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 3.1](#openssl-31)
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 3.0 and 3.1 [xx XXX xxxx]
27 * s_client and s_server apps now explicitly say when the TLS version
28 does not include the renegotiation mechanism. This avoids confusion
29 between that scenario versus when the TLS version includes secure
30 renegotiation but the peer lacks support for it.
34 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
35 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
36 of 160 bits and above and less than 224 bits were previously accepted by
37 default but are now no longer allowed. By default TLS compression was
38 already disabled in previous OpenSSL versions. At security level 2 it cannot
43 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
48 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
49 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
50 will need to load the legacy crypto provider.
54 * The various OBJ_* functions have been made thread safe.
58 * CCM8 cipher suites in TLS have been downgraded to security level zero
59 because they use a short authentication tag which lowers their strength.
63 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
68 * Parallel dual-prime 1536/2048-bit modular exponentiation for
69 AVX512_IFMA capable processors.
71 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
76 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
77 listed here are only a brief description.
78 The migration guide contains more detailed information related to new features,
79 breaking changes, and mappings for the large list of deprecated functions.
81 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
83 ### Changes between 3.0.0 and 3.0.1 [14 dec 2021]
85 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
86 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
87 verify a certificate supplied by a server. That function may return a
88 negative return value to indicate an internal error (for example out of
89 memory). Such a negative return value is mishandled by OpenSSL and will cause
90 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
91 success and a subsequent call to SSL_get_error() to return the value
92 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
93 returned by OpenSSL if the application has previously called
94 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
95 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
96 totally unexpected and applications may not behave correctly as a result. The
97 exact behaviour will depend on the application but it could result in
98 crashes, infinite loops or other similar incorrect responses.
100 This issue is made more serious in combination with a separate bug in OpenSSL
101 3.0 that will cause X509_verify_cert() to indicate an internal error when
102 processing a certificate chain. This will occur where a certificate does not
103 include the Subject Alternative Name extension but where a Certificate
104 Authority has enforced name constraints. This issue can occur even with valid
110 * Corrected a few file name and file reference bugs in the build,
111 installation and setup scripts, which lead to installation verification
112 failures. Slightly enhanced the installation verification script.
116 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
121 * Fixed PVK encoder to properly query for the passphrase.
125 * Multiple fixes in the OSSL_HTTP API functions.
129 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
130 OSSL_PARAM_INTEGER data type and return error on negative numbers
131 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
132 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
136 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
140 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
144 * Multiple threading fixes.
148 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
152 * Allow fetching an operation from the provider that owns an unexportable key
153 as a fallback if that is still allowed by the property query.
157 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
159 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
164 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
165 S390X capability vector to zero. This simplifies testing of different code
166 paths on S390X architecture.
170 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
171 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
172 SP 800-38D". The communication will fail at this point.
176 * The EC_GROUP_clear_free() function is deprecated as there is nothing
177 confidential in EC_GROUP data.
181 * The byte order mark (BOM) character is ignored if encountered at the
182 beginning of a PEM-formatted file.
186 * Added CMS support for the Russian GOST algorithms.
190 * Due to move of the implementation of cryptographic operations
191 to the providers, validation of various operation parameters can
192 be postponed until the actual operation is executed where previously
193 it happened immediately when an operation parameter was set.
195 For example when setting an unsupported curve with
196 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
197 fail but later keygen operations with the EVP_PKEY_CTX will fail.
199 *OpenSSL team members and many third party contributors*
201 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
202 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
203 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
204 instead to retrieve these algorithms from a provider.
208 * On build targets where the multilib postfix is set in the build
209 configuration the libdir directory was changing based on whether
210 the lib directory with the multilib postfix exists on the system
211 or not. This unpredictable behavior was removed and eventual
212 multilib postfix is now always added to the default libdir. Use
213 `--libdir=lib` to override the libdir if adding the postfix is
218 * The triple DES key wrap functionality now conforms to RFC 3217 but is
219 no longer interoperable with OpenSSL 1.1.1.
223 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
224 function codes, this function can only cause problems for calling
229 * Add a configurable flag to output date formats as ISO 8601. Does not
230 change the default date format.
234 * Version of MSVC earlier than 1300 could get link warnings, which could
235 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
236 Support for this flag has been removed.
240 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
241 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
242 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
243 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
244 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
248 * The signatures of the functions to get and set options on SSL and
249 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
250 Some source code changes may be required.
254 * The public definitions of conf_method_st and conf_st have been
255 deprecated. They will be made opaque in a future release.
257 *Rich Salz and Tomáš Mráz*
259 * Client-initiated renegotiation is disabled by default. To allow it, use
260 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
261 flag, or the "ClientRenegotiation" config parameter as appropriate.
265 * Add "abspath" and "includedir" pragma's to config files, to prevent,
266 or modify relative pathname inclusion.
270 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
271 validated. Please consult the README-FIPS and
272 README-PROVIDERS files, as well as the migration guide.
274 *OpenSSL team members and many third party contributors*
276 * For the key types DH and DHX the allowed settable parameters are now different.
280 * The openssl commands that read keys, certificates, and CRLs now
281 automatically detect the PEM or DER format of the input files.
283 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
285 * Added enhanced PKCS#12 APIs which accept a library context.
289 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
293 * Added support for Kernel TLS (KTLS).
295 *Boris Pismenny, John Baldwin and Andrew Gallatin*
297 * Support for RFC 5746 secure renegotiation is now required by default for
298 SSL or TLS connections to succeed.
302 * The signature of the `copy` functional parameter of the
303 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
304 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
305 the signature of the `pub_decode` functional parameter of the
306 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
307 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
311 * The error return values from some control calls (ctrl) have changed.
315 * A public key check is now performed during EVP_PKEY_derive_set_peer().
319 * Many functions in the EVP_ namespace that are getters of values from
320 implementations or contexts were renamed to include get or get0 in their
321 names. Old names are provided as macro aliases for compatibility and
326 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
327 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
328 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
333 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
336 * The output from the command line applications may have minor
341 * The output from numerous "printing" may have minor changes.
345 * Windows thread synchronization uses read/write primitives (SRWLock) when
346 supported by the OS, otherwise CriticalSection continues to be used.
350 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
351 work on read only BIO source/sinks that do not support these functions.
352 This allows piping or redirection of a file BIO using stdin to be buffered
353 into memory. This is used internally in OSSL_DECODER_from_bio().
357 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
358 this function would return one of the values OSSL_STORE_INFO_NAME,
359 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
360 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
361 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
362 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
363 using this function should be amended to handle the changed return value.
367 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
368 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
369 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
370 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
371 contains more than one certificate identifier: This means that all
372 certificates referenced there MUST be part of the validation chain.
376 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
377 RC5, DESX and DES have been moved to the legacy provider.
381 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
382 RIPEMD-160 have been moved to the legacy provider.
386 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
391 * The deprecated functions EVP_PKEY_get0_RSA(),
392 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
393 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
394 well as the similarly named "get1" functions behave differently in
399 * A number of functions handling low-level keys or engines were deprecated
400 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
401 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
402 EVP_PKEY_get0_siphash().
406 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
407 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
408 will need to load the legacy crypto provider. This includes these PBE
409 algorithms which use this KDF:
410 - NID_pbeWithMD2AndDES_CBC
411 - NID_pbeWithMD5AndDES_CBC
412 - NID_pbeWithSHA1AndRC2_CBC
413 - NID_pbeWithMD2AndRC2_CBC
414 - NID_pbeWithMD5AndRC2_CBC
415 - NID_pbeWithSHA1AndDES_CBC
419 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
420 BIO_debug_callback() functions.
424 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
425 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
429 * The RAND_METHOD APIs have been deprecated.
433 * The SRP APIs have been deprecated.
437 * Add a compile time option to prevent the caching of provider fetched
438 algorithms. This is enabled by including the no-cached-fetch option
439 at configuration time.
443 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
444 count of PKCS12_DEFAULT_ITER.
446 *Tomáš Mráz and Sahana Prasad*
448 * The openssl speed command does not use low-level API calls anymore.
452 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
455 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
457 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
461 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
462 providers may supply their own group implementations (using either the "key
463 exchange" or the "key encapsulation" methods) which will automatically be
464 detected and used by libssl.
466 *Matt Caswell, Nicola Tuveri*
468 * The undocumented function X509_certificate_type() has been deprecated;
472 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
476 * Removed RSA padding mode for SSLv23 (which was only used for
477 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
478 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
483 * Deprecated the obsolete X9.31 RSA key generation related functions.
485 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
486 is not allowed to return a value > 1, this is no more taken as failure.
488 *Viktor Dukhovni and David von Oheimb*
490 * Deprecated the obsolete X9.31 RSA key generation related functions
491 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
492 BN_X931_generate_prime_ex().
496 * The default key generation method for the regular 2-prime RSA keys was
497 changed to the FIPS 186-4 B.3.6 method.
501 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
505 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
509 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
510 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
512 *Rich Salz, Richard Levitte, and David von Oheimb*
514 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
518 * Deprecated `OCSP_parse_url()`.
522 * Validation of SM2 keys has been separated from the validation of regular EC
527 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
528 switches: a validation failure triggers an early exit, returning a failure
529 exit status to the parent process.
533 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
534 to ignore unknown ciphers.
538 * The `-cipher-commands` and `-digest-commands` options
539 of the command line utility `list` have been deprecated.
540 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
544 * Added convenience functions for generating asymmetric key pairs:
545 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
546 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
550 * All of the low level EC_KEY functions have been deprecated.
552 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
554 * Deprecated all the libcrypto and libssl error string loading
559 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
560 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
565 * The `-crypt` option to the `passwd` command line tool has been removed.
569 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
574 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
578 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
579 EVP_PKEY_get1_tls_encodedpoint().
583 * The security callback, which can be customised by application code, supports
584 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
585 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
589 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
590 interface. Their functionality remains unchanged.
594 * Added new option for 'openssl list', '-providers', which will display the
595 list of loaded providers, their names, version and status. It optionally
596 displays their gettable parameters.
600 * Removed EVP_PKEY_set_alias_type().
604 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
605 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
609 * Changed all "STACK" functions to be macros instead of inline functions. Macro
610 parameters are still checked for type safety at compile time via helper
615 * Remove the RAND_DRBG API
617 *Paul Dale and Matthias St. Pierre*
619 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
620 as well as actual hostnames.
624 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
625 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
626 conversely, silently ignore DTLS protocol version bounds when configuring
627 TLS-based contexts. The commands can be repeated to set bounds of both
628 types. The same applies with the corresponding "min_protocol" and
629 "max_protocol" command-line switches, in case some application uses both TLS
632 SSL_CTX instances that are created for a fixed protocol version (e.g.
633 `TLSv1_server_method()`) also silently ignore version bounds. Previously
634 attempts to apply bounds to these protocol versions would result in an
635 error. Now only the "version-flexible" SSL_CTX instances are subject to
636 limits in configuration files in command-line options.
640 * Deprecated the `ENGINE` API. Engines should be replaced with providers
645 * Reworked the recorded ERR codes to make better space for system errors.
646 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
647 given code is a system error (true) or an OpenSSL error (false).
651 * Reworked the test perl framework to better allow parallel testing.
653 *Nicola Tuveri and David von Oheimb*
655 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
656 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
660 * 'Configure' has been changed to figure out the configuration target if
661 none is given on the command line. Consequently, the 'config' script is
662 now only a mere wrapper. All documentation is changed to only mention
665 *Rich Salz and Richard Levitte*
667 * Added a library context `OSSL_LIB_CTX` that applications as well as
668 other libraries can use to form a separate context within which
669 libcrypto operations are performed.
673 * Added various `_ex` functions to the OpenSSL API that support using
674 a non-default `OSSL_LIB_CTX`.
678 * Handshake now fails if Extended Master Secret extension is dropped
683 * Dropped interactive mode from the `openssl` program.
687 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
689 *David von Oheimb and Shane Lontis*
691 * Deprecated `EC_METHOD_get_field_type()`.
695 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
696 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
697 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
701 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
705 * Add CAdES-BES signature verification support, mostly derived
706 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
708 *Filipe Raimundo da Silva*
710 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
714 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
715 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
719 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
723 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
724 EC_KEY_precompute_mult().
728 * Deprecated EC_POINTs_mul().
732 * Removed FIPS_mode() and FIPS_mode_set().
736 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
740 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
741 EC_POINT_get_Jprojective_coordinates_GFp().
745 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
746 arrays to be more easily constructed via a series of utility functions.
747 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
748 the various push functions and finally convert to a passable OSSL_PARAM
749 array using OSSL_PARAM_BLD_to_param().
753 * The security strength of SHA1 and MD5 based signatures in TLS has been
758 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
759 contain a provider side internal key.
763 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
767 * Project text documents not yet having a proper file name extension
768 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
769 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
773 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
774 have been converted to Markdown with the goal to produce documents
775 which not only look pretty when viewed online in the browser, but
776 remain well readable inside a plain text editor.
778 To achieve this goal, a 'minimalistic' Markdown style has been applied
779 which avoids formatting elements that interfere too much with the
780 reading flow in the text file. For example, it
782 * avoids [ATX headings][] and uses [setext headings][] instead
783 (which works for `<h1>` and `<h2>` headings only).
784 * avoids [inline links][] and uses [reference links][] instead.
785 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
787 [ATX headings]: https://github.github.com/gfm/#atx-headings
788 [setext headings]: https://github.github.com/gfm/#setext-headings
789 [inline links]: https://github.github.com/gfm/#inline-link
790 [reference links]: https://github.github.com/gfm/#reference-link
791 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
792 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
794 *Matthias St. Pierre*
796 * The test suite is changed to preserve results of each test recipe.
797 A new directory test-runs/ with subdirectories named like the
798 test recipes are created in the build tree for this purpose.
802 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
803 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
804 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
806 *David von Oheimb, Martin Peylo*
808 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
809 It supports arbitrary request and response content types, GET redirection,
810 TLS, connections via HTTP(S) proxies, connections and exchange via
811 user-defined BIOs (allowing implicit connections), persistent connections,
812 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
813 The legacy OCSP-focused (and only partly documented) API
814 is retained for backward compatibility, while most of it is deprecated.
818 * Added `util/check-format.pl`, a tool for checking adherence to the
819 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
820 The checks performed are incomplete and yield some false positives.
821 Still the tool should be useful for detecting most typical glitches.
825 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
826 If domain name resolution yields multiple IP addresses all of them are tried
827 after `connect()` failures.
831 * All of the low level RSA functions have been deprecated.
835 * X509 certificates signed using SHA1 are no longer allowed at security
840 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
841 modified to use PKEY APIs. These commands are now in maintenance mode
842 and no new features will be added to them.
846 * The command line utility rsautl has been deprecated.
850 * The command line utilities genrsa and rsa have been modified to use PKEY
851 APIs. They now write PKCS#8 keys by default. These commands are now in
852 maintenance mode and no new features will be added to them.
856 * All of the low level DH functions have been deprecated.
858 *Paul Dale and Matt Caswell*
860 * All of the low level DSA functions have been deprecated.
864 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
865 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
869 * Deprecated low level ECDH and ECDSA functions.
873 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
877 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
878 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
879 a new formulation to include all the things it can be used for,
880 as well as words of caution.
884 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
888 * All of the low level HMAC functions have been deprecated.
890 *Paul Dale and David von Oheimb*
892 * Over two thousand fixes were made to the documentation, including:
893 - Common options (such as -rand/-writerand, TLS version control, etc)
894 were refactored and point to newly-enhanced descriptions in openssl.pod.
895 - Added style conformance for all options (with help from Richard Levitte),
896 documented all reported missing options, added a CI build to check
897 that all options are documented and that no unimplemented options
899 - Documented some internals, such as all use of environment variables.
900 - Addressed all internal broken L<> references.
904 * All of the low level CMAC functions have been deprecated.
908 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
909 functions have been deprecated.
911 *Paul Dale and David von Oheimb*
913 * Corrected the documentation of the return values from the `EVP_DigestSign*`
914 set of functions. The documentation mentioned negative values for some
915 errors, but this was never the case, so the mention of negative values
918 Code that followed the documentation and thereby check with something
919 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
923 * All of the low level cipher functions have been deprecated.
925 *Matt Caswell and Paul Dale*
927 * Removed include/openssl/opensslconf.h.in and replaced it with
928 include/openssl/configuration.h.in, which differs in not including
929 <openssl/macros.h>. A short header include/openssl/opensslconf.h
930 was added to include both.
932 This allows internal hacks where one might need to modify the set
933 of configured macros, for example this if deprecated symbols are
934 still supposed to be available internally:
936 #include <openssl/configuration.h>
938 #undef OPENSSL_NO_DEPRECATED
939 #define OPENSSL_SUPPRESS_DEPRECATED
941 #include <openssl/macros.h>
943 This should not be used by applications that use the exported
944 symbols, as that will lead to linking errors.
948 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
949 used in exponentiation with 512-bit moduli. No EC algorithms are
950 affected. Analysis suggests that attacks against 2-prime RSA1024,
951 3-prime RSA1536, and DSA1024 as a result of this defect would be very
952 difficult to perform and are not believed likely. Attacks against DH512
953 are considered just feasible. However, for an attack the target would
954 have to re-use the DH512 private key, which is not recommended anyway.
955 Also applications directly using the low-level API BN_mod_exp may be
956 affected if they use BN_FLG_CONSTTIME.
961 * Most memory-debug features have been deprecated, and the functionality
962 replaced with no-ops.
966 * Added documentation for the STACK API.
970 * Introduced a new method type and API, OSSL_ENCODER, to represent
971 generic encoders. These do the same sort of job that PEM writers
972 and d2i functions do, but with support for methods supplied by
973 providers, and the possibility for providers to support other
978 * Introduced a new method type and API, OSSL_DECODER, to represent
979 generic decoders. These do the same sort of job that PEM readers
980 and i2d functions do, but with support for methods supplied by
981 providers, and the possibility for providers to support other
986 * Added a .pragma directive to the syntax of configuration files, to
987 allow varying behavior in a supported and predictable manner.
988 Currently added pragma:
992 This allows dollar signs to be a keyword character unless it's
993 followed by a opening brace or parenthesis. This is useful for
994 platforms where dollar signs are commonly used in names, such as
995 volume names and system directory names on VMS.
999 * Added functionality to create an EVP_PKEY from user data.
1003 * Change the interpretation of the '--api' configuration option to
1004 mean that this is a desired API compatibility level with no
1005 further meaning. The previous interpretation, that this would
1006 also mean to remove all deprecated symbols up to and including
1007 the given version, no requires that 'no-deprecated' is also used
1008 in the configuration.
1010 When building applications, the desired API compatibility level
1011 can be set with the OPENSSL_API_COMPAT macro like before. For
1012 API compatibility version below 3.0, the old style numerical
1013 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1014 For version 3.0 and on, the value is expected to be the decimal
1015 value calculated from the major and minor version like this:
1017 MAJOR * 10000 + MINOR * 100
1021 -DOPENSSL_API_COMPAT=30000 For 3.0
1022 -DOPENSSL_API_COMPAT=30200 For 3.2
1024 To hide declarations that are deprecated up to and including the
1025 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1026 given when building the application as well.
1030 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1031 access to certificate and CRL stores via URIs and OSSL_STORE
1034 This adds the following functions:
1036 - X509_LOOKUP_store()
1037 - X509_STORE_load_file()
1038 - X509_STORE_load_path()
1039 - X509_STORE_load_store()
1040 - SSL_add_store_cert_subjects_to_stack()
1041 - SSL_CTX_set_default_verify_store()
1042 - SSL_CTX_load_verify_file()
1043 - SSL_CTX_load_verify_dir()
1044 - SSL_CTX_load_verify_store()
1048 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1049 The presence of this system service is determined at run-time.
1053 * Added functionality to create an EVP_PKEY context based on data
1054 for methods from providers. This takes an algorithm name and a
1055 property query string and simply stores them, with the intent
1056 that any operation that uses this context will use those strings
1057 to fetch the needed methods implicitly, thereby making the port
1058 of application written for pre-3.0 OpenSSL easier.
1062 * The undocumented function NCONF_WIN32() has been deprecated; for
1063 conversion details see the HISTORY section of doc/man5/config.pod
1067 * Introduced the new functions EVP_DigestSignInit_ex() and
1068 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1069 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1070 pages for further details.
1074 * Over two thousand fixes were made to the documentation, including:
1075 adding missing command flags, better style conformance, documentation
1078 *Rich Salz, Richard Levitte*
1080 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1081 X25519, X448, Ed25519 and Ed448.
1085 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1090 * Deprecated the public definition of `ERR_STATE` as well as the function
1091 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1096 * Added ERR functionality to give callers access to the stored function
1097 names that have replaced the older function code based functions.
1099 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1100 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1101 ERR_peek_error_all() and ERR_peek_last_error_all().
1103 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1104 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1105 ERR_func_error_string().
1109 * Extended testing to be verbose for failing tests only. The make variables
1110 VERBOSE_FAILURE or VF can be used to enable this:
1112 $ make VF=1 test # Unix
1113 $ mms /macro=(VF=1) test ! OpenVMS
1114 $ nmake VF=1 test # Windows
1118 * Added the `-copy_extensions` option to the `x509` command for use with
1119 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1120 all extensions in the request are copied to the certificate or vice versa.
1122 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1124 * Added the `-copy_extensions` option to the `req` command for use with
1125 `-x509`. When given with the `copy` or `copyall` argument,
1126 all extensions in the certification request are copied to the certificate.
1130 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1131 they generate are by default RFC 5280 compliant in the following sense:
1132 There is a subjectKeyIdentifier extension with a hash value of the public key
1133 and for not self-signed certs there is an authorityKeyIdentifier extension
1134 with a keyIdentifier field or issuer information identifying the signing key.
1135 This is done unless some configuration overrides the new default behavior,
1136 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1140 * Added several checks to `X509_verify_cert()` according to requirements in
1141 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1142 (which may be done by using the CLI option `-x509_strict`):
1143 * The basicConstraints of CA certificates must be marked critical.
1144 * CA certificates must explicitly include the keyUsage extension.
1145 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1146 * The issuer name of any certificate must not be empty.
1147 * The subject name of CA certs, certs with keyUsage crlSign,
1148 and certs without subjectAlternativeName must not be empty.
1149 * If a subjectAlternativeName extension is given it must not be empty.
1150 * The signatureAlgorithm field and the cert signature must be consistent.
1151 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1152 must not be marked critical.
1153 * The authorityKeyIdentifier must be given for X.509v3 certs
1154 unless they are self-signed.
1155 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1159 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1160 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1164 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1165 used even when parsing explicit parameters, when loading a encoded key
1166 or calling `EC_GROUP_new_from_ecpkparameters()`/
1167 `EC_GROUP_new_from_ecparameters()`.
