336 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13337 uses the special reorder version of SET OF to sort the attributes
13338 and reorder them to match the encoded order. This resolves a long
13339 standing problem: a verify on a PKCS7 structure just after signing
13340 it used to fail because the attribute order did not match the
13341 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13342 it uses the received order. This is necessary to tolerate some broken
13343 software that does not order SET OF. This is handled by encoding
13344 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13345 to produce the required SET OF.
13349 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13350 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13351 files to get correct declarations of the ASN.1 item variables.
13355 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13356 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13357 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13358 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13359 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13360 ASN1_ITEM and no wrapper functions.
13364 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13365 replace the old function pointer based I/O routines. Change most of
13366 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13370 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13371 lines, recognize more "algorithms" that can be deselected, and make
13372 it complain about algorithm deselection that isn't recognised.
13376 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13377 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13378 to use new functions. Add NO_ASN1_OLD which can be set to remove
13379 some old style ASN1 functions: this can be used to determine if old
13380 code will still work when these eventually go away.
13384 * New extension functions for OCSP structures, these follow the
13385 same conventions as certificates and CRLs.
13389 * New function X509V3_add1_i2d(). This automatically encodes and
13390 adds an extension. Its behaviour can be customised with various
13391 flags to append, replace or delete. Various wrappers added for
13392 certificates and CRLs.
13396 * Fix to avoid calling the underlying ASN1 print routine when
13397 an extension cannot be parsed. Correct a typo in the
13398 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13402 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13403 entries for variables.
13407 * Add functionality to `apps/openssl.c` for detecting locking
13408 problems: As the program is single-threaded, all we have
13409 to do is register a locking callback using an array for
13410 storing which locks are currently held by the program.
13414 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13415 SSL_get_ex_data_X509_STORE_idx(), which is used in
13416 ssl_verify_cert_chain() and thus can be called at any time
13417 during TLS/SSL handshakes so that thread-safety is essential.
13418 Unfortunately, the ex_data design is not at all suited
13419 for multi-threaded use, so it probably should be abolished.
13423 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13425 *Broadcom, tweaked and integrated by Geoff Thorpe*
13427 * Move common extension printing code to new function
13428 X509V3_print_extensions(). Reorganise OCSP print routines and
13429 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13433 * New function X509_signature_print() to remove duplication in some
13438 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13439 set (this was treated exactly the same as SET OF previously). This
13440 is used to reorder the STACK representing the structure to match the
13441 encoding. This will be used to get round a problem where a PKCS7
13442 structure which was signed could not be verified because the STACK
13443 order did not reflect the encoded order.
13447 * Reimplement the OCSP ASN1 module using the new code.
13451 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13452 for its ASN1 operations. The old style function pointers still exist
13453 for now but they will eventually go away.
13457 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13458 completely replaces the old ASN1 functionality with a table driven
13459 encoder and decoder which interprets an ASN1_ITEM structure describing
13460 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13461 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13462 has also been converted to the new form.
13466 * Change BN_mod_exp_recp so that negative moduli are tolerated
13467 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13468 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13469 for negative moduli.
13473 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13474 of not touching the result's sign bit.
13478 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13483 * Changed the LHASH code to use prototypes for callbacks, and created
13484 macros to declare and implement thin (optionally static) functions
13485 that provide type-safety and avoid function pointer casting for the
13486 type-specific callbacks.
13490 * Added Kerberos Cipher Suites to be used with TLS, as written in
13492 *Veers Staats <staatsvr@asc.hpc.mil>,
13493 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13495 * Reformat the FAQ so the different questions and answers can be divided
13496 in sections depending on the subject.
13500 * Have the zlib compression code load ZLIB.DLL dynamically under
13505 * New function BN_mod_sqrt for computing square roots modulo a prime
13506 (using the probabilistic Tonelli-Shanks algorithm unless
13507 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13508 be handled deterministically).
13510 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13512 * Make BN_mod_inverse faster by explicitly handling small quotients
13513 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13514 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13518 * New function BN_kronecker.
13522 * Fix BN_gcd so that it works on negative inputs; the result is
13523 positive unless both parameters are zero.
13524 Previously something reasonably close to an infinite loop was
13525 possible because numbers could be growing instead of shrinking
13526 in the implementation of Euclid's algorithm.
13530 * Fix BN_is_word() and BN_is_one() macros to take into account the
13531 sign of the number in question.
13533 Fix BN_is_word(a,w) to work correctly for w == 0.
13535 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13536 because its test if the absolute value of 'a' equals 'w'.
13537 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13538 it exists mostly for use in the implementations of BN_is_zero(),
13539 BN_is_one(), and BN_is_word().
13543 * New function BN_swap.
13547 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13548 the exponentiation functions are more likely to produce reasonable
13549 results on negative inputs.
13553 * Change BN_mod_mul so that the result is always non-negative.
13554 Previously, it could be negative if one of the factors was negative;
13555 I don't think anyone really wanted that behaviour.
13559 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13560 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13561 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13562 and add new functions:
13571 BN_mod_lshift1_quick
13573 BN_mod_lshift_quick
13575 These functions always generate non-negative results.
13577 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13578 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13580 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13581 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13582 be reduced modulo `m`.
13584 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13587 The following entry accidentally appeared in the CHANGES file
13588 distributed with OpenSSL 0.9.7. The modifications described in
13589 it do *not* apply to OpenSSL 0.9.7.
13591 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13592 was actually never needed) and in BN_mul(). The removal in BN_mul()
13593 required a small change in bn_mul_part_recursive() and the addition
13594 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13595 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13596 bn_sub_words() and bn_add_words() except they take arrays with
13602 * In 'openssl passwd', verify passwords read from the terminal
13603 unless the '-salt' option is used (which usually means that
13604 verification would just waste user's time since the resulting
13605 hash is going to be compared with some given password hash)
13606 or the new '-noverify' option is used.
13608 This is an incompatible change, but it does not affect
13609 non-interactive use of 'openssl passwd' (passwords on the command
13610 line, '-stdin' option, '-in ...' option) and thus should not
13611 cause any problems.
13615 * Remove all references to RSAref, since there's no more need for it.
13619 * Make DSO load along a path given through an environment variable
13620 (SHLIB_PATH) with shl_load().
13624 * Constify the ENGINE code as a result of BIGNUM constification.
13625 Also constify the RSA code and most things related to it. In a
13626 few places, most notable in the depth of the ASN.1 code, ugly
13627 casts back to non-const were required (to be solved at a later
13632 * Make it so the openssl application has all engines loaded by default.
13636 * Constify the BIGNUM routines a little more.
13640 * Add the following functions:
13642 ENGINE_load_cswift()
13644 ENGINE_load_atalla()
13645 ENGINE_load_nuron()
13646 ENGINE_load_builtin_engines()
13648 That way, an application can itself choose if external engines that
13649 are built-in in OpenSSL shall ever be used or not. The benefit is
13650 that applications won't have to be linked with libdl or other dso
13651 libraries unless it's really needed.
13653 Changed 'openssl engine' to load all engines on demand.
13654 Changed the engine header files to avoid the duplication of some
13655 declarations (they differed!).
13659 * 'openssl engine' can now list capabilities.
13663 * Better error reporting in 'openssl engine'.
13667 * Never call load_dh_param(NULL) in s_server.
13671 * Add engine application. It can currently list engines by name and
13672 identity, and test if they are actually available.
13676 * Improve RPM specification file by forcing symbolic linking and making
13677 sure the installed documentation is also owned by root.root.
13679 *Damien Miller <djm@mindrot.org>*
13681 * Give the OpenSSL applications more possibilities to make use of
13682 keys (public as well as private) handled by engines.
13686 * Add OCSP code that comes from CertCo.
13690 * Add VMS support for the Rijndael code.
13694 * Added untested support for Nuron crypto accelerator.
13698 * Add support for external cryptographic devices. This code was
13699 previously distributed separately as the "engine" branch.
13701 *Geoff Thorpe, Richard Levitte*
13703 * Rework the filename-translation in the DSO code. It is now possible to
13704 have far greater control over how a "name" is turned into a filename
13705 depending on the operating environment and any oddities about the
13706 different shared library filenames on each system.
13710 * Support threads on FreeBSD-elf in Configure.
13714 * Fix for SHA1 assembly problem with MASM: it produces
13715 warnings about corrupt line number information when assembling
13716 with debugging information. This is caused by the overlapping
13719 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13722 NCONF_get_number() has no error checking at all. As a replacement,
13723 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13724 promoted strongly. The old NCONF_get_number is kept around for
13725 binary backward compatibility.
13726 Make it possible for methods to load from something other than a BIO,
13727 by providing a function pointer that is given a name instead of a BIO.
13728 For example, this could be used to load configuration data from an
13733 * Fix for non blocking accept BIOs. Added new I/O special reason
13734 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13735 with non blocking I/O was not possible because no retry code was
13736 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13741 * Added the beginnings of Rijndael support.
13745 * Fix for bug in DirectoryString mask setting. Add support for
13746 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13747 to allow certificate printing to more controllable, additional
13748 'certopt' option to 'x509' to allow new printing options to be
13753 * Clean old EAY MD5 hack from e_os.h.
13757 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13759 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13760 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13762 *Joe Orton, Steve Henson*
13764 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13766 * Fix additional bug revealed by the NISCC test suite:
13768 Stop bug triggering large recursion when presented with
13769 certain ASN.1 tags ([CVE-2003-0851])
13773 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13775 * Fix various bugs revealed by running the NISCC test suite:
13777 Stop out of bounds reads in the ASN1 code when presented with
13778 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13780 If verify callback ignores invalid public key errors don't try to check
13781 certificate signature with the NULL public key.
13785 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13786 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13791 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13792 extra data after the compression methods not only for TLS 1.0
13793 but also for SSL 3.0 (as required by the specification).
13795 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13797 * Change X509_certificate_type() to mark the key as exported/exportable
13798 when it's 512 *bits* long, not 512 bytes.
13802 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13804 * Countermeasure against the Klima-Pokorny-Rosa extension of
13805 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13806 a protocol version number mismatch like a decryption error
13807 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13811 * Turn on RSA blinding by default in the default implementation
13812 to avoid a timing attack. Applications that don't want it can call
13813 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13814 They would be ill-advised to do so in most cases.
13816 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13818 * Change RSA blinding code so that it works when the PRNG is not
13819 seeded (in this case, the secret RSA exponent is abused as
13820 an unpredictable seed -- if it is not unpredictable, there
13821 is no point in blinding anyway). Make RSA blinding thread-safe
13822 by remembering the creator's thread ID in rsa->blinding and
13823 having all other threads use local one-time blinding factors
13824 (this requires more computation than sharing rsa->blinding, but
13825 avoids excessive locking; and if an RSA object is not shared
13826 between threads, blinding will still be very fast).
13830 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13832 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13833 via timing by performing a MAC computation even if incorrect
13834 block cipher padding has been found. This is a countermeasure
13835 against active attacks where the attacker has to distinguish
13836 between bad padding and a MAC verification error. ([CVE-2003-0078])
13838 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13839 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13840 Martin Vuagnoux (EPFL, Ilion)*
13842 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13844 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13845 memory from its contents. This is done with a counter that will
13846 place alternating values in each byte. This can be used to solve
13847 two issues: 1) the removal of calls to memset() by highly optimizing
13848 compilers, and 2) cleansing with other values than 0, since those can
13849 be read through on certain media, for example a swap space on disk.
13853 * Bugfix: client side session caching did not work with external caching,
13854 because the session->cipher setting was not restored when reloading
13855 from the external cache. This problem was masked, when
13856 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13857 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13861 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13862 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13864 *Zeev Lieber <zeev-l@yahoo.com>*
13866 * Undo an undocumented change introduced in 0.9.6e which caused
13867 repeated calls to OpenSSL_add_all_ciphers() and
13868 OpenSSL_add_all_digests() to be ignored, even after calling
13873 * Change the default configuration reader to deal with last line not
13874 being properly terminated.
13878 * Change X509_NAME_cmp() so it applies the special rules on handling
13879 DN values that are of type PrintableString, as well as RDNs of type
13880 emailAddress where the value has the type ia5String.
13882 *stefank@valicert.com via Richard Levitte*
13884 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13885 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13886 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13887 the bitwise-OR of the two for use by the majority of applications
13888 wanting this behaviour, and update the docs. The documented
13889 behaviour and actual behaviour were inconsistent and had been
13890 changing anyway, so this is more a bug-fix than a behavioural
13893 *Geoff Thorpe, diagnosed by Nadav Har'El*
13895 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13896 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13900 * Fix initialization code race conditions in
13901 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13902 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13903 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13904 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13905 ssl2_get_cipher_by_char(),
13906 ssl3_get_cipher_by_char().
13908 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13910 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13911 the cached sessions are flushed, as the remove_cb() might use ex_data
13912 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13913 (see [openssl.org #212]).
13915 *Geoff Thorpe, Lutz Jaenicke*
13917 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13918 length, instead of the encoding length to d2i_ASN1_OBJECT.
13922 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13924 * [In 0.9.6g-engine release:]
13925 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13927 *Lynn Gazis <lgazis@rainbow.com>*
13929 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13931 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13932 and get fix the header length calculation.
13933 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13934 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13936 * Use proper error handling instead of 'assertions' in buffer
13937 overflow checks added in 0.9.6e. This prevents DoS (the
13938 assertions could call abort()).
13940 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13942 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13944 * Add various sanity checks to asn1_get_length() to reject
13945 the ASN1 length bytes if they exceed sizeof(long), will appear
13946 negative or the content length exceeds the length of the
13949 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13951 * Fix cipher selection routines: ciphers without encryption had no flags
13952 for the cipher strength set and where therefore not handled correctly
13953 by the selection routines (PR #130).
13957 * Fix EVP_dsa_sha macro.
13962 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13963 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13964 that was added in OpenSSL 0.9.6d.
13966 As the countermeasure turned out to be incompatible with some
13967 broken SSL implementations, the new option is part of SSL_OP_ALL.
13968 SSL_OP_ALL is usually employed when compatibility with weird SSL
13969 implementations is desired (e.g. '-bugs' option to 's_client' and
13970 's_server'), so the new option is automatically set in many
13975 * Changes in security patch:
13977 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13978 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13979 Air Force Materiel Command, USAF, under agreement number
13982 * Add various sanity checks to asn1_get_length() to reject
13983 the ASN1 length bytes if they exceed sizeof(long), will appear
13984 negative or the content length exceeds the length of the
13985 supplied buffer. ([CVE-2002-0659])
13987 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13989 * Assertions for various potential buffer overflows, not known to
13990 happen in practice.
13992 *Ben Laurie (CHATS)*
13994 * Various temporary buffers to hold ASCII versions of integers were
13995 too small for 64 bit platforms. ([CVE-2002-0655])
13996 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13998 * Remote buffer overflow in SSL3 protocol - an attacker could
13999 supply an oversized session ID to a client. ([CVE-2002-0656])
14001 *Ben Laurie (CHATS)*
14003 * Remote buffer overflow in SSL2 protocol - an attacker could
14004 supply an oversized client master key. ([CVE-2002-0656])
14006 *Ben Laurie (CHATS)*
14008 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14010 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14011 encoded as NULL) with id-dsa-with-sha1.
14013 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14015 * Check various `X509_...()` return values in `apps/req.c`.
14017 *Nils Larsch <nla@trustcenter.de>*
14019 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14020 an end-of-file condition would erroneously be flagged, when the CRLF
14021 was just at the end of a processed block. The bug was discovered when
14022 processing data through a buffering memory BIO handing the data to a
14023 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14024 <ptsekov@syntrex.com> and Nedelcho Stanev.
14028 * Implement a countermeasure against a vulnerability recently found
14029 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14030 before application data chunks to avoid the use of known IVs
14031 with data potentially chosen by the attacker.
14035 * Fix length checks in ssl3_get_client_hello().
14039 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14040 to prevent ssl3_read_internal() from incorrectly assuming that
14041 ssl3_read_bytes() found application data while handshake
14042 processing was enabled when in fact s->s3->in_read_app_data was
14043 merely automatically cleared during the initial handshake.
14045 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14047 * Fix object definitions for Private and Enterprise: they were not
14048 recognized in their shortname (=lowercase) representation. Extend
14049 obj_dat.pl to issue an error when using undefined keywords instead
14050 of silently ignoring the problem (Svenning Sorensen
14051 <sss@sss.dnsalias.net>).
14055 * Fix DH_generate_parameters() so that it works for 'non-standard'
14056 generators, i.e. generators other than 2 and 5. (Previously, the
14057 code did not properly initialise the 'add' and 'rem' values to
14058 BN_generate_prime().)
14060 In the new general case, we do not insist that 'generator' is
14061 actually a primitive root: This requirement is rather pointless;
14062 a generator of the order-q subgroup is just as good, if not
14067 * Map new X509 verification errors to alerts. Discovered and submitted by
14068 Tom Wu <tom@arcot.com>.
14072 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14073 returning non-zero before the data has been completely received
14074 when using non-blocking I/O.
14076 *Bodo Moeller; problem pointed out by John Hughes*
14078 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14080 *Ben Laurie, Lutz Jaenicke*
14082 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14083 Yoram Zahavi <YoramZ@gilian.com>).