1168 This prevents bypass of security hardening and performance gains,
1169 especially for curves with specialized EC_METHODs.
1170 By default, if a key encoded with explicit parameters is loaded and later
1171 encoded, the output is still encoded with explicit parameters, even if
1172 internally a "named" EC_GROUP is used for computation.
1176 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1177 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1178 NULL. After this change, only the cofactor parameter can be NULL. It also
1179 does some minimal sanity checks on the passed order.
1184 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1185 An attack is simple, if the first CMS_recipientInfo is valid but the
1186 second CMS_recipientInfo is chosen ciphertext. If the second
1187 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1188 encryption key will be replaced by garbage, and the message cannot be
1189 decoded, but if the RSA decryption fails, the correct encryption key is
1190 used and the recipient will not notice the attack.
1191 As a work around for this potential attack the length of the decrypted
1192 key must be equal to the cipher default key length, in case the
1193 certifiate is not given and all recipientInfo are tried out.
1194 The old behaviour can be re-enabled in the CMS code by setting the
1195 CMS_DEBUG_DECRYPT flag.
1199 * Early start up entropy quality from the DEVRANDOM seed source has been
1200 improved for older Linux systems. The RAND subsystem will wait for
1201 /dev/random to be producing output before seeding from /dev/urandom.
1202 The seeded state is stored for future library initialisations using
1203 a system global shared memory segment. The shared memory identifier
1204 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1205 the desired value. The default identifier is 114.
1209 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1210 when primes for RSA keys are computed.
1211 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1212 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1213 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1214 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1215 This avoids possible fingerprinting of newly generated RSA modules.
1219 * Correct the extended master secret constant on EBCDIC systems. Without this
1220 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1221 negotiate EMS will fail. Unfortunately this also means that TLS connections
1222 between EBCDIC systems with this fix, and EBCDIC systems without this
1223 fix will fail if they negotiate EMS.
1227 * Changed the library initialisation so that the config file is now loaded
1228 by default. This was already the case for libssl. It now occurs for both
1229 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1230 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1234 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1235 where the former acts as a replacement for `ERR_put_error()`, and the
1236 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1237 `ERR_raise_data()` adds more flexibility by taking a format string and
1238 an arbitrary number of arguments following it, to be processed with
1243 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1244 to check if a named provider is loaded and available. When called, it
1245 will also activate all fallback providers if such are still present.
1249 * Enforce a minimum DH modulus size of 512 bits.
1253 * Changed DH parameters to generate the order q subgroup instead of 2q.
1254 Previously generated DH parameters are still accepted by DH_check
1255 but DH_generate_key works around that by clearing bit 0 of the
1256 private key for those. This avoids leaking bit 0 of the private key.
1260 * Significantly reduce secure memory usage by the randomness pools.
1264 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1269 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1270 algorithms. An implementation of a key exchange algorithm can be obtained
1271 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1272 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1273 the older EVP_PKEY_derive_init() function. See the man pages for the new
1274 functions for further details.
1278 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1282 * Removed the function names from error messages and deprecated the
1287 * Removed NextStep support and the macro OPENSSL_UNISTD
1291 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1292 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1293 Also removed "export var as function" capability; we do not export
1294 variables, only functions.
1298 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1299 an error and 1 indicating success. In previous versions of OpenSSL this
1300 was a void type. If a key was set longer than the maximum possible this
1305 * Support SM2 signing and verification schemes with X509 certificate.
1309 * Use SHA256 as the default digest for TS query in the `ts` app.
1313 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1317 * Default cipher lists/suites are now available via a function, the
1318 #defines are deprecated.
1322 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1323 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1324 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1328 * Join the directories crypto/x509 and crypto/x509v3
1332 * Added command 'openssl kdf' that uses the EVP_KDF API.
1336 * Added command 'openssl mac' that uses the EVP_MAC API.
1340 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1341 as default directories. Also added the command 'openssl info'
1342 for scripting purposes.
1346 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1351 * Add prediction resistance to the DRBG reseeding process.
1355 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1356 mandated by IEEE Std 1619-2018.
1360 * Added newline escaping functionality to a filename when using openssl dgst.
1361 This output format is to replicate the output format found in the `*sum`
1362 checksum programs. This aims to preserve backward compatibility.
1364 *Matt Eaton, Richard Levitte, and Paul Dale*
1366 * Removed the heartbeat message in DTLS feature, as it has very
1367 little usage and doesn't seem to fulfill a valuable purpose.
1368 The configuration option is now deprecated.
1372 * Changed the output of 'openssl {digestname} < file' to display the
1373 digest name in its output.
1377 * Added a new generic trace API which provides support for enabling
1378 instrumentation through trace output.
1380 *Richard Levitte & Matthias St. Pierre*
1382 * Added build tests for C++. These are generated files that only do one
1383 thing, to include one public OpenSSL head file each. This tests that
1384 the public header files can be usefully included in a C++ application.
1386 This test isn't enabled by default. It can be enabled with the option
1387 'enable-buildtest-c++'.
1391 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1395 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1399 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1403 * Added KMAC to EVP_MAC.
1407 * Added property based algorithm implementation selection framework to
1412 * Added SCA hardening for modular field inversion in EC_GROUP through
1413 a new dedicated field_inv() pointer in EC_METHOD.
1414 This also addresses a leakage affecting conversions from projective
1415 to affine coordinates.
1417 *Billy Bob Brumley, Nicola Tuveri*
1419 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1420 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1421 those algorithms that were already supported through the EVP_PKEY API
1422 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1423 and scrypt are now wrappers that call EVP_KDF.
1427 * Build devcrypto engine as a dynamic engine.
1429 *Eneas U de Queiroz*
1431 * Add keyed BLAKE2 to EVP_MAC.
1435 * Fix a bug in the computation of the endpoint-pair shared secret used
1436 by DTLS over SCTP. This breaks interoperability with older versions
1437 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1438 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1439 interoperability with such broken implementations. However, enabling
1440 this switch breaks interoperability with correct implementations.
1442 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1443 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1447 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1451 * Changed the license to the Apache License v2.0.
1455 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1457 - Major releases (indicated by incrementing the MAJOR release number)
1458 may introduce incompatible API/ABI changes.
1459 - Minor releases (indicated by incrementing the MINOR release number)
1460 may introduce new features but retain API/ABI compatibility.
1461 - Patch releases (indicated by incrementing the PATCH number)
1462 are intended for bug fixes and other improvements of existing
1463 features only (like improving performance or adding documentation)
1464 and retain API/ABI compatibility.
1468 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1472 * Remove the 'dist' target and add a tarball building script. The
1473 'dist' target has fallen out of use, and it shouldn't be
1474 necessary to configure just to create a source distribution.
1478 * Recreate the OS390-Unix config target. It no longer relies on a
1479 special script like it did for OpenSSL pre-1.1.0.
1483 * Instead of having the source directories listed in Configure, add
1484 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1489 * Add GMAC to EVP_MAC.
1493 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1497 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1498 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1499 to facilitate the continued use of MACs through raw private keys in
1500 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1504 * Deprecate ECDH_KDF_X9_62().
1508 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1509 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1510 are retained for backwards compatibility.
1514 * AES-XTS mode now enforces that its two keys are different to mitigate
1515 the attacked described in "Efficient Instantiations of Tweakable
1516 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1517 Details of this attack can be obtained from:
1518 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1522 * Rename the object files, i.e. give them other names than in previous
1523 versions. Their names now include the name of the final product, as
1524 well as its type mnemonic (bin, lib, shlib).
1528 * Added new option for 'openssl list', '-objects', which will display the
1529 list of built in objects, i.e. OIDs with names.
1533 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1534 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1539 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1540 improves application performance by removing data copies and providing
1541 applications with zero-copy system calls such as sendfile and splice.
1545 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1549 * `PKCS12_parse` now maintains the order of the parsed certificates
1550 when outputting them via `*ca` (rather than reversing it).
1554 * Deprecated pthread fork support methods.
1558 * Added support for FFDHE key exchange in TLS 1.3.
1562 * Added a new concept for OpenSSL plugability: providers. This
1563 functionality is designed to replace the ENGINE API and ENGINE
1564 implementations, and to be much more dynamic, allowing provider
1565 authors to introduce new algorithms among other things, as long as
1566 there's an API that supports the algorithm type.
1568 With this concept comes a new core API for interaction between
1569 libcrypto and provider implementations. Public libcrypto functions
1570 that want to use providers do so through this core API.
1572 The main documentation for this core API is found in
1573 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1574 refer to other manuals describing the API specific for supported
1575 algorithm types (also called operations).
1582 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
1584 * Avoid loading of a dynamic engine twice.
1588 * Prioritise DANE TLSA issuer certs over peer certs
1592 * Fixed random API for MacOS prior to 10.12
1594 These MacOS versions don't support the CommonCrypto APIs
1598 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1600 * Fixed an SM2 Decryption Buffer Overflow.
1602 In order to decrypt SM2 encrypted data an application is expected to
1603 call the API function EVP_PKEY_decrypt(). Typically an application will
1604 call this function twice. The first time, on entry, the "out" parameter
1605 can be NULL and, on exit, the "outlen" parameter is populated with the
1606 buffer size required to hold the decrypted plaintext. The application
1607 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1608 again, but this time passing a non-NULL value for the "out" parameter.
1610 A bug in the implementation of the SM2 decryption code means that the
1611 calculation of the buffer size required to hold the plaintext returned
1612 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1613 size required by the second call. This can lead to a buffer overflow
1614 when EVP_PKEY_decrypt() is called by the application a second time with
1615 a buffer that is too small.
1617 A malicious attacker who is able present SM2 content for decryption to
1618 an application could cause attacker chosen data to overflow the buffer
1619 by up to a maximum of 62 bytes altering the contents of other data held
1620 after the buffer, possibly changing application behaviour or causing
1621 the application to crash. The location of the buffer is application
1622 dependent but is typically heap allocated.
1627 * Fixed various read buffer overruns processing ASN.1 strings
1629 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1630 structure which contains a buffer holding the string data and a field
1631 holding the buffer length. This contrasts with normal C strings which
1632 are repesented as a buffer for the string data which is terminated
1633 with a NUL (0) byte.
1635 Although not a strict requirement, ASN.1 strings that are parsed using
1636 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1637 well as any string whose value has been set with the ASN1_STRING_set()
1638 function will additionally NUL terminate the byte array in the
1639 ASN1_STRING structure.
1641 However, it is possible for applications to directly construct valid
1642 ASN1_STRING structures which do not NUL terminate the byte array by
1643 directly setting the "data" and "length" fields in the ASN1_STRING
1644 array. This can also happen by using the ASN1_STRING_set0() function.
1646 Numerous OpenSSL functions that print ASN.1 data have been found to
1647 assume that the ASN1_STRING byte array will be NUL terminated, even
1648 though this is not guaranteed for strings that have been directly
1649 constructed. Where an application requests an ASN.1 structure to be
1650 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1651 been directly constructed by the application without NUL terminating
1652 the "data" field, then a read buffer overrun can occur.
1654 The same thing can also occur during name constraints processing
1655 of certificates (for example if a certificate has been directly
1656 constructed by the application instead of loading it via the OpenSSL
1657 parsing functions, and the certificate contains non NUL terminated
1658 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1659 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1661 If a malicious actor can cause an application to directly construct an
1662 ASN1_STRING and then process it through one of the affected OpenSSL
1663 functions then this issue could be hit. This might result in a crash
1664 (causing a Denial of Service attack). It could also result in the
1665 disclosure of private memory contents (such as private keys, or
1666 sensitive plaintext).
1671 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1673 * Fixed a problem with verifying a certificate chain when using the
1674 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1675 the certificates present in a certificate chain. It is not set by default.
1677 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1678 the chain that have explicitly encoded elliptic curve parameters was added
1679 as an additional strict check.
1681 An error in the implementation of this check meant that the result of a
1682 previous check to confirm that certificates in the chain are valid CA
1683 certificates was overwritten. This effectively bypasses the check
1684 that non-CA certificates must not be able to issue other certificates.
1686 If a "purpose" has been configured then there is a subsequent opportunity
1687 for checks that the certificate is a valid CA. All of the named "purpose"
1688 values implemented in libcrypto perform this check. Therefore, where
1689 a purpose is set the certificate chain will still be rejected even when the
1690 strict flag has been used. A purpose is set by default in libssl client and
1691 server certificate verification routines, but it can be overridden or
1692 removed by an application.
1694 In order to be affected, an application must explicitly set the
1695 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1696 for the certificate verification or, in the case of TLS client or server
1697 applications, override the default purpose.
1702 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1703 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1704 renegotiation ClientHello omits the signature_algorithms extension (where it
1705 was present in the initial ClientHello), but includes a
1706 signature_algorithms_cert extension then a NULL pointer dereference will
1707 result, leading to a crash and a denial of service attack.
1709 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1710 (which is the default configuration). OpenSSL TLS clients are not impacted by
1714 *Peter Kästle and Samuel Sapalski*
1716 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1718 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1719 create a unique hash value based on the issuer and serial number data
1720 contained within an X509 certificate. However it was failing to correctly
1721 handle any errors that may occur while parsing the issuer field (which might
1722 occur if the issuer field is maliciously constructed). This may subsequently
1723 result in a NULL pointer deref and a crash leading to a potential denial of
1729 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1730 padding mode to correctly check for rollback attacks. This is considered a
1731 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1736 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1737 functions. Previously they could overflow the output length argument in some
1738 cases where the input length is close to the maximum permissable length for
1739 an integer on the platform. In such cases the return value from the function
1740 call would be 1 (indicating success), but the output length value would be
1741 negative. This could cause applications to behave incorrectly or crash.
1746 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1747 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1748 could be exploited in a side channel attack to recover the password. Since
1749 the attack is local host only this is outside of the current OpenSSL
1750 threat model and therefore no CVE is assigned.
1752 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1757 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1759 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1760 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1761 If an attacker can control both items being compared then this could lead
1762 to a possible denial of service attack. OpenSSL itself uses the
1763 GENERAL_NAME_cmp function for two purposes:
1764 1) Comparing CRL distribution point names between an available CRL and a
1765 CRL distribution point embedded in an X509 certificate
1766 2) When verifying that a timestamp response token signer matches the
1767 timestamp authority name (exposed via the API functions
1768 TS_RESP_verify_response and TS_RESP_verify_token)
1773 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1775 * Certificates with explicit curve parameters are now disallowed in
1776 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1780 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1781 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1782 conversely, silently ignore DTLS protocol version bounds when configuring
1783 TLS-based contexts. The commands can be repeated to set bounds of both
1784 types. The same applies with the corresponding "min_protocol" and
1785 "max_protocol" command-line switches, in case some application uses both TLS
1788 SSL_CTX instances that are created for a fixed protocol version (e.g.
1789 TLSv1_server_method()) also silently ignore version bounds. Previously
1790 attempts to apply bounds to these protocol versions would result in an
1791 error. Now only the "version-flexible" SSL_CTX instances are subject to
1792 limits in configuration files in command-line options.
1796 * Handshake now fails if Extended Master Secret extension is dropped
1801 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1803 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1805 * Fixed segmentation fault in SSL_check_chain()
1806 Server or client applications that call the SSL_check_chain() function
1807 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1808 dereference as a result of incorrect handling of the
1809 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1810 or unrecognised signature algorithm is received from the peer. This could
1811 be exploited by a malicious peer in a Denial of Service attack.
1816 * Added AES consttime code for no-asm configurations
1817 an optional constant time support for AES was added
1818 when building openssl for no-asm.
1819 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1820 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1821 At this time this feature is by default disabled.
1822 It will be enabled by default in 3.0.
1826 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1828 * Revert the change of EOF detection while reading in libssl to avoid
1829 regressions in applications depending on the current way of reporting
1830 the EOF. As the existing method is not fully accurate the change to
1831 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1832 branch and will be present in the 3.0 release.
1836 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1837 when primes for RSA keys are computed.
1838 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1839 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1840 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1841 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1842 This avoids possible fingerprinting of newly generated RSA modules.
1846 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1848 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1849 while reading in libssl then we would report an error back to the
1850 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1851 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1852 therefore give a hint as to what went wrong.
1856 * Check that ed25519 and ed448 are allowed by the security level. Previously
1857 signature algorithms not using an MD were not being checked that they were
1858 allowed by the security level.
1862 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1863 was not quite right. The behaviour was not consistent between resumption
1864 and normal handshakes, and also not quite consistent with historical
1865 behaviour. The behaviour in various scenarios has been clarified and
1866 it has been updated to make it match historical behaviour as closely as
1871 * *[VMS only]* The header files that the VMS compilers include automatically,
1872 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1873 that the C++ compiler doesn't understand. This is a shortcoming in the
1874 compiler, but can be worked around with `__cplusplus` guards.
1876 C++ applications that use OpenSSL libraries must be compiled using the
1877 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1878 functions. Otherwise, only functions with symbols of less than 31
1879 characters can be used, as the linker will not be able to successfully
1880 resolve symbols with longer names.
1884 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1885 The presence of this system service is determined at run-time.
1889 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1894 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1896 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1897 number generator (RNG). This was intended to include protection in the
1898 event of a fork() system call in order to ensure that the parent and child
1899 processes did not share the same RNG state. However this protection was not
1900 being used in the default case.
1902 A partial mitigation for this issue is that the output from a high
1903 precision timer is mixed into the RNG state so the likelihood of a parent
1904 and child process sharing state is significantly reduced.
1906 If an application already calls OPENSSL_init_crypto() explicitly using
1907 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1910 *Matthias St. Pierre*
1912 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1913 used even when parsing explicit parameters, when loading a encoded key
1914 or calling `EC_GROUP_new_from_ecpkparameters()`/
1915 `EC_GROUP_new_from_ecparameters()`.
1916 This prevents bypass of security hardening and performance gains,
1917 especially for curves with specialized EC_METHODs.
1918 By default, if a key encoded with explicit parameters is loaded and later
1919 encoded, the output is still encoded with explicit parameters, even if
1920 internally a "named" EC_GROUP is used for computation.
1924 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1925 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1926 NULL. After this change, only the cofactor parameter can be NULL. It also
1927 does some minimal sanity checks on the passed order.
1932 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1933 An attack is simple, if the first CMS_recipientInfo is valid but the
1934 second CMS_recipientInfo is chosen ciphertext. If the second
1935 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1936 encryption key will be replaced by garbage, and the message cannot be
1937 decoded, but if the RSA decryption fails, the correct encryption key is
1938 used and the recipient will not notice the attack.
1939 As a work around for this potential attack the length of the decrypted
1940 key must be equal to the cipher default key length, in case the
1941 certifiate is not given and all recipientInfo are tried out.
1942 The old behaviour can be re-enabled in the CMS code by setting the
1943 CMS_DEBUG_DECRYPT flag.
1948 * Early start up entropy quality from the DEVRANDOM seed source has been
1949 improved for older Linux systems. The RAND subsystem will wait for
1950 /dev/random to be producing output before seeding from /dev/urandom.
1951 The seeded state is stored for future library initialisations using
1952 a system global shared memory segment. The shared memory identifier
1953 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1954 the desired value. The default identifier is 114.
1958 * Correct the extended master secret constant on EBCDIC systems. Without this
1959 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1960 negotiate EMS will fail. Unfortunately this also means that TLS connections
1961 between EBCDIC systems with this fix, and EBCDIC systems without this
1962 fix will fail if they negotiate EMS.
1966 * Use Windows installation paths in the mingw builds
1968 Mingw isn't a POSIX environment per se, which means that Windows
1969 paths should be used for installation.
1974 * Changed DH_check to accept parameters with order q and 2q subgroups.
1975 With order 2q subgroups the bit 0 of the private key is not secret
1976 but DH_generate_key works around that by clearing bit 0 of the
1977 private key for those. This avoids leaking bit 0 of the private key.
1981 * Significantly reduce secure memory usage by the randomness pools.
1985 * Revert the DEVRANDOM_WAIT feature for Linux systems
1987 The DEVRANDOM_WAIT feature added a select() call to wait for the
1988 /dev/random device to become readable before reading from the
1989 /dev/urandom device.
1991 It turned out that this change had negative side effects on
1992 performance which were not acceptable. After some discussion it
1993 was decided to revert this feature and leave it up to the OS
1994 resp. the platform maintainer to ensure a proper initialization
1995 during early boot time.
1997 *Matthias St. Pierre*
1999 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2001 * Add build tests for C++. These are generated files that only do one
2002 thing, to include one public OpenSSL head file each. This tests that
2003 the public header files can be usefully included in a C++ application.
2005 This test isn't enabled by default. It can be enabled with the option
2006 'enable-buildtest-c++'.
2010 * Enable SHA3 pre-hashing for ECDSA and DSA.
2014 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2015 This changes the size when using the `genpkey` command when no size is given.
2016 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2017 generation commands to use 2048 bits by default.
2021 * Reorganize the manual pages to consistently have RETURN VALUES,
2022 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2023 util/fix-doc-nits accordingly.
2025 *Paul Yang, Joshua Lock*
2027 * Add the missing accessor EVP_PKEY_get0_engine()
2031 * Have commands like `s_client` and `s_server` output the signature scheme
2032 along with other cipher suite parameters when debugging.
2036 * Make OPENSSL_config() error agnostic again.
2040 * Do the error handling in RSA decryption constant time.
2044 * Prevent over long nonces in ChaCha20-Poly1305.
2046 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2047 for every encryption operation. RFC 7539 specifies that the nonce value
2048 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2049 and front pads the nonce with 0 bytes if it is less than 12
2050 bytes. However it also incorrectly allows a nonce to be set of up to 16
2051 bytes. In this case only the last 12 bytes are significant and any
2052 additional leading bytes are ignored.
2054 It is a requirement of using this cipher that nonce values are
2055 unique. Messages encrypted using a reused nonce value are susceptible to
2056 serious confidentiality and integrity attacks. If an application changes
2057 the default nonce length to be longer than 12 bytes and then makes a
2058 change to the leading bytes of the nonce expecting the new value to be a
2059 new unique nonce then such an application could inadvertently encrypt
2060 messages with a reused nonce.
2062 Additionally the ignored bytes in a long nonce are not covered by the
2063 integrity guarantee of this cipher. Any application that relies on the
2064 integrity of these ignored leading bytes of a long nonce may be further
2065 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2066 is safe because no such use sets such a long nonce value. However user
2067 applications that use this cipher directly and set a non-default nonce
2068 length to be longer than 12 bytes may be vulnerable.
2070 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2076 * Add DEVRANDOM_WAIT feature for Linux systems
2078 On older Linux systems where the getrandom() system call is not available,
2079 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2080 Contrary to getrandom(), the /dev/urandom device will not block during
2081 early boot when the kernel CSPRNG has not been seeded yet.
2083 To mitigate this known weakness, use select() to wait for /dev/random to
2084 become readable before reading from /dev/urandom.
2086 * Ensure that SM2 only uses SM3 as digest algorithm
2090 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2092 * Change the info callback signals for the start and end of a post-handshake
2093 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2094 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2095 confused by this and assume that a TLSv1.2 renegotiation has started. This
2096 can break KeyUpdate handling. Instead we no longer signal the start and end
2097 of a post handshake message exchange (although the messages themselves are
2098 still signalled). This could break some applications that were expecting
2099 the old signals. However without this KeyUpdate is not usable for many
2104 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2106 * Timing vulnerability in DSA signature generation
2108 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2109 timing side channel attack. An attacker could use variations in the signing
2110 algorithm to recover the private key.
2112 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2117 * Timing vulnerability in ECDSA signature generation
2119 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2120 timing side channel attack. An attacker could use variations in the signing
2121 algorithm to recover the private key.
2123 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2128 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2129 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2130 of two gigabytes and the error handling improved.
2132 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2133 categorized as a normal bug, not a security issue, because the DRBG reseeds
2134 automatically and is fully functional even without additional randomness
2135 provided by the application.
2137 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2139 * Add a new ClientHello callback. Provides a callback interface that gives
2140 the application the ability to adjust the nascent SSL object at the
2141 earliest stage of ClientHello processing, immediately after extensions have
2142 been collected but before they have been processed. In particular, this
2143 callback can adjust the supported TLS versions in response to the contents
2148 * Add SM2 base algorithm support.
2152 * s390x assembly pack: add (improved) hardware-support for the following
2153 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2154 aes-cfb/cfb8, aes-ecb.
2158 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2159 parameter is no longer accepted, as it leads to a corrupt table. NULL
2160 pem_str is reserved for alias entries only.
2164 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2165 step for prime curves. The new implementation is based on formulae from
2166 differential addition-and-doubling in homogeneous projective coordinates
2167 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2168 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2169 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2170 to work in projective coordinates.
2172 *Billy Bob Brumley, Nicola Tuveri*
2174 * Change generating and checking of primes so that the error rate of not
2175 being prime depends on the intended use based on the size of the input.
2176 For larger primes this will result in more rounds of Miller-Rabin.
2177 The maximal error rate for primes with more than 1080 bits is lowered
2180 *Kurt Roeckx, Annie Yousar*
2182 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2186 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2187 moving between systems, and to avoid confusion when a Windows build is
2188 done with mingw vs with MSVC. For POSIX installs, there's still a
2189 symlink or copy named 'tsget' to avoid that confusion as well.
2193 * Revert blinding in ECDSA sign and instead make problematic addition
2194 length-invariant. Switch even to fixed-length Montgomery multiplication.
2198 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2199 step for binary curves. The new implementation is based on formulae from
2200 differential addition-and-doubling in mixed Lopez-Dahab projective
2201 coordinates, modified to independently blind the operands.
2203 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2205 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2206 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2207 EC_METHODs to implement their own specialized "ladder step", to take
2208 advantage of more favorable coordinate systems or more efficient
2209 differential addition-and-doubling algorithms.
2211 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2213 * Modified the random device based seed sources to keep the relevant
2214 file descriptors open rather than reopening them on each access.
2215 This allows such sources to operate in a chroot() jail without
2216 the associated device nodes being available. This behaviour can be
2217 controlled using RAND_keep_random_devices_open().
2221 * Numerous side-channel attack mitigations have been applied. This may have
2222 performance impacts for some algorithms for the benefit of improved
2223 security. Specific changes are noted in this change log by their respective
2228 * AIX shared library support overhaul. Switch to AIX "natural" way of
2229 handling shared libraries, which means collecting shared objects of
2230 different versions and bitnesses in one common archive. This allows to
2231 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2232 doesn't affect the way 3rd party applications are linked, only how
2233 multi-version installation is managed.
2237 * Make ec_group_do_inverse_ord() more robust and available to other
2238 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2239 mitigations are applied to the fallback BN_mod_inverse().
2240 When using this function rather than BN_mod_inverse() directly, new
2241 EC cryptosystem implementations are then safer-by-default.
2245 * Add coordinate blinding for EC_POINT and implement projective
2246 coordinate blinding for generic prime curves as a countermeasure to
2247 chosen point SCA attacks.
2249 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2251 * Add blinding to ECDSA and DSA signatures to protect against side channel
2252 attacks discovered by Keegan Ryan (NCC Group).
2256 * Enforce checking in the `pkeyutl` command to ensure that the input
2257 length does not exceed the maximum supported digest length when performing
2258 a sign, verify or verifyrecover operation.
2262 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2263 I/O in combination with something like select() or poll() will hang. This
2264 can be turned off again using SSL_CTX_clear_mode().
2265 Many applications do not properly handle non-application data records, and
2266 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2267 around the problems in those applications, but can also break some.
2268 It's recommended to read the manpages about SSL_read(), SSL_write(),
2269 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2270 SSL_CTX_set_read_ahead() again.
2274 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2275 now allow empty (zero character) pass phrases.
2279 * Apply blinding to binary field modular inversion and remove patent
2280 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2284 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2285 binary and prime elliptic curves.
2289 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2290 constant time fixed point multiplication.
2294 * Revise elliptic curve scalar multiplication with timing attack
2295 defenses: ec_wNAF_mul redirects to a constant time implementation
2296 when computing fixed point and variable point multiplication (which
2297 in OpenSSL are mostly used with secret scalars in keygen, sign,
2298 ECDH derive operations).
2299 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2302 * Updated CONTRIBUTING
2306 * Updated DRBG / RAND to request nonce and additional low entropy
2307 randomness from the system.
2309 *Matthias St. Pierre*
2311 * Updated 'openssl rehash' to use OpenSSL consistent default.
2315 * Moved the load of the ssl_conf module to libcrypto, which helps
2316 loading engines that libssl uses before libssl is initialised.
2320 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2324 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2326 *Ingo Schwarze, Rich Salz*
2328 * Added output of accepting IP address and port for 'openssl s_server'
2332 * Added a new API for TLSv1.3 ciphersuites:
2333 SSL_CTX_set_ciphersuites()
2334 SSL_set_ciphersuites()
2338 * Memory allocation failures consistently add an error to the error
2343 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2344 in libcrypto when run as setuid/setgid.
2348 * Load any config file by default when libssl is used.
2352 * Added new public header file <openssl/rand_drbg.h> and documentation
2353 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2355 *Matthias St. Pierre*
2357 * QNX support removed (cannot find contributors to get their approval
2358 for the license change).
2362 * TLSv1.3 replay protection for early data has been implemented. See the
2363 SSL_read_early_data() man page for further details.
2367 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2368 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2369 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2370 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2371 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2372 configuration has been separated out. See the ciphers man page or the
2373 SSL_CTX_set_ciphersuites() man page for more information.
2377 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2378 in responder mode now supports the new "-multi" option, which
2379 spawns the specified number of child processes to handle OCSP
2380 requests. The "-timeout" option now also limits the OCSP
2381 responder's patience to wait to receive the full client request
2382 on a newly accepted connection. Child processes are respawned
2383 as needed, and the CA index file is automatically reloaded
2384 when changed. This makes it possible to run the "ocsp" responder
2385 as a long-running service, making the OpenSSL CA somewhat more
2386 feature-complete. In this mode, most diagnostic messages logged
2387 after entering the event loop are logged via syslog(3) rather than
2392 * Added support for X448 and Ed448. Heavily based on original work by
2397 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2398 objects loaded. This adds the functions OSSL_STORE_expect() and
2399 OSSL_STORE_find() as well as needed tools to construct searches and
2400 get the search data out of them.
2404 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2405 version of OpenSSL should review their configuration settings to ensure
2406 that they are still appropriate for TLSv1.3. For further information see:
2407 <https://wiki.openssl.org/index.php/TLS1.3>
2411 * Grand redesign of the OpenSSL random generator
2413 The default RAND method now utilizes an AES-CTR DRBG according to
2414 NIST standard SP 800-90Ar1. The new random generator is essentially
2415 a port of the default random generator from the OpenSSL FIPS 2.0
2416 object module. It is a hybrid deterministic random bit generator
2417 using an AES-CTR bit stream and which seeds and reseeds itself
2418 automatically using trusted system entropy sources.
2420 Some of its new features are:
2421 - Support for multiple DRBG instances with seed chaining.
2422 - The default RAND method makes use of a DRBG.
2423 - There is a public and private DRBG instance.
2424 - The DRBG instances are fork-safe.
2425 - Keep all global DRBG instances on the secure heap if it is enabled.
2426 - The public and private DRBG instance are per thread for lock free
2429 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2431 * Changed Configure so it only says what it does and doesn't dump
2432 so much data. Instead, ./configdata.pm should be used as a script
2433 to display all sorts of configuration data.
2437 * Added processing of "make variables" to Configure.
2441 * Added SHA512/224 and SHA512/256 algorithm support.
2445 * The last traces of Netware support, first removed in 1.1.0, have
2450 * Get rid of Makefile.shared, and in the process, make the processing
2451 of certain files (rc.obj, or the .def/.map/.opt files produced from
2452 the ordinal files) more visible and hopefully easier to trace and
2453 debug (or make silent).
2457 * Make it possible to have environment variable assignments as
2458 arguments to config / Configure.
2462 * Add multi-prime RSA (RFC 8017) support.
2466 * Add SM3 implemented according to GB/T 32905-2016
2467 *Jack Lloyd <jack.lloyd@ribose.com>,*
2468 *Ronald Tse <ronald.tse@ribose.com>,*
2469 *Erick Borsboom <erick.borsboom@ribose.com>*
2471 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2472 as documented in RFC6066.
2473 Based on a patch from Tomasz Moń
2475 *Filipe Raimundo da Silva*
2477 * Add SM4 implemented according to GB/T 32907-2016.
2478 *Jack Lloyd <jack.lloyd@ribose.com>,*
2479 *Ronald Tse <ronald.tse@ribose.com>,*
2480 *Erick Borsboom <erick.borsboom@ribose.com>*
2482 * Reimplement -newreq-nodes and ERR_error_string_n; the
2483 original author does not agree with the license change.
2487 * Add ARIA AEAD TLS support.
2491 * Some macro definitions to support VS6 have been removed. Visual
2492 Studio 6 has not worked since 1.1.0
2496 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2497 without clearing the errors.
2501 * Add "atfork" functions. If building on a system that without
2502 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2503 requirements. The RAND facility now uses/requires this.
2511 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2512 not possible to disable entirely. However, it's still possible to
2513 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2516 To disable, configure with 'no-ui-console'. 'no-ui' is still
2517 possible to use as an alias. Check at compile time with the
2518 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2519 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2523 * Add a STORE module, which implements a uniform and URI based reader of
2524 stores that can contain keys, certificates, CRLs and numerous other
2525 objects. The main API is loosely based on a few stdio functions,
2526 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2527 OSSL_STORE_error and OSSL_STORE_close.
2528 The implementation uses backends called "loaders" to implement arbitrary
2529 URI schemes. There is one built in "loader" for the 'file' scheme.
2533 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2534 then adjusted to work on FreeBSD 8.4 as well.
2535 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2536 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2540 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2541 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2542 error code calls like this:
2544 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2546 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2547 that can be encoded in C. For the foreseeable future, this will only
2550 *Richard Levitte and Tim Hudson*
2552 * Removed BSD cryptodev engine.
2556 * Add a build target 'build_all_generated', to build all generated files
2557 and only that. This can be used to prepare everything that requires
2558 things like perl for a system that lacks perl and then move everything
2559 to that system and do the rest of the build there.
2563 * In the UI interface, make it possible to duplicate the user data. This
2564 can be used by engines that need to retain the data for a longer time
2565 than just the call where this user data is passed.
2569 * Ignore the '-named_curve auto' value for compatibility of applications
2572 *Tomáš Mráz <tmraz@fedoraproject.org>*
2574 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2575 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2576 alerts across multiple records (some of which could be empty). In practice
2577 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2578 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2579 support this at all. Supporting it adds significant complexity to the
2580 record layer, and its removal is unlikely to cause interoperability
2585 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2586 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2587 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2592 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2593 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2595 *Richard Levitte, Andy Polyakov*
2597 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2602 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2603 platform rather than 'mingw'.
2607 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2608 success if they are asked to add an object which already exists
2609 in the store. This change cascades to other functions which load
2610 certificates and CRLs.
2614 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2615 facilitate stack unwinding even from assembly subroutines.
2619 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2620 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2624 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2625 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2626 which is the minimum version we support.
2630 * Certificate time validation (X509_cmp_time) enforces stricter
2631 compliance with RFC 5280. Fractional seconds and timezone offsets
2632 are no longer allowed.
2636 * Add support for ARIA
2640 * s_client will now send the Server Name Indication (SNI) extension by
2641 default unless the new "-noservername" option is used. The server name is
2642 based on the host provided to the "-connect" option unless overridden by
2643 using "-servername".
2647 * Add support for SipHash
2651 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2652 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2653 prevent issues where no progress is being made and the peer continually
2654 sends unrecognised record types, using up resources processing them.
2658 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2659 using the algorithm defined in
2660 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2664 * Heartbeat support has been removed; the ABI is changed for now.
2666 *Richard Levitte, Rich Salz*
2668 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2672 * The RSA "null" method, which was partially supported to avoid patent
2673 issues, has been replaced to always returns NULL.
2680 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2682 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2683 used even when parsing explicit parameters, when loading a encoded key
2684 or calling `EC_GROUP_new_from_ecpkparameters()`/
2685 `EC_GROUP_new_from_ecparameters()`.
2686 This prevents bypass of security hardening and performance gains,
2687 especially for curves with specialized EC_METHODs.
2688 By default, if a key encoded with explicit parameters is loaded and later
2689 encoded, the output is still encoded with explicit parameters, even if
2690 internally a "named" EC_GROUP is used for computation.
2694 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2695 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2696 NULL. After this change, only the cofactor parameter can be NULL. It also
2697 does some minimal sanity checks on the passed order.
2702 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2703 An attack is simple, if the first CMS_recipientInfo is valid but the
2704 second CMS_recipientInfo is chosen ciphertext. If the second
2705 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2706 encryption key will be replaced by garbage, and the message cannot be
2707 decoded, but if the RSA decryption fails, the correct encryption key is
2708 used and the recipient will not notice the attack.
2709 As a work around for this potential attack the length of the decrypted
2710 key must be equal to the cipher default key length, in case the
2711 certifiate is not given and all recipientInfo are tried out.
2712 The old behaviour can be re-enabled in the CMS code by setting the
2713 CMS_DEBUG_DECRYPT flag.
2718 * Use Windows installation paths in the mingw builds
2720 Mingw isn't a POSIX environment per se, which means that Windows
2721 paths should be used for installation.
2726 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2728 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2729 This changes the size when using the `genpkey` command when no size is given.
2730 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2731 generation commands to use 2048 bits by default.
2735 * Prevent over long nonces in ChaCha20-Poly1305.
2737 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2738 for every encryption operation. RFC 7539 specifies that the nonce value
2739 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2740 and front pads the nonce with 0 bytes if it is less than 12
2741 bytes. However it also incorrectly allows a nonce to be set of up to 16
2742 bytes. In this case only the last 12 bytes are significant and any
2743 additional leading bytes are ignored.
2745 It is a requirement of using this cipher that nonce values are
2746 unique. Messages encrypted using a reused nonce value are susceptible to
2747 serious confidentiality and integrity attacks. If an application changes
2748 the default nonce length to be longer than 12 bytes and then makes a
2749 change to the leading bytes of the nonce expecting the new value to be a
2750 new unique nonce then such an application could inadvertently encrypt
2751 messages with a reused nonce.
2753 Additionally the ignored bytes in a long nonce are not covered by the
2754 integrity guarantee of this cipher. Any application that relies on the
2755 integrity of these ignored leading bytes of a long nonce may be further
2756 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2757 is safe because no such use sets such a long nonce value. However user
2758 applications that use this cipher directly and set a non-default nonce
2759 length to be longer than 12 bytes may be vulnerable.
2761 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2767 * Added SCA hardening for modular field inversion in EC_GROUP through
2768 a new dedicated field_inv() pointer in EC_METHOD.
2769 This also addresses a leakage affecting conversions from projective
2770 to affine coordinates.
2772 *Billy Bob Brumley, Nicola Tuveri*
2774 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2775 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2779 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2783 * Remove the 'dist' target and add a tarball building script. The
2784 'dist' target has fallen out of use, and it shouldn't be
2785 necessary to configure just to create a source distribution.
2789 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2791 * Timing vulnerability in DSA signature generation
2793 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2794 timing side channel attack. An attacker could use variations in the signing
2795 algorithm to recover the private key.
2797 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2802 * Timing vulnerability in ECDSA signature generation
2804 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2805 timing side channel attack. An attacker could use variations in the signing
2806 algorithm to recover the private key.
2808 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2813 * Add coordinate blinding for EC_POINT and implement projective
2814 coordinate blinding for generic prime curves as a countermeasure to
2815 chosen point SCA attacks.
2817 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2819 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2821 * Client DoS due to large DH parameter
2823 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2824 malicious server can send a very large prime value to the client. This will
2825 cause the client to spend an unreasonably long period of time generating a
2826 key for this prime resulting in a hang until the client has finished. This
2827 could be exploited in a Denial Of Service attack.
2829 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2834 * Cache timing vulnerability in RSA Key Generation
2836 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2837 a cache timing side channel attack. An attacker with sufficient access to
2838 mount cache timing attacks during the RSA key generation process could
2839 recover the private key.
2841 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2842 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2847 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2848 parameter is no longer accepted, as it leads to a corrupt table. NULL
2849 pem_str is reserved for alias entries only.
2853 * Revert blinding in ECDSA sign and instead make problematic addition
2854 length-invariant. Switch even to fixed-length Montgomery multiplication.
2858 * Change generating and checking of primes so that the error rate of not
2859 being prime depends on the intended use based on the size of the input.
2860 For larger primes this will result in more rounds of Miller-Rabin.
2861 The maximal error rate for primes with more than 1080 bits is lowered
2864 *Kurt Roeckx, Annie Yousar*
2866 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2870 * Add blinding to ECDSA and DSA signatures to protect against side channel
2871 attacks discovered by Keegan Ryan (NCC Group).
2875 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2876 now allow empty (zero character) pass phrases.
2880 * Certificate time validation (X509_cmp_time) enforces stricter
2881 compliance with RFC 5280. Fractional seconds and timezone offsets
2882 are no longer allowed.
2886 * Fixed a text canonicalisation bug in CMS
2888 Where a CMS detached signature is used with text content the text goes
2889 through a canonicalisation process first prior to signing or verifying a
2890 signature. This process strips trailing space at the end of lines, converts
2891 line terminators to CRLF and removes additional trailing line terminators
2892 at the end of a file. A bug in the canonicalisation process meant that
2893 some characters, such as form-feed, were incorrectly treated as whitespace
2894 and removed. This is contrary to the specification (RFC5485). This fix
2895 could mean that detached text data signed with an earlier version of
2896 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2897 signed with a fixed OpenSSL may fail to verify with an earlier version of
2898 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2899 and use the "-binary" flag (for the "cms" command line application) or set
2900 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2904 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2906 * Constructed ASN.1 types with a recursive definition could exceed the stack
2908 Constructed ASN.1 types with a recursive definition (such as can be found
2909 in PKCS7) could eventually exceed the stack given malicious input with
2910 excessive recursion. This could result in a Denial Of Service attack. There
2911 are no such structures used within SSL/TLS that come from untrusted sources
2912 so this is considered safe.
2914 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2920 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2922 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2923 effectively reduced to only comparing the least significant bit of each
2924 byte. This allows an attacker to forge messages that would be considered as
2925 authenticated in an amount of tries lower than that guaranteed by the
2926 security claims of the scheme. The module can only be compiled by the
2927 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2929 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2935 * Add a build target 'build_all_generated', to build all generated files
2936 and only that. This can be used to prepare everything that requires
2937 things like perl for a system that lacks perl and then move everything
2938 to that system and do the rest of the build there.
2942 * Backport SSL_OP_NO_RENGOTIATION
2944 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2945 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2946 changes this is no longer possible in 1.1.0. Therefore the new
2947 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2948 1.1.0 to provide equivalent functionality.
2950 Note that if an application built against 1.1.0h headers (or above) is run
2951 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2952 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2956 * Removed the OS390-Unix config target. It relied on a script that doesn't
2961 * rsaz_1024_mul_avx2 overflow bug on x86_64
2963 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2964 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2965 Analysis suggests that attacks against RSA and DSA as a result of this
2966 defect would be very difficult to perform and are not believed likely.
2967 Attacks against DH1024 are considered just feasible, because most of the
2968 work necessary to deduce information about a private key may be performed
2969 offline. The amount of resources required for such an attack would be
2970 significant. However, for an attack on TLS to be meaningful, the server
2971 would have to share the DH1024 private key among multiple clients, which is
2972 no longer an option since CVE-2016-0701.
2974 This only affects processors that support the AVX2 but not ADX extensions
2975 like Intel Haswell (4th generation).
2977 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2978 was originally found via the OSS-Fuzz project.
2983 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2985 * bn_sqrx8x_internal carry bug on x86_64
2987 There is a carry propagating bug in the x86_64 Montgomery squaring
2988 procedure. No EC algorithms are affected. Analysis suggests that attacks
2989 against RSA and DSA as a result of this defect would be very difficult to
2990 perform and are not believed likely. Attacks against DH are considered just
2991 feasible (although very difficult) because most of the work necessary to
2992 deduce information about a private key may be performed offline. The amount
2993 of resources required for such an attack would be very significant and
2994 likely only accessible to a limited number of attackers. An attacker would
2995 additionally need online access to an unpatched system using the target
2996 private key in a scenario with persistent DH parameters and a private
2997 key that is shared between multiple clients.
2999 This only affects processors that support the BMI1, BMI2 and ADX extensions
3000 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3002 This issue was reported to OpenSSL by the OSS-Fuzz project.
3007 * Malformed X.509 IPAddressFamily could cause OOB read
3009 If an X.509 certificate has a malformed IPAddressFamily extension,
3010 OpenSSL could do a one-byte buffer overread. The most likely result
3011 would be an erroneous display of the certificate in text format.
3013 This issue was reported to OpenSSL by the OSS-Fuzz project.
3018 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3020 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3021 platform rather than 'mingw'.
3025 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3026 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3027 which is the minimum version we support.
3031 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3033 * Encrypt-Then-Mac renegotiation crash
3035 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3036 negotiated where it was not in the original handshake (or vice-versa) then
3037 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3038 and servers are affected.
3040 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3045 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3047 * Truncated packet could crash via OOB read
3049 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3050 cipher is being used, then a truncated packet can cause that host to
3051 perform an out-of-bounds read, usually resulting in a crash.