14087 * Add information about CygWin 1.3 and on, and preserve proper
14088 configuration for the versions before that.
14090 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14092 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14093 check whether we deal with a copy of a session and do not delete from
14094 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14095 <izhar@checkpoint.com>.
14099 * Do not store session data into the internal session cache, if it
14100 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14101 flag is set). Proposed by Aslam <aslam@funk.com>.
14105 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14110 * [In 0.9.6d-engine release:]
14111 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14113 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14115 * Add the configuration target linux-s390x.
14117 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14119 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14120 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14121 variable as an indication that a ClientHello message has been
14122 received. As the flag value will be lost between multiple
14123 invocations of ssl3_accept when using non-blocking I/O, the
14124 function may not be aware that a handshake has actually taken
14125 place, thus preventing a new session from being added to the
14128 To avoid this problem, we now set s->new_session to 2 instead of
14129 using a local variable.
14131 *Lutz Jaenicke, Bodo Moeller*
14133 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14134 if the SSL_R_LENGTH_MISMATCH error is detected.
14136 *Geoff Thorpe, Bodo Moeller*
14138 * New 'shared_ldflag' column in Configure platform table.
14142 * Fix EVP_CIPHER_mode macro.
14144 *"Dan S. Camper" <dan@bti.net>*
14146 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14147 type, we must throw them away by setting rr->length to 0.
14149 *D P Chang <dpc@qualys.com>*
14151 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14153 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14154 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14155 worked incorrectly for those cases where range = `10..._2` and
14156 `3*range` is two bits longer than range.)
14160 * Only add signing time to PKCS7 structures if it is not already
14165 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14166 OBJ_ld_ce should be OBJ_id_ce.
14167 Also some ip-pda OIDs in crypto/objects/objects.txt were
14168 incorrect (cf. RFC 3039).
14170 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14172 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14173 returns early because it has nothing to do.
14175 *Andy Schneider <andy.schneider@bjss.co.uk>*
14177 * [In 0.9.6c-engine release:]
14178 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14180 *Andy Schneider <andy.schneider@bjss.co.uk>*
14182 * [In 0.9.6c-engine release:]
14183 Add support for Cryptographic Appliance's keyserver technology.
14184 (Use engine 'keyclient')
14186 *Cryptographic Appliances and Geoff Thorpe*
14188 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14189 is called via tools/c89.sh because arguments have to be
14190 rearranged (all '-L' options must appear before the first object
14193 *Richard Shapiro <rshapiro@abinitio.com>*
14195 * [In 0.9.6c-engine release:]
14196 Add support for Broadcom crypto accelerator cards, backported
14199 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14201 * [In 0.9.6c-engine release:]
14202 Add support for SureWare crypto accelerator cards from
14203 Baltimore Technologies. (Use engine 'sureware')
14205 *Baltimore Technologies and Mark Cox*
14207 * [In 0.9.6c-engine release:]
14208 Add support for crypto accelerator cards from Accelerated
14209 Encryption Processing, www.aep.ie. (Use engine 'aep')
14211 *AEP Inc. and Mark Cox*
14213 * Add a configuration entry for gcc on UnixWare.
14215 *Gary Benson <gbenson@redhat.com>*
14217 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14218 messages are stored in a single piece (fixed-length part and
14219 variable-length part combined) and fix various bugs found on the way.
14223 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14224 instead. BIO_gethostbyname() does not know what timeouts are
14225 appropriate, so entries would stay in cache even when they have
14227 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14229 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14230 faced with a pathologically small ClientHello fragment that does
14231 not contain client_version: Instead of aborting with an error,
14232 simply choose the highest available protocol version (i.e.,
14233 TLS 1.0 unless it is disabled). In practice, ClientHello
14234 messages are never sent like this, but this change gives us
14235 strictly correct behaviour at least for TLS.
14239 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14240 never resets s->method to s->ctx->method when called from within
14241 one of the SSL handshake functions.
14243 *Bodo Moeller; problem pointed out by Niko Baric*
14245 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14246 (sent using the client's version number) if client_version is
14247 smaller than the protocol version in use. Also change
14248 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14249 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14250 the client will at least see that alert.
14254 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14259 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14260 client receives HelloRequest while in a handshake.
14262 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14264 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14265 should end in 'break', not 'goto end' which circumvents various
14266 cleanups done in state SSL_ST_OK. But session related stuff
14267 must be disabled for SSL_ST_OK in the case that we just sent a
14270 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14271 before just sending a HelloRequest.
14273 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14275 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14276 reveal whether illegal block cipher padding was found or a MAC
14277 verification error occurred. (Neither SSLerr() codes nor alerts
14278 are directly visible to potential attackers, but the information
14279 may leak via logfiles.)
14281 Similar changes are not required for the SSL 2.0 implementation
14282 because the number of padding bytes is sent in clear for SSL 2.0,
14283 and the extra bytes are just ignored. However ssl/s2_pkt.c
14284 failed to verify that the purported number of padding bytes is in
14289 * Add OpenUNIX-8 support including shared libraries
14290 (Boyd Lynn Gerber <gerberb@zenez.com>).
14294 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14295 'wristwatch attack' using huge encoding parameters (cf.
14296 James H. Manger's CRYPTO 2001 paper). Note that the
14297 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14298 encoding parameters and hence was not vulnerable.
14302 * BN_sqr() bug fix.
14304 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14306 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14307 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14308 followed by modular reduction.
14310 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14312 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14313 equivalent based on BN_pseudo_rand() instead of BN_rand().
14317 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14318 This function was broken, as the check for a new client hello message
14319 to handle SGC did not allow these large messages.
14320 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14324 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14328 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14329 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14333 * Rework the configuration and shared library support for Tru64 Unix.
14334 The configuration part makes use of modern compiler features and
14335 still retains old compiler behavior for those that run older versions
14336 of the OS. The shared library support part includes a variant that
14337 uses the RPATH feature, and is available through the special
14338 configuration target "alpha-cc-rpath", which will never be selected
14341 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14343 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14344 with the same message size as in ssl3_get_certificate_request().
14345 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14346 messages might inadvertently be reject as too long.
14348 *Petr Lampa <lampa@fee.vutbr.cz>*
14350 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14354 * Modified SSL library such that the verify_callback that has been set
14355 specifically for an SSL object with SSL_set_verify() is actually being
14356 used. Before the change, a verify_callback set with this function was
14357 ignored and the verify_callback() set in the SSL_CTX at the time of
14358 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14359 to allow the necessary settings.
14363 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14364 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14365 done automatically (in contradiction to the requirements of the C
14366 standard). This made problems when used from OpenSSH.
14370 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14371 dh->length and always used
14373 BN_rand_range(priv_key, dh->p).
14375 BN_rand_range() is not necessary for Diffie-Hellman, and this
14376 specific range makes Diffie-Hellman unnecessarily inefficient if
14377 dh->length (recommended exponent length) is much smaller than the
14378 length of dh->p. We could use BN_rand_range() if the order of
14379 the subgroup was stored in the DH structure, but we only have
14384 BN_rand(priv_key, l, ...)
14386 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14393 RSA_eay_public_encrypt
14394 RSA_eay_private_decrypt
14395 RSA_eay_private_encrypt (signing)
14396 RSA_eay_public_decrypt (signature verification)
14398 (default implementations for RSA_public_encrypt,
14399 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14400 always reject numbers >= n.
14404 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14405 to synchronize access to 'locking_thread'. This is necessary on
14406 systems where access to 'locking_thread' (an 'unsigned long'
14407 variable) is not atomic.
14411 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14412 *before* setting the 'crypto_lock_rand' flag. The previous code had
14413 a race condition if 0 is a valid thread ID.
14415 *Travis Vitek <vitek@roguewave.com>*
14417 * Add support for shared libraries under Irix.
14419 *Albert Chin-A-Young <china@thewrittenword.com>*
14421 * Add configuration option to build on Linux on both big-endian and
14422 little-endian MIPS.
14424 *Ralf Baechle <ralf@uni-koblenz.de>*
14426 * Add the possibility to create shared libraries on HP-UX.
14430 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14432 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14433 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14434 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14435 PRNG state recovery was possible based on the output of
14436 one PRNG request appropriately sized to gain knowledge on
14437 'md' followed by enough consecutive 1-byte PRNG requests
14438 to traverse all of 'state'.
14440 1. When updating 'md_local' (the current thread's copy of 'md')
14441 during PRNG output generation, hash all of the previous
14442 'md_local' value, not just the half used for PRNG output.
14444 2. Make the number of bytes from 'state' included into the hash
14445 independent from the number of PRNG bytes requested.
14447 The first measure alone would be sufficient to avoid
14448 Markku-Juhani's attack. (Actually it had never occurred
14449 to me that the half of 'md_local' used for chaining was the
14450 half from which PRNG output bytes were taken -- I had always
14451 assumed that the secret half would be used.) The second
14452 measure makes sure that additional data from 'state' is never
14453 mixed into 'md_local' in small portions; this heuristically
14454 further strengthens the PRNG.
14458 * Fix crypto/bn/asm/mips3.s.
14462 * When only the key is given to "enc", the IV is undefined. Print out
14463 an error message in this case.
14467 * Handle special case when X509_NAME is empty in X509 printing routines.
14471 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14472 positive and less than q.
14476 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14477 used: it isn't thread safe and the add_lock_callback should handle
14480 *Paul Rose <Paul.Rose@bridge.com>*
14482 * Verify that incoming data obeys the block size in
14483 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14489 *Ulf Möller, Bodo Möller*
14491 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14492 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14493 when fixing the server behaviour for backwards-compatible 'client
14494 hello' messages. (Note that the attack is impractical against
14495 SSL 3.0 and TLS 1.0 anyway because length and version checking
14496 means that the probability of guessing a valid ciphertext is
14497 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14500 Before 0.9.5, the countermeasure (hide the error by generating a
14501 random 'decryption result') did not work properly because
14502 ERR_clear_error() was missing, meaning that SSL_get_error() would
14503 detect the supposedly ignored error.
14505 Both problems are now fixed.
14509 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14510 (previously it was 1024).
14514 * Fix for compatibility mode trust settings: ignore trust settings
14515 unless some valid trust or reject settings are present.
14519 * Fix for blowfish EVP: its a variable length cipher.
14523 * Fix various bugs related to DSA S/MIME verification. Handle missing
14524 parameters in DSA public key structures and return an error in the
14525 DSA routines if parameters are absent.
14529 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14530 in the current directory if neither $RANDFILE nor $HOME was set.
14531 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14532 caused some confusion to Windows users who haven't defined $HOME.
14533 Thus RAND_file_name() is changed again: e_os.h can define a
14534 DEFAULT_HOME, which will be used if $HOME is not set.
14535 For Windows, we use "C:"; on other platforms, we still require
14536 environment variables.
14538 * Move 'if (!initialized) RAND_poll()' into regions protected by
14539 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14540 having multiple threads call RAND_poll() concurrently.
14544 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14545 combination of a flag and a thread ID variable.
14546 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14547 flag), *other* threads can enter ssleay_add_bytes without obeying
14548 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14549 that they do not hold after the first thread unsets add_do_not_lock).
14553 * Change bctest again: '-x' expressions are not available in all
14554 versions of 'test'.
14558 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14560 * Fix a couple of memory leaks in PKCS7_dataDecode()
14562 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14564 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14565 the default extension for executables, if any. Also, make the perl
14566 scripts that use symlink() to test if it really exists and use "cp"
14567 if it doesn't. All this made OpenSSL compilable and installable in
14572 * Fix for asn1_GetSequence() for indefinite length constructed data.
14573 If SEQUENCE is length is indefinite just set c->slen to the total
14574 amount of data available.
14576 *Steve Henson, reported by shige@FreeBSD.org*
14578 *This change does not apply to 0.9.7.*
14580 * Change bctest to avoid here-documents inside command substitution
14581 (workaround for FreeBSD /bin/sh bug).
14582 For compatibility with Ultrix, avoid shell functions (introduced
14583 in the bctest version that searches along $PATH).
14587 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14588 with des_encrypt() defined on some operating systems, like Solaris
14593 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14594 On the Importance of Eliminating Errors in Cryptographic
14595 Computations, J. Cryptology 14 (2001) 2, 101-119,
14596 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14600 * MIPS assembler BIGNUM division bug fix.
14604 * Disabled incorrect Alpha assembler code.
14608 * Fix PKCS#7 decode routines so they correctly update the length
14609 after reading an EOC for the EXPLICIT tag.
14613 *This change does not apply to 0.9.7.*
14615 * Fix bug in PKCS#12 key generation routines. This was triggered
14616 if a 3DES key was generated with a 0 initial byte. Include
14617 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14618 (but broken) behaviour.
14622 * Enhance bctest to search for a working bc along $PATH and print
14625 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14627 * Fix memory leaks in err.c: free err_data string if necessary;
14628 don't write to the wrong index in ERR_set_error_data.
14632 * Implement ssl23_peek (analogous to ssl23_read), which previously
14637 * Replace rdtsc with `_emit` statements for VC++ version 5.
14639 *Jeremy Cooper <jeremy@baymoo.org>*
14641 * Make it possible to reuse SSLv2 sessions.
14645 * In copy_email() check for >= 0 as a return value for
14646 X509_NAME_get_index_by_NID() since 0 is a valid index.
14648 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14650 * Avoid coredump with unsupported or invalid public keys by checking if
14651 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14652 PKCS7_verify() fails with non detached data.
14656 * Don't use getenv in library functions when run as setuid/setgid.
14657 New function OPENSSL_issetugid().
14661 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14662 due to incorrect handling of multi-threading:
14664 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14666 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14668 3. Count how many times MemCheck_off() has been called so that
14669 nested use can be treated correctly. This also avoids
14670 inband-signalling in the previous code (which relied on the
14671 assumption that thread ID 0 is impossible).
14675 * Add "-rand" option also to s_client and s_server.
14679 * Fix CPU detection on Irix 6.x.
14680 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14681 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14683 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14688 *This change does not apply to 0.9.7.*
14690 * Use the cached encoding of an X509_NAME structure rather than
14691 copying it. This is apparently the reason for the libsafe "errors"
14692 but the code is actually correct.
14696 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14697 Bleichenbacher's DSA attack.
14698 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14699 to be set and top=0 forces the highest bit to be set; top=-1 is new
14700 and leaves the highest bit random.
14702 *Ulf Moeller, Bodo Moeller*
14704 * In the `NCONF_...`-based implementations for `CONF_...` queries
14705 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14706 a temporary CONF structure with the data component set to NULL
14707 (which gives segmentation faults in lh_retrieve).
14708 Instead, use NULL for the CONF pointer in CONF_get_string and
14709 CONF_get_number (which may use environment variables) and directly
14710 return NULL from CONF_get_section.
14714 * Fix potential buffer overrun for EBCDIC.
14718 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14719 keyUsage if basicConstraints absent for a CA.
14723 * Make SMIME_write_PKCS7() write mail header values with a format that
14724 is more generally accepted (no spaces before the semicolon), since
14725 some programs can't parse those values properly otherwise. Also make
14726 sure BIO's that break lines after each write do not create invalid
14731 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14732 macros previously used would not encode an empty SEQUENCE OF
14733 and break the signature.
14737 *This change does not apply to 0.9.7.*
14739 * Zero the premaster secret after deriving the master secret in
14744 * Add some EVP_add_digest_alias registrations (as found in
14745 OpenSSL_add_all_digests()) to SSL_library_init()
14746 aka OpenSSL_add_ssl_algorithms(). This provides improved
14747 compatibility with peers using X.509 certificates
14748 with unconventional AlgorithmIdentifier OIDs.
14752 * Fix for Irix with NO_ASM.
14754 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14756 * ./config script fixes.
14758 *Ulf Moeller, Richard Levitte*
14760 * Fix 'openssl passwd -1'.
14764 * Change PKCS12_key_gen_asc() so it can cope with non null
14765 terminated strings whose length is passed in the passlen
14766 parameter, for example from PEM callbacks. This was done
14767 by adding an extra length parameter to asc2uni().
14769 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14771 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14772 call failed, free the DSA structure.
14776 * Fix to uni2asc() to cope with zero length Unicode strings.
14777 These are present in some PKCS#12 files.
14781 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14782 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14783 when writing a 32767 byte record.
14785 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14787 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14788 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14790 (RSA objects have a reference count access to which is protected
14791 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14792 so they are meant to be shared between threads.)
14793 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14794 "Reddie, Steven" <Steven.Reddie@ca.com>*
14796 * Fix a deadlock in CRYPTO_mem_leaks().
14800 * Use better test patterns in bntest.
14804 * rand_win.c fix for Borland C.
14808 * BN_rshift bugfix for n == 0.
14812 * Add a 'bctest' script that checks for some known 'bc' bugs
14813 so that 'make test' does not abort just because 'bc' is broken.
14817 * Store verify_result within SSL_SESSION also for client side to
14818 avoid potential security hole. (Re-used sessions on the client side
14819 always resulted in verify_result==X509_V_OK, not using the original
14820 result of the server certificate verification.)
14824 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14825 SSL3_RT_APPLICATION_DATA, return 0.
14826 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14831 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14832 releases, have been re-implemented by renaming the previous
14833 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14834 and ssl3_read_internal, respectively, and adding 'peek' parameters
14835 to them. The new ssl[23]_{read,peek} functions are calls to
14836 ssl[23]_read_internal with the 'peek' flag set appropriately.