3053 This issue was reported to OpenSSL by Robert Święcki of Google.
3058 * Bad (EC)DHE parameters cause a client crash
3060 If a malicious server supplies bad parameters for a DHE or ECDHE key
3061 exchange then this can result in the client attempting to dereference a
3062 NULL pointer leading to a client crash. This could be exploited in a Denial
3065 This issue was reported to OpenSSL by Guido Vranken.
3070 * BN_mod_exp may produce incorrect results on x86_64
3072 There is a carry propagating bug in the x86_64 Montgomery squaring
3073 procedure. No EC algorithms are affected. Analysis suggests that attacks
3074 against RSA and DSA as a result of this defect would be very difficult to
3075 perform and are not believed likely. Attacks against DH are considered just
3076 feasible (although very difficult) because most of the work necessary to
3077 deduce information about a private key may be performed offline. The amount
3078 of resources required for such an attack would be very significant and
3079 likely only accessible to a limited number of attackers. An attacker would
3080 additionally need online access to an unpatched system using the target
3081 private key in a scenario with persistent DH parameters and a private
3082 key that is shared between multiple clients. For example this can occur by
3083 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3084 similar to CVE-2015-3193 but must be treated as a separate problem.
3086 This issue was reported to OpenSSL by the OSS-Fuzz project.
3091 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3093 * ChaCha20/Poly1305 heap-buffer-overflow
3095 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3096 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3097 crash. This issue is not considered to be exploitable beyond a DoS.
3099 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3104 * CMS Null dereference
3106 Applications parsing invalid CMS structures can crash with a NULL pointer
3107 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3108 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3109 structure callback if an attempt is made to free certain invalid encodings.
3110 Only CHOICE structures using a callback which do not handle NULL value are
3113 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3118 * Montgomery multiplication may produce incorrect results
3120 There is a carry propagating bug in the Broadwell-specific Montgomery
3121 multiplication procedure that handles input lengths divisible by, but
3122 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3123 and DH private keys are impossible. This is because the subroutine in
3124 question is not used in operations with the private key itself and an input
3125 of the attacker's direct choice. Otherwise the bug can manifest itself as
3126 transient authentication and key negotiation failures or reproducible
3127 erroneous outcome of public-key operations with specially crafted input.
3128 Among EC algorithms only Brainpool P-512 curves are affected and one
3129 presumably can attack ECDH key negotiation. Impact was not analyzed in
3130 detail, because pre-requisites for attack are considered unlikely. Namely
3131 multiple clients have to choose the curve in question and the server has to
3132 share the private key among them, neither of which is default behaviour.
3133 Even then only clients that chose the curve will be affected.
3135 This issue was publicly reported as transient failures and was not
3136 initially recognized as a security issue. Thanks to Richard Morgan for
3137 providing reproducible case.
3142 * Removed automatic addition of RPATH in shared libraries and executables,
3143 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3147 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3149 * Fix Use After Free for large message sizes
3151 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3152 message larger than approx 16k is received then the underlying buffer to
3153 store the incoming message is reallocated and moved. Unfortunately a
3154 dangling pointer to the old location is left which results in an attempt to
3155 write to the previously freed location. This is likely to result in a
3156 crash, however it could potentially lead to execution of arbitrary code.
3158 This issue only affects OpenSSL 1.1.0a.
3160 This issue was reported to OpenSSL by Robert Święcki.
3165 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3167 * OCSP Status Request extension unbounded memory growth
3169 A malicious client can send an excessively large OCSP Status Request
3170 extension. If that client continually requests renegotiation, sending a
3171 large OCSP Status Request extension each time, then there will be unbounded
3172 memory growth on the server. This will eventually lead to a Denial Of
3173 Service attack through memory exhaustion. Servers with a default
3174 configuration are vulnerable even if they do not support OCSP. Builds using
3175 the "no-ocsp" build time option are not affected.
3177 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3182 * SSL_peek() hang on empty record
3184 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3185 sends an empty record. This could be exploited by a malicious peer in a
3186 Denial Of Service attack.
3188 This issue was reported to OpenSSL by Alex Gaynor.
3193 * Excessive allocation of memory in tls_get_message_header() and
3194 dtls1_preprocess_fragment()
3196 A (D)TLS message includes 3 bytes for its length in the header for the
3197 message. This would allow for messages up to 16Mb in length. Messages of
3198 this length are excessive and OpenSSL includes a check to ensure that a
3199 peer is sending reasonably sized messages in order to avoid too much memory
3200 being consumed to service a connection. A flaw in the logic of version
3201 1.1.0 means that memory for the message is allocated too early, prior to
3202 the excessive message length check. Due to way memory is allocated in
3203 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3204 to service a connection. This could lead to a Denial of Service through
3205 memory exhaustion. However, the excessive message length check still takes
3206 place, and this would cause the connection to immediately fail. Assuming
3207 that the application calls SSL_free() on the failed connection in a timely
3208 manner then the 21Mb of allocated memory will then be immediately freed
3209 again. Therefore the excessive memory allocation will be transitory in
3210 nature. This then means that there is only a security impact if:
3212 1) The application does not call SSL_free() in a timely manner in the event
3213 that the connection fails
3215 2) The application is working in a constrained environment where there is
3216 very little free memory
3218 3) The attacker initiates multiple connection attempts such that there are
3219 multiple connections in a state where memory has been allocated for the
3220 connection; SSL_free() has not yet been called; and there is insufficient
3221 memory to service the multiple requests.
3223 Except in the instance of (1) above any Denial Of Service is likely to be
3224 transitory because as soon as the connection fails the memory is
3225 subsequently freed again in the SSL_free() call. However there is an
3226 increased risk during this period of application crashes due to the lack of
3227 memory - which would then mean a more serious Denial of Service.
3229 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3230 (CVE-2016-6307 and CVE-2016-6308)
3234 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3235 had to be removed. Primary reason is that vendor assembler can't
3236 assemble our modules with -KPIC flag. As result it, assembly
3237 support, was not even available as option. But its lack means
3238 lack of side-channel resistant code, which is incompatible with
3239 security by todays standards. Fortunately gcc is readily available
3240 prepackaged option, which we firmly point at...
3244 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3246 * Windows command-line tool supports UTF-8 opt-in option for arguments
3247 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3248 (to any value) allows Windows user to access PKCS#12 file generated
3249 with Windows CryptoAPI and protected with non-ASCII password, as well
3250 as files generated under UTF-8 locale on Linux also protected with
3255 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3256 have been disabled by default and removed from DEFAULT, just like RC4.
3257 See the RC4 item below to re-enable both.
3261 * The method for finding the storage location for the Windows RAND seed file
3262 has changed. First we check %RANDFILE%. If that is not set then we check
3263 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3264 all else fails we fall back to C:\.
3268 * The EVP_EncryptUpdate() function has had its return type changed from void
3269 to int. A return of 0 indicates and error while a return of 1 indicates
3274 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3275 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3276 off the constant time implementation for RSA, DSA and DH have been made
3277 no-ops and deprecated.
3281 * Windows RAND implementation was simplified to only get entropy by
3282 calling CryptGenRandom(). Various other RAND-related tickets
3285 *Joseph Wylie Yandle, Rich Salz*
3287 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3288 and `OPENSSL_LH_`, respectively. The old names are available
3289 with API compatibility. They new names are now completely documented.
3293 * Unify TYPE_up_ref(obj) methods signature.
3294 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3295 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3296 int (instead of void) like all others TYPE_up_ref() methods.
3297 So now these methods also check the return value of CRYPTO_atomic_add(),
3298 and the validity of object reference counter.
3300 *fdasilvayy@gmail.com*
3302 * With Windows Visual Studio builds, the .pdb files are installed
3303 alongside the installed libraries and executables. For a static
3304 library installation, ossl_static.pdb is the associate compiler
3305 generated .pdb file to be used when linking programs.
3309 * Remove openssl.spec. Packaging files belong with the packagers.
3313 * Automatic Darwin/OSX configuration has had a refresh, it will now
3314 recognise x86_64 architectures automatically. You can still decide
3315 to build for a different bitness with the environment variable
3316 KERNEL_BITS (can be 32 or 64), for example:
3318 KERNEL_BITS=32 ./config
3322 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3323 256 bit AES and HMAC with SHA256.
3327 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3331 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3335 * To enable users to have their own config files and build file templates,
3336 Configure looks in the directory indicated by the environment variable
3337 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3338 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3339 name and is used as is.
3343 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3344 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3345 X509_CERT_FILE_CTX was removed.
3349 * "shared" builds are now the default. To create only static libraries use
3350 the "no-shared" Configure option.
3354 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3355 All of these option have not worked for some while and are fundamental
3360 * Make various cleanup routines no-ops and mark them as deprecated. Most
3361 global cleanup functions are no longer required because they are handled
3362 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3363 Explicitly de-initing can cause problems (e.g. where a library that uses
3364 OpenSSL de-inits, but an application is still using it). The affected
3365 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3366 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3367 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3368 COMP_zlib_cleanup().
3372 * --strict-warnings no longer enables runtime debugging options
3373 such as REF_DEBUG. Instead, debug options are automatically
3374 enabled with '--debug' builds.
3376 *Andy Polyakov, Emilia Käsper*
3378 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3379 have been moved out of the public header files. New functions for managing
3380 these have been added.
3384 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3385 objects have been moved out of the public header files. New
3386 functions for managing these have been added.
3390 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3391 have been moved out of the public header files. New functions for managing
3392 these have been added.
3396 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3397 moved out of the public header files. New functions for managing these
3402 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3406 * Removed the mk1mf build scripts.
3410 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3411 it is always safe to #include a header now.
3415 * Removed the aged BC-32 config and all its supporting scripts
3419 * Removed support for Ultrix, Netware, and OS/2.
3423 * Add support for HKDF.
3425 *Alessandro Ghedini*
3427 * Add support for blake2b and blake2s
3431 * Added support for "pipelining". Ciphers that have the
3432 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3433 encryptions/decryptions simultaneously. There are currently no built-in
3434 ciphers with this property but the expectation is that engines will be able
3435 to offer it to significantly improve throughput. Support has been extended
3436 into libssl so that multiple records for a single connection can be
3437 processed in one go (for >=TLS 1.1).
3441 * Added the AFALG engine. This is an async capable engine which is able to
3442 offload work to the Linux kernel. In this initial version it only supports
3443 AES128-CBC. The kernel must be version 4.1.0 or greater.
3447 * OpenSSL now uses a new threading API. It is no longer necessary to
3448 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3449 are two supported threading models: pthreads and windows threads. It is
3450 also possible to configure OpenSSL at compile time for "no-threads". The
3451 old threading API should no longer be used. The functions have been
3452 replaced with "no-op" compatibility macros.
3454 *Alessandro Ghedini, Matt Caswell*
3456 * Modify behavior of ALPN to invoke callback after SNI/servername
3457 callback, such that updates to the SSL_CTX affect ALPN.
3461 * Add SSL_CIPHER queries for authentication and key-exchange.
3465 * Changes to the DEFAULT cipherlist:
3466 - Prefer (EC)DHE handshakes over plain RSA.
3467 - Prefer AEAD ciphers over legacy ciphers.
3468 - Prefer ECDSA over RSA when both certificates are available.
3469 - Prefer TLSv1.2 ciphers/PRF.
3470 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3475 * Change the ECC default curve list to be this, in order: x25519,
3476 secp256r1, secp521r1, secp384r1.
3480 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3481 disabled by default. They can be re-enabled using the
3482 enable-weak-ssl-ciphers option to Configure.
3486 * If the server has ALPN configured, but supports no protocols that the
3487 client advertises, send a fatal "no_application_protocol" alert.
3488 This behaviour is SHALL in RFC 7301, though it isn't universally
3489 implemented by other servers.
3493 * Add X25519 support.
3494 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3495 for public and private key encoding using the format documented in
3496 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3497 key generation and key derivation.
3499 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3504 * Deprecate SRP_VBASE_get_by_user.
3505 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3506 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3507 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3508 seed, even if the seed is configured.
3510 Users should use SRP_VBASE_get1_by_user instead. Note that in
3511 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3512 also that even though configuring the SRP seed attempts to hide
3513 invalid usernames by continuing the handshake with fake
3514 credentials, this behaviour is not constant time and no strong
3515 guarantees are made that the handshake is indistinguishable from
3516 that of a valid user.
3520 * Configuration change; it's now possible to build dynamic engines
3521 without having to build shared libraries and vice versa. This
3522 only applies to the engines in `engines/`, those in `crypto/engine/`
3523 will always be built into libcrypto (i.e. "static").
3525 Building dynamic engines is enabled by default; to disable, use
3526 the configuration option "disable-dynamic-engine".
3528 The only requirements for building dynamic engines are the
3529 presence of the DSO module and building with position independent
3530 code, so they will also automatically be disabled if configuring
3531 with "disable-dso" or "disable-pic".
3533 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3534 are also taken away from openssl/opensslconf.h, as they are
3539 * Configuration change; if there is a known flag to compile
3540 position independent code, it will always be applied on the
3541 libcrypto and libssl object files, and never on the application
3542 object files. This means other libraries that use routines from
3543 libcrypto / libssl can be made into shared libraries regardless
3544 of how OpenSSL was configured.
3546 If this isn't desirable, the configuration options "disable-pic"
3547 or "no-pic" can be used to disable the use of PIC. This will
3548 also disable building shared libraries and dynamic engines.
3552 * Removed JPAKE code. It was experimental and has no wide use.
3556 * The INSTALL_PREFIX Makefile variable has been renamed to
3557 DESTDIR. That makes for less confusion on what this variable
3558 is for. Also, the configuration option --install_prefix is
3563 * Heartbeat for TLS has been removed and is disabled by default
3564 for DTLS; configure with enable-heartbeats. Code that uses the
3565 old #define's might need to be updated.
3567 *Emilia Käsper, Rich Salz*
3569 * Rename REF_CHECK to REF_DEBUG.
3573 * New "unified" build system
3575 The "unified" build system is aimed to be a common system for all
3576 platforms we support. With it comes new support for VMS.
3578 This system builds supports building in a different directory tree
3579 than the source tree. It produces one Makefile (for unix family
3580 or lookalikes), or one descrip.mms (for VMS).
3582 The source of information to make the Makefile / descrip.mms is
3583 small files called 'build.info', holding the necessary
3584 information for each directory with source to compile, and a
3585 template in Configurations, like unix-Makefile.tmpl or
3588 With this change, the library names were also renamed on Windows
3589 and on VMS. They now have names that are closer to the standard
3590 on Unix, and include the major version number, and in certain
3591 cases, the architecture they are built for. See "Notes on shared
3592 libraries" in INSTALL.
3594 We rely heavily on the perl module Text::Template.
3598 * Added support for auto-initialisation and de-initialisation of the library.
3599 OpenSSL no longer requires explicit init or deinit routines to be called,
3600 except in certain circumstances. See the OPENSSL_init_crypto() and
3601 OPENSSL_init_ssl() man pages for further information.
3605 * The arguments to the DTLSv1_listen function have changed. Specifically the
3606 "peer" argument is now expected to be a BIO_ADDR object.
3608 * Rewrite of BIO networking library. The BIO library lacked consistent
3609 support of IPv6, and adding it required some more extensive
3610 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3611 which hold all types of addresses and chains of address information.
3612 It also introduces a new API, with functions like BIO_socket,
3613 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3614 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3615 have been adapted accordingly.
3619 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3624 * CRIME protection: disable compression by default, even if OpenSSL is
3625 compiled with zlib enabled. Applications can still enable compression
3626 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3627 using the SSL_CONF library to configure compression.
3631 * The signature of the session callback configured with
3632 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3633 was explicitly marked as `const unsigned char*` instead of
3638 * Always DPURIFY. Remove the use of uninitialized memory in the
3639 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3643 * Removed many obsolete configuration items, including
3644 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3645 MD2_CHAR, MD2_INT, MD2_LONG
3647 IDEA_SHORT, IDEA_LONG
3648 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3650 *Rich Salz, with advice from Andy Polyakov*
3652 * Many BN internals have been moved to an internal header file.
3654 *Rich Salz with help from Andy Polyakov*
3656 * Configuration and writing out the results from it has changed.
3657 Files such as Makefile include/openssl/opensslconf.h and are now
3658 produced through general templates, such as Makefile.in and
3659 crypto/opensslconf.h.in and some help from the perl module
3662 Also, the center of configuration information is no longer
3663 Makefile. Instead, Configure produces a perl module in
3664 configdata.pm which holds most of the config data (in the hash
3665 table %config), the target data that comes from the target
3666 configuration in one of the `Configurations/*.conf` files (in
3671 * To clarify their intended purposes, the Configure options
3672 --prefix and --openssldir change their semantics, and become more
3673 straightforward and less interdependent.
3675 --prefix shall be used exclusively to give the location INSTALLTOP
3676 where programs, scripts, libraries, include files and manuals are
3677 going to be installed. The default is now /usr/local.
3679 --openssldir shall be used exclusively to give the default
3680 location OPENSSLDIR where certificates, private keys, CRLs are
3681 managed. This is also where the default openssl.cnf gets
3683 If the directory given with this option is a relative path, the
3684 values of both the --prefix value and the --openssldir value will
3685 be combined to become OPENSSLDIR.
3686 The default for --openssldir is INSTALLTOP/ssl.
3688 Anyone who uses --openssldir to specify where OpenSSL is to be
3689 installed MUST change to use --prefix instead.
3693 * The GOST engine was out of date and therefore it has been removed. An up
3694 to date GOST engine is now being maintained in an external repository.
3695 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3696 support for GOST ciphersuites (these are only activated if a GOST engine
3701 * EGD is no longer supported by default; use enable-egd when
3704 *Ben Kaduk and Rich Salz*
3706 * The distribution now has Makefile.in files, which are used to
3707 create Makefile's when Configure is run. *Configure must be run
3708 before trying to build now.*
3712 * The return value for SSL_CIPHER_description() for error conditions
3717 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3719 Obtaining and performing DNSSEC validation of TLSA records is
3720 the application's responsibility. The application provides
3721 the TLSA records of its choice to OpenSSL, and these are then
3722 used to authenticate the peer.
3724 The TLSA records need not even come from DNS. They can, for
3725 example, be used to implement local end-entity certificate or
3726 trust-anchor "pinning", where the "pin" data takes the form
3727 of TLSA records, which can augment or replace verification
3728 based on the usual WebPKI public certification authorities.
3732 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3733 continues to support deprecated interfaces in default builds.
3734 However, applications are strongly advised to compile their
3735 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3736 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3737 or the 1.1.0 releases.
3739 In environments in which all applications have been ported to
3740 not use any deprecated interfaces OpenSSL's Configure script
3741 should be used with the --api=1.1.0 option to entirely remove
3742 support for the deprecated features from the library and
3743 unconditionally disable them in the installed headers.
3744 Essentially the same effect can be achieved with the "no-deprecated"
3745 argument to Configure, except that this will always restrict
3746 the build to just the latest API, rather than a fixed API
3749 As applications are ported to future revisions of the API,
3750 they should update their compile-time OPENSSL_API_COMPAT define
3751 accordingly, but in most cases should be able to continue to
3752 compile with later releases.
3754 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3755 0x10000000L and 0x00908000L, respectively. However those
3756 versions did not support the OPENSSL_API_COMPAT feature, and
3757 so applications are not typically tested for explicit support
3758 of just the undeprecated features of either release.
3762 * Add support for setting the minimum and maximum supported protocol.
3763 It can bet set via the SSL_set_min_proto_version() and
3764 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3765 MaxProtocol. It's recommended to use the new APIs to disable
3766 protocols instead of disabling individual protocols using
3767 SSL_set_options() or SSL_CONF's Protocol. This change also
3768 removes support for disabling TLS 1.2 in the OpenSSL TLS
3769 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3773 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3777 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3778 and integrates ECDSA and ECDH functionality into EC. Implementations can
3779 now redirect key generation and no longer need to convert to or from
3782 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3783 include the ec.h header file instead.
3787 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3788 ciphers who are no longer supported and drops support the ephemeral RSA key
3789 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3793 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3794 opaque. For HMAC_CTX, the following constructors and destructors
3797 HMAC_CTX *HMAC_CTX_new(void);
3798 void HMAC_CTX_free(HMAC_CTX *ctx);
3800 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3801 destroy such methods has been added. See EVP_MD_meth_new(3) and
3802 EVP_CIPHER_meth_new(3) for documentation.
3805 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3806 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3807 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3808 an already created structure.
3809 2) For consistency with the majority of our object creators and
3810 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3811 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3812 for deprecated builds.
3816 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3817 cryptographic operations to be performed asynchronously as long as an
3818 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3819 further details. Libssl has also had this capability integrated with the
3820 introduction of the new mode SSL_MODE_ASYNC and associated error
3821 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3822 pages. This work was developed in partnership with Intel Corp.
3826 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3827 always enabled now. If you want to disable the support you should
3828 exclude it using the list of supported ciphers. This also means that the
3829 "-no_ecdhe" option has been removed from s_server.
3833 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3834 SSL_{CTX_}set1_curves() which can set a list.
3838 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3839 curve you want to support using SSL_{CTX_}set1_curves().
3843 * State machine rewrite. The state machine code has been significantly
3844 refactored in order to remove much duplication of code and solve issues
3845 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3846 further details). This change does have some associated API changes.
3847 Notably the SSL_state() function has been removed and replaced by
3848 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3849 SSL_set_state() has been removed altogether. The previous handshake states
3850 defined in ssl.h and ssl3.h have also been removed.
3854 * All instances of the string "ssleay" in the public API were replaced
3855 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3856 Some error codes related to internal RSA_eay API's were renamed.
3860 * The demo files in crypto/threads were moved to demo/threads.
3864 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3867 *Matt Caswell, Rich Salz*
3869 * New ASN.1 embed macro.
3871 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3872 structure is not allocated: it is part of the parent. That is instead of
3880 This reduces memory fragmentation and make it impossible to accidentally
3881 set a mandatory field to NULL.
3883 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3884 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3885 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3890 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3894 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3895 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3896 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3897 DES and RC4 ciphersuites.
3901 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3902 This changes the decoding behaviour for some invalid messages,
3903 though the change is mostly in the more lenient direction, and
3904 legacy behaviour is preserved as much as possible.
3908 * Fix no-stdio build.
3909 *David Woodhouse <David.Woodhouse@intel.com> and also*
3910 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3912 * New testing framework
3913 The testing framework has been largely rewritten and is now using
3914 perl and the perl modules Test::Harness and an extended variant of
3915 Test::More called OpenSSL::Test to do its work. All test scripts in
3916 test/ have been rewritten into test recipes, and all direct calls to
3917 executables in test/Makefile have become individual recipes using the
3918 simplified testing OpenSSL::Test::Simple.
3920 For documentation on our testing modules, do:
3922 perldoc test/testlib/OpenSSL/Test/Simple.pm
3923 perldoc test/testlib/OpenSSL/Test.pm
3927 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3928 are used; the latter aborts on memory leaks (usually checked on exit).
3929 Some undocumented "set malloc, etc., hooks" functions were removed
3930 and others were changed. All are now documented.
3934 * In DSA_generate_parameters_ex, if the provided seed is too short,
3937 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3939 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3940 from RFC4279, RFC4785, RFC5487, RFC5489.
3942 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3943 original RSA_PSK patch.
3947 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3948 era flag was never set throughout the codebase (only read). Also removed
3949 SSL3_FLAGS_POP_BUFFER which was only used if
3950 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3954 * Changed the default name options in the "ca", "crl", "req" and "x509"
3955 to be "oneline" instead of "compat".
3959 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3960 not aware of clients that still exhibit this bug, and the workaround
3961 hasn't been working properly for a while.
3965 * The return type of BIO_number_read() and BIO_number_written() as well as
3966 the corresponding num_read and num_write members in the BIO structure has
3967 changed from unsigned long to uint64_t. On platforms where an unsigned
3968 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3973 * Given the pervasive nature of TLS extensions it is inadvisable to run
3974 OpenSSL without support for them. It also means that maintaining
3975 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3976 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3980 * Removed support for the two export grade static DH ciphersuites
3981 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3982 were newly added (along with a number of other static DH ciphersuites) to
3983 1.0.2. However the two export ones have *never* worked since they were
3984 introduced. It seems strange in any case to be adding new export
3985 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3989 * Version negotiation has been rewritten. In particular SSLv23_method(),
3990 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3991 and turned into macros which simply call the new preferred function names
3992 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3993 should use the new names instead. Also as part of this change the ssl23.h
3994 header file has been removed.
3998 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3999 code and the associated standard is no longer considered fit-for-purpose.
4003 * RT2547 was closed. When generating a private key, try to make the
4004 output file readable only by the owner. This behavior change might
4005 be noticeable when interacting with other software.
4007 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4012 * Added HTTP GET support to the ocsp command.
4016 * Changed default digest for the dgst and enc commands from MD5 to
4021 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4025 * Added support for TLS extended master secret from
4026 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4027 initial patch which was a great help during development.
4031 * All libssl internal structures have been removed from the public header
4032 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4033 now redundant). Users should not attempt to access internal structures
4034 directly. Instead they should use the provided API functions.
4038 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4039 Access to deprecated functions can be re-enabled by running config with
4040 "enable-deprecated". In addition applications wishing to use deprecated
4041 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4042 will, by default, disable some transitive includes that previously existed
4043 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4047 * Added support for OCB mode. OpenSSL has been granted a patent license
4048 compatible with the OpenSSL license for use of OCB. Details are available
4049 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4050 for OCB can be removed by calling config with no-ocb.
4054 * SSLv2 support has been removed. It still supports receiving a SSLv2
4055 compatible client hello.
4059 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4060 done while fixing the error code for the key-too-small case.
4062 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4064 * CA.sh has been removed; use CA.pl instead.
4068 * Removed old DES API.
4072 * Remove various unsupported platforms:
4078 Sinix/ReliantUNIX RM400
4083 16-bit platforms such as WIN16
4087 * Clean up OPENSSL_NO_xxx #define's
4088 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4089 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4090 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4091 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4092 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4093 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4094 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4095 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4096 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4097 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4101 * Cleaned up dead code
4102 Remove all but one '#ifdef undef' which is to be looked at.
4106 * Clean up calling of xxx_free routines.
4107 Just like free(), fix most of the xxx_free routines to accept
4108 NULL. Remove the non-null checks from callers. Save much code.
4112 * Add secure heap for storage of private keys (when possible).
4113 Add BIO_s_secmem(), CBIGNUM, etc.
4114 Contributed by Akamai Technologies under our Corporate CLA.
4118 * Experimental support for a new, fast, unbiased prime candidate generator,
4119 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4121 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4123 * New output format NSS in the sess_id command line tool. This allows
4124 exporting the session id and the master key in NSS keylog format.
4126 *Martin Kaiser <martin@kaiser.cx>*
4128 * Harmonize version and its documentation. -f flag is used to display
4131 *mancha <mancha1@zoho.com>*
4133 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4134 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4136 *mancha <mancha1@zoho.com>*
4138 * Fix some double frees. These are not thought to be exploitable.
4140 *mancha <mancha1@zoho.com>*
4142 * A missing bounds check in the handling of the TLS heartbeat extension
4143 can be used to reveal up to 64k of memory to a connected client or
4146 Thanks for Neel Mehta of Google Security for discovering this bug and to
4147 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4148 preparing the fix ([CVE-2014-0160])
4150 *Adam Langley, Bodo Moeller*
4152 * Fix for the attack described in the paper "Recovering OpenSSL
4153 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4154 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4155 <http://eprint.iacr.org/2014/140>
4157 Thanks to Yuval Yarom and Naomi Benger for discovering this
4158 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4160 *Yuval Yarom and Naomi Benger*
4162 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4163 this fixes a limitation in previous versions of OpenSSL.
4167 * Experimental encrypt-then-mac support.
4169 Experimental support for encrypt then mac from
4170 draft-gutmann-tls-encrypt-then-mac-02.txt
4172 To enable it set the appropriate extension number (0x42 for the test
4173 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4175 For non-compliant peers (i.e. just about everything) this should have no
4178 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4182 * Add EVP support for key wrapping algorithms, to avoid problems with
4183 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4184 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4185 algorithms and include tests cases.
4189 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4194 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4195 MGF1 digest and OAEP label.
4199 * Make openssl verify return errors.
4201 *Chris Palmer <palmer@google.com> and Ben Laurie*
4203 * New function ASN1_TIME_diff to calculate the difference between two
4204 ASN1_TIME structures or one structure and the current time.
4208 * Update fips_test_suite to support multiple command line options. New
4209 test to induce all self test errors in sequence and check expected
4214 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4215 sign or verify all in one operation.
4219 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4220 test programs and fips_test_suite. Includes functionality to parse
4221 the minimal script output of fipsalgest.pl directly.
4225 * Add authorisation parameter to FIPS_module_mode_set().
4229 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4233 * Use separate DRBG fields for internal and external flags. New function
4234 FIPS_drbg_health_check() to perform on demand health checking. Add
4235 generation tests to fips_test_suite with reduced health check interval to
4236 demonstrate periodic health checking. Add "nodh" option to
4237 fips_test_suite to skip very slow DH test.
4241 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4246 * More extensive health check for DRBG checking many more failure modes.
4247 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4248 combination: call this in fips_test_suite.
4252 * Add support for canonical generation of DSA parameter 'g'. See
4255 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4256 POST to handle HMAC cases.
4260 * Add functions FIPS_module_version() and FIPS_module_version_text()
4261 to return numerical and string versions of the FIPS module number.
4265 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4266 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4267 outside the validated module in the FIPS capable OpenSSL.
4271 * Minor change to DRBG entropy callback semantics. In some cases
4272 there is no multiple of the block length between min_len and
4273 max_len. Allow the callback to return more than max_len bytes
4274 of entropy but discard any extra: it is the callback's responsibility
4275 to ensure that the extra data discarded does not impact the
4276 requested amount of entropy.
4280 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4281 information in FIPS186-3, SP800-57 and SP800-131A.
4285 * CCM support via EVP. Interface is very similar to GCM case except we
4286 must supply all data in one chunk (i.e. no update, final) and the
4287 message length must be supplied if AAD is used. Add algorithm test
4292 * Initial version of POST overhaul. Add POST callback to allow the status
4293 of POST to be monitored and/or failures induced. Modify fips_test_suite
4294 to use callback. Always run all selftests even if one fails.
4298 * XTS support including algorithm test driver in the fips_gcmtest program.
4299 Note: this does increase the maximum key length from 32 to 64 bytes but
4300 there should be no binary compatibility issues as existing applications
4301 will never use XTS mode.
4305 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4306 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4307 performs algorithm blocking for unapproved PRNG types. Also do not
4308 set PRNG type in FIPS_mode_set(): leave this to the application.
4309 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4310 the standard OpenSSL PRNG: set additional data to a date time vector.
4314 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4315 This shouldn't present any incompatibility problems because applications
4316 shouldn't be using these directly and any that are will need to rethink
4317 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4321 * Extensive self tests and health checking required by SP800-90 DRBG.
4322 Remove strength parameter from FIPS_drbg_instantiate and always
4323 instantiate at maximum supported strength.
4327 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4331 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4335 * New function DH_compute_key_padded() to compute a DH key and pad with
4336 leading zeroes if needed: this complies with SP800-56A et al.
4340 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4341 anything, incomplete, subject to change and largely untested at present.
4345 * Modify fipscanisteronly build option to only build the necessary object
4346 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4350 * Add experimental option FIPSSYMS to give all symbols in
4351 fipscanister.o and FIPS or fips prefix. This will avoid
4352 conflicts with future versions of OpenSSL. Add perl script
4353 util/fipsas.pl to preprocess assembly language source files
4354 and rename any affected symbols.
4358 * Add selftest checks and algorithm block of non-fips algorithms in
4359 FIPS mode. Remove DES2 from selftests.
4363 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4364 return internal method without any ENGINE dependencies. Add new
4365 tiny fips sign and verify functions.
4369 * New build option no-ec2m to disable characteristic 2 code.
4373 * New build option "fipscanisteronly". This only builds fipscanister.o
4374 and (currently) associated fips utilities. Uses the file Makefile.fips
4375 instead of Makefile.org as the prototype.
4379 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4380 Update fips_gcmtest to use IV generator.
4384 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4385 setting output buffer to NULL. The `*Final` function must be
4386 called although it will not retrieve any additional data. The tag
4387 can be set or retrieved with a ctrl. The IV length is by default 12
4388 bytes (96 bits) but can be set to an alternative value. If the IV
4389 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4394 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4395 underlying do_cipher function handles all cipher semantics itself
4396 including padding and finalisation. This is useful if (for example)
4397 an ENGINE cipher handles block padding itself. The behaviour of
4398 do_cipher is subtly changed if this flag is set: the return value
4399 is the number of characters written to the output buffer (zero is
4400 no longer an error code) or a negative error code. Also if the
4401 input buffer is NULL and length 0 finalisation should be performed.
4405 * If a candidate issuer certificate is already part of the constructed
4406 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4410 * Improve forward-security support: add functions
4412 void SSL_CTX_set_not_resumable_session_callback(
4413 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4414 void SSL_set_not_resumable_session_callback(
4415 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4417 for use by SSL/TLS servers; the callback function will be called whenever a
4418 new session is created, and gets to decide whether the session may be
4419 cached to make it resumable (return 0) or not (return 1). (As by the
4420 SSL/TLS protocol specifications, the session_id sent by the server will be
4421 empty to indicate that the session is not resumable; also, the server will
4422 not generate RFC 4507 (RFC 5077) session tickets.)
4424 A simple reasonable callback implementation is to return is_forward_secure.
4425 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4426 by the SSL/TLS server library, indicating whether it can provide forward
4429 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4431 * New -verify_name option in command line utilities to set verification
4436 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4437 Add CMAC pkey methods.
4441 * Experimental renegotiation in s_server -www mode. If the client
4442 browses /reneg connection is renegotiated. If /renegcert it is
4443 renegotiated requesting a certificate.
4447 * Add an "external" session cache for debugging purposes to s_server. This
4448 should help trace issues which normally are only apparent in deployed
4449 multi-process servers.
4453 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4454 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4455 BIO_set_cipher() and some obscure PEM functions were changed so they
4456 can now return an error. The RAND changes required a change to the
4457 RAND_METHOD structure.
4461 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4462 a gcc attribute to warn if the result of a function is ignored. This
4463 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4464 whose return value is often ignored.
4468 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4469 These allow SCTs (signed certificate timestamps) to be requested and
4470 validated when establishing a connection.
4472 *Rob Percival <robpercival@google.com>*
4477 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4479 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4480 used even when parsing explicit parameters, when loading a encoded key
4481 or calling `EC_GROUP_new_from_ecpkparameters()`/
4482 `EC_GROUP_new_from_ecparameters()`.
4483 This prevents bypass of security hardening and performance gains,
4484 especially for curves with specialized EC_METHODs.
4485 By default, if a key encoded with explicit parameters is loaded and later
4486 encoded, the output is still encoded with explicit parameters, even if
4487 internally a "named" EC_GROUP is used for computation.
4491 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4492 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4493 NULL. After this change, only the cofactor parameter can be NULL. It also
4494 does some minimal sanity checks on the passed order.
4499 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4500 An attack is simple, if the first CMS_recipientInfo is valid but the
4501 second CMS_recipientInfo is chosen ciphertext. If the second
4502 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4503 encryption key will be replaced by garbage, and the message cannot be
4504 decoded, but if the RSA decryption fails, the correct encryption key is
4505 used and the recipient will not notice the attack.
4506 As a work around for this potential attack the length of the decrypted
4507 key must be equal to the cipher default key length, in case the
4508 certifiate is not given and all recipientInfo are tried out.
4509 The old behaviour can be re-enabled in the CMS code by setting the
4510 CMS_DEBUG_DECRYPT flag.
4515 * Document issue with installation paths in diverse Windows builds
4517 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4518 binaries and run-time config file.
4523 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4525 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4526 This changes the size when using the `genpkey` command when no size is given.
4527 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4528 generation commands to use 2048 bits by default.
4532 * Add FIPS support for Android Arm 64-bit
4534 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4535 Module in Version 2.0.10. For some reason, the corresponding target
4536 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4537 built with FIPS support on Android Arm 64-bit. This omission has been
4540 *Matthias St. Pierre*
4542 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4544 * 0-byte record padding oracle
4546 If an application encounters a fatal protocol error and then calls
4547 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4548 then OpenSSL can respond differently to the calling application if a 0 byte
4549 record is received with invalid padding compared to if a 0 byte record is
4550 received with an invalid MAC. If the application then behaves differently
4551 based on that in a way that is detectable to the remote peer, then this
4552 amounts to a padding oracle that could be used to decrypt data.
4554 In order for this to be exploitable "non-stitched" ciphersuites must be in
4555 use. Stitched ciphersuites are optimised implementations of certain
4556 commonly used ciphersuites. Also the application must call SSL_shutdown()
4557 twice even if a protocol error has occurred (applications should not do
4558 this but some do anyway).
4560 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4561 Aviram, with additional investigation by Steven Collison and Andrew
4562 Hourselt. It was reported to OpenSSL on 10th December 2018.
4567 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4571 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4573 * Microarchitecture timing vulnerability in ECC scalar multiplication
4575 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4576 shown to be vulnerable to a microarchitecture timing side channel attack.
4577 An attacker with sufficient access to mount local timing attacks during
4578 ECDSA signature generation could recover the private key.
4580 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4581 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4587 * Timing vulnerability in DSA signature generation
4589 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4590 timing side channel attack. An attacker could use variations in the signing
4591 algorithm to recover the private key.
4593 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4598 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4599 Module, accidentally introduced while backporting security fixes from the
4600 development branch and hindering the use of ECC in FIPS mode.
4604 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4606 * Client DoS due to large DH parameter
4608 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4609 malicious server can send a very large prime value to the client. This will
4610 cause the client to spend an unreasonably long period of time generating a
4611 key for this prime resulting in a hang until the client has finished. This
4612 could be exploited in a Denial Of Service attack.
4614 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4619 * Cache timing vulnerability in RSA Key Generation
4621 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4622 a cache timing side channel attack. An attacker with sufficient access to
4623 mount cache timing attacks during the RSA key generation process could
4624 recover the private key.
4626 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4627 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4632 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4633 parameter is no longer accepted, as it leads to a corrupt table. NULL
4634 pem_str is reserved for alias entries only.
4638 * Revert blinding in ECDSA sign and instead make problematic addition
4639 length-invariant. Switch even to fixed-length Montgomery multiplication.
4643 * Change generating and checking of primes so that the error rate of not
4644 being prime depends on the intended use based on the size of the input.
4645 For larger primes this will result in more rounds of Miller-Rabin.
4646 The maximal error rate for primes with more than 1080 bits is lowered
4649 *Kurt Roeckx, Annie Yousar*
4651 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4655 * Add blinding to ECDSA and DSA signatures to protect against side channel
4656 attacks discovered by Keegan Ryan (NCC Group).
4660 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4661 now allow empty (zero character) pass phrases.
4665 * Certificate time validation (X509_cmp_time) enforces stricter
4666 compliance with RFC 5280. Fractional seconds and timezone offsets
4667 are no longer allowed.
4671 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4673 * Constructed ASN.1 types with a recursive definition could exceed the stack
4675 Constructed ASN.1 types with a recursive definition (such as can be found
4676 in PKCS7) could eventually exceed the stack given malicious input with
4677 excessive recursion. This could result in a Denial Of Service attack. There
4678 are no such structures used within SSL/TLS that come from untrusted sources
4679 so this is considered safe.
4681 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4687 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4689 * Read/write after SSL object in error state
4691 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4692 mechanism. The intent was that if a fatal error occurred during a handshake
4693 then OpenSSL would move into the error state and would immediately fail if
4694 you attempted to continue the handshake. This works as designed for the
4695 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4696 SSL_connect()), however due to a bug it does not work correctly if
4697 SSL_read() or SSL_write() is called directly. In that scenario, if the
4698 handshake fails then a fatal error will be returned in the initial function
4699 call. If SSL_read()/SSL_write() is subsequently called by the application
4700 for the same SSL object then it will succeed and the data is passed without
4701 being decrypted/encrypted directly from the SSL/TLS record layer.
4703 In order to exploit this issue an application bug would have to be present
4704 that resulted in a call to SSL_read()/SSL_write() being issued after having
4705 already received a fatal error.
4707 This issue was reported to OpenSSL by David Benjamin (Google).
4712 * rsaz_1024_mul_avx2 overflow bug on x86_64
4714 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4715 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4716 Analysis suggests that attacks against RSA and DSA as a result of this
4717 defect would be very difficult to perform and are not believed likely.
4718 Attacks against DH1024 are considered just feasible, because most of the
4719 work necessary to deduce information about a private key may be performed
4720 offline. The amount of resources required for such an attack would be
4721 significant. However, for an attack on TLS to be meaningful, the server
4722 would have to share the DH1024 private key among multiple clients, which is
4723 no longer an option since CVE-2016-0701.
4725 This only affects processors that support the AVX2 but not ADX extensions
4726 like Intel Haswell (4th generation).
4728 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4729 was originally found via the OSS-Fuzz project.
4734 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4736 * bn_sqrx8x_internal carry bug on x86_64
4738 There is a carry propagating bug in the x86_64 Montgomery squaring
4739 procedure. No EC algorithms are affected. Analysis suggests that attacks
4740 against RSA and DSA as a result of this defect would be very difficult to
4741 perform and are not believed likely. Attacks against DH are considered just
4742 feasible (although very difficult) because most of the work necessary to
4743 deduce information about a private key may be performed offline. The amount
4744 of resources required for such an attack would be very significant and
4745 likely only accessible to a limited number of attackers. An attacker would
4746 additionally need online access to an unpatched system using the target
4747 private key in a scenario with persistent DH parameters and a private
4748 key that is shared between multiple clients.
4750 This only affects processors that support the BMI1, BMI2 and ADX extensions
4751 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4753 This issue was reported to OpenSSL by the OSS-Fuzz project.
4758 * Malformed X.509 IPAddressFamily could cause OOB read
4760 If an X.509 certificate has a malformed IPAddressFamily extension,
4761 OpenSSL could do a one-byte buffer overread. The most likely result
4762 would be an erroneous display of the certificate in text format.
4764 This issue was reported to OpenSSL by the OSS-Fuzz project.
4768 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4770 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4771 platform rather than 'mingw'.
4775 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4777 * Truncated packet could crash via OOB read
4779 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4780 cipher is being used, then a truncated packet can cause that host to
4781 perform an out-of-bounds read, usually resulting in a crash.
4783 This issue was reported to OpenSSL by Robert Święcki of Google.
4788 * BN_mod_exp may produce incorrect results on x86_64
4790 There is a carry propagating bug in the x86_64 Montgomery squaring
4791 procedure. No EC algorithms are affected. Analysis suggests that attacks
4792 against RSA and DSA as a result of this defect would be very difficult to
4793 perform and are not believed likely. Attacks against DH are considered just
4794 feasible (although very difficult) because most of the work necessary to
4795 deduce information about a private key may be performed offline. The amount
4796 of resources required for such an attack would be very significant and
4797 likely only accessible to a limited number of attackers. An attacker would
4798 additionally need online access to an unpatched system using the target
4799 private key in a scenario with persistent DH parameters and a private
4800 key that is shared between multiple clients. For example this can occur by
4801 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4802 similar to CVE-2015-3193 but must be treated as a separate problem.
4804 This issue was reported to OpenSSL by the OSS-Fuzz project.
4809 * Montgomery multiplication may produce incorrect results
4811 There is a carry propagating bug in the Broadwell-specific Montgomery
4812 multiplication procedure that handles input lengths divisible by, but
4813 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4814 and DH private keys are impossible. This is because the subroutine in
4815 question is not used in operations with the private key itself and an input
4816 of the attacker's direct choice. Otherwise the bug can manifest itself as
4817 transient authentication and key negotiation failures or reproducible
4818 erroneous outcome of public-key operations with specially crafted input.
4819 Among EC algorithms only Brainpool P-512 curves are affected and one
4820 presumably can attack ECDH key negotiation. Impact was not analyzed in
4821 detail, because pre-requisites for attack are considered unlikely. Namely
4822 multiple clients have to choose the curve in question and the server has to
4823 share the private key among them, neither of which is default behaviour.
4824 Even then only clients that chose the curve will be affected.
4826 This issue was publicly reported as transient failures and was not
4827 initially recognized as a security issue. Thanks to Richard Morgan for
4828 providing reproducible case.
4833 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4834 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4835 prevent issues where no progress is being made and the peer continually
4836 sends unrecognised record types, using up resources processing them.
4840 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4842 * Missing CRL sanity check
4844 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4845 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4846 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4848 This issue only affects the OpenSSL 1.0.2i
4853 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4855 * OCSP Status Request extension unbounded memory growth
4857 A malicious client can send an excessively large OCSP Status Request
4858 extension. If that client continually requests renegotiation, sending a
4859 large OCSP Status Request extension each time, then there will be unbounded
4860 memory growth on the server. This will eventually lead to a Denial Of
4861 Service attack through memory exhaustion. Servers with a default
4862 configuration are vulnerable even if they do not support OCSP. Builds using
4863 the "no-ocsp" build time option are not affected.