14837 A 'peek' parameter has also been added to ssl3_read_bytes, which
14838 does the actual work for ssl3_read_internal.
14842 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14843 the method-specific "init()" handler. Also clean up ex_data after
14844 calling the method-specific "finish()" handler. Previously, this was
14845 happening the other way round.
14849 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14850 The previous value, 12, was not always sufficient for BN_mod_exp().
14854 * Make sure that shared libraries get the internal name engine with
14855 the full version number and not just 0. This should mark the
14856 shared libraries as not backward compatible. Of course, this should
14857 be changed again when we can guarantee backward binary compatibility.
14861 * Fix typo in get_cert_by_subject() in by_dir.c
14863 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14865 * Rework the system to generate shared libraries:
14867 - Make note of the expected extension for the shared libraries and
14868 if there is a need for symbolic links from for example libcrypto.so.0
14869 to libcrypto.so.0.9.7. There is extended info in Configure for
14872 - Make as few rebuilds of the shared libraries as possible.
14874 - Still avoid linking the OpenSSL programs with the shared libraries.
14876 - When installing, install the shared libraries separately from the
14881 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14883 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14884 and not in SSL_clear because the latter is also used by the
14885 accept/connect functions; previously, the settings made by
14886 SSL_set_read_ahead would be lost during the handshake.
14888 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14890 * Correct util/mkdef.pl to be selective about disabled algorithms.
14891 Previously, it would create entries for disabled algorithms no
14896 * Added several new manual pages for SSL_* function.
14900 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14902 * In ssl23_get_client_hello, generate an error message when faced
14903 with an initial SSL 3.0/TLS record that is too small to contain the
14904 first two bytes of the ClientHello message, i.e. client_version.
14905 (Note that this is a pathologic case that probably has never happened
14906 in real life.) The previous approach was to use the version number
14907 from the record header as a substitute; but our protocol choice
14908 should not depend on that one because it is not authenticated
14909 by the Finished messages.
14913 * More robust randomness gathering functions for Windows.
14915 *Jeffrey Altman <jaltman@columbia.edu>*
14917 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14918 not set then we don't setup the error code for issuer check errors
14919 to avoid possibly overwriting other errors which the callback does
14920 handle. If an application does set the flag then we assume it knows
14921 what it is doing and can handle the new informational codes
14926 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14927 a general "ANY" type, as such it should be able to decode anything
14928 including tagged types. However it didn't check the class so it would
14929 wrongly interpret tagged types in the same way as their universal
14930 counterpart and unknown types were just rejected. Changed so that the
14931 tagged and unknown types are handled in the same way as a SEQUENCE:
14932 that is the encoding is stored intact. There is also a new type
14933 "V_ASN1_OTHER" which is used when the class is not universal, in this
14934 case we have no idea what the actual type is so we just lump them all
14939 * On VMS, stdout may very well lead to a file that is written to
14940 in a record-oriented fashion. That means that every write() will
14941 write a separate record, which will be read separately by the
14942 programs trying to read from it. This can be very confusing.
14944 The solution is to put a BIO filter in the way that will buffer
14945 text until a linefeed is reached, and then write everything a
14946 line at a time, so every record written will be an actual line,
14947 not chunks of lines and not (usually doesn't happen, but I've
14948 seen it once) several lines in one record. BIO_f_linebuffer() is
14951 Currently, it's a VMS-only method, because that's where it has
14952 been tested well enough.
14956 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14957 it can return incorrect results.
14958 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14959 but it was in 0.9.6-beta[12].)
14963 * Disable the check for content being present when verifying detached
14964 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14965 include zero length content when signing messages.
14969 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14970 BIO_ctrl (for BIO pairs).
14974 * Add DSO method for VMS.
14978 * Bug fix: Montgomery multiplication could produce results with the
14983 * Add RPM specification openssl.spec and modify it to build three
14984 packages. The default package contains applications, application
14985 documentation and run-time libraries. The devel package contains
14986 include files, static libraries and function documentation. The
14987 doc package contains the contents of the doc directory. The original
14988 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14992 * Add a large number of documentation files for many SSL routines.
14994 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14996 * Add a configuration entry for Sony News 4.
14998 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15000 * Don't set the two most significant bits to one when generating a
15001 random number < q in the DSA library.
15005 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15006 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15007 the underlying transport is blocking) if a handshake took place.
15008 (The default behaviour is needed by applications such as s_client
15009 and s_server that use select() to determine when to use SSL_read;
15010 but for applications that know in advance when to expect data, it
15011 just makes things more complicated.)
15015 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15020 * Add a few more EBCDIC conditionals that make `req` and `x509`
15021 work better on such systems.
15023 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15025 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15026 Update PKCS12_parse() so it copies the friendlyName and the
15027 keyid to the certificates aux info.
15031 * Fix bug in PKCS7_verify() which caused an infinite loop
15032 if there was more than one signature.
15034 *Sven Uszpelkat <su@celocom.de>*
15036 * Major change in util/mkdef.pl to include extra information
15037 about each symbol, as well as presenting variables as well
15038 as functions. This change means that there's n more need
15039 to rebuild the .num files when some algorithms are excluded.
15043 * Allow the verify time to be set by an application,
15044 rather than always using the current time.
15048 * Phase 2 verify code reorganisation. The certificate
15049 verify code now looks up an issuer certificate by a
15050 number of criteria: subject name, authority key id
15051 and key usage. It also verifies self signed certificates
15052 by the same criteria. The main comparison function is
15053 X509_check_issued() which performs these checks.
15055 Lot of changes were necessary in order to support this
15056 without completely rewriting the lookup code.
15058 Authority and subject key identifier are now cached.
15060 The LHASH 'certs' is X509_STORE has now been replaced
15061 by a STACK_OF(X509_OBJECT). This is mainly because an
15062 LHASH can't store or retrieve multiple objects with
15063 the same hash value.
15065 As a result various functions (which were all internal
15066 use only) have changed to handle the new X509_STORE
15067 structure. This will break anything that messed round
15068 with X509_STORE internally.
15070 The functions X509_STORE_add_cert() now checks for an
15071 exact match, rather than just subject name.
15073 The X509_STORE API doesn't directly support the retrieval
15074 of multiple certificates matching a given criteria, however
15075 this can be worked round by performing a lookup first
15076 (which will fill the cache with candidate certificates)
15077 and then examining the cache for matches. This is probably
15078 the best we can do without throwing out X509_LOOKUP
15079 entirely (maybe later...).
15081 The X509_VERIFY_CTX structure has been enhanced considerably.
15083 All certificate lookup operations now go via a get_issuer()
15084 callback. Although this currently uses an X509_STORE it
15085 can be replaced by custom lookups. This is a simple way
15086 to bypass the X509_STORE hackery necessary to make this
15087 work and makes it possible to use more efficient techniques
15088 in future. A very simple version which uses a simple
15089 STACK for its trusted certificate store is also provided
15090 using X509_STORE_CTX_trusted_stack().
15092 The verify_cb() and verify() callbacks now have equivalents
15093 in the X509_STORE_CTX structure.
15095 X509_STORE_CTX also has a 'flags' field which can be used
15096 to customise the verify behaviour.
15100 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15101 excludes S/MIME capabilities.
15105 * When a certificate request is read in keep a copy of the
15106 original encoding of the signed data and use it when outputting
15107 again. Signatures then use the original encoding rather than
15108 a decoded, encoded version which may cause problems if the
15109 request is improperly encoded.
15113 * For consistency with other BIO_puts implementations, call
15114 buffer_write(b, ...) directly in buffer_puts instead of calling
15117 In BIO_puts, increment b->num_write as in BIO_write.
15119 *Peter.Sylvester@EdelWeb.fr*
15121 * Fix BN_mul_word for the case where the word is 0. (We have to use
15122 BN_zero, we may not return a BIGNUM with an array consisting of
15123 words set to zero.)
15127 * Avoid calling abort() from within the library when problems are
15128 detected, except if preprocessor symbols have been defined
15129 (such as REF_CHECK, BN_DEBUG etc.).
15133 * New openssl application 'rsautl'. This utility can be
15134 used for low-level RSA operations. DER public key
15135 BIO/fp routines also added.
15139 * New Configure entry and patches for compiling on QNX 4.
15141 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15143 * A demo state-machine implementation was sponsored by
15144 Nuron (<http://www.nuron.com/>) and is now available in
15145 demos/state_machine.
15149 * New options added to the 'dgst' utility for signature
15150 generation and verification.
15154 * Unrecognized PKCS#7 content types are now handled via a
15155 catch all ASN1_TYPE structure. This allows unsupported
15156 types to be stored as a "blob" and an application can
15157 encode and decode it manually.
15161 * Fix various signed/unsigned issues to make a_strex.c
15162 compile under VC++.
15164 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15166 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15167 length if passed a buffer. ASN1_INTEGER_to_BN failed
15168 if passed a NULL BN and its argument was negative.
15170 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15172 * Modification to PKCS#7 encoding routines to output definite
15173 length encoding. Since currently the whole structures are in
15174 memory there's not real point in using indefinite length
15175 constructed encoding. However if OpenSSL is compiled with
15176 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15180 * Added BIO_vprintf() and BIO_vsnprintf().
15184 * Added more prefixes to parse for in the strings written
15185 through a logging bio, to cover all the levels that are available
15186 through syslog. The prefixes are now:
15188 PANIC, EMERG, EMR => LOG_EMERG
15189 ALERT, ALR => LOG_ALERT
15190 CRIT, CRI => LOG_CRIT
15191 ERROR, ERR => LOG_ERR
15192 WARNING, WARN, WAR => LOG_WARNING
15193 NOTICE, NOTE, NOT => LOG_NOTICE
15194 INFO, INF => LOG_INFO
15195 DEBUG, DBG => LOG_DEBUG
15197 and as before, if none of those prefixes are present at the
15198 beginning of the string, LOG_ERR is chosen.
15200 On Win32, the `LOG_*` levels are mapped according to this:
15202 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15203 LOG_WARNING => EVENTLOG_WARNING_TYPE
15204 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15208 * Made it possible to reconfigure with just the configuration
15209 argument "reconf" or "reconfigure". The command line arguments
15210 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15211 and are retrieved from there when reconfiguring.
15217 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15219 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15223 * The obj_dat.pl script was messing up the sorting of object
15224 names. The reason was that it compared the quoted version
15225 of strings as a result "OCSP" > "OCSP Signing" because
15226 " > SPACE. Changed script to store unquoted versions of
15227 names and add quotes on output. It was also omitting some
15228 names from the lookup table if they were given a default
15229 value (that is if SN is missing it is given the same
15230 value as LN and vice versa), these are now added on the
15231 grounds that if an object has a name we should be able to
15232 look it up. Finally added warning output when duplicate
15233 short or long names are found.
15237 * Changes needed for Tandem NSK.
15239 *Scott Uroff <scott@xypro.com>*
15241 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15242 RSA_padding_check_SSLv23(), special padding was never detected
15243 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15244 version rollback attacks was not effective.
15246 In s23_clnt.c, don't use special rollback-attack detection padding
15247 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15248 client; similarly, in s23_srvr.c, don't do the rollback check if
15249 SSL 2.0 is the only protocol enabled in the server.
15253 * Make it possible to get hexdumps of unprintable data with 'openssl
15254 asn1parse'. By implication, the functions ASN1_parse_dump() and
15255 BIO_dump_indent() are added.
15259 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15260 these print out strings and name structures based on various
15261 flags including RFC2253 support and proper handling of
15262 multibyte characters. Added options to the 'x509' utility
15263 to allow the various flags to be set.
15267 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15268 Also change the functions X509_cmp_current_time() and
15269 X509_gmtime_adj() work with an ASN1_TIME structure,
15270 this will enable certificates using GeneralizedTime in validity
15271 dates to be checked.
15275 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15276 negative public key encodings) on by default,
15277 NO_NEG_PUBKEY_BUG can be set to disable it.
15281 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15282 content octets. An i2c_ASN1_OBJECT is unnecessary because
15283 the encoding can be trivially obtained from the structure.
15287 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15288 not read locks (`CRYPTO_r_[un]lock`).
15292 * A first attempt at creating official support for shared
15293 libraries through configuration. I've kept it so the
15294 default is static libraries only, and the OpenSSL programs
15295 are always statically linked for now, but there are
15296 preparations for dynamic linking in place.
15297 This has been tested on Linux and Tru64.
15301 * Randomness polling function for Win9x, as described in:
15302 Peter Gutmann, Software Generation of Practically Strong
15307 * Fix so PRNG is seeded in req if using an already existing
15312 * New options to smime application. -inform and -outform
15313 allow alternative formats for the S/MIME message including
15314 PEM and DER. The -content option allows the content to be
15315 specified separately. This should allow things like Netscape
15316 form signing output easier to verify.
15320 * Fix the ASN1 encoding of tags using the 'long form'.
15324 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15325 STRING types. These convert content octets to and from the
15326 underlying type. The actual tag and length octets are
15327 already assumed to have been read in and checked. These
15328 are needed because all other string types have virtually
15329 identical handling apart from the tag. By having versions
15330 of the ASN1 functions that just operate on content octets
15331 IMPLICIT tagging can be handled properly. It also allows
15332 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15333 and ASN1_INTEGER are identical apart from the tag.
15337 * Change the handling of OID objects as follows:
15339 - New object identifiers are inserted in objects.txt, following
15340 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15341 - objects.pl is used to process obj_mac.num and create a new
15343 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15346 This is currently kind of a hack, and the perl code in objects.pl
15347 isn't very elegant, but it works as I intended. The simplest way
15348 to check that it worked correctly is to look in obj_dat.h and
15349 check the array nid_objs and make sure the objects haven't moved
15350 around (this is important!). Additions are OK, as well as
15351 consistent name changes.
15355 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15359 * Addition of the command line parameter '-rand file' to 'openssl req'.
15360 The given file adds to whatever has already been seeded into the
15361 random pool through the RANDFILE configuration file option or
15362 environment variable, or the default random state file.
15366 * mkstack.pl now sorts each macro group into lexical order.
15367 Previously the output order depended on the order the files
15368 appeared in the directory, resulting in needless rewriting
15373 * Patches to make OpenSSL compile under Win32 again. Mostly
15374 work arounds for the VC++ problem that it treats func() as
15375 func(void). Also stripped out the parts of mkdef.pl that
15376 added extra typesafe functions: these no longer exist.
15380 * Reorganisation of the stack code. The macros are now all
15381 collected in safestack.h . Each macro is defined in terms of
15382 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15383 DEBUG_SAFESTACK is now handled in terms of function casts,
15384 this has the advantage of retaining type safety without the
15385 use of additional functions. If DEBUG_SAFESTACK is not defined
15386 then the non typesafe macros are used instead. Also modified the
15387 mkstack.pl script to handle the new form. Needs testing to see
15388 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15389 the default if no major problems. Similar behaviour for ASN1_SET_OF
15390 and PKCS12_STACK_OF.
15394 * When some versions of IIS use the 'NET' form of private key the
15395 key derivation algorithm is different. Normally MD5(password) is
15396 used as a 128 bit RC4 key. In the modified case
15397 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15398 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15399 as the old Netscape_RSA functions except they have an additional
15400 'sgckey' parameter which uses the modified algorithm. Also added
15401 an -sgckey command line option to the rsa utility. Thanks to
15402 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15403 algorithm to openssl-dev.
15407 * The evp_local.h macros were using 'c.##kname' which resulted in
15408 invalid expansion on some systems (SCO 5.0.5 for example).
15409 Corrected to 'c.kname'.
15411 *Phillip Porch <root@theporch.com>*
15413 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15414 a STACK of email addresses from a certificate or request, these look
15415 in the subject name and the subject alternative name extensions and
15416 omit any duplicate addresses.
15420 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15421 This makes DSA verification about 2 % faster.
15425 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15426 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15427 plus overhead for 1024 bit moduli).
15428 This makes exponentiations about 0.5 % faster for 1024 bit
15429 exponents (as measured by "openssl speed rsa2048").
15433 * Rename memory handling macros to avoid conflicts with other
15435 Malloc => OPENSSL_malloc
15436 Malloc_locked => OPENSSL_malloc_locked
15437 Realloc => OPENSSL_realloc
15438 Free => OPENSSL_free
15442 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15443 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15447 * CygWin32 support.
15449 *John Jarvie <jjarvie@newsguy.com>*
15451 * The type-safe stack code has been rejigged. It is now only compiled
15452 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15453 by default all type-specific stack functions are "#define"d back to
15454 standard stack functions. This results in more streamlined output
15455 but retains the type-safety checking possibilities of the original
15460 * The STACK code has been cleaned up, and certain type declarations
15461 that didn't make a lot of sense have been brought in line. This has
15462 also involved a cleanup of sorts in safestack.h to more correctly
15463 map type-safe stack functions onto their plain stack counterparts.
15464 This work has also resulted in a variety of "const"ifications of
15465 lots of the code, especially `_cmp` operations which should normally
15466 be prototyped with "const" parameters anyway.