4865 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4870 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4873 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4879 * OOB write in MDC2_Update()
4881 An overflow can occur in MDC2_Update() either if called directly or
4882 through the EVP_DigestUpdate() function using MDC2. If an attacker
4883 is able to supply very large amounts of input data after a previous
4884 call to EVP_EncryptUpdate() with a partial block then a length check
4885 can overflow resulting in a heap corruption.
4887 The amount of data needed is comparable to SIZE_MAX which is impractical
4890 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4895 * Malformed SHA512 ticket DoS
4897 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4898 DoS attack where a malformed ticket will result in an OOB read which will
4901 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4902 a custom server callback and ticket lookup mechanism.
4904 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4909 * OOB write in BN_bn2dec()
4911 The function BN_bn2dec() does not check the return value of BN_div_word().
4912 This can cause an OOB write if an application uses this function with an
4913 overly large BIGNUM. This could be a problem if an overly large certificate
4914 or CRL is printed out from an untrusted source. TLS is not affected because
4915 record limits will reject an oversized certificate before it is parsed.
4917 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4922 * OOB read in TS_OBJ_print_bio()
4924 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4925 the total length the OID text representation would use and not the amount
4926 of data written. This will result in OOB reads when large OIDs are
4929 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4934 * Pointer arithmetic undefined behaviour
4936 Avoid some undefined pointer arithmetic
4938 A common idiom in the codebase is to check limits in the following manner:
4941 Where "p" points to some malloc'd data of SIZE bytes and
4944 "len" here could be from some externally supplied data (e.g. from a TLS
4947 The rules of C pointer arithmetic are such that "p + len" is only well
4948 defined where len <= SIZE. Therefore the above idiom is actually
4949 undefined behaviour.
4951 For example this could cause problems if some malloc implementation
4952 provides an address for "p" such that "p + len" actually overflows for
4953 values of len that are too big and therefore p + len < limit.
4955 This issue was reported to OpenSSL by Guido Vranken
4960 * Constant time flag not preserved in DSA signing
4962 Operations in the DSA signing algorithm should run in constant time in
4963 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4964 implementation means that a non-constant time codepath is followed for
4965 certain operations. This has been demonstrated through a cache-timing
4966 attack to be sufficient for an attacker to recover the private DSA key.
4968 This issue was reported by César Pereida (Aalto University), Billy Brumley
4969 (Tampere University of Technology), and Yuval Yarom (The University of
4970 Adelaide and NICTA).
4975 * DTLS buffered message DoS
4977 In a DTLS connection where handshake messages are delivered out-of-order
4978 those messages that OpenSSL is not yet ready to process will be buffered
4979 for later use. Under certain circumstances, a flaw in the logic means that
4980 those messages do not get removed from the buffer even though the handshake
4981 has been completed. An attacker could force up to approx. 15 messages to
4982 remain in the buffer when they are no longer required. These messages will
4983 be cleared when the DTLS connection is closed. The default maximum size for
4984 a message is 100k. Therefore the attacker could force an additional 1500k
4985 to be consumed per connection. By opening many simulataneous connections an
4986 attacker could cause a DoS attack through memory exhaustion.
4988 This issue was reported to OpenSSL by Quan Luo.
4993 * DTLS replay protection DoS
4995 A flaw in the DTLS replay attack protection mechanism means that records
4996 that arrive for future epochs update the replay protection "window" before
4997 the MAC for the record has been validated. This could be exploited by an
4998 attacker by sending a record for the next epoch (which does not have to
4999 decrypt or have a valid MAC), with a very large sequence number. This means
5000 that all subsequent legitimate packets are dropped causing a denial of
5001 service for a specific DTLS connection.
5003 This issue was reported to OpenSSL by the OCAP audit team.
5008 * Certificate message OOB reads
5010 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5011 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5012 theoretical DoS risk but this has not been observed in practice on common
5015 The messages affected are client certificate, client certificate request
5016 and server certificate. As a result the attack can only be performed
5017 against a client or a server which enables client authentication.
5019 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5024 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5026 * Prevent padding oracle in AES-NI CBC MAC check
5028 A MITM attacker can use a padding oracle attack to decrypt traffic
5029 when the connection uses an AES CBC cipher and the server support
5032 This issue was introduced as part of the fix for Lucky 13 padding
5033 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5034 constant time by making sure that always the same bytes are read and
5035 compared against either the MAC or padding bytes. But it no longer
5036 checked that there was enough data to have both the MAC and padding
5039 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5043 * Fix EVP_EncodeUpdate overflow
5045 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5046 Base64 encoding of binary data. If an attacker is able to supply very large
5047 amounts of input data then a length check can overflow resulting in a heap
5050 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5051 the `PEM_write_bio*` family of functions. These are mainly used within the
5052 OpenSSL command line applications, so any application which processes data
5053 from an untrusted source and outputs it as a PEM file should be considered
5054 vulnerable to this issue. User applications that call these APIs directly
5055 with large amounts of untrusted data may also be vulnerable.
5057 This issue was reported by Guido Vranken.
5062 * Fix EVP_EncryptUpdate overflow
5064 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5065 is able to supply very large amounts of input data after a previous call to
5066 EVP_EncryptUpdate() with a partial block then a length check can overflow
5067 resulting in a heap corruption. Following an analysis of all OpenSSL
5068 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5069 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5070 the first called function after an EVP_EncryptInit(), and therefore that
5071 specific call must be safe. The second form is where the length passed to
5072 EVP_EncryptUpdate() can be seen from the code to be some small value and
5073 therefore there is no possibility of an overflow. Since all instances are
5074 one of these two forms, it is believed that there can be no overflows in
5075 internal code due to this problem. It should be noted that
5076 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5077 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5078 of these calls have also been analysed too and it is believed there are no
5079 instances in internal usage where an overflow could occur.
5081 This issue was reported by Guido Vranken.
5086 * Prevent ASN.1 BIO excessive memory allocation
5088 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5089 a short invalid encoding can cause allocation of large amounts of memory
5090 potentially consuming excessive resources or exhausting memory.
5092 Any application parsing untrusted data through d2i BIO functions is
5093 affected. The memory based functions such as d2i_X509() are *not* affected.
5094 Since the memory based functions are used by the TLS library, TLS
5095 applications are not affected.
5097 This issue was reported by Brian Carpenter.
5104 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5105 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5106 in arbitrary stack data being returned in the buffer.
5108 This issue was reported by Guido Vranken.
5113 * Modify behavior of ALPN to invoke callback after SNI/servername
5114 callback, such that updates to the SSL_CTX affect ALPN.
5118 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5123 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5124 methods are enabled and ssl2 is disabled the methods return NULL.
5128 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5130 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5131 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5132 provide any "EXPORT" or "LOW" strength ciphers.
5136 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5137 is by default disabled at build-time. Builds that are not configured with
5138 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5139 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5140 will need to explicitly call either of:
5142 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5144 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5146 as appropriate. Even if either of those is used, or the application
5147 explicitly uses the version-specific SSLv2_method() or its client and
5148 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5149 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5150 ciphers, and SSLv2 56-bit DES are no longer available.
5155 * Fix a double-free in DSA code
5157 A double free bug was discovered when OpenSSL parses malformed DSA private
5158 keys and could lead to a DoS attack or memory corruption for applications
5159 that receive DSA private keys from untrusted sources. This scenario is
5162 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5168 * Disable SRP fake user seed to address a server memory leak.
5170 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5172 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5173 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5174 was changed to ignore the "fake user" SRP seed, even if the seed
5177 Users should use SRP_VBASE_get1_by_user instead. Note that in
5178 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5179 also that even though configuring the SRP seed attempts to hide
5180 invalid usernames by continuing the handshake with fake
5181 credentials, this behaviour is not constant time and no strong
5182 guarantees are made that the handshake is indistinguishable from
5183 that of a valid user.
5188 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5190 In the BN_hex2bn function the number of hex digits is calculated using an
5191 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5192 large values of `i` this can result in `bn_expand` not allocating any
5193 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5194 field as NULL leading to a subsequent NULL ptr deref. For very large values
5195 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5196 In this case memory is allocated to the internal BIGNUM data field, but it
5197 is insufficiently sized leading to heap corruption. A similar issue exists
5198 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5199 is ever called by user applications with very large untrusted hex/dec data.
5200 This is anticipated to be a rare occurrence.
5202 All OpenSSL internal usage of these functions use data that is not expected
5203 to be untrusted, e.g. config file data or application command line
5204 arguments. If user developed applications generate config file data based
5205 on untrusted data then it is possible that this could also lead to security
5206 consequences. This is also anticipated to be rare.
5208 This issue was reported to OpenSSL by Guido Vranken.
5213 * Fix memory issues in `BIO_*printf` functions
5215 The internal `fmtstr` function used in processing a "%s" format string in
5216 the `BIO_*printf` functions could overflow while calculating the length of a
5217 string and cause an OOB read when printing very long strings.
5219 Additionally the internal `doapr_outch` function can attempt to write to an
5220 OOB memory location (at an offset from the NULL pointer) in the event of a
5221 memory allocation failure. In 1.0.2 and below this could be caused where
5222 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5223 could be in processing a very long "%s" format string. Memory leaks can
5226 The first issue may mask the second issue dependent on compiler behaviour.
5227 These problems could enable attacks where large amounts of untrusted data
5228 is passed to the `BIO_*printf` functions. If applications use these functions
5229 in this way then they could be vulnerable. OpenSSL itself uses these
5230 functions when printing out human-readable dumps of ASN.1 data. Therefore
5231 applications that print this data could be vulnerable if the data is from
5232 untrusted sources. OpenSSL command line applications could also be
5233 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5234 as command line arguments.
5236 Libssl is not considered directly vulnerable. Additionally certificates etc
5237 received via remote connections via libssl are also unlikely to be able to
5238 trigger these issues because of message size limits enforced within libssl.
5240 This issue was reported to OpenSSL Guido Vranken.
5245 * Side channel attack on modular exponentiation
5247 A side-channel attack was found which makes use of cache-bank conflicts on
5248 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5249 of RSA keys. The ability to exploit this issue is limited as it relies on
5250 an attacker who has control of code in a thread running on the same
5251 hyper-threaded core as the victim thread which is performing decryptions.
5253 This issue was reported to OpenSSL by Yuval Yarom, The University of
5254 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5255 Nadia Heninger, University of Pennsylvania with more information at
5256 <http://cachebleed.info>.
5261 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5262 if no keysize is specified with default_bits. This fixes an
5263 omission in an earlier change that changed all RSA/DSA key generation
5264 commands to use 2048 bits by default.
5268 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5270 * DH small subgroups
5272 Historically OpenSSL only ever generated DH parameters based on "safe"
5273 primes. More recently (in version 1.0.2) support was provided for
5274 generating X9.42 style parameter files such as those required for RFC 5114
5275 support. The primes used in such files may not be "safe". Where an
5276 application is using DH configured with parameters based on primes that are
5277 not "safe" then an attacker could use this fact to find a peer's private
5278 DH exponent. This attack requires that the attacker complete multiple
5279 handshakes in which the peer uses the same private DH exponent. For example
5280 this could be used to discover a TLS server's private DH exponent if it's
5281 reusing the private DH exponent or it's using a static DH ciphersuite.
5283 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5284 TLS. It is not on by default. If the option is not set then the server
5285 reuses the same private DH exponent for the life of the server process and
5286 would be vulnerable to this attack. It is believed that many popular
5287 applications do set this option and would therefore not be at risk.
5289 The fix for this issue adds an additional check where a "q" parameter is
5290 available (as is the case in X9.42 based parameters). This detects the
5291 only known attack, and is the only possible defense for static DH
5292 ciphersuites. This could have some performance impact.
5294 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5295 default and cannot be disabled. This could have some performance impact.
5297 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5302 * SSLv2 doesn't block disabled ciphers
5304 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5305 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5306 been disabled, provided that the SSLv2 protocol was not also disabled via
5309 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5310 and Sebastian Schinzel.
5315 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5317 * BN_mod_exp may produce incorrect results on x86_64
5319 There is a carry propagating bug in the x86_64 Montgomery squaring
5320 procedure. No EC algorithms are affected. Analysis suggests that attacks
5321 against RSA and DSA as a result of this defect would be very difficult to
5322 perform and are not believed likely. Attacks against DH are considered just
5323 feasible (although very difficult) because most of the work necessary to
5324 deduce information about a private key may be performed offline. The amount
5325 of resources required for such an attack would be very significant and
5326 likely only accessible to a limited number of attackers. An attacker would
5327 additionally need online access to an unpatched system using the target
5328 private key in a scenario with persistent DH parameters and a private
5329 key that is shared between multiple clients. For example this can occur by
5330 default in OpenSSL DHE based SSL/TLS ciphersuites.
5332 This issue was reported to OpenSSL by Hanno Böck.
5337 * Certificate verify crash with missing PSS parameter
5339 The signature verification routines will crash with a NULL pointer
5340 dereference if presented with an ASN.1 signature using the RSA PSS
5341 algorithm and absent mask generation function parameter. Since these
5342 routines are used to verify certificate signature algorithms this can be
5343 used to crash any certificate verification operation and exploited in a
5344 DoS attack. Any application which performs certificate verification is
5345 vulnerable including OpenSSL clients and servers which enable client
5348 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5353 * X509_ATTRIBUTE memory leak
5355 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5356 memory. This structure is used by the PKCS#7 and CMS routines so any
5357 application which reads PKCS#7 or CMS data from untrusted sources is
5358 affected. SSL/TLS is not affected.
5360 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5366 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5367 This changes the decoding behaviour for some invalid messages,
5368 though the change is mostly in the more lenient direction, and
5369 legacy behaviour is preserved as much as possible.
5373 * In DSA_generate_parameters_ex, if the provided seed is too short,
5376 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5378 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5380 * Alternate chains certificate forgery
5382 During certificate verification, OpenSSL will attempt to find an
5383 alternative certificate chain if the first attempt to build such a chain
5384 fails. An error in the implementation of this logic can mean that an
5385 attacker could cause certain checks on untrusted certificates to be
5386 bypassed, such as the CA flag, enabling them to use a valid leaf
5387 certificate to act as a CA and "issue" an invalid certificate.
5389 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5394 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5396 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5397 incompatibility in the handling of HMAC. The previous ABI has now been
5402 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5404 * Malformed ECParameters causes infinite loop
5406 When processing an ECParameters structure OpenSSL enters an infinite loop
5407 if the curve specified is over a specially malformed binary polynomial
5410 This can be used to perform denial of service against any
5411 system which processes public keys, certificate requests or
5412 certificates. This includes TLS clients and TLS servers with
5413 client authentication enabled.
5415 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5420 * Exploitable out-of-bounds read in X509_cmp_time
5422 X509_cmp_time does not properly check the length of the ASN1_TIME
5423 string and can read a few bytes out of bounds. In addition,
5424 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5427 An attacker can use this to craft malformed certificates and CRLs of
5428 various sizes and potentially cause a segmentation fault, resulting in
5429 a DoS on applications that verify certificates or CRLs. TLS clients
5430 that verify CRLs are affected. TLS clients and servers with client
5431 authentication enabled may be affected if they use custom verification
5434 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5435 independently by Hanno Böck.
5440 * PKCS7 crash with missing EnvelopedContent
5442 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5443 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5444 with missing content and trigger a NULL pointer dereference on parsing.
5446 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5447 structures from untrusted sources are affected. OpenSSL clients and
5448 servers are not affected.
5450 This issue was reported to OpenSSL by Michal Zalewski (Google).
5455 * CMS verify infinite loop with unknown hash function
5457 When verifying a signedData message the CMS code can enter an infinite loop
5458 if presented with an unknown hash function OID. This can be used to perform
5459 denial of service against any system which verifies signedData messages using
5461 This issue was reported to OpenSSL by Johannes Bauer.
5466 * Race condition handling NewSessionTicket
5468 If a NewSessionTicket is received by a multi-threaded client when attempting to
5469 reuse a previous ticket then a race condition can occur potentially leading to
5470 a double free of the ticket data.
5475 * Only support 256-bit or stronger elliptic curves with the
5476 'ecdh_auto' setting (server) or by default (client). Of supported
5477 curves, prefer P-256 (both).
5481 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5483 * ClientHello sigalgs DoS fix
5485 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5486 invalid signature algorithms extension a NULL pointer dereference will
5487 occur. This can be exploited in a DoS attack against the server.
5489 This issue was was reported to OpenSSL by David Ramos of Stanford
5493 *Stephen Henson and Matt Caswell*
5495 * Multiblock corrupted pointer fix
5497 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5498 feature only applies on 64 bit x86 architecture platforms that support AES
5499 NI instructions. A defect in the implementation of "multiblock" can cause
5500 OpenSSL's internal write buffer to become incorrectly set to NULL when
5501 using non-blocking IO. Typically, when the user application is using a
5502 socket BIO for writing, this will only result in a failed connection.
5503 However if some other BIO is used then it is likely that a segmentation
5504 fault will be triggered, thus enabling a potential DoS attack.
5506 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5511 * Segmentation fault in DTLSv1_listen fix
5513 The DTLSv1_listen function is intended to be stateless and processes the
5514 initial ClientHello from many peers. It is common for user code to loop
5515 over the call to DTLSv1_listen until a valid ClientHello is received with
5516 an associated cookie. A defect in the implementation of DTLSv1_listen means
5517 that state is preserved in the SSL object from one invocation to the next
5518 that can lead to a segmentation fault. Errors processing the initial
5519 ClientHello can trigger this scenario. An example of such an error could be
5520 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5523 This issue was reported to OpenSSL by Per Allansson.
5528 * Segmentation fault in ASN1_TYPE_cmp fix
5530 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5531 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5532 certificate signature algorithm consistency this can be used to crash any
5533 certificate verification operation and exploited in a DoS attack. Any
5534 application which performs certificate verification is vulnerable including
5535 OpenSSL clients and servers which enable client authentication.
5540 * Segmentation fault for invalid PSS parameters fix
5542 The signature verification routines will crash with a NULL pointer
5543 dereference if presented with an ASN.1 signature using the RSA PSS
5544 algorithm and invalid parameters. Since these routines are used to verify
5545 certificate signature algorithms this can be used to crash any
5546 certificate verification operation and exploited in a DoS attack. Any
5547 application which performs certificate verification is vulnerable including
5548 OpenSSL clients and servers which enable client authentication.
5550 This issue was was reported to OpenSSL by Brian Carpenter.
5555 * ASN.1 structure reuse memory corruption fix
5557 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5558 memory corruption via an invalid write. Such reuse is and has been
5559 strongly discouraged and is believed to be rare.
5561 Applications that parse structures containing CHOICE or ANY DEFINED BY
5562 components may be affected. Certificate parsing (d2i_X509 and related
5563 functions) are however not affected. OpenSSL clients and servers are
5569 * PKCS7 NULL pointer dereferences fix
5571 The PKCS#7 parsing code does not handle missing outer ContentInfo
5572 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5573 missing content and trigger a NULL pointer dereference on parsing.
5575 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5576 otherwise parse PKCS#7 structures from untrusted sources are
5577 affected. OpenSSL clients and servers are not affected.
5579 This issue was reported to OpenSSL by Michal Zalewski (Google).
5584 * DoS via reachable assert in SSLv2 servers fix
5586 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5587 servers that both support SSLv2 and enable export cipher suites by sending
5588 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5590 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5591 (OpenSSL development team).
5596 * Empty CKE with client auth and DHE fix
5598 If client auth is used then a server can seg fault in the event of a DHE
5599 ciphersuite being selected and a zero length ClientKeyExchange message
5600 being sent by the client. This could be exploited in a DoS attack.
5605 * Handshake with unseeded PRNG fix
5607 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5608 with an unseeded PRNG. The conditions are:
5609 - The client is on a platform where the PRNG has not been seeded
5610 automatically, and the user has not seeded manually
5611 - A protocol specific client method version has been used (i.e. not
5612 SSL_client_methodv23)
5613 - A ciphersuite is used that does not require additional random data from
5614 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5616 If the handshake succeeds then the client random that has been used will
5617 have been generated from a PRNG with insufficient entropy and therefore the
5618 output may be predictable.
5620 For example using the following command with an unseeded openssl will
5621 succeed on an unpatched platform:
5623 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5628 * Use After Free following d2i_ECPrivatekey error fix
5630 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5631 could cause a use after free condition. This, in turn, could cause a double
5632 free in several private key parsing functions (such as d2i_PrivateKey
5633 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5634 for applications that receive EC private keys from untrusted
5635 sources. This scenario is considered rare.
5637 This issue was discovered by the BoringSSL project and fixed in their
5643 * X509_to_X509_REQ NULL pointer deref fix
5645 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5646 the certificate key is invalid. This function is rarely used in practice.
5648 This issue was discovered by Brian Carpenter.
5653 * Removed the export ciphers from the DEFAULT ciphers
5657 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5659 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5660 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5661 So far those who have to target multiple platforms would compromise
5662 and argue that binary targeting say ARMv5 would still execute on
5663 ARMv8. "Universal" build resolves this compromise by providing
5664 near-optimal performance even on newer platforms.
5668 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5669 (other platforms pending).
5671 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5673 * Add support for the SignedCertificateTimestampList certificate and
5674 OCSP response extensions from RFC6962.
5678 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5679 for corner cases. (Certain input points at infinity could lead to
5680 bogus results, with non-infinity inputs mapped to infinity too.)
5684 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5685 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5686 common cases are optimized and there still is room for further
5687 improvements. Vector Permutation AES for Altivec is also added.
5691 * Add support for little-endian ppc64 Linux target.
5693 *Marcelo Cerri (IBM)*
5695 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5696 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5697 are optimized and there still is room for further improvements.
5698 Both 32- and 64-bit modes are supported.
5700 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5702 * Improved ARMv7 NEON support.
5706 * Support for SPARC Architecture 2011 crypto extensions, first
5707 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5708 SHA256/512, MD5, GHASH and modular exponentiation.
5710 *Andy Polyakov, David Miller*
5712 * Accelerated modular exponentiation for Intel processors, a.k.a.
5715 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5717 * Support for new and upcoming Intel processors, including AVX2,
5718 BMI and SHA ISA extensions. This includes additional "stitched"
5719 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5722 This work was sponsored by Intel Corp.
5726 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5727 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5728 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5732 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5733 this fixes a limitation in previous versions of OpenSSL.
5737 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5738 MGF1 digest and OAEP label.
5742 * Add EVP support for key wrapping algorithms, to avoid problems with
5743 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5744 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5745 algorithms and include tests cases.
5749 * Add functions to allocate and set the fields of an ECDSA_METHOD
5752 *Douglas E. Engert, Steve Henson*
5754 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5755 difference in days and seconds between two tm or ASN1_TIME structures.
5759 * Add -rev test option to s_server to just reverse order of characters
5760 received by client and send back to server. Also prints an abbreviated
5761 summary of the connection parameters.
5765 * New option -brief for s_client and s_server to print out a brief summary
5766 of connection parameters.
5770 * Add callbacks for arbitrary TLS extensions.
5772 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5774 * New option -crl_download in several openssl utilities to download CRLs
5775 from CRLDP extension in certificates.
5779 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5783 * New function X509_CRL_diff to generate a delta CRL from the difference
5784 of two full CRLs. Add support to "crl" utility.
5788 * New functions to set lookup_crls function and to retrieve
5789 X509_STORE from X509_STORE_CTX.
5793 * Print out deprecated issuer and subject unique ID fields in
5798 * Extend OCSP I/O functions so they can be used for simple general purpose
5799 HTTP as well as OCSP. New wrapper function which can be used to download
5800 CRLs using the OCSP API.
5804 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5808 * `SSL_CONF*` functions. These provide a common framework for application
5809 configuration using configuration files or command lines.
5813 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5814 message callback and prints the results. Needs compile time option
5815 "enable-ssl-trace". New options to s_client and s_server to enable
5820 * New ctrl and macro to retrieve supported points extensions.
5821 Print out extension in s_server and s_client.
5825 * New functions to retrieve certificate signature and signature
5830 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5835 * New Suite B modes for TLS code. These use and enforce the requirements
5836 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5837 only use Suite B curves. The Suite B modes can be set by using the
5838 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5842 * New chain verification flags for Suite B levels of security. Check
5843 algorithms are acceptable when flags are set in X509_verify_cert.
5847 * Make tls1_check_chain return a set of flags indicating checks passed
5848 by a certificate chain. Add additional tests to handle client
5849 certificates: checks for matching certificate type and issuer name
5854 * If an attempt is made to use a signature algorithm not in the peer
5855 preference list abort the handshake. If client has no suitable
5856 signature algorithms in response to a certificate request do not
5857 use the certificate.
5861 * If server EC tmp key is not in client preference list abort handshake.
5865 * Add support for certificate stores in CERT structure. This makes it
5866 possible to have different stores per SSL structure or one store in
5867 the parent SSL_CTX. Include distinct stores for certificate chain
5868 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5869 to build and store a certificate chain in CERT structure: returning
5870 an error if the chain cannot be built: this will allow applications
5871 to test if a chain is correctly configured.
5873 Note: if the CERT based stores are not set then the parent SSL_CTX
5874 store is used to retain compatibility with existing behaviour.
5878 * New function ssl_set_client_disabled to set a ciphersuite disabled
5879 mask based on the current session, check mask when sending client
5880 hello and checking the requested ciphersuite.
5884 * New ctrls to retrieve and set certificate types in a certificate
5885 request message. Print out received values in s_client. If certificate
5886 types is not set with custom values set sensible values based on
5887 supported signature algorithms.
5891 * Support for distinct client and server supported signature algorithms.
5895 * Add certificate callback. If set this is called whenever a certificate
5896 is required by client or server. An application can decide which
5897 certificate chain to present based on arbitrary criteria: for example
5898 supported signature algorithms. Add very simple example to s_server.
5899 This fixes many of the problems and restrictions of the existing client
5900 certificate callback: for example you can now clear an existing
5901 certificate and specify the whole chain.
5905 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5906 the certificate can be used for (if anything). Set valid_flags field
5907 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5908 to have similar checks in it.
5910 Add new "cert_flags" field to CERT structure and include a "strict mode".
5911 This enforces some TLS certificate requirements (such as only permitting
5912 certificate signature algorithms contained in the supported algorithms
5913 extension) which some implementations ignore: this option should be used
5914 with caution as it could cause interoperability issues.
5918 * Update and tidy signature algorithm extension processing. Work out
5919 shared signature algorithms based on preferences and peer algorithms
5920 and print them out in s_client and s_server. Abort handshake if no
5921 shared signature algorithms.
5925 * Add new functions to allow customised supported signature algorithms
5926 for SSL and SSL_CTX structures. Add options to s_client and s_server
5931 * New function SSL_certs_clear() to delete all references to certificates
5932 from an SSL structure. Before this once a certificate had been added
5933 it couldn't be removed.
5937 * Integrate hostname, email address and IP address checking with certificate
5938 verification. New verify options supporting checking in openssl utility.
5942 * Fixes and wildcard matching support to hostname and email checking
5943 functions. Add manual page.
5945 *Florian Weimer (Red Hat Product Security Team)*
5947 * New functions to check a hostname email or IP address against a
5948 certificate. Add options x509 utility to print results of checks against
5953 * Fix OCSP checking.
5955 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5957 * Initial experimental support for explicitly trusted non-root CAs.
5958 OpenSSL still tries to build a complete chain to a root but if an
5959 intermediate CA has a trust setting included that is used. The first
5960 setting is used: whether to trust (e.g., -addtrust option to the x509
5965 * Add -trusted_first option which attempts to find certificates in the
5966 trusted store even if an untrusted chain is also supplied.
5970 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5971 platform support for Linux and Android.
5975 * Support for linux-x32, ILP32 environment in x86_64 framework.
5979 * Experimental multi-implementation support for FIPS capable OpenSSL.
5980 When in FIPS mode the approved implementations are used as normal,
5981 when not in FIPS mode the internal unapproved versions are used instead.
5982 This means that the FIPS capable OpenSSL isn't forced to use the
5983 (often lower performance) FIPS implementations outside FIPS mode.
5987 * Transparently support X9.42 DH parameters when calling
5988 PEM_read_bio_DHparameters. This means existing applications can handle
5989 the new parameter format automatically.
5993 * Initial experimental support for X9.42 DH parameter format: mainly
5994 to support use of 'q' parameter for RFC5114 parameters.
5998 * Add DH parameters from RFC5114 including test data to dhtest.
6002 * Support for automatic EC temporary key parameter selection. If enabled
6003 the most preferred EC parameters are automatically used instead of
6004 hardcoded fixed parameters. Now a server just has to call:
6005 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6006 support ECDH and use the most appropriate parameters.
6010 * Enhance and tidy EC curve and point format TLS extension code. Use
6011 static structures instead of allocation if default values are used.
6012 New ctrls to set curves we wish to support and to retrieve shared curves.
6013 Print out shared curves in s_server. New options to s_server and s_client
6014 to set list of supported curves.
6018 * New ctrls to retrieve supported signature algorithms and
6019 supported curve values as an array of NIDs. Extend openssl utility
6020 to print out received values.
6024 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6025 between NIDs and the more common NIST names such as "P-256". Enhance
6026 ecparam utility and ECC method to recognise the NIST names for curves.
6030 * Enhance SSL/TLS certificate chain handling to support different
6031 chains for each certificate instead of one chain in the parent SSL_CTX.
6035 * Support for fixed DH ciphersuite client authentication: where both
6036 server and client use DH certificates with common parameters.
6040 * Support for fixed DH ciphersuites: those requiring DH server
6045 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6047 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6048 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6049 X509_CINF_get_signature were reverted post internal team review.
6054 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6056 * OCSP Status Request extension unbounded memory growth
6058 A malicious client can send an excessively large OCSP Status Request
6059 extension. If that client continually requests renegotiation, sending a
6060 large OCSP Status Request extension each time, then there will be unbounded
6061 memory growth on the server. This will eventually lead to a Denial Of
6062 Service attack through memory exhaustion. Servers with a default
6063 configuration are vulnerable even if they do not support OCSP. Builds using
6064 the "no-ocsp" build time option are not affected.
6066 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6071 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6074 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6080 * OOB write in MDC2_Update()
6082 An overflow can occur in MDC2_Update() either if called directly or
6083 through the EVP_DigestUpdate() function using MDC2. If an attacker
6084 is able to supply very large amounts of input data after a previous
6085 call to EVP_EncryptUpdate() with a partial block then a length check
6086 can overflow resulting in a heap corruption.
6088 The amount of data needed is comparable to SIZE_MAX which is impractical
6091 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6096 * Malformed SHA512 ticket DoS
6098 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6099 DoS attack where a malformed ticket will result in an OOB read which will
6102 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6103 a custom server callback and ticket lookup mechanism.
6105 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6110 * OOB write in BN_bn2dec()
6112 The function BN_bn2dec() does not check the return value of BN_div_word().
6113 This can cause an OOB write if an application uses this function with an
6114 overly large BIGNUM. This could be a problem if an overly large certificate
6115 or CRL is printed out from an untrusted source. TLS is not affected because
6116 record limits will reject an oversized certificate before it is parsed.
6118 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6123 * OOB read in TS_OBJ_print_bio()
6125 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6126 the total length the OID text representation would use and not the amount
6127 of data written. This will result in OOB reads when large OIDs are
6130 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6135 * Pointer arithmetic undefined behaviour
6137 Avoid some undefined pointer arithmetic
6139 A common idiom in the codebase is to check limits in the following manner:
6142 Where "p" points to some malloc'd data of SIZE bytes and
6145 "len" here could be from some externally supplied data (e.g. from a TLS
6148 The rules of C pointer arithmetic are such that "p + len" is only well
6149 defined where len <= SIZE. Therefore the above idiom is actually
6150 undefined behaviour.
6152 For example this could cause problems if some malloc implementation
6153 provides an address for "p" such that "p + len" actually overflows for
6154 values of len that are too big and therefore p + len < limit.
6156 This issue was reported to OpenSSL by Guido Vranken
6161 * Constant time flag not preserved in DSA signing
6163 Operations in the DSA signing algorithm should run in constant time in
6164 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6165 implementation means that a non-constant time codepath is followed for
6166 certain operations. This has been demonstrated through a cache-timing
6167 attack to be sufficient for an attacker to recover the private DSA key.
6169 This issue was reported by César Pereida (Aalto University), Billy Brumley
6170 (Tampere University of Technology), and Yuval Yarom (The University of
6171 Adelaide and NICTA).
6176 * DTLS buffered message DoS
6178 In a DTLS connection where handshake messages are delivered out-of-order
6179 those messages that OpenSSL is not yet ready to process will be buffered
6180 for later use. Under certain circumstances, a flaw in the logic means that
6181 those messages do not get removed from the buffer even though the handshake
6182 has been completed. An attacker could force up to approx. 15 messages to
6183 remain in the buffer when they are no longer required. These messages will
6184 be cleared when the DTLS connection is closed. The default maximum size for
6185 a message is 100k. Therefore the attacker could force an additional 1500k
6186 to be consumed per connection. By opening many simulataneous connections an
6187 attacker could cause a DoS attack through memory exhaustion.
6189 This issue was reported to OpenSSL by Quan Luo.
6194 * DTLS replay protection DoS
6196 A flaw in the DTLS replay attack protection mechanism means that records
6197 that arrive for future epochs update the replay protection "window" before
6198 the MAC for the record has been validated. This could be exploited by an
6199 attacker by sending a record for the next epoch (which does not have to
6200 decrypt or have a valid MAC), with a very large sequence number. This means
6201 that all subsequent legitimate packets are dropped causing a denial of
6202 service for a specific DTLS connection.
6204 This issue was reported to OpenSSL by the OCAP audit team.
6209 * Certificate message OOB reads
6211 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6212 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6213 theoretical DoS risk but this has not been observed in practice on common
6216 The messages affected are client certificate, client certificate request
6217 and server certificate. As a result the attack can only be performed
6218 against a client or a server which enables client authentication.
6220 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6225 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6227 * Prevent padding oracle in AES-NI CBC MAC check
6229 A MITM attacker can use a padding oracle attack to decrypt traffic
6230 when the connection uses an AES CBC cipher and the server support
6233 This issue was introduced as part of the fix for Lucky 13 padding
6234 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6235 constant time by making sure that always the same bytes are read and
6236 compared against either the MAC or padding bytes. But it no longer
6237 checked that there was enough data to have both the MAC and padding
6240 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6245 * Fix EVP_EncodeUpdate overflow
6247 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6248 Base64 encoding of binary data. If an attacker is able to supply very large
6249 amounts of input data then a length check can overflow resulting in a heap
6252 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6253 the `PEM_write_bio*` family of functions. These are mainly used within the
6254 OpenSSL command line applications, so any application which processes data
6255 from an untrusted source and outputs it as a PEM file should be considered
6256 vulnerable to this issue. User applications that call these APIs directly
6257 with large amounts of untrusted data may also be vulnerable.
6259 This issue was reported by Guido Vranken.
6264 * Fix EVP_EncryptUpdate overflow
6266 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6267 is able to supply very large amounts of input data after a previous call to
6268 EVP_EncryptUpdate() with a partial block then a length check can overflow
6269 resulting in a heap corruption. Following an analysis of all OpenSSL
6270 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6271 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6272 the first called function after an EVP_EncryptInit(), and therefore that
6273 specific call must be safe. The second form is where the length passed to
6274 EVP_EncryptUpdate() can be seen from the code to be some small value and
6275 therefore there is no possibility of an overflow. Since all instances are
6276 one of these two forms, it is believed that there can be no overflows in
6277 internal code due to this problem. It should be noted that
6278 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6279 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6280 of these calls have also been analysed too and it is believed there are no
6281 instances in internal usage where an overflow could occur.
6283 This issue was reported by Guido Vranken.
6288 * Prevent ASN.1 BIO excessive memory allocation
6290 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6291 a short invalid encoding can casuse allocation of large amounts of memory
6292 potentially consuming excessive resources or exhausting memory.
6294 Any application parsing untrusted data through d2i BIO functions is
6295 affected. The memory based functions such as d2i_X509() are *not* affected.
6296 Since the memory based functions are used by the TLS library, TLS
6297 applications are not affected.
6299 This issue was reported by Brian Carpenter.
6306 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6307 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6308 in arbitrary stack data being returned in the buffer.
6310 This issue was reported by Guido Vranken.
6315 * Modify behavior of ALPN to invoke callback after SNI/servername
6316 callback, such that updates to the SSL_CTX affect ALPN.
6320 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6325 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6326 methods are enabled and ssl2 is disabled the methods return NULL.
6330 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6332 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6333 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6334 provide any "EXPORT" or "LOW" strength ciphers.
6338 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6339 is by default disabled at build-time. Builds that are not configured with
6340 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6341 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6342 will need to explicitly call either of:
6344 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6346 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6348 as appropriate. Even if either of those is used, or the application
6349 explicitly uses the version-specific SSLv2_method() or its client and
6350 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6351 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6352 ciphers, and SSLv2 56-bit DES are no longer available.
6357 * Fix a double-free in DSA code
6359 A double free bug was discovered when OpenSSL parses malformed DSA private
6360 keys and could lead to a DoS attack or memory corruption for applications
6361 that receive DSA private keys from untrusted sources. This scenario is
6364 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6370 * Disable SRP fake user seed to address a server memory leak.
6372 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6374 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6375 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6376 was changed to ignore the "fake user" SRP seed, even if the seed
6379 Users should use SRP_VBASE_get1_by_user instead. Note that in
6380 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6381 also that even though configuring the SRP seed attempts to hide
6382 invalid usernames by continuing the handshake with fake
6383 credentials, this behaviour is not constant time and no strong
6384 guarantees are made that the handshake is indistinguishable from
6385 that of a valid user.
6390 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6392 In the BN_hex2bn function the number of hex digits is calculated using an
6393 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6394 large values of `i` this can result in `bn_expand` not allocating any
6395 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6396 field as NULL leading to a subsequent NULL ptr deref. For very large values
6397 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6398 In this case memory is allocated to the internal BIGNUM data field, but it
6399 is insufficiently sized leading to heap corruption. A similar issue exists
6400 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6401 is ever called by user applications with very large untrusted hex/dec data.
6402 This is anticipated to be a rare occurrence.
6404 All OpenSSL internal usage of these functions use data that is not expected
6405 to be untrusted, e.g. config file data or application command line
6406 arguments. If user developed applications generate config file data based
6407 on untrusted data then it is possible that this could also lead to security
6408 consequences. This is also anticipated to be rare.
6410 This issue was reported to OpenSSL by Guido Vranken.
6415 * Fix memory issues in `BIO_*printf` functions
6417 The internal `fmtstr` function used in processing a "%s" format string in
6418 the `BIO_*printf` functions could overflow while calculating the length of a
6419 string and cause an OOB read when printing very long strings.
6421 Additionally the internal `doapr_outch` function can attempt to write to an
6422 OOB memory location (at an offset from the NULL pointer) in the event of a
6423 memory allocation failure. In 1.0.2 and below this could be caused where
6424 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6425 could be in processing a very long "%s" format string. Memory leaks can
6428 The first issue may mask the second issue dependent on compiler behaviour.
6429 These problems could enable attacks where large amounts of untrusted data
6430 is passed to the `BIO_*printf` functions. If applications use these functions
6431 in this way then they could be vulnerable. OpenSSL itself uses these
6432 functions when printing out human-readable dumps of ASN.1 data. Therefore
6433 applications that print this data could be vulnerable if the data is from
6434 untrusted sources. OpenSSL command line applications could also be
6435 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6436 as command line arguments.
6438 Libssl is not considered directly vulnerable. Additionally certificates etc
6439 received via remote connections via libssl are also unlikely to be able to
6440 trigger these issues because of message size limits enforced within libssl.
6442 This issue was reported to OpenSSL Guido Vranken.
6447 * Side channel attack on modular exponentiation
6449 A side-channel attack was found which makes use of cache-bank conflicts on
6450 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6451 of RSA keys. The ability to exploit this issue is limited as it relies on
6452 an attacker who has control of code in a thread running on the same
6453 hyper-threaded core as the victim thread which is performing decryptions.
6455 This issue was reported to OpenSSL by Yuval Yarom, The University of
6456 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6457 Nadia Heninger, University of Pennsylvania with more information at
6458 <http://cachebleed.info>.
6463 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6464 if no keysize is specified with default_bits. This fixes an
6465 omission in an earlier change that changed all RSA/DSA key generation
6466 commands to use 2048 bits by default.
6470 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6472 * Protection for DH small subgroup attacks
6474 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6475 switched on by default and cannot be disabled. This could have some
6480 * SSLv2 doesn't block disabled ciphers
6482 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6483 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6484 been disabled, provided that the SSLv2 protocol was not also disabled via
6487 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6488 and Sebastian Schinzel.
6493 * Reject DH handshakes with parameters shorter than 1024 bits.
6497 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6499 * Certificate verify crash with missing PSS parameter
6501 The signature verification routines will crash with a NULL pointer
6502 dereference if presented with an ASN.1 signature using the RSA PSS
6503 algorithm and absent mask generation function parameter. Since these
6504 routines are used to verify certificate signature algorithms this can be
6505 used to crash any certificate verification operation and exploited in a
6506 DoS attack. Any application which performs certificate verification is
6507 vulnerable including OpenSSL clients and servers which enable client
6510 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6515 * X509_ATTRIBUTE memory leak
6517 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6518 memory. This structure is used by the PKCS#7 and CMS routines so any
6519 application which reads PKCS#7 or CMS data from untrusted sources is
6520 affected. SSL/TLS is not affected.
6522 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6528 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6529 This changes the decoding behaviour for some invalid messages,
6530 though the change is mostly in the more lenient direction, and
6531 legacy behaviour is preserved as much as possible.
6535 * In DSA_generate_parameters_ex, if the provided seed is too short,
6536 use a random seed, as already documented.
6538 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6540 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6542 * Alternate chains certificate forgery
6544 During certificate verfification, OpenSSL will attempt to find an
6545 alternative certificate chain if the first attempt to build such a chain
6546 fails. An error in the implementation of this logic can mean that an
6547 attacker could cause certain checks on untrusted certificates to be
6548 bypassed, such as the CA flag, enabling them to use a valid leaf
6549 certificate to act as a CA and "issue" an invalid certificate.
6551 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6557 * Race condition handling PSK identify hint
6559 If PSK identity hints are received by a multi-threaded client then
6560 the values are wrongly updated in the parent SSL_CTX structure. This can
6561 result in a race condition potentially leading to a double free of the
6567 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6569 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6570 incompatibility in the handling of HMAC. The previous ABI has now been
6573 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6575 * Malformed ECParameters causes infinite loop
6577 When processing an ECParameters structure OpenSSL enters an infinite loop
6578 if the curve specified is over a specially malformed binary polynomial
6581 This can be used to perform denial of service against any
6582 system which processes public keys, certificate requests or
6583 certificates. This includes TLS clients and TLS servers with
6584 client authentication enabled.
6586 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6591 * Exploitable out-of-bounds read in X509_cmp_time
6593 X509_cmp_time does not properly check the length of the ASN1_TIME
6594 string and can read a few bytes out of bounds. In addition,
6595 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6598 An attacker can use this to craft malformed certificates and CRLs of
6599 various sizes and potentially cause a segmentation fault, resulting in
6600 a DoS on applications that verify certificates or CRLs. TLS clients
6601 that verify CRLs are affected. TLS clients and servers with client
6602 authentication enabled may be affected if they use custom verification
6605 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6606 independently by Hanno Böck.
6611 * PKCS7 crash with missing EnvelopedContent
6613 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6614 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6615 with missing content and trigger a NULL pointer dereference on parsing.
6617 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6618 structures from untrusted sources are affected. OpenSSL clients and
6619 servers are not affected.
6621 This issue was reported to OpenSSL by Michal Zalewski (Google).
6626 * CMS verify infinite loop with unknown hash function
6628 When verifying a signedData message the CMS code can enter an infinite loop
6629 if presented with an unknown hash function OID. This can be used to perform
6630 denial of service against any system which verifies signedData messages using
6632 This issue was reported to OpenSSL by Johannes Bauer.
6637 * Race condition handling NewSessionTicket
6639 If a NewSessionTicket is received by a multi-threaded client when attempting to
6640 reuse a previous ticket then a race condition can occur potentially leading to
6641 a double free of the ticket data.
6646 * Reject DH handshakes with parameters shorter than 768 bits.
6648 *Kurt Roeckx and Emilia Kasper*
6650 * dhparam: generate 2048-bit parameters by default.
6652 *Kurt Roeckx and Emilia Kasper*
6654 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6656 * Segmentation fault in ASN1_TYPE_cmp fix
6658 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6659 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6660 certificate signature algorithm consistency this can be used to crash any
6661 certificate verification operation and exploited in a DoS attack. Any
6662 application which performs certificate verification is vulnerable including
6663 OpenSSL clients and servers which enable client authentication.
6668 * ASN.1 structure reuse memory corruption fix
6670 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6671 memory corruption via an invalid write. Such reuse is and has been
6672 strongly discouraged and is believed to be rare.