15470 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15471 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15472 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15473 where all of 'md' is used each time the PRNG is used, but 'state'
15474 is used only indexed by a cyclic counter. As entropy may not be
15475 well distributed from the beginning, 'md' is important as a
15476 chaining variable. However, the output function chains only half
15477 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15478 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15479 in all of 'state' being rewritten, with the new values depending
15480 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15484 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15485 the handshake is continued after ssl_verify_cert_chain();
15486 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15487 can lead to 'unexplainable' connection aborts later.
15489 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15491 * Major EVP API cipher revision.
15492 Add hooks for extra EVP features. This allows various cipher
15493 parameters to be set in the EVP interface. Support added for variable
15494 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15495 setting of RC2 and RC5 parameters.
15497 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15500 Remove lots of duplicated code from the EVP library. For example *every*
15501 cipher init() function handles the 'iv' in the same way according to the
15502 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15503 for CFB and OFB modes they zero ctx->num.
15505 New functionality allows removal of S/MIME code RC2 hack.
15507 Most of the routines have the same form and so can be declared in terms
15510 By shifting this to the top level EVP_CipherInit() it can be removed from
15511 all individual ciphers. If the cipher wants to handle IVs or keys
15512 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15515 Change lots of functions like EVP_EncryptUpdate() to now return a
15516 value: although software versions of the algorithms cannot fail
15517 any installed hardware versions can.
15521 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15522 this option is set, tolerate broken clients that send the negotiated
15523 protocol version number instead of the requested protocol version
15528 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15529 i.e. non-zero for export ciphersuites, zero otherwise.
15530 Previous versions had this flag inverted, inconsistent with
15531 rsa_tmp_cb (..._TMP_RSA_CB).
15533 *Bodo Moeller; problem reported by Amit Chopra*
15535 * Add missing DSA library text string. Work around for some IIS
15536 key files with invalid SEQUENCE encoding.
15540 * Add a document (doc/standards.txt) that list all kinds of standards
15541 and so on that are implemented in OpenSSL.
15545 * Enhance c_rehash script. Old version would mishandle certificates
15546 with the same subject name hash and wouldn't handle CRLs at all.
15547 Added -fingerprint option to crl utility, to support new c_rehash
15552 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15556 * Fix for SSL server purpose checking. Server checking was
15557 rejecting certificates which had extended key usage present
15558 but no ssl client purpose.
15560 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15562 * Make PKCS#12 code work with no password. The PKCS#12 spec
15563 is a little unclear about how a blank password is handled.
15564 Since the password in encoded as a BMPString with terminating
15565 double NULL a zero length password would end up as just the
15566 double NULL. However no password at all is different and is
15567 handled differently in the PKCS#12 key generation code. NS
15568 treats a blank password as zero length. MSIE treats it as no
15569 password on export: but it will try both on import. We now do
15570 the same: PKCS12_parse() tries zero length and no password if
15571 the password is set to "" or NULL (NULL is now a valid password:
15572 it wasn't before) as does the pkcs12 application.
15576 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15577 perror when PEM_read_bio_X509_REQ fails, the error message must
15578 be obtained from the error queue.
15582 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15583 it in ERR_remove_state if appropriate, and change ERR_get_state
15584 accordingly to avoid race conditions (this is necessary because
15585 thread_hash is no longer constant once set).
15589 * Bugfix for linux-elf makefile.one.
15593 * RSA_get_default_method() will now cause a default
15594 RSA_METHOD to be chosen if one doesn't exist already.
15595 Previously this was only set during a call to RSA_new()
15596 or RSA_new_method(NULL) meaning it was possible for
15597 RSA_get_default_method() to return NULL.
15601 * Added native name translation to the existing DSO code
15602 that will convert (if the flag to do so is set) filenames
15603 that are sufficiently small and have no path information
15604 into a canonical native form. Eg. "blah" converted to
15605 "libblah.so" or "blah.dll" etc.
15609 * New function ERR_error_string_n(e, buf, len) which is like
15610 ERR_error_string(e, buf), but writes at most 'len' bytes
15611 including the 0 terminator. For ERR_error_string_n, 'buf'
15614 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15616 * CONF library reworked to become more general. A new CONF
15617 configuration file reader "class" is implemented as well as a
15618 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15619 old `CONF_*` functions are still there, but are reimplemented to
15620 work in terms of the new functions. Also, a set of functions
15621 to handle the internal storage of the configuration data is
15622 provided to make it easier to write new configuration file
15623 reader "classes" (I can definitely see something reading a
15624 configuration file in XML format, for example), called `_CONF_*`,
15625 or "the configuration storage API"...
15627 The new configuration file reading functions are:
15629 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15630 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15632 NCONF_default, NCONF_WIN32
15634 NCONF_dump_fp, NCONF_dump_bio
15636 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15637 NCONF_new creates a new CONF object. This works in the same way
15638 as other interfaces in OpenSSL, like the BIO interface.
15639 `NCONF_dump_*` dump the internal storage of the configuration file,
15640 which is useful for debugging. All other functions take the same
15641 arguments as the old `CONF_*` functions with the exception of the
15642 first that must be a `CONF *` instead of a `LHASH *`.
15644 To make it easier to use the new classes with the old `CONF_*` functions,
15645 the function CONF_set_default_method is provided.
15649 * Add '-tls1' option to 'openssl ciphers', which was already
15650 mentioned in the documentation but had not been implemented.
15651 (This option is not yet really useful because even the additional
15652 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15656 * Initial DSO code added into libcrypto for letting OpenSSL (and
15657 OpenSSL-based applications) load shared libraries and bind to
15658 them in a portable way.
15660 *Geoff Thorpe, with contributions from Richard Levitte*
15662 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15664 * Make sure _lrotl and _lrotr are only used with MSVC.
15666 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15667 (the default implementation of RAND_status).
15669 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15670 to '-clrext' (= clear extensions), as intended and documented.
15671 *Bodo Moeller; inconsistency pointed out by Michael Attili
15672 <attili@amaxo.com>*
15674 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15675 was larger than the MD block size.
15677 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15679 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15680 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15681 using the passed key: if the passed key was a private key the result
15682 of X509_print(), for example, would be to print out all the private key
15687 * des_quad_cksum() byte order bug fix.
15688 *Ulf Möller, using the problem description in krb4-0.9.7, where
15689 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15691 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15694 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15696 * For easily testing in shell scripts whether some command
15697 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15698 returns with exit code 0 iff no command of the given name is available.
15699 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15700 the output goes to stdout and nothing is printed to stderr.
15701 Additional arguments are always ignored.
15703 Since for each cipher there is a command of the same name,
15704 the 'no-cipher' compilation switches can be tested this way.
15706 ('openssl no-XXX' is not able to detect pseudo-commands such
15707 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15711 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15715 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15716 is set; it will be thrown away anyway because each handshake creates
15718 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15719 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15720 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15721 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15725 * New s_client option -ign_eof: EOF at stdin is ignored, and
15726 'Q' and 'R' lose their special meanings (quit/renegotiate).
15727 This is part of what -quiet does; unlike -quiet, -ign_eof
15728 does not suppress any output.
15732 * Add compatibility options to the purpose and trust code. The
15733 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15734 accepts a certificate or CA, this was the previous behaviour,
15735 with all the associated security issues.
15737 X509_TRUST_COMPAT is the old trust behaviour: only and
15738 automatically trust self signed roots in certificate store. A
15739 new trust setting X509_TRUST_DEFAULT is used to specify that
15740 a purpose has no associated trust setting and it should instead
15741 use the value in the default purpose.
15745 * Fix the PKCS#8 DSA private key code so it decodes keys again
15746 and fix a memory leak.
15750 * In util/mkerr.pl (which implements 'make errors'), preserve
15751 reason strings from the previous version of the .c file, as
15752 the default to have only downcase letters (and digits) in
15753 automatically generated reasons codes is not always appropriate.
15757 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15758 using strerror. Previously, ERR_reason_error_string() returned
15759 library names as reason strings for SYSerr; but SYSerr is a special
15760 case where small numbers are errno values, not library numbers.
15764 * Add '-dsaparam' option to 'openssl dhparam' application. This
15765 converts DSA parameters into DH parameters. (When creating parameters,
15766 DSA_generate_parameters is used.)
15770 * Include 'length' (recommended exponent length) in C code generated
15771 by 'openssl dhparam -C'.
15775 * The second argument to set_label in perlasm was already being used
15776 so couldn't be used as a "file scope" flag. Moved to third argument
15781 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15782 instead of RAND_bytes for encryption IVs and salts.
15786 * Include RAND_status() into RAND_METHOD instead of implementing
15787 it only for md_rand.c Otherwise replacing the PRNG by calling
15788 RAND_set_rand_method would be impossible.
15792 * Don't let DSA_generate_key() enter an infinite loop if the random
15793 number generation fails.
15797 * New 'rand' application for creating pseudo-random output.
15801 * Added configuration support for Linux/IA64
15803 *Rolf Haberrecker <rolf@suse.de>*
15805 * Assembler module support for Mingw32.
15809 * Shared library support for HPUX (in shlib/).
15811 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15813 * Shared library support for Solaris gcc.
15815 *Lutz Behnke <behnke@trustcenter.de>*
15817 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15819 * PKCS7_encrypt() was adding text MIME headers twice because they
15820 were added manually and by SMIME_crlf_copy().
15824 * In bntest.c don't call BN_rand with zero bits argument.
15826 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15828 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15829 case was implemented. This caused BN_div_recp() to fail occasionally.
15833 * Add an optional second argument to the set_label() in the perl
15834 assembly language builder. If this argument exists and is set
15835 to 1 it signals that the assembler should use a symbol whose
15836 scope is the entire file, not just the current function. This
15837 is needed with MASM which uses the format label:: for this scope.
15839 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15841 * Change the ASN1 types so they are typedefs by default. Before
15842 almost all types were #define'd to ASN1_STRING which was causing
15843 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15848 * Change names of new functions to the new get1/get0 naming
15849 convention: After 'get1', the caller owns a reference count
15850 and has to call `..._free`; 'get0' returns a pointer to some
15851 data structure without incrementing reference counters.
15852 (Some of the existing 'get' functions increment a reference
15853 counter, some don't.)
15854 Similarly, 'set1' and 'add1' functions increase reference
15855 counters or duplicate objects.
15859 * Allow for the possibility of temp RSA key generation failure:
15860 the code used to assume it always worked and crashed on failure.
15864 * Fix potential buffer overrun problem in BIO_printf().
15865 *Ulf Möller, using public domain code by Patrick Powell; problem
15866 pointed out by David Sacerdote <das33@cornell.edu>*
15868 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15869 RAND_egd() and RAND_status(). In the command line application,
15870 the EGD socket can be specified like a seed file using RANDFILE
15875 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15876 Some CAs (e.g. Verisign) distribute certificates in this form.
15880 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15881 list to exclude them. This means that no special compilation option
15882 is needed to use anonymous DH: it just needs to be included in the
15887 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15888 EVP_MD_type. The old functionality is available in a new macro called
15889 EVP_MD_md(). Change code that uses it and update docs.
15893 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15894 where the `void *` argument is replaced by a function pointer argument.
15895 Previously `void *` was abused to point to functions, which works on
15896 many platforms, but is not correct. As these functions are usually
15897 called by macros defined in OpenSSL header files, most source code
15898 should work without changes.
15902 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15903 sections with information on -D... compiler switches used for
15904 compiling the library so that applications can see them. To enable
15905 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15906 must be defined. E.g.,
15907 #define OPENSSL_ALGORITHM_DEFINES
15908 #include <openssl/opensslconf.h>
15909 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15911 *Richard Levitte, Ulf and Bodo Möller*
15913 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15918 * Change the 'other' type in certificate aux info to a STACK_OF
15919 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15920 the required ASN1 format: arbitrary types determined by an OID.
15924 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15925 argument to 'req'. This is not because the function is newer or
15926 better than others it just uses the work 'NEW' in the certificate
15927 request header lines. Some software needs this.
15931 * Reorganise password command line arguments: now passwords can be
15932 obtained from various sources. Delete the PEM_cb function and make
15933 it the default behaviour: i.e. if the callback is NULL and the
15934 usrdata argument is not NULL interpret it as a null terminated pass
15935 phrase. If usrdata and the callback are NULL then the pass phrase
15936 is prompted for as usual.
15940 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15941 the support is automatically enabled. The resulting binaries will
15942 autodetect the card and use it if present.
15944 *Ben Laurie and Compaq Inc.*
15946 * Work around for Netscape hang bug. This sends certificate request
15947 and server done in one record. Since this is perfectly legal in the
15948 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15949 the bugs/SSLv3 entry for more info.
15953 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15957 * Add -rand argument to smime and pkcs12 applications and read/write
15962 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15966 * Add command line password options to the remaining applications.
15970 * Bug fix for BN_div_recp() for numerators with an even number of
15975 * More tests in bntest.c, and changed test_bn output.
15979 * ./config recognizes MacOS X now.
15983 * Bug fix for BN_div() when the first words of num and divisor are
15984 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15988 * Add support for various broken PKCS#8 formats, and command line
15989 options to produce them.
15993 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15994 get temporary BIGNUMs from a BN_CTX.
15998 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16003 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16004 include a #define from the old name to the new. The original intent
16005 was that statically linked binaries could for example just call
16006 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16007 link with digests. This never worked because SSLeay_add_all_digests()
16008 and SSLeay_add_all_ciphers() were in the same source file so calling
16009 one would link with the other. They are now in separate source files.
16013 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16017 * Use a less unusual form of the Miller-Rabin primality test (it used
16018 a binary algorithm for exponentiation integrated into the Miller-Rabin
16019 loop, our standard modexp algorithms are faster).
16023 * Support for the EBCDIC character set completed.
16025 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16027 * Source code cleanups: use const where appropriate, eliminate casts,
16028 use `void *` instead of `char *` in lhash.
16032 * Bugfix: ssl3_send_server_key_exchange was not restartable
16033 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16034 this the server could overwrite ephemeral keys that the client
16039 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16040 using 50 iterations of the Rabin-Miller test.
16042 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16043 iterations of the Rabin-Miller test as required by the appendix
16044 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16045 As BN_is_prime_fasttest includes trial division, DSA parameter
16046 generation becomes much faster.
16048 This implies a change for the callback functions in DSA_is_prime
16049 and DSA_generate_parameters: The callback function is called once
16050 for each positive witness in the Rabin-Miller test, not just
16051 occasionally in the inner loop; and the parameters to the
16052 callback function now provide an iteration count for the outer
16053 loop rather than for the current invocation of the inner loop.
16054 DSA_generate_parameters additionally can call the callback
16055 function with an 'iteration count' of -1, meaning that a
16056 candidate has passed the trial division test (when q is generated
16057 from an application-provided seed, trial division is skipped).
16061 * New function BN_is_prime_fasttest that optionally does trial
16062 division before starting the Rabin-Miller test and has
16063 an additional BN_CTX * argument (whereas BN_is_prime always
16064 has to allocate at least one BN_CTX).
16065 'callback(1, -1, cb_arg)' is called when a number has passed the
16066 trial division stage.
16070 * Fix for bug in CRL encoding. The validity dates weren't being handled
16075 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16079 * New function BN_pseudo_rand().
16083 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16084 bignum version of BN_from_montgomery() with the working code from
16085 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16090 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16091 made it impossible to use the same SSL_SESSION data structure in
16092 SSL2 clients in multiple threads.
16096 * The return value of RAND_load_file() no longer counts bytes obtained
16097 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16098 to seed the PRNG (previously an explicit byte count was required).
16100 *Ulf Möller, Bodo Möller*
16102 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16103 used `char *` instead of `void *` and had casts all over the place.
16107 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16111 * Retain source code compatibility for BN_prime_checks macro:
16112 BN_is_prime(..., BN_prime_checks, ...) now uses
16113 BN_prime_checks_for_size to determine the appropriate number of
16114 Rabin-Miller iterations.
16118 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16119 DH_CHECK_P_NOT_SAFE_PRIME.
16120 (Check if this is true? OpenPGP calls them "strong".)
16124 * Merge the functionality of "dh" and "gendh" programs into a new program
16125 "dhparam". The old programs are retained for now but will handle DH keys
16126 (instead of parameters) in future.
16130 * Make the ciphers, s_server and s_client programs check the return values
16131 when a new cipher list is set.
16135 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16136 ciphers. Before when the 56bit ciphers were enabled the sorting was
16139 The syntax for the cipher sorting has been extended to support sorting by
16140 cipher-strength (using the strength_bits hard coded in the tables).
16141 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16143 Fix a bug in the cipher-command parser: when supplying a cipher command
16144 string with an "undefined" symbol (neither command nor alphanumeric
16145 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16146 an error is flagged.
16148 Due to the strength-sorting extension, the code of the
16149 ssl_create_cipher_list() function was completely rearranged. I hope that
16150 the readability was also increased :-)
16152 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16154 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16155 for the first serial number and places 2 in the serial number file. This
16156 avoids problems when the root CA is created with serial number zero and
16157 the first user certificate has the same issuer name and serial number
16162 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16163 the new code. Add documentation for this stuff.
16167 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16168 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16169 structures and behave in an analogous way to the X509v3 functions:
16170 they shouldn't be called directly but wrapper functions should be used
16173 So we also now have some wrapper functions that call the X509at functions
16174 when passed certificate requests. (TO DO: similar things can be done with
16175 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16176 things. Some of these need some d2i or i2d and print functionality
16177 because they handle more complex structures.)
16181 * Add missing #ifndefs that caused missing symbols when building libssl
16182 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16183 NO_RSA in `ssl/s2*.c`.
16185 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16187 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16188 has a return value which indicates the quality of the random data
16189 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16190 error queue. New function RAND_pseudo_bytes() generates output that is
16191 guaranteed to be unique but not unpredictable. RAND_add is like
16192 RAND_seed, but takes an extra argument for an entropy estimate
16193 (RAND_seed always assumes full entropy).
16197 * Do more iterations of Rabin-Miller probable prime test (specifically,
16198 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16199 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16200 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16201 false-positive rate of at most 2^-80 for random input.
16205 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16209 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16210 in the 0.9.5 release), this returns the chain
16211 from an X509_CTX structure with a dup of the stack and all
16212 the X509 reference counts upped: so the stack will exist
16213 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16216 Also make SSL_SESSION_print() print out the verify return
16221 * Add manpage for the pkcs12 command. Also change the default
16222 behaviour so MAC iteration counts are used unless the new
16223 -nomaciter option is used. This improves file security and
16224 only older versions of MSIE (4.0 for example) need it.
16228 * Honor the no-xxx Configure options when creating .DEF files.
16232 * Add PKCS#10 attributes to field table: challengePassword,
16233 unstructuredName and unstructuredAddress. These are taken from
16234 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16235 international characters are used.
16237 More changes to X509_ATTRIBUTE code: allow the setting of types
16238 based on strings. Remove the 'loc' parameter when adding
16239 attributes because these will be a SET OF encoding which is sorted
16244 * Initial changes to the 'req' utility to allow request generation
16245 automation. This will allow an application to just generate a template
16246 file containing all the field values and have req construct the
16249 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16250 used all over the place including certificate requests and PKCS#7
16251 structures. They are currently handled manually where necessary with
16252 some primitive wrappers for PKCS#7. The new functions behave in a
16253 manner analogous to the X509 extension functions: they allow
16254 attributes to be looked up by NID and added.
16256 Later something similar to the X509V3 code would be desirable to
16257 automatically handle the encoding, decoding and printing of the
16258 more complex types. The string types like challengePassword can
16259 be handled by the string table functions.
16261 Also modified the multi byte string table handling. Now there is
16262 a 'global mask' which masks out certain types. The table itself
16263 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16264 is useful when for example there is only one permissible type
16265 (as in countryName) and using the mask might result in no valid
16270 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16271 SSL_get_peer_finished to allow applications to obtain the latest
16272 Finished messages sent to the peer or expected from the peer,
16273 respectively. (SSL_get_peer_finished is usually the Finished message
16274 actually received from the peer, otherwise the protocol will be aborted.)
16276 As the Finished message are message digests of the complete handshake
16277 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16278 be used for external authentication procedures when the authentication
16279 provided by SSL/TLS is not desired or is not enough.
16283 * Enhanced support for Alpha Linux is added. Now ./config checks if
16284 the host supports BWX extension and if Compaq C is present on the
16285 $PATH. Just exploiting of the BWX extension results in 20-30%
16286 performance kick for some algorithms, e.g. DES and RC4 to mention
16287 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16292 * Add support for MS "fast SGC". This is arguably a violation of the
16293 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16294 weak crypto and after checking the certificate is SGC a second one
16295 with strong crypto. MS SGC stops the first handshake after receiving
16296 the server certificate message and sends a second client hello. Since
16297 a server will typically do all the time consuming operations before
16298 expecting any further messages from the client (server key exchange
16299 is the most expensive) there is little difference between the two.
16301 To get OpenSSL to support MS SGC we have to permit a second client
16302 hello message after we have sent server done. In addition we have to
16303 reset the MAC if we do get this second client hello.
16307 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16308 if a DER encoded private key is RSA or DSA traditional format. Changed
16309 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16310 format DER encoded private key. Newer code should use PKCS#8 format which
16311 has the key type encoded in the ASN1 structure. Added DER private key
16312 support to pkcs8 application.
16316 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16317 ciphersuites has been selected (as required by the SSL 3/TLS 1
16318 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16319 is set, we interpret this as a request to violate the specification
16320 (the worst that can happen is a handshake failure, and 'correct'
16321 behaviour would result in a handshake failure anyway).
16325 * In SSL_CTX_add_session, take into account that there might be multiple
16326 SSL_SESSION structures with the same session ID (e.g. when two threads
16327 concurrently obtain them from an external cache).
16328 The internal cache can handle only one SSL_SESSION with a given ID,
16329 so if there's a conflict, we now throw out the old one to achieve
16334 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16335 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16336 some routines that use cipher OIDs: some ciphers do not have OIDs
16337 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16342 * Simplify the trust setting structure and code. Now we just have
16343 two sequences of OIDs for trusted and rejected settings. These will
16344 typically have values the same as the extended key usage extension
16345 and any application specific purposes.
16347 The trust checking code now has a default behaviour: it will just
16348 check for an object with the same NID as the passed id. Functions can
16349 be provided to override either the default behaviour or the behaviour
16350 for a given id. SSL client, server and email already have functions
16351 in place for compatibility: they check the NID and also return "trusted"
16352 if the certificate is self signed.
16356 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16357 traditional format into an EVP_PKEY structure.
16361 * Add a password callback function PEM_cb() which either prompts for
16362 a password if usr_data is NULL or otherwise assumes it is a null
16363 terminated password. Allow passwords to be passed on command line
16364 environment or config files in a few more utilities.
16368 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16369 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16370 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16371 Update documentation.
16375 * Support for ASN1 "NULL" type. This could be handled before by using
16376 ASN1_TYPE but there wasn't any function that would try to read a NULL
16377 and produce an error if it couldn't. For compatibility we also have
16378 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16379 don't allocate anything because they don't need to.
16383 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16386 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16388 * Rebuild of the memory allocation routines used by OpenSSL code and
16389 possibly others as well. The purpose is to make an interface that
16390 provide hooks so anyone can build a separate set of allocation and
16391 deallocation routines to be used by OpenSSL, for example memory
16392 pool implementations, or something else, which was previously hard
16393 since Malloc(), Realloc() and Free() were defined as macros having
16394 the values malloc, realloc and free, respectively (except for Win32
16395 compilations). The same is provided for memory debugging code.
16396 OpenSSL already comes with functionality to find memory leaks, but
16397 this gives people a chance to debug other memory problems.
16399 With these changes, a new set of functions and macros have appeared:
16401 CRYPTO_set_mem_debug_functions() [F]
16402 CRYPTO_get_mem_debug_functions() [F]
16403 CRYPTO_dbg_set_options() [F]
16404 CRYPTO_dbg_get_options() [F]
16405 CRYPTO_malloc_debug_init() [M]
16407 The memory debug functions are NULL by default, unless the library
16408 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16409 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16410 gives the standard debugging functions that come with OpenSSL) or
16411 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16412 provided by the library user) must be used. When the standard
16413 debugging functions are used, CRYPTO_dbg_set_options can be used to
16414 request additional information:
16415 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16416 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16418 Also, things like CRYPTO_set_mem_functions will always give the
16419 expected result (the new set of functions is used for allocation
16420 and deallocation) at all times, regardless of platform and compiler
16423 To finish it up, some functions that were never use in any other
16424 way than through macros have a new API and new semantic:
16426 CRYPTO_dbg_malloc()
16427 CRYPTO_dbg_realloc()
16430 All macros of value have retained their old syntax.
16432 *Richard Levitte and Bodo Moeller*
16434 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16435 ordering of SMIMECapabilities wasn't in "strength order" and there
16436 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16441 * Some ASN1 types with illegal zero length encoding (INTEGER,
16442 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16444 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16446 * Merge in my S/MIME library for OpenSSL. This provides a simple
16447 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16448 functionality to handle multipart/signed properly) and a utility
16449 called 'smime' to call all this stuff. This is based on code I
16450 originally wrote for Celo who have kindly allowed it to be
16451 included in OpenSSL.
16455 * Add variants des_set_key_checked and des_set_key_unchecked of
16456 des_set_key (aka des_key_sched). Global variable des_check_key
16457 decides which of these is called by des_set_key; this way
16458 des_check_key behaves as it always did, but applications and
16459 the library itself, which was buggy for des_check_key == 1,
16460 have a cleaner way to pick the version they need.
16464 * New function PKCS12_newpass() which changes the password of a
16469 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16470 dynamic mix. In both cases the ids can be used as an index into the
16471 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16472 functions so they accept a list of the field values and the
16473 application doesn't need to directly manipulate the X509_TRUST
16478 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16483 * Modify the way the V3 extension code looks up extensions. This now
16484 works in a similar way to the object code: we have some "standard"
16485 extensions in a static table which is searched with OBJ_bsearch()
16486 and the application can add dynamic ones if needed. The file
16487 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16488 updated whenever a new extension is added to the core code and kept
16489 in ext_nid order. There is a simple program 'tabtest.c' which checks
16490 this. New extensions are not added too often so this file can readily
16491 be maintained manually.
16493 There are two big advantages in doing things this way. The extensions
16494 can be looked up immediately and no longer need to be "added" using
16495 X509V3_add_standard_extensions(): this function now does nothing.
16496 Side note: I get *lots* of email saying the extension code doesn't
16497 work because people forget to call this function.
16498 Also no dynamic allocation is done unless new extensions are added:
16499 so if we don't add custom extensions there is no need to call
16500 X509V3_EXT_cleanup().
16504 * Modify enc utility's salting as follows: make salting the default. Add a
16505 magic header, so unsalted files fail gracefully instead of just decrypting
16506 to garbage. This is because not salting is a big security hole, so people
16507 should be discouraged from doing it.
16511 * Fixes and enhancements to the 'x509' utility. It allowed a message
16512 digest to be passed on the command line but it only used this
16513 parameter when signing a certificate. Modified so all relevant
16514 operations are affected by the digest parameter including the
16515 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16516 DSA key was used because it didn't fix the digest.
16520 * Initial certificate chain verify code. Currently tests the untrusted
16521 certificates for consistency with the verify purpose (which is set
16522 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16524 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16525 this is because it will reject chains with invalid extensions whereas
16526 every previous version of OpenSSL and SSLeay made no checks at all.
16528 Trust code: checks the root CA for the relevant trust settings. Trust
16529 settings have an initial value consistent with the verify purpose: e.g.
16530 if the verify purpose is for SSL client use it expects the CA to be
16531 trusted for SSL client use. However the default value can be changed to
16532 permit custom trust settings: one example of this would be to only trust
16533 certificates from a specific "secure" set of CAs.
16535 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16536 which should be used for version portability: especially since the
16537 verify structure is likely to change more often now.
16539 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16540 to set them. If not set then assume SSL clients will verify SSL servers
16543 Two new options to the verify program: -untrusted allows a set of
16544 untrusted certificates to be passed in and -purpose which sets the
16545 intended purpose of the certificate. If a purpose is set then the
16546 new chain verify code is used to check extension consistency.
16550 * Support for the authority information access extension.
16554 * Modify RSA and DSA PEM read routines to transparently handle
16555 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16556 public keys in a format compatible with certificate
16557 SubjectPublicKeyInfo structures. Unfortunately there were already
16558 functions called *_PublicKey_* which used various odd formats so
16559 these are retained for compatibility: however the DSA variants were
16560 never in a public release so they have been deleted. Changed dsa/rsa
16561 utilities to handle the new format: note no releases ever handled public
16562 keys so we should be OK.
16564 The primary motivation for this change is to avoid the same fiasco
16565 that dogs private keys: there are several incompatible private key
16566 formats some of which are standard and some OpenSSL specific and
16567 require various evil hacks to allow partial transparent handling and
16568 even then it doesn't work with DER formats. Given the option anything
16569 other than PKCS#8 should be dumped: but the other formats have to
16570 stay in the name of compatibility.
16572 With public keys and the benefit of hindsight one standard format
16573 is used which works with EVP_PKEY, RSA or DSA structures: though
16574 it clearly returns an error if you try to read the wrong kind of key.
16576 Added a -pubkey option to the 'x509' utility to output the public key.
16577 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16578 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16579 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16580 that do the same as the `EVP_PKEY_assign_*()` except they up the
16581 reference count of the added key (they don't "swallow" the
16586 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16587 CRLs would fail if the file contained no certificates or no CRLs:
16588 added a new function to read in both types and return the number
16589 read: this means that if none are read it will be an error. The
16590 DER versions of the certificate and CRL reader would always fail
16591 because it isn't possible to mix certificates and CRLs in DER format
16592 without choking one or the other routine. Changed this to just read
16593 a certificate: this is the best we can do. Also modified the code
16594 in `apps/verify.c` to take notice of return codes: it was previously
16595 attempting to read in certificates from NULL pointers and ignoring
16596 any errors: this is one reason why the cert and CRL reader seemed
16597 to work. It doesn't check return codes from the default certificate
16598 routines: these may well fail if the certificates aren't installed.
16602 * Code to support otherName option in GeneralName.
16606 * First update to verify code. Change the verify utility
16607 so it warns if it is passed a self signed certificate:
16608 for consistency with the normal behaviour. X509_verify
16609 has been modified to it will now verify a self signed
16610 certificate if *exactly* the same certificate appears
16611 in the store: it was previously impossible to trust a
16612 single self signed certificate. This means that:
16613 openssl verify ss.pem
16614 now gives a warning about a self signed certificate but
16615 openssl verify -CAfile ss.pem ss.pem
16620 * For servers, store verify_result in SSL_SESSION data structure
16621 (and add it to external session representation).
16622 This is needed when client certificate verifications fails,
16623 but an application-provided verification callback (set by
16624 SSL_CTX_set_cert_verify_callback) allows accepting the session
16625 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16626 but returns 1): When the session is reused, we have to set
16627 ssl->verify_result to the appropriate error code to avoid
16630 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16632 * Fix a bug in the new PKCS#7 code: it didn't consider the
16633 case in PKCS7_dataInit() where the signed PKCS7 structure
16634 didn't contain any existing data because it was being created.
16636 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16638 * Add a salt to the key derivation routines in enc.c. This
16639 forms the first 8 bytes of the encrypted file. Also add a
16640 -S option to allow a salt to be input on the command line.
16644 * New function X509_cmp(). Oddly enough there wasn't a function
16645 to compare two certificates. We do this by working out the SHA1
16646 hash and comparing that. X509_cmp() will be needed by the trust
16651 * SSL_get1_session() is like SSL_get_session(), but increments
16652 the reference count in the SSL_SESSION returned.
16654 *Geoff Thorpe <geoff@eu.c2.net>*
16656 * Fix for 'req': it was adding a null to request attributes.
16657 Also change the X509_LOOKUP and X509_INFO code to handle
16658 certificate auxiliary information.
16662 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16667 * Add the possibility to add extra information to the memory leak
16668 detecting output, to form tracebacks, showing from where each
16669 allocation was originated: CRYPTO_push_info("constant string") adds
16670 the string plus current file name and line number to a per-thread
16671 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16672 is like calling CYRPTO_pop_info() until the stack is empty.
16673 Also updated memory leak detection code to be multi-thread-safe.
16677 * Add options -text and -noout to pkcs7 utility and delete the
16678 encryption options which never did anything. Update docs.
16682 * Add options to some of the utilities to allow the pass phrase
16683 to be included on either the command line (not recommended on
16684 OSes like Unix) or read from the environment. Update the
16685 manpages and fix a few bugs.
16689 * Add a few manpages for some of the openssl commands.
16693 * Fix the -revoke option in ca. It was freeing up memory twice,
16694 leaking and not finding already revoked certificates.
16698 * Extensive changes to support certificate auxiliary information.
16699 This involves the use of X509_CERT_AUX structure and X509_AUX
16700 functions. An X509_AUX function such as PEM_read_X509_AUX()
16701 can still read in a certificate file in the usual way but it
16702 will also read in any additional "auxiliary information". By
16703 doing things this way a fair degree of compatibility can be
16704 retained: existing certificates can have this information added
16705 using the new 'x509' options.
16707 Current auxiliary information includes an "alias" and some trust
16708 settings. The trust settings will ultimately be used in enhanced
16709 certificate chain verification routines: currently a certificate
16710 can only be trusted if it is self signed and then it is trusted
16715 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16716 The problem was that one of the replacement routines had not been working
16717 since SSLeay releases. For now the offending routine has been replaced
16718 with non-optimised assembler. Even so, this now gives around 95%
16719 performance improvement for 1024 bit RSA signs.
16723 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16724 handling. Most clients have the effective key size in bits equal to
16725 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16726 A few however don't do this and instead use the size of the decrypted key
16727 to determine the RC2 key length and the AlgorithmIdentifier to determine
16728 the effective key length. In this case the effective key length can still
16729 be 40 bits but the key length can be 168 bits for example. This is fixed
16730 by manually forcing an RC2 key into the EVP_PKEY structure because the
16731 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16732 the key length and effective key length are equal.
16736 * Add a bunch of functions that should simplify the creation of
16737 X509_NAME structures. Now you should be able to do:
16738 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16739 and have it automatically work out the correct field type and fill in
16740 the structures. The more adventurous can try:
16741 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16742 and it will (hopefully) work out the correct multibyte encoding.
16746 * Change the 'req' utility to use the new field handling and multibyte
16747 copy routines. Before the DN field creation was handled in an ad hoc
16748 way in req, ca, and x509 which was rather broken and didn't support
16749 BMPStrings or UTF8Strings. Since some software doesn't implement
16750 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16751 using the dirstring_type option. See the new comment in the default
16752 openssl.cnf for more info.
16756 * Make crypto/rand/md_rand.c more robust:
16757 - Assure unique random numbers after fork().
16758 - Make sure that concurrent threads access the global counter and
16759 md serializably so that we never lose entropy in them
16760 or use exactly the same state in multiple threads.
16761 Access to the large state is not always serializable because
16762 the additional locking could be a performance killer, and
16763 md should be large enough anyway.
16767 * New file `apps/app_rand.c` with commonly needed functionality
16768 for handling the random seed file.
16770 Use the random seed file in some applications that previously did not:
16772 dsaparam -genkey (which also ignored its '-rand' option),
16775 x509 (when signing).
16776 Except on systems with /dev/urandom, it is crucial to have a random
16777 seed file at least for key creation, DSA signing, and for DH exchanges;
16778 for RSA signatures we could do without one.
16780 gendh and gendsa (unlike genrsa) used to read only the first byte
16781 of each file listed in the '-rand' option. The function as previously
16782 found in genrsa is now in app_rand.c and is used by all programs
16783 that support '-rand'.
16787 * In RAND_write_file, use mode 0600 for creating files;
16788 don't just chmod when it may be too late.
16792 * Report an error from X509_STORE_load_locations
16793 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16797 * New function ASN1_mbstring_copy() this copies a string in either
16798 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16799 into an ASN1_STRING type. A mask of permissible types is passed
16800 and it chooses the "minimal" type to use or an error if not type
16805 * Add function equivalents to the various macros in asn1.h. The old
16806 macros are retained with an `M_` prefix. Code inside the library can
16807 use the `M_` macros. External code (including the openssl utility)
16808 should *NOT* in order to be "shared library friendly".
16812 * Add various functions that can check a certificate's extensions
16813 to see if it usable for various purposes such as SSL client,
16814 server or S/MIME and CAs of these types. This is currently
16815 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16816 verification. Also added a -purpose flag to x509 utility to
16817 print out all the purposes.
16821 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16826 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16827 for, obtain and decode and extension and obtain its critical flag.
16828 This allows all the necessary extension code to be handled in a
16829 single function call.
16833 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16834 platforms. See crypto/rc4/rc4_enc.c for further details.
16838 * New -noout option to asn1parse. This causes no output to be produced
16839 its main use is when combined with -strparse and -out to extract data
16840 from a file (which may not be in ASN.1 format).
16844 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16845 when producing the local key id.
16847 *Richard Levitte <levitte@stacken.kth.se>*
16849 * New option -dhparam in s_server. This allows a DH parameter file to be
16850 stated explicitly. If it is not stated then it tries the first server
16851 certificate file. The previous behaviour hard coded the filename
16856 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16857 a public key to be input or output. For example:
16858 openssl rsa -in key.pem -pubout -out pubkey.pem
16859 Also added necessary DSA public key functions to handle this.
16863 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16864 in the message. This was handled by allowing
16865 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16867 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16869 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16870 to the end of the strings whereas this didn't. This would cause problems
16871 if strings read with d2i_ASN1_bytes() were later modified.
16873 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16875 * Fix for base64 decode bug. When a base64 bio reads only one line of
16876 data and it contains EOF it will end up returning an error. This is
16877 caused by input 46 bytes long. The cause is due to the way base64
16878 BIOs find the start of base64 encoded data. They do this by trying a
16879 trial decode on each line until they find one that works. When they
16880 do a flag is set and it starts again knowing it can pass all the
16881 data directly through the decoder. Unfortunately it doesn't reset
16882 the context it uses. This means that if EOF is reached an attempt
16883 is made to pass two EOFs through the context and this causes the
16884 resulting error. This can also cause other problems as well. As is
16885 usual with these problems it takes *ages* to find and the fix is
16886 trivial: move one line.
16888 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16890 * Ugly workaround to get s_client and s_server working under Windows. The
16891 old code wouldn't work because it needed to select() on sockets and the
16892 tty (for keypresses and to see if data could be written). Win32 only
16893 supports select() on sockets so we select() with a 1s timeout on the
16894 sockets and then see if any characters are waiting to be read, if none
16895 are present then we retry, we also assume we can always write data to
16896 the tty. This isn't nice because the code then blocks until we've
16897 received a complete line of data and it is effectively polling the
16898 keyboard at 1s intervals: however it's quite a bit better than not
16899 working at all :-) A dedicated Windows application might handle this
16900 with an event loop for example.
16904 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16905 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16906 will be called when RSA_sign() and RSA_verify() are used. This is useful
16907 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16908 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16909 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16910 This necessitated the support of an extra signature type NID_md5_sha1
16911 for SSL signatures and modifications to the SSL library to use it instead
16912 of calling RSA_public_decrypt() and RSA_private_encrypt().
16916 * Add new -verify -CAfile and -CApath options to the crl program, these
16917 will lookup a CRL issuers certificate and verify the signature in a
16918 similar way to the verify program. Tidy up the crl program so it
16919 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16920 less strict. It will now permit CRL extensions even if it is not
16921 a V2 CRL: this will allow it to tolerate some broken CRLs.
16925 * Initialize all non-automatic variables each time one of the openssl
16926 sub-programs is started (this is necessary as they may be started
16927 multiple times from the "OpenSSL>" prompt).
16929 *Lennart Bang, Bodo Moeller*
16931 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16932 removing all other RSA functionality (this is what NO_RSA does). This
16933 is so (for example) those in the US can disable those operations covered
16934 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16939 * Non-copying interface to BIO pairs.
16940 (still largely untested)
16944 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16945 ASCII string. This was handled independently in various places before.
16949 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16950 UTF8 strings a character at a time.
16954 * Use client_version from client hello to select the protocol
16955 (s23_srvr.c) and for RSA client key exchange verification
16956 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16960 * Add various utility functions to handle SPKACs, these were previously
16961 handled by poking round in the structure internals. Added new function
16962 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16963 print, verify and generate SPKACs. Based on an original idea from
16964 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16968 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16972 * Allow the config file extension section to be overwritten on the
16973 command line. Based on an original idea from Massimiliano Pala
16974 <madwolf@comune.modena.it>. The new option is called -extensions
16975 and can be applied to ca, req and x509. Also -reqexts to override
16976 the request extensions in req and -crlexts to override the crl extensions
16981 * Add new feature to the SPKAC handling in ca. Now you can include
16982 the same field multiple times by preceding it by "XXXX." for example:
16985 this is the same syntax as used in the req config file.
16989 * Allow certificate extensions to be added to certificate requests. These
16990 are specified in a 'req_extensions' option of the req section of the
16991 config file. They can be printed out with the -text option to req but
16992 are otherwise ignored at present.
16996 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16997 data read consists of only the final block it would not decrypted because
16998 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16999 A misplaced 'break' also meant the decrypted final block might not be
17000 copied until the next read.
17004 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17005 a few extra parameters to the DH structure: these will be useful if
17006 for example we want the value of 'q' or implement X9.42 DH.
17010 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17011 provides hooks that allow the default DSA functions or functions on a
17012 "per key" basis to be replaced. This allows hardware acceleration and
17013 hardware key storage to be handled without major modification to the
17014 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17015 associated functions.
17019 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17020 as "read only": it can't be written to and the buffer it points to will
17021 not be freed. Reading from a read only BIO is much more efficient than
17022 a normal memory BIO. This was added because there are several times when
17023 an area of memory needs to be read from a BIO. The previous method was
17024 to create a memory BIO and write the data to it, this results in two
17025 copies of the data and an O(n^2) reading algorithm. There is a new
17026 function BIO_new_mem_buf() which creates a read only memory BIO from
17027 an area of memory. Also modified the PKCS#7 routines to use read only
17032 * Bugfix: ssl23_get_client_hello did not work properly when called in
17033 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17034 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17035 but a retry condition occurred while trying to read the rest.
17039 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17040 NID_pkcs7_encrypted by default: this was wrong since this should almost
17041 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17042 the encrypted data type: this is a more sensible place to put it and it
17043 allows the PKCS#12 code to be tidied up that duplicated this
17048 * Changed obj_dat.pl script so it takes its input and output files on
17049 the command line. This should avoid shell escape redirection problems
17054 * Initial support for certificate extension requests, these are included
17055 in things like Xenroll certificate requests. Included functions to allow
17056 extensions to be obtained and added.
17060 * -crlf option to s_client and s_server for sending newlines as
17061 CRLF (as required by many protocols).
17065 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17067 * Install libRSAglue.a when OpenSSL is built with RSAref.
17069 *Ralf S. Engelschall*
17071 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17073 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17075 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17080 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17081 DH parameters/keys (q is lost during that conversion, but the resulting
17082 DH parameters contain its length).
17084 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17085 much faster than DH_generate_parameters (which creates parameters
17086 where `p = 2*q + 1`), and also the smaller q makes DH computations
17087 much more efficient (160-bit exponentiation instead of 1024-bit
17088 exponentiation); so this provides a convenient way to support DHE
17089 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17090 utter importance to use
17091 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17093 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17094 when such DH parameters are used, because otherwise small subgroup
17095 attacks may become possible!
17099 * Avoid memory leak in i2d_DHparams.
17103 * Allow the -k option to be used more than once in the enc program:
17104 this allows the same encrypted message to be read by multiple recipients.
17108 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17109 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17110 it will always use the numerical form of the OID, even if it has a short
17115 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17116 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17117 otherwise bn_mod_exp was called. In the case of hardware keys for example
17118 no private key components need be present and it might store extra data
17119 in the RSA structure, which cannot be accessed from bn_mod_exp.
17120 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17121 private key operations.
17125 * Added support for SPARC Linux.
17129 * pem_password_cb function type incompatibly changed from
17130 typedef int pem_password_cb(char *buf, int size, int rwflag);
17132 ....(char *buf, int size, int rwflag, void *userdata);
17133 so that applications can pass data to their callbacks:
17134 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17135 additional void * argument, which is just handed through whenever
17136 the password callback is called.
17138 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17140 New function SSL_CTX_set_default_passwd_cb_userdata.
17142 Compatibility note: As many C implementations push function arguments
17143 onto the stack in reverse order, the new library version is likely to
17144 interoperate with programs that have been compiled with the old
17145 pem_password_cb definition (PEM_whatever takes some data that
17146 happens to be on the stack as its last argument, and the callback
17147 just ignores this garbage); but there is no guarantee whatsoever that
17150 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17151 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17152 problems not only on Windows, but also on some Unix platforms.
17153 To avoid problematic command lines, these definitions are now in an
17154 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17155 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17159 * MIPS III/IV assembler module is reimplemented.
17163 * More DES library cleanups: remove references to srand/rand and
17164 delete an unused file.
17168 * Add support for the free Netwide assembler (NASM) under Win32,
17169 since not many people have MASM (ml) and it can be hard to obtain.
17170 This is currently experimental but it seems to work OK and pass all
17171 the tests. Check out INSTALL.W32 for info.
17175 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17176 without temporary keys kept an extra copy of the server key,
17177 and connections with temporary keys did not free everything in case
17182 * New function RSA_check_key and new openssl rsa option -check
17183 for verifying the consistency of RSA keys.
17185 *Ulf Moeller, Bodo Moeller*
17187 * Various changes to make Win32 compile work:
17188 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17189 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17190 comparison" warnings.
17191 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17195 * Add a debugging option to PKCS#5 v2 key generation function: when
17196 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17197 derived keys are printed to stderr.
17201 * Copy the flags in ASN1_STRING_dup().
17203 *Roman E. Pavlov <pre@mo.msk.ru>*
17205 * The x509 application mishandled signing requests containing DSA
17206 keys when the signing key was also DSA and the parameters didn't match.
17208 It was supposed to omit the parameters when they matched the signing key:
17209 the verifying software was then supposed to automatically use the CA's
17210 parameters if they were absent from the end user certificate.
17212 Omitting parameters is no longer recommended. The test was also
17213 the wrong way round! This was probably due to unusual behaviour in
17214 EVP_cmp_parameters() which returns 1 if the parameters match.
17215 This meant that parameters were omitted when they *didn't* match and
17216 the certificate was useless. Certificates signed with 'ca' didn't have
17219 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17221 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17222 The interface is as follows:
17223 Applications can use
17224 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17225 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17226 "off" is now the default.
17227 The library internally uses
17228 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17229 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17230 to disable memory-checking temporarily.
17232 Some inconsistent states that previously were possible (and were
17233 even the default) are now avoided.
17235 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17236 with each memory chunk allocated; this is occasionally more helpful
17237 than just having a counter.
17239 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17241 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17246 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17247 which largely parallels "options", but is for changing API behaviour,
17248 whereas "options" are about protocol behaviour.
17249 Initial "mode" flags are:
17251 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17252 a single record has been written.
17253 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17254 retries use the same buffer location.
17255 (But all of the contents must be
17260 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17263 * Fix problems with no-hmac etc.
17265 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17267 * New functions RSA_get_default_method(), RSA_set_method() and
17268 RSA_get_method(). These allows replacement of RSA_METHODs without having
17269 to mess around with the internals of an RSA structure.
17273 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17274 Also really enable memory leak checks in openssl.c and in some
17277 *Chad C. Mulligan, Bodo Moeller*
17279 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17280 up the length of negative integers. This has now been simplified to just
17281 store the length when it is first determined and use it later, rather
17282 than trying to keep track of where data is copied and updating it to
17284 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17286 * Add a new function PKCS7_signatureVerify. This allows the verification
17287 of a PKCS#7 signature but with the signing certificate passed to the
17288 function itself. This contrasts with PKCS7_dataVerify which assumes the
17289 certificate is present in the PKCS#7 structure. This isn't always the
17290 case: certificates can be omitted from a PKCS#7 structure and be
17291 distributed by "out of band" means (such as a certificate database).
17295 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17296 function prototypes in pem.h, also change util/mkdef.pl to add the
17297 necessary function names.
17301 * mk1mf.pl (used by Windows builds) did not properly read the
17302 options set by Configure in the top level Makefile, and Configure
17303 was not even able to write more than one option correctly.
17304 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17308 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17309 file to be loaded from a BIO or FILE pointer. The BIO version will
17310 for example allow memory BIOs to contain config info.
17314 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17315 Whoever hopes to achieve shared-library compatibility across versions
17316 must use this, not the compile-time macro.
17317 (Exercise 0.9.4: Which is the minimum library version required by
17319 Note: All this applies only to multi-threaded programs, others don't
17324 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17325 through a BIO pair triggered the default case, i.e.
17326 SSLerr(...,SSL_R_UNKNOWN_STATE).
17330 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17331 can use the SSL library even if none of the specific BIOs is
17336 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17337 for the encoded length.
17339 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17341 * Add initial documentation of the X509V3 functions.
17345 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17346 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17347 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17348 secure PKCS#8 private key format with a high iteration count.
17352 * Fix determination of Perl interpreter: A perl or perl5
17353 *directory* in $PATH was also accepted as the interpreter.
17355 *Ralf S. Engelschall*
17357 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17358 wrong with it but it was very old and did things like calling
17359 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17360 unusual formatting.
17364 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17365 to use the new extension code.
17369 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17370 with macros. This should make it easier to change their form, add extra
17371 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17376 * Add to configuration table a new entry that can specify an alternative
17377 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17378 according to Mark Crispin <MRC@Panda.COM>.
17382 * DES CBC did not update the IV. Weird.
17386 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17387 Changing the behaviour of the former might break existing programs --
17388 where IV updating is needed, des_ncbc_encrypt can be used.
17391 * When bntest is run from "make test" it drives bc to check its
17392 calculations, as well as internally checking them. If an internal check
17393 fails, it needs to cause bc to give a non-zero result or make test carries
17394 on without noticing the failure. Fixed.
17398 * DES library cleanups.
17402 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17403 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17404 ciphers. NOTE: although the key derivation function has been verified
17405 against some published test vectors it has not been extensively tested
17406 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17411 * Instead of "mkdir -p", which is not fully portable, use new
17412 Perl script "util/mkdir-p.pl".
17416 * Rewrite the way password based encryption (PBE) is handled. It used to
17417 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17418 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17419 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17420 the 'parameter' field of the AlgorithmIdentifier is passed to the
17421 underlying key generation function so it must do its own ASN1 parsing.
17422 This has also changed the EVP_PBE_CipherInit() function which now has a
17423 'parameter' argument instead of literal salt and iteration count values
17424 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17428 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17429 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17430 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17431 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17432 value was just used as a "magic string" and not used directly its
17433 value doesn't matter.
17437 * Introduce some semblance of const correctness to BN. Shame C doesn't
17442 * "linux-sparc64" configuration (ultrapenguin).
17444 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17445 "linux-sparc" configuration.
17447 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17449 * config now generates no-xxx options for missing ciphers.
17453 * Support the EBCDIC character set (work in progress).
17454 File ebcdic.c not yet included because it has a different license.
17456 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17458 * Support BS2000/OSD-POSIX.
17460 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17462 * Make callbacks for key generation use `void *` instead of `char *`.
17466 * Make S/MIME samples compile (not yet tested).
17470 * Additional typesafe stacks.
17474 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17478 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17480 * New configuration variant "sco5-gcc".
17482 * Updated some demos.
17484 *Sean O Riordain, Wade Scholine*
17486 * Add missing BIO_free at exit of pkcs12 application.
17490 * Fix memory leak in conf.c.
17494 * Updates for Win32 to assembler version of MD5.
17498 * Set #! path to perl in `apps/der_chop` to where we found it
17499 instead of using a fixed path.
17503 * SHA library changes for irix64-mips4-cc.
17507 * Improvements for VMS support.
17511 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17513 * Bignum library bug fix. IRIX 6 passes "make test" now!
17514 This also avoids the problems with SC4.2 and unpatched SC5.
17516 *Andy Polyakov <appro@fy.chalmers.se>*
17518 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17519 These are required because of the typesafe stack would otherwise break
17520 existing code. If old code used a structure member which used to be STACK
17521 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17522 sk_num or sk_value it would produce an error because the num, data members
17523 are not present in STACK_OF. Now it just produces a warning. sk_set
17524 replaces the old method of assigning a value to sk_value
17525 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17526 that does this will no longer work (and should use sk_set instead) but
17527 this could be regarded as a "questionable" behaviour anyway.
17531 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17532 correctly handle encrypted S/MIME data.
17536 * Change type of various DES function arguments from des_cblock
17537 (which means, in function argument declarations, pointer to char)
17538 to des_cblock * (meaning pointer to array with 8 char elements),
17539 which allows the compiler to do more typechecking; it was like
17540 that back in SSLeay, but with lots of ugly casts.
17542 Introduce new type const_des_cblock.
17546 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17547 problems: find RecipientInfo structure that matches recipient certificate
17548 and initialise the ASN1 structures properly based on passed cipher.
17552 * Belatedly make the BN tests actually check the results.
17556 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17557 to and from BNs: it was completely broken. New compilation option
17558 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17559 key elements as negative integers.
17563 * Reorganize and speed up MD5.
17565 *Andy Polyakov <appro@fy.chalmers.se>*
17569 *Richard Levitte <richard@levitte.org>*
17571 * New option -out to asn1parse to allow the parsed structure to be
17572 output to a file. This is most useful when combined with the -strparse
17573 option to examine the output of things like OCTET STRINGS.
17577 * Make SSL library a little more fool-proof by not requiring any longer
17578 that `SSL_set_{accept,connect}_state` be called before
17579 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17580 in many applications because usually everything *appeared* to work as
17581 intended anyway -- now it really works as intended).
17585 * Move openssl.cnf out of lib/.
17589 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17590 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17591 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17593 *Ralf S. Engelschall*
17595 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17596 handle PKCS#7 enveloped data properly.
17598 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17600 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17601 copying pointers. The cert_st handling is changed by this in
17602 various ways (and thus what used to be known as ctx->default_cert
17603 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17604 any longer when s->cert does not give us what we need).
17605 ssl_cert_instantiate becomes obsolete by this change.
17606 As soon as we've got the new code right (possibly it already is?),
17607 we have solved a couple of bugs of the earlier code where s->cert
17608 was used as if it could not have been shared with other SSL structures.
17610 Note that using the SSL API in certain dirty ways now will result
17611 in different behaviour than observed with earlier library versions:
17612 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17613 does not influence s as it used to.
17615 In order to clean up things more thoroughly, inside SSL_SESSION
17616 we don't use CERT any longer, but a new structure SESS_CERT
17617 that holds per-session data (if available); currently, this is
17618 the peer's certificate chain and, for clients, the server's certificate
17619 and temporary key. CERT holds only those values that can have
17620 meaningful defaults in an SSL_CTX.
17624 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17625 from the internal representation. Various PKCS#7 fixes: remove some
17626 evil casts and set the enc_dig_alg field properly based on the signing
17631 * Allow PKCS#12 password to be set from the command line or the
17632 environment. Let 'ca' get its config file name from the environment
17633 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17638 * Allow certificate policies extension to use an IA5STRING for the
17639 organization field. This is contrary to the PKIX definition but
17640 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17645 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17646 without disallowing inline assembler and the like for non-pedantic builds.
17650 * Support Borland C++ builder.
17652 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17658 * SHA-1 cleanups and performance enhancements.
17660 *Andy Polyakov <appro@fy.chalmers.se>*
17662 * Sparc v8plus assembler for the bignum library.
17664 *Andy Polyakov <appro@fy.chalmers.se>*
17666 * Accept any -xxx and +xxx compiler options in Configure.
17670 * Update HPUX configuration.
17674 * Add missing `sk_<type>_unshift()` function to safestack.h
17676 *Ralf S. Engelschall*
17678 * New function SSL_CTX_use_certificate_chain_file that sets the
17679 "extra_cert"s in addition to the certificate. (This makes sense
17680 only for "PEM" format files, as chains as a whole are not
17685 * Support verify_depth from the SSL API.
17686 x509_vfy.c had what can be considered an off-by-one-error:
17687 Its depth (which was not part of the external interface)
17688 was actually counting the number of certificates in a chain;
17689 now it really counts the depth.
17693 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17694 instead of X509err, which often resulted in confusing error
17695 messages since the error codes are not globally unique
17696 (e.g. an alleged error in ssl3_accept when a certificate
17697 didn't match the private key).
17699 * New function SSL_CTX_set_session_id_context that allows to set a default
17700 value (so that you don't need SSL_set_session_id_context for each
17701 connection using the SSL_CTX).
17705 * OAEP decoding bug fix.
17709 * Support INSTALL_PREFIX for package builders, as proposed by
17714 * New Configure options "threads" and "no-threads". For systems
17715 where the proper compiler options are known (currently Solaris
17716 and Linux), "threads" is the default.
17720 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17724 * Install various scripts to $(OPENSSLDIR)/misc, not to
17725 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17726 such as /usr/local/bin.
17730 * "make linux-shared" to build shared libraries.
17732 *Niels Poppe <niels@netbox.org>*
17734 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17738 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17739 extension adding in x509 utility.
17743 * Remove NOPROTO sections and error code comments.
17747 * Partial rewrite of the DEF file generator to now parse the ANSI
17752 * New Configure options --prefix=DIR and --openssldir=DIR.
17756 * Complete rewrite of the error code script(s). It is all now handled
17757 by one script at the top level which handles error code gathering,
17758 header rewriting and C source file generation. It should be much better
17759 than the old method: it now uses a modified version of Ulf's parser to
17760 read the ANSI prototypes in all header files (thus the old K&R definitions
17761 aren't needed for error creation any more) and do a better job of
17762 translating function codes into names. The old 'ASN1 error code embedded
17763 in a comment' is no longer necessary and it doesn't use .err files which
17764 have now been deleted. Also the error code call doesn't have to appear all
17765 on one line (which resulted in some large lines...).
17769 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17773 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17774 0 (which usually indicates a closed connection), but continue reading.
17778 * Fix some race conditions.
17782 * Add support for CRL distribution points extension. Add Certificate
17783 Policies and CRL distribution points documentation.
17787 * Move the autogenerated header file parts to crypto/opensslconf.h.
17791 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17792 8 of keying material. Merlin has also confirmed interop with this fix
17793 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17795 *Merlin Hughes <merlin@baltimore.ie>*
17797 * Fix lots of warnings.
17799 *Richard Levitte <levitte@stacken.kth.se>*
17801 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17802 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17804 *Richard Levitte <levitte@stacken.kth.se>*
17806 * Fix problems with sizeof(long) == 8.
17808 *Andy Polyakov <appro@fy.chalmers.se>*
17810 * Change functions to ANSI C.
17814 * Fix typos in error codes.
17816 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17818 * Remove defunct assembler files from Configure.
17822 * SPARC v8 assembler BIGNUM implementation.
17824 *Andy Polyakov <appro@fy.chalmers.se>*
17826 * Support for Certificate Policies extension: both print and set.
17827 Various additions to support the r2i method this uses.
17831 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17832 return a const string when you are expecting an allocated buffer.
17836 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17837 types DirectoryString and DisplayText.
17841 * Add code to allow r2i extensions to access the configuration database,
17842 add an LHASH database driver and add several ctx helper functions.
17846 * Fix an evil bug in bn_expand2() which caused various BN functions to
17847 fail when they extended the size of a BIGNUM.
17851 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17852 support typesafe stack.
17856 * Fix typo in SSL_[gs]et_options().
17858 *Nils Frostberg <nils@medcom.se>*
17860 * Delete various functions and files that belonged to the (now obsolete)
17861 old X509V3 handling code.
17865 * New Configure option "rsaref".
17869 * Don't auto-generate pem.h.
17873 * Introduce type-safe ASN.1 SETs.
17877 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17879 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17881 * Introduce type-safe STACKs. This will almost certainly break lots of code
17882 that links with OpenSSL (well at least cause lots of warnings), but fear
17883 not: the conversion is trivial, and it eliminates loads of evil casts. A
17884 few STACKed things have been converted already. Feel free to convert more.
17885 In the fullness of time, I'll do away with the STACK type altogether.
17889 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17890 specified in `<certfile>` by updating the entry in the index.txt file.
17891 This way one no longer has to edit the index.txt file manually for
17892 revoking a certificate. The -revoke option does the gory details now.
17894 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17896 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17897 `-text` option at all and this way the `-noout -text` combination was
17898 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17900 *Ralf S. Engelschall*
17902 * Make sure a corresponding plain text error message exists for the
17903 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17904 verify callback function determined that a certificate was revoked.
17906 *Ralf S. Engelschall*
17908 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17909 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17910 all available ciphers including rc5, which was forgotten until now.
17911 In order to let the testing shell script know which algorithms
17912 are available, a new (up to now undocumented) command
17913 `openssl list-cipher-commands` is used.
17917 * Bugfix: s_client occasionally would sleep in select() when
17918 it should have checked SSL_pending() first.
17922 * New functions DSA_do_sign and DSA_do_verify to provide access to
17923 the raw DSA values prior to ASN.1 encoding.
17927 * Tweaks to Configure
17929 *Niels Poppe <niels@netbox.org>*
17931 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17936 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17940 * New config option to avoid instructions that are illegal on the 80386.
17941 The default code is faster, but requires at least a 486.
17945 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17946 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17947 same as SSL2_VERSION anyway.
17951 * New "-showcerts" option for s_client.
17955 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17956 application. Various cleanups and fixes.
17960 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17961 modify error routines to work internally. Add error codes and PBE init
17962 to library startup routines.
17966 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17967 packing functions to asn1 and evp. Changed function names and error
17968 codes along the way.
17972 * PKCS12 integration: and so it begins... First of several patches to
17973 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17974 objects to objects.h
17978 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17979 and display support for Thawte strong extranet extension.
17983 * Add LinuxPPC support.
17985 *Jeff Dubrule <igor@pobox.org>*
17987 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17988 bn_div_words in alpha.s.
17990 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17992 * Make sure the RSA OAEP test is skipped under -DRSAref because
17993 OAEP isn't supported when OpenSSL is built with RSAref.
17995 *Ulf Moeller <ulf@fitug.de>*
17997 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17998 so they no longer are missing under -DNOPROTO.
18000 *Soren S. Jorvang <soren@t.dk>*
18002 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18004 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18005 doesn't work when the session is reused. Coming soon!
18009 * Fix a security hole, that allows sessions to be reused in the wrong
18010 context thus bypassing client cert protection! All software that uses
18011 client certs and session caches in multiple contexts NEEDS PATCHING to
18012 allow session reuse! A fuller solution is in the works.
18014 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18016 * Some more source tree cleanups (removed obsolete files
18017 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18018 permission on "config" script to be executable) and a fix for the INSTALL
18021 *Ulf Moeller <ulf@fitug.de>*
18023 * Remove some legacy and erroneous uses of malloc, free instead of
18026 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18028 * Make rsa_oaep_test return non-zero on error.
18030 *Ulf Moeller <ulf@fitug.de>*
18032 * Add support for native Solaris shared libraries. Configure
18033 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18034 if someone would make that last step automatic.
18036 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18038 * ctx_size was not built with the right compiler during "make links". Fixed.
18042 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18043 except NULL ciphers". This means the default cipher list will no longer
18044 enable NULL ciphers. They need to be specifically enabled e.g. with
18045 the string "DEFAULT:eNULL".
18049 * Fix to RSA private encryption routines: if p < q then it would
18050 occasionally produce an invalid result. This will only happen with
18051 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18055 * Be less restrictive and allow also `perl util/perlpath.pl
18056 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18057 because this way one can also use an interpreter named `perl5` (which is
18058 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18059 installed as `perl`).
18061 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18063 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18065 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18067 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18068 advapi32.lib to Win32 build and change the pem test comparison
18069 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18070 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18071 and crypto/des/ede_cbcm_enc.c.
18075 * DES quad checksum was broken on big-endian architectures. Fixed.
18079 * Comment out two functions in bio.h that aren't implemented. Fix up the
18080 Win32 test batch file so it (might) work again. The Win32 test batch file
18081 is horrible: I feel ill....
18085 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18086 in e_os.h. Audit of header files to check ANSI and non ANSI
18087 sections: 10 functions were absent from non ANSI section and not exported
18088 from Windows DLLs. Fixed up libeay.num for new functions.
18092 * Make `openssl version` output lines consistent.
18094 *Ralf S. Engelschall*
18096 * Fix Win32 symbol export lists for BIO functions: Added
18097 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18098 to ms/libeay{16,32}.def.
18100 *Ralf S. Engelschall*
18102 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18103 fine under Unix and passes some trivial tests I've now added. But the
18104 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18105 added to make sure no one expects that this stuff really works in the
18106 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18107 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18110 *Ralf S. Engelschall*
18112 * Fix the generation of two part addresses in perl.
18114 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18116 * Add config entry for Linux on MIPS.
18118 *John Tobey <jtobey@channel1.com>*
18120 * Make links whenever Configure is run, unless we are on Windoze.
18124 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18125 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18130 * Add a useful kludge to allow package maintainers to specify compiler and
18131 other platforms details on the command line without having to patch the
18132 Configure script every time: One now can use
18133 `perl Configure <id>:<details>`,
18134 i.e. platform ids are allowed to have details appended
18135 to them (separated by colons). This is treated as there would be a static
18136 pre-configured entry in Configure's %table under key `<id>` with value
18137 `<details>` and `perl Configure <id>` is called. So, when you want to
18138 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18139 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18140 now, which overrides the FreeBSD-elf entry on-the-fly.
18142 *Ralf S. Engelschall*
18144 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18148 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18149 on the `perl Configure ...` command line. This way one can compile
18150 OpenSSL libraries with Position Independent Code (PIC) which is needed
18151 for linking it into DSOs.
18153 *Ralf S. Engelschall*
18155 * Remarkably, export ciphers were totally broken and no-one had noticed!
18160 * Cleaned up the LICENSE document: The official contact for any license
18161 questions now is the OpenSSL core team under openssl-core@openssl.org.
18162 And add a paragraph about the dual-license situation to make sure people
18163 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18164 to the OpenSSL toolkit.
18166 *Ralf S. Engelschall*
18168 * General source tree makefile cleanups: Made `making xxx in yyy...`
18169 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18170 Additionally cleaned up the `make links` target: Remove unnecessary
18171 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18172 to speed processing and no longer clutter the display with confusing
18173 stuff. Instead only the actually done links are displayed.
18175 *Ralf S. Engelschall*
18177 * Permit null encryption ciphersuites, used for authentication only. It used
18178 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18179 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18184 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18185 signed attributes when verifying signatures (this would break them),
18186 the detached data encoding was wrong and public keys obtained using
18187 X509_get_pubkey() weren't freed.
18191 * Add text documentation for the BUFFER functions. Also added a work around
18192 to a Win95 console bug. This was triggered by the password read stuff: the
18193 last character typed gets carried over to the next fread(). If you were
18194 generating a new cert request using 'req' for example then the last
18195 character of the passphrase would be CR which would then enter the first
18200 * Added the new 'Includes OpenSSL Cryptography Software' button as
18201 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18202 button and can be used by applications based on OpenSSL to show the
18203 relationship to the OpenSSL project.
18205 *Ralf S. Engelschall*
18207 * Remove confusing variables in function signatures in files
18208 ssl/ssl_lib.c and ssl/ssl.h.
18210 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18212 * Don't install bss_file.c under PREFIX/include/
18214 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18216 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18217 functions that return function pointers and has support for NT specific
18218 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18219 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18220 unsigned to signed types: this was killing the Win32 compile.
18224 * Add new certificate file to stack functions,
18225 SSL_add_dir_cert_subjects_to_stack() and
18226 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18227 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18228 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18229 This means that Apache-SSL and similar packages don't have to mess around
18230 to add as many CAs as they want to the preferred list.
18234 * Experiment with doxygen documentation. Currently only partially applied to
18236 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18237 openssl.doxy as the configuration file.
18241 * Get rid of remaining C++-style comments which strict C compilers hate.
18243 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18245 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18246 compiled in by default: it has problems with large keys.
18250 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18251 DH private keys and/or callback functions which directly correspond to
18252 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18253 is needed for applications which have to configure certificates on a
18254 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18256 For the RSA certificate situation is makes no difference, but
18257 for the DSA certificate situation this fixes the "no shared cipher"
18258 problem where the OpenSSL cipher selection procedure failed because the
18259 temporary keys were not overtaken from the context and the API provided
18260 no way to reconfigure them.
18261 The new functions now let applications reconfigure the stuff and they
18262 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18263 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18264 non-public-API function ssl_cert_instantiate() is used as a helper
18265 function and also to reduce code redundancy inside ssl_rsa.c.
18267 *Ralf S. Engelschall*
18269 * Move s_server -dcert and -dkey options out of the undocumented feature
18270 area because they are useful for the DSA situation and should be
18271 recognized by the users.
18273 *Ralf S. Engelschall*
18275 * Fix the cipher decision scheme for export ciphers: the export bits are
18276 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18277 SSL_EXP_MASK. So, the original variable has to be used instead of the
18278 already masked variable.
18280 *Richard Levitte <levitte@stacken.kth.se>*
18282 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18284 *Richard Levitte <levitte@stacken.kth.se>*
18286 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18287 from `int` to `unsigned int` because it is a length and initialized by
18288 EVP_DigestFinal() which expects an `unsigned int *`.
18290 *Richard Levitte <levitte@stacken.kth.se>*
18292 * Don't hard-code path to Perl interpreter on shebang line of Configure
18293 script. Instead use the usual Shell->Perl transition trick.
18295 *Ralf S. Engelschall*
18297 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18298 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18299 -noout -modulus` as it's already the case for `openssl rsa -noout
18300 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18301 currently the public key is printed (a decision which was already done by
18302 `openssl dsa -modulus` in the past) which serves a similar purpose.
18303 Additionally the NO_RSA no longer completely removes the whole -modulus
18304 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18307 *Ralf S. Engelschall*
18309 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18310 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18312 *Arne Ansper <arne@ats.cyber.ee>*
18314 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18315 to be added. Now both 'req' and 'ca' can use new objects defined in the
18320 * Add cool BIO that does syslog (or event log on NT).
18322 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18324 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18325 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18326 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18327 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18331 * Add preliminary config info for new extension code.
18335 * Make RSA_NO_PADDING really use no padding.
18337 *Ulf Moeller <ulf@fitug.de>*
18339 * Generate errors when private/public key check is done.
18343 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18344 for some CRL extensions and new objects added.
18348 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18349 key usage extension and fuller support for authority key id.
18353 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18354 padding method for RSA, which is recommended for new applications in PKCS
18355 #1 v2.0 (RFC 2437, October 1998).
18356 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18357 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18358 against Bleichbacher's attack on RSA.
18359 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18362 * Updates to the new SSL compression code
18364 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18366 * Fix so that the version number in the master secret, when passed
18367 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18368 (because the server will not accept higher), that the version number
18369 is 0x03,0x01, not 0x03,0x00
18371 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18373 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18374 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18375 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18379 * Support for RAW extensions where an arbitrary extension can be
18380 created by including its DER encoding. See `apps/openssl.cnf` for
18385 * Make sure latest Perl versions don't interpret some generated C array
18386 code as Perl array code in the crypto/err/err_genc.pl script.
18388 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18390 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18391 not many people have the assembler. Various Win32 compilation fixes and
18392 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18393 build instructions.
18397 * Modify configure script 'Configure' to automatically create crypto/date.h
18398 file under Win32 and also build pem.h from pem.org. New script
18399 util/mkfiles.pl to create the MINFO file on environments that can't do a
18400 'make files': perl util/mkfiles.pl >MINFO should work.
18404 * Major rework of DES function declarations, in the pursuit of correctness
18405 and purity. As a result, many evil casts evaporated, and some weirdness,
18406 too. You may find this causes warnings in your code. Zapping your evil
18407 casts will probably fix them. Mostly.
18411 * Fix for a typo in asn1.h. Bug fix to object creation script
18412 obj_dat.pl. It considered a zero in an object definition to mean
18413 "end of object": none of the objects in objects.h have any zeros
18414 so it wasn't spotted.
18416 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18418 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18419 Masking (CBCM). In the absence of test vectors, the best I have been able
18420 to do is check that the decrypt undoes the encrypt, so far. Send me test
18421 vectors if you have them.
18425 * Correct calculation of key length for export ciphers (too much space was
18426 allocated for null ciphers). This has not been tested!
18430 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18431 message is now correct (it understands "crypto" and "ssl" on its
18432 command line). There is also now an "update" option. This will update
18433 the util/ssleay.num and util/libeay.num files with any new functions.
18435 perl util/mkdef.pl crypto ssl update
18436 it will update them.
18440 * Overhauled the Perl interface:
18441 - ported BN stuff to OpenSSL's different BN library
18442 - made the perl/ source tree CVS-aware
18443 - renamed the package from SSLeay to OpenSSL (the files still contain
18444 their history because I've copied them in the repository)
18445 - removed obsolete files (the test scripts will be replaced
18446 by better Test::Harness variants in the future)
18448 *Ralf S. Engelschall*
18450 * First cut for a very conservative source tree cleanup:
18451 1. merge various obsolete readme texts into doc/ssleay.txt
18452 where we collect the old documents and readme texts.
18453 2. remove the first part of files where I'm already sure that we no
18454 longer need them because of three reasons: either they are just temporary
18455 files which were left by Eric or they are preserved original files where
18456 I've verified that the diff is also available in the CVS via "cvs diff
18457 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18458 the crypto/md/ stuff).
18460 *Ralf S. Engelschall*
18462 * More extension code. Incomplete support for subject and issuer alt
18463 name, issuer and authority key id. Change the i2v function parameters
18464 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18465 what that's for :-) Fix to ASN1 macro which messed up
18466 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18470 * Preliminary support for ENUMERATED type. This is largely copied from the
18475 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18477 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18479 * Make sure `make rehash` target really finds the `openssl` program.
18481 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18483 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18484 like to hear about it if this slows down other processors.
18488 * Add CygWin32 platform information to Configure script.
18490 *Alan Batie <batie@aahz.jf.intel.com>*
18492 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18494 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18496 * New program nseq to manipulate netscape certificate sequences
18500 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18505 * Fixes to BN code. Previously the default was to define BN_RECURSION
18506 but the BN code had some problems that would cause failures when
18507 doing certificate verification and some other functions.
18509 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18511 * Add ASN1 and PEM code to support netscape certificate sequences.
18515 * Add ASN1 and PEM code to support netscape certificate sequences.
18519 * Add several PKIX and private extended key usage OIDs.
18523 * Modify the 'ca' program to handle the new extension code. Modify
18524 openssl.cnf for new extension format, add comments.
18528 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18529 and add a sample to openssl.cnf so req -x509 now adds appropriate
18534 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18535 error code, add initial support to X509_print() and x509 application.
18539 * Takes a deep breath and start adding X509 V3 extension support code. Add
18540 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18541 stuff is currently isolated and isn't even compiled yet.
18545 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18546 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18547 Removed the versions check from X509 routines when loading extensions:
18548 this allows certain broken certificates that don't set the version
18549 properly to be processed.
18553 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18554 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18555 can still be regenerated with "make depend".
18559 * Spelling mistake in C version of CAST-128.
18561 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18563 * Changes to the error generation code. The perl script err-code.pl
18564 now reads in the old error codes and retains the old numbers, only
18565 adding new ones if necessary. It also only changes the .err files if new
18566 codes are added. The makefiles have been modified to only insert errors
18567 when needed (to avoid needlessly modifying header files). This is done
18568 by only inserting errors if the .err file is newer than the auto generated
18569 C file. To rebuild all the error codes from scratch (the old behaviour)
18570 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18571 or delete all the .err files.
18575 * CAST-128 was incorrectly implemented for short keys. The C version has
18576 been fixed, but is untested. The assembler versions are also fixed, but
18577 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18578 to regenerate it if needed.
18579 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18580 Hagino <itojun@kame.net>*
18582 * File was opened incorrectly in randfile.c.
18584 *Ulf Möller <ulf@fitug.de>*
18586 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18587 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18588 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18589 al: it's just almost always a UTCTime. Note this patch adds new error
18590 codes so do a "make errors" if there are problems.
18594 * Correct Linux 1 recognition in config.
18596 *Ulf Möller <ulf@fitug.de>*
18598 * Remove pointless MD5 hash when using DSA keys in ca.
18600 *Anonymous <nobody@replay.com>*
18602 * Generate an error if given an empty string as a cert directory. Also
18603 generate an error if handed NULL (previously returned 0 to indicate an
18604 error, but didn't set one).
18606 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18608 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18612 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18613 parameters. This was causing a warning which killed off the Win32 compile.
18617 * Remove C++ style comments from crypto/bn/bn_local.h.
18619 *Neil Costigan <neil.costigan@celocom.com>*
18621 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18622 based on a text string, looking up short and long names and finally
18623 "dot" format. The "dot" format stuff didn't work. Added new function
18624 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18625 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18626 OID is not part of the table.
18630 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18631 X509_LOOKUP_by_alias().
18635 * Sort openssl functions by name.
18639 * Get the `gendsa` command working and add it to the `list` command. Remove
18640 encryption from sample DSA keys (in case anyone is interested the password
18645 * Make *all* `*_free` functions accept a NULL pointer.
18647 *Frans Heymans <fheymans@isaserver.be>*
18649 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18652 *Anonymous <nobody@replay.com>*
18654 * s_server should send the CAfile as acceptable CAs, not its own cert.
18656 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18658 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18660 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18662 * Temp key "for export" tests were wrong in s3_srvr.c.
18664 *Anonymous <nobody@replay.com>*
18666 * Add prototype for temp key callback functions
18667 SSL_CTX_set_tmp_{rsa,dh}_callback().
18671 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18672 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18676 * X509_name_add_entry() freed the wrong thing after an error.
18678 *Arne Ansper <arne@ats.cyber.ee>*
18680 * rsa_eay.c would attempt to free a NULL context.
18682 *Arne Ansper <arne@ats.cyber.ee>*
18684 * BIO_s_socket() had a broken should_retry() on Windoze.
18686 *Arne Ansper <arne@ats.cyber.ee>*
18688 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18690 *Arne Ansper <arne@ats.cyber.ee>*
18692 * Make sure the already existing X509_STORE->depth variable is initialized
18693 in X509_STORE_new(), but document the fact that this variable is still
18694 unused in the certificate verification process.
18696 *Ralf S. Engelschall*
18698 * Fix the various library and `apps/` files to free up pkeys obtained from
18699 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18703 * Fix reference counting in X509_PUBKEY_get(). This makes
18704 demos/maurice/example2.c work, amongst others, probably.
18706 *Steve Henson and Ben Laurie*
18708 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18709 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18710 are no longer created. This way we have a single and consistent command
18711 line interface `openssl <command>`, similar to `cvs <command>`.
18713 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18715 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18716 BIT STRING wrapper always have zero unused bits.
18720 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18724 * Make the top-level INSTALL documentation easier to understand.
18728 * Makefiles updated to exit if an error occurs in a sub-directory
18729 make (including if user presses ^C) [Paul Sutton]
18731 * Make Montgomery context stuff explicit in RSA data structure.
18735 * Fix build order of pem and err to allow for generated pem.h.
18739 * Fix renumbering bug in X509_NAME_delete_entry().
18743 * Enhanced the err-ins.pl script so it makes the error library number
18744 global and can add a library name. This is needed for external ASN1 and
18745 other error libraries.
18749 * Fixed sk_insert which never worked properly.
18753 * Fix ASN1 macros so they can handle indefinite length constructed
18754 EXPLICIT tags. Some non standard certificates use these: they can now
18759 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18760 into a single doc/ssleay.txt bundle. This way the information is still
18761 preserved but no longer messes up this directory. Now it's new room for
18762 the new set of documentation files.
18764 *Ralf S. Engelschall*
18766 * SETs were incorrectly DER encoded. This was a major pain, because they
18767 shared code with SEQUENCEs, which aren't coded the same. This means that
18768 almost everything to do with SETs or SEQUENCEs has either changed name or
18769 number of arguments.
18771 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18773 * Fix test data to work with the above.
18777 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18778 was already fixed by Eric for 0.9.1 it seems.
18780 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18782 * Autodetect FreeBSD3.
18786 * Fix various bugs in Configure. This affects the following platforms:
18790 unixware-2.0-pentium
18795 * Eliminate generated files from CVS. Reorder tests to regenerate files
18796 before they are needed.
18800 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18804 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18806 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18807 changed SSLeay to OpenSSL in version strings.
18809 *Ralf S. Engelschall*
18811 * Some fixups to the top-level documents.
18815 * Fixed the nasty bug where rsaref.h was not found under compile-time
18816 because the symlink to include/ was missing.
18818 *Ralf S. Engelschall*
18820 * Incorporated the popular no-RSA/DSA-only patches
18821 which allow to compile a RSA-free SSLeay.
18823 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18825 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18826 when "ssleay" is still not found.
18828 *Ralf S. Engelschall*
18830 * Added more platforms to Configure: Cray T3E, HPUX 11,
18832 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18834 * Updated the README file.
18836 *Ralf S. Engelschall*
18838 * Added various .cvsignore files in the CVS repository subdirs
18839 to make a "cvs update" really silent.
18841 *Ralf S. Engelschall*
18843 * Recompiled the error-definition header files and added
18844 missing symbols to the Win32 linker tables.
18846 *Ralf S. Engelschall*
18848 * Cleaned up the top-level documents;
18849 o new files: CHANGES and LICENSE
18850 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18851 o merged COPYRIGHT into LICENSE
18852 o removed obsolete TODO file
18853 o renamed MICROSOFT to INSTALL.W32
18855 *Ralf S. Engelschall*
18857 * Removed dummy files from the 0.9.1b source tree:
18858 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18859 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18860 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18861 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18862 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18864 *Ralf S. Engelschall*
18866 * Added various platform portability fixes.
18870 * The Genesis of the OpenSSL rpject:
18871 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18872 Young and Tim J. Hudson created while they were working for C2Net until
18875 *The OpenSSL Project*
18877 ### Changes between 0.9.0b and 0.9.1b [not released]
18879 * Updated a few CA certificates under certs/
18883 * Changed some BIGNUM api stuff.
18887 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18888 DGUX x86, Linux Alpha, etc.
18892 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18893 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18898 * Add -strparse option to asn1pars program which parses nested
18901 *Dr Stephen Henson <shenson@bigfoot.com>*
18903 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18907 * DSA fix for "ca" program.
18911 * Added "-genkey" option to "dsaparam" program.
18915 * Added RIPE MD160 (rmd160) message digest.
18919 * Added -a (all) option to "ssleay version" command.
18923 * Added PLATFORM define which is the id given to Configure.
18927 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18931 * Extended the ASN.1 parser routines.
18935 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18939 * Added a BN_CTX to the BN library.
18943 * Fixed the weak key values in DES library
18947 * Changed API in EVP library for cipher aliases.
18951 * Added support for RC2/64bit cipher.
18955 * Converted the lhash library to the crypto/mem.c functions.
18959 * Added more recognized ASN.1 object ids.
18963 * Added more RSA padding checks for SSL/TLS.
18967 * Added BIO proxy/filter functionality.
18971 * Added extra_certs to SSL_CTX which can be used
18972 send extra CA certificates to the client in the CA cert chain sending
18973 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18977 * Now Fortezza is denied in the authentication phase because
18978 this is key exchange mechanism is not supported by SSLeay at all.
18982 * Additional PKCS1 checks.
18986 * Support the string "TLSv1" for all TLS v1 ciphers.
18990 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18991 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18995 * Fixed a few memory leaks.
18999 * Fixed various code and comment typos.
19003 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19004 bytes sent in the client random.
19006 *Edward Bishop <ebishop@spyglass.com>*
19010 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19011 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19012 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19013 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19014 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19015 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19016 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19017 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19018 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19019 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19020 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19021 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19022 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19023 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19024 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19025 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19026 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19027 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19028 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19029 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19030 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19031 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19032 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19033 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19034 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19035 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19036 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19037 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19038 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19039 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19040 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19041 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19042 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19043 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19044 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19045 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19046 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19047 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19048 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19049 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19050 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19051 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19052 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19053 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19054 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19055 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19056 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19057 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19058 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19059 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19060 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19061 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19062 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19063 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19064 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19065 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19066 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19067 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19068 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19069 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19070 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19071 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19072 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19073 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19074 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19075 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19076 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19077 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19078 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19079 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19080 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19081 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19082 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19083 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19084 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19085 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19086 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19087 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19088 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19089 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19090 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19091 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19092 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19093 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19094 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19095 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19096 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19097 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19098 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19099 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19100 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19101 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19102 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19103 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19104 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19105 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19106 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19107 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19108 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19109 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19110 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19111 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19112 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19113 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19114 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19115 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19116 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19117 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19118 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19119 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19120 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19121 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19122 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19123 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19124 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19125 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19126 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19127 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19128 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19129 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19130 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19131 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19132 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19133 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19134 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19135 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19136 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19137 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19138 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19139 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19140 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19141 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19142 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19143 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19144 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19145 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19146 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19147 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19148 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19149 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19150 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19151 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19152 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19153 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19154 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19155 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19156 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19157 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19158 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19159 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19160 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19161 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19162 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19163 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19164 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19165 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19166 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19167 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19168 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19169 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19170 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19171 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655