6674 Applications that parse structures containing CHOICE or ANY DEFINED BY
6675 components may be affected. Certificate parsing (d2i_X509 and related
6676 functions) are however not affected. OpenSSL clients and servers are
6682 * PKCS7 NULL pointer dereferences fix
6684 The PKCS#7 parsing code does not handle missing outer ContentInfo
6685 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6686 missing content and trigger a NULL pointer dereference on parsing.
6688 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6689 otherwise parse PKCS#7 structures from untrusted sources are
6690 affected. OpenSSL clients and servers are not affected.
6692 This issue was reported to OpenSSL by Michal Zalewski (Google).
6697 * DoS via reachable assert in SSLv2 servers fix
6699 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6700 servers that both support SSLv2 and enable export cipher suites by sending
6701 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6703 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6704 (OpenSSL development team).
6709 * Use After Free following d2i_ECPrivatekey error fix
6711 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6712 could cause a use after free condition. This, in turn, could cause a double
6713 free in several private key parsing functions (such as d2i_PrivateKey
6714 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6715 for applications that receive EC private keys from untrusted
6716 sources. This scenario is considered rare.
6718 This issue was discovered by the BoringSSL project and fixed in their
6724 * X509_to_X509_REQ NULL pointer deref fix
6726 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6727 the certificate key is invalid. This function is rarely used in practice.
6729 This issue was discovered by Brian Carpenter.
6734 * Removed the export ciphers from the DEFAULT ciphers
6738 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6740 * Build fixes for the Windows and OpenVMS platforms
6742 *Matt Caswell and Richard Levitte*
6744 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6746 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6747 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6748 dereference. This could lead to a Denial Of Service attack. Thanks to
6749 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6754 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6755 dtls1_buffer_record function under certain conditions. In particular this
6756 could occur if an attacker sent repeated DTLS records with the same
6757 sequence number but for the next epoch. The memory leak could be exploited
6758 by an attacker in a Denial of Service attack through memory exhaustion.
6759 Thanks to Chris Mueller for reporting this issue.
6764 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6765 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6766 method would be set to NULL which could later result in a NULL pointer
6767 dereference. Thanks to Frank Schmirler for reporting this issue.
6772 * Abort handshake if server key exchange message is omitted for ephemeral
6775 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6776 reporting this issue.
6781 * Remove non-export ephemeral RSA code on client and server. This code
6782 violated the TLS standard by allowing the use of temporary RSA keys in
6783 non-export ciphersuites and could be used by a server to effectively
6784 downgrade the RSA key length used to a value smaller than the server
6785 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6786 INRIA or reporting this issue.
6791 * Fixed issue where DH client certificates are accepted without verification.
6792 An OpenSSL server will accept a DH certificate for client authentication
6793 without the certificate verify message. This effectively allows a client to
6794 authenticate without the use of a private key. This only affects servers
6795 which trust a client certificate authority which issues certificates
6796 containing DH keys: these are extremely rare and hardly ever encountered.
6797 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6803 * Ensure that the session ID context of an SSL is updated when its
6804 SSL_CTX is updated via SSL_set_SSL_CTX.
6806 The session ID context is typically set from the parent SSL_CTX,
6807 and can vary with the CTX.
6811 * Fix various certificate fingerprint issues.
6813 By using non-DER or invalid encodings outside the signed portion of a
6814 certificate the fingerprint can be changed without breaking the signature.
6815 Although no details of the signed portion of the certificate can be changed
6816 this can cause problems with some applications: e.g. those using the
6817 certificate fingerprint for blacklists.
6819 1. Reject signatures with non zero unused bits.
6821 If the BIT STRING containing the signature has non zero unused bits reject
6822 the signature. All current signature algorithms require zero unused bits.
6824 2. Check certificate algorithm consistency.
6826 Check the AlgorithmIdentifier inside TBS matches the one in the
6827 certificate signature. NB: this will result in signature failure
6828 errors for some broken certificates.
6830 Thanks to Konrad Kraszewski from Google for reporting this issue.
6832 3. Check DSA/ECDSA signatures use DER.
6834 Re-encode DSA/ECDSA signatures and compare with the original received
6835 signature. Return an error if there is a mismatch.
6837 This will reject various cases including garbage after signature
6838 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6839 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6840 (negative or with leading zeroes).
6842 Further analysis was conducted and fixes were developed by Stephen Henson
6843 of the OpenSSL core team.
6849 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6850 results on some platforms, including x86_64. This bug occurs at random
6851 with a very low probability, and is not known to be exploitable in any
6852 way, though its exact impact is difficult to determine. Thanks to Pieter
6853 Wuille (Blockstream) who reported this issue and also suggested an initial
6854 fix. Further analysis was conducted by the OpenSSL development team and
6855 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6856 the OpenSSL core team.
6861 * Do not resume sessions on the server if the negotiated protocol
6862 version does not match the session's version. Resuming with a different
6863 version, while not strictly forbidden by the RFC, is of questionable
6864 sanity and breaks all known clients.
6866 *David Benjamin, Emilia Käsper*
6868 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6869 early CCS messages during renegotiation. (Note that because
6870 renegotiation is encrypted, this early CCS was not exploitable.)
6874 * Tighten client-side session ticket handling during renegotiation:
6875 ensure that the client only accepts a session ticket if the server sends
6876 the extension anew in the ServerHello. Previously, a TLS client would
6877 reuse the old extension state and thus accept a session ticket if one was
6878 announced in the initial ServerHello.
6880 Similarly, ensure that the client requires a session ticket if one
6881 was advertised in the ServerHello. Previously, a TLS client would
6882 ignore a missing NewSessionTicket message.
6886 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6890 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6891 sends a carefully crafted handshake message, to cause OpenSSL to fail
6892 to free up to 64k of memory causing a memory leak. This could be
6893 exploited in a Denial Of Service attack. This issue affects OpenSSL
6894 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6895 whether SRTP is used or configured. Implementations of OpenSSL that
6896 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6898 The fix was developed by the OpenSSL team.
6903 * Session Ticket Memory Leak.
6905 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6906 integrity of that ticket is first verified. In the event of a session
6907 ticket integrity check failing, OpenSSL will fail to free memory
6908 causing a memory leak. By sending a large number of invalid session
6909 tickets an attacker could exploit this issue in a Denial Of Service
6915 * Build option no-ssl3 is incomplete.
6917 When OpenSSL is configured with "no-ssl3" as a build option, servers
6918 could accept and complete a SSL 3.0 handshake, and clients could be
6919 configured to send them.
6922 *Akamai and the OpenSSL team*
6924 * Add support for TLS_FALLBACK_SCSV.
6925 Client applications doing fallback retries should call
6926 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6929 *Adam Langley, Bodo Moeller*
6931 * Add additional DigestInfo checks.
6933 Re-encode DigestInto in DER and check against the original when
6934 verifying RSA signature: this will reject any improperly encoded
6935 DigestInfo structures.
6937 Note: this is a precautionary measure and no attacks are currently known.
6941 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6943 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6944 SRP code can be overrun an internal buffer. Add sanity check that
6945 g, A, B < N to SRP code.
6947 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6948 Group for discovering this issue.
6953 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6954 TLS 1.0 instead of higher protocol versions when the ClientHello message
6955 is badly fragmented. This allows a man-in-the-middle attacker to force a
6956 downgrade to TLS 1.0 even if both the server and the client support a
6957 higher protocol version, by modifying the client's TLS records.
6959 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6960 researching this issue.
6965 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6966 to a denial of service attack. A malicious server can crash the client
6967 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6968 ciphersuite and sending carefully crafted handshake messages.
6970 Thanks to Felix Gröbert (Google) for discovering and researching this
6976 * By sending carefully crafted DTLS packets an attacker could cause openssl
6977 to leak memory. This can be exploited through a Denial of Service attack.
6978 Thanks to Adam Langley for discovering and researching this issue.
6983 * An attacker can force openssl to consume large amounts of memory whilst
6984 processing DTLS handshake messages. This can be exploited through a
6985 Denial of Service attack.
6986 Thanks to Adam Langley for discovering and researching this issue.
6991 * An attacker can force an error condition which causes openssl to crash
6992 whilst processing DTLS packets due to memory being freed twice. This
6993 can be exploited through a Denial of Service attack.
6994 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7000 * If a multithreaded client connects to a malicious server using a resumed
7001 session and the server sends an ec point format extension it could write
7002 up to 255 bytes to freed memory.
7004 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7010 * A malicious server can crash an OpenSSL client with a null pointer
7011 dereference (read) by specifying an SRP ciphersuite even though it was not
7012 properly negotiated with the client. This can be exploited through a
7013 Denial of Service attack.
7015 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7016 discovering and researching this issue.
7021 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7022 X509_name_oneline, X509_name_print_ex et al. to leak some information
7023 from the stack. Applications may be affected if they echo pretty printing
7024 output to the attacker.
7026 Thanks to Ivan Fratric (Google) for discovering this issue.
7029 *Emilia Käsper, and Steve Henson*
7031 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7032 for corner cases. (Certain input points at infinity could lead to
7033 bogus results, with non-infinity inputs mapped to infinity too.)
7037 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7039 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7040 handshake can force the use of weak keying material in OpenSSL
7041 SSL/TLS clients and servers.
7043 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7044 researching this issue. ([CVE-2014-0224])
7046 *KIKUCHI Masashi, Steve Henson*
7048 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7049 OpenSSL DTLS client the code can be made to recurse eventually crashing
7052 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7055 *Imre Rad, Steve Henson*
7057 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7058 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7059 client or server. This is potentially exploitable to run arbitrary
7060 code on a vulnerable client or server.
7062 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7064 *Jüri Aedla, Steve Henson*
7066 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7067 are subject to a denial of service attack.
7069 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7070 this issue. ([CVE-2014-3470])
7072 *Felix Gröbert, Ivan Fratric, Steve Henson*
7074 * Harmonize version and its documentation. -f flag is used to display
7077 *mancha <mancha1@zoho.com>*
7079 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7080 in i2d_ECPrivateKey.
7082 *mancha <mancha1@zoho.com>*
7084 * Fix some double frees. These are not thought to be exploitable.
7086 *mancha <mancha1@zoho.com>*
7088 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7090 * A missing bounds check in the handling of the TLS heartbeat extension
7091 can be used to reveal up to 64k of memory to a connected client or
7094 Thanks for Neel Mehta of Google Security for discovering this bug and to
7095 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7096 preparing the fix ([CVE-2014-0160])
7098 *Adam Langley, Bodo Moeller*
7100 * Fix for the attack described in the paper "Recovering OpenSSL
7101 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7102 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7103 <http://eprint.iacr.org/2014/140>
7105 Thanks to Yuval Yarom and Naomi Benger for discovering this
7106 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7108 *Yuval Yarom and Naomi Benger*
7110 * TLS pad extension: draft-agl-tls-padding-03
7112 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7113 TLS client Hello record length value would otherwise be > 255 and
7114 less that 512 pad with a dummy extension containing zeroes so it
7115 is at least 512 bytes long.
7117 *Adam Langley, Steve Henson*
7119 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7121 * Fix for TLS record tampering bug. A carefully crafted invalid
7122 handshake could crash OpenSSL with a NULL pointer exception.
7123 Thanks to Anton Johansson for reporting this issues.
7126 * Keep original DTLS digest and encryption contexts in retransmission
7127 structures so we can use the previous session parameters if they need
7128 to be resent. ([CVE-2013-6450])
7132 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7133 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7134 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7135 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7136 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7137 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7139 *Rob Stradling, Adam Langley*
7141 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7143 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7144 supporting platforms or when small records were transferred.
7146 *Andy Polyakov, Steve Henson*
7148 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7150 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7152 This addresses the flaw in CBC record processing discovered by
7153 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7154 at: <http://www.isg.rhul.ac.uk/tls/>
7156 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7157 Security Group at Royal Holloway, University of London
7158 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7159 Emilia Käsper for the initial patch.
7162 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7164 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7165 ciphersuites which can be exploited in a denial of service attack.
7166 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7167 and detecting this bug and to Wolfgang Ettlinger
7168 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7173 * Return an error when checking OCSP signatures when key is NULL.
7174 This fixes a DoS attack. ([CVE-2013-0166])
7178 * Make openssl verify return errors.
7180 *Chris Palmer <palmer@google.com> and Ben Laurie*
7182 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7183 the right response is stapled. Also change SSL_get_certificate()
7184 so it returns the certificate actually sent.
7185 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7187 *Rob Stradling <rob.stradling@comodo.com>*
7189 * Fix possible deadlock when decoding public keys.
7193 * Don't use TLS 1.0 record version number in initial client hello
7198 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7200 * Sanity check record length before skipping explicit IV in TLS
7201 1.2, 1.1 and DTLS to fix DoS attack.
7203 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7204 fuzzing as a service testing platform.
7209 * Initialise tkeylen properly when encrypting CMS messages.
7210 Thanks to Solar Designer of Openwall for reporting this issue.
7214 * In FIPS mode don't try to use composite ciphers as they are not
7219 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7221 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7222 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7223 mean any application compiled against OpenSSL 1.0.0 headers setting
7224 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7225 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7226 0x10000000L Any application which was previously compiled against
7227 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7228 will need to be recompiled as a result. Letting be results in
7229 inability to disable specifically TLS 1.1 and in client context,
7230 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7234 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7235 disable just protocol X, but all protocols above X *if* there are
7236 protocols *below* X still enabled. In more practical terms it means
7237 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7238 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7239 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7244 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7246 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7247 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7248 in CRYPTO_realloc_clean.
7250 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7251 issue and to Adam Langley <agl@chromium.org> for fixing it.
7254 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7256 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7260 * Workarounds for some broken servers that "hang" if a client hello
7261 record length exceeds 255 bytes.
7263 1. Do not use record version number > TLS 1.0 in initial client
7264 hello: some (but not all) hanging servers will now work.
7265 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7266 the number of ciphers sent in the client hello. This should be
7267 set to an even number, such as 50, for example by passing:
7268 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7269 Most broken servers should now work.
7270 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7271 TLS 1.2 client support entirely.
7275 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7279 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7281 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7282 STRING form instead of a DigestInfo.
7286 * The format used for MDC2 RSA signatures is inconsistent between EVP
7287 and the RSA_sign/RSA_verify functions. This was made more apparent when
7288 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7289 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7290 the correct format in RSA_verify so both forms transparently work.
7294 * Some servers which support TLS 1.0 can choke if we initially indicate
7295 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7296 encrypted premaster secret. As a workaround use the maximum permitted
7297 client version in client hello, this should keep such servers happy
7298 and still work with previous versions of OpenSSL.
7302 * Add support for TLS/DTLS heartbeats.
7304 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7306 * Add support for SCTP.
7308 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7310 * Improved PRNG seeding for VOS.
7312 *Paul Green <Paul.Green@stratus.com>*
7314 * Extensive assembler packs updates, most notably:
7316 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7317 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7318 - x86_64: bit-sliced AES implementation;
7319 - ARM: NEON support, contemporary platforms optimizations;
7320 - s390x: z196 support;
7321 - `*`: GHASH and GF(2^m) multiplication implementations;
7325 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7326 (removal of unnecessary code)
7328 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7330 * Add TLS key material exporter from RFC 5705.
7334 * Add DTLS-SRTP negotiation from RFC 5764.
7338 * Add Next Protocol Negotiation,
7339 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7340 disabled with a no-npn flag to config or Configure. Code donated
7343 *Adam Langley <agl@google.com> and Ben Laurie*
7345 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7346 NIST-P256, NIST-P521, with constant-time single point multiplication on
7347 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7348 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7349 Code made available under Apache License version 2.0.
7351 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7352 line to include this in your build of OpenSSL, and run "make depend" (or
7353 "make update"). This enables the following EC_METHODs:
7355 EC_GFp_nistp224_method()
7356 EC_GFp_nistp256_method()
7357 EC_GFp_nistp521_method()
7359 EC_GROUP_new_by_curve_name() will automatically use these (while
7360 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7363 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7365 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7366 all platforms. Move ssize_t definition from e_os.h to the public
7367 header file e_os2.h as it now appears in public header file cms.h
7371 * New -sigopt option to the ca, req and x509 utilities. Additional
7372 signature parameters can be passed using this option and in
7377 * Add RSA PSS signing function. This will generate and set the
7378 appropriate AlgorithmIdentifiers for PSS based on those in the
7379 corresponding EVP_MD_CTX structure. No application support yet.
7383 * Support for companion algorithm specific ASN1 signing routines.
7384 New function ASN1_item_sign_ctx() signs a pre-initialised
7385 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7386 the appropriate parameters.
7390 * Add new algorithm specific ASN1 verification initialisation function
7391 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7392 handling will be the same no matter what EVP_PKEY_METHOD is used.
7393 Add a PSS handler to support verification of PSS signatures: checked
7394 against a number of sample certificates.
7398 * Add signature printing for PSS. Add PSS OIDs.
7400 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7402 * Add algorithm specific signature printing. An individual ASN1 method
7403 can now print out signatures instead of the standard hex dump.
7405 More complex signatures (e.g. PSS) can print out more meaningful
7406 information. Include DSA version that prints out the signature
7411 * Password based recipient info support for CMS library: implementing
7416 * Split password based encryption into PBES2 and PBKDF2 functions. This
7417 neatly separates the code into cipher and PBE sections and is required
7418 for some algorithms that split PBES2 into separate pieces (such as
7419 password based CMS).
7423 * Session-handling fixes:
7424 - Fix handling of connections that are resuming with a session ID,
7425 but also support Session Tickets.
7426 - Fix a bug that suppressed issuing of a new ticket if the client
7427 presented a ticket with an expired session.
7428 - Try to set the ticket lifetime hint to something reasonable.
7429 - Make tickets shorter by excluding irrelevant information.
7430 - On the client side, don't ignore renewed tickets.
7432 *Adam Langley, Bodo Moeller (Google)*
7434 * Fix PSK session representation.
7438 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7440 This work was sponsored by Intel.
7444 * Add GCM support to TLS library. Some custom code is needed to split
7445 the IV between the fixed (from PRF) and explicit (from TLS record)
7446 portions. This adds all GCM ciphersuites supported by RFC5288 and
7447 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7448 add a special AESGCM string for GCM only.
7452 * Expand range of ctrls for AES GCM. Permit setting invocation
7453 field on decrypt and retrieval of invocation field only on encrypt.
7457 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7458 As required by RFC5289 these ciphersuites cannot be used if for
7459 versions of TLS earlier than 1.2.
7463 * For FIPS capable OpenSSL interpret a NULL default public key method
7464 as unset and return the appropriate default but do *not* set the default.
7465 This means we can return the appropriate method in applications that
7466 switch between FIPS and non-FIPS modes.
7470 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7471 ENGINE is used then we cannot handle that in the FIPS module so we
7472 keep original code iff non-FIPS operations are allowed.
7476 * Add -attime option to openssl utilities.
7478 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7480 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7484 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7485 FIPS EC methods unconditionally for now.
7489 * New build option no-ec2m to disable characteristic 2 code.
7493 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7494 all cases can be covered as some introduce binary incompatibilities.
7498 * Redirect RSA operations to FIPS module including keygen,
7499 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7503 * Add similar low-level API blocking to ciphers.
7507 * low-level digest APIs are not approved in FIPS mode: any attempt
7508 to use these will cause a fatal error. Applications that *really* want
7509 to use them can use the `private_*` version instead.
7513 * Redirect cipher operations to FIPS module for FIPS builds.
7517 * Redirect digest operations to FIPS module for FIPS builds.
7521 * Update build system to add "fips" flag which will link in fipscanister.o
7522 for static and shared library builds embedding a signature if needed.
7526 * Output TLS supported curves in preference order instead of numerical
7527 order. This is currently hardcoded for the highest order curves first.
7528 This should be configurable so applications can judge speed vs strength.
7532 * Add TLS v1.2 server support for client authentication.
7536 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7541 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7542 FIPS modules versions.
7546 * Add TLS v1.2 client side support for client authentication. Keep cache
7547 of handshake records longer as we don't know the hash algorithm to use
7548 until after the certificate request message is received.
7552 * Initial TLS v1.2 client support. Add a default signature algorithms
7553 extension including all the algorithms we support. Parse new signature
7554 format in client key exchange. Relax some ECC signing restrictions for
7555 TLS v1.2 as indicated in RFC5246.
7559 * Add server support for TLS v1.2 signature algorithms extension. Switch
7560 to new signature format when needed using client digest preference.
7561 All server ciphersuites should now work correctly in TLS v1.2. No client
7562 support yet and no support for client certificates.
7566 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7567 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7568 ciphersuites. At present only RSA key exchange ciphersuites work with
7569 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7570 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7571 and version checking.
7575 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7576 with this defined it will not be affected by any changes to ssl internal
7577 structures. Add several utility functions to allow openssl application
7578 to work with OPENSSL_NO_SSL_INTERN defined.
7582 * A long standing patch to add support for SRP from EdelWeb (Peter
7583 Sylvester and Christophe Renou) was integrated.
7584 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7585 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7588 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7592 * Permit abbreviated handshakes when renegotiating using the function
7593 SSL_renegotiate_abbreviated().
7595 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7597 * Add call to ENGINE_register_all_complete() to
7598 ENGINE_load_builtin_engines(), so some implementations get used
7599 automatically instead of needing explicit application support.
7603 * Add support for TLS key exporter as described in RFC5705.
7605 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7607 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7608 a few changes are required:
7610 Add SSL_OP_NO_TLSv1_1 flag.
7611 Add TLSv1_1 methods.
7612 Update version checking logic to handle version 1.1.
7613 Add explicit IV handling (ported from DTLS code).
7614 Add command line options to s_client/s_server.
7621 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7623 * X509_ATTRIBUTE memory leak
7625 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7626 memory. This structure is used by the PKCS#7 and CMS routines so any
7627 application which reads PKCS#7 or CMS data from untrusted sources is
7628 affected. SSL/TLS is not affected.
7630 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7636 * Race condition handling PSK identify hint
7638 If PSK identity hints are received by a multi-threaded client then
7639 the values are wrongly updated in the parent SSL_CTX structure. This can
7640 result in a race condition potentially leading to a double free of the
7646 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7648 * Malformed ECParameters causes infinite loop
7650 When processing an ECParameters structure OpenSSL enters an infinite loop
7651 if the curve specified is over a specially malformed binary polynomial
7654 This can be used to perform denial of service against any
7655 system which processes public keys, certificate requests or
7656 certificates. This includes TLS clients and TLS servers with
7657 client authentication enabled.
7659 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7664 * Exploitable out-of-bounds read in X509_cmp_time
7666 X509_cmp_time does not properly check the length of the ASN1_TIME
7667 string and can read a few bytes out of bounds. In addition,
7668 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7671 An attacker can use this to craft malformed certificates and CRLs of
7672 various sizes and potentially cause a segmentation fault, resulting in
7673 a DoS on applications that verify certificates or CRLs. TLS clients
7674 that verify CRLs are affected. TLS clients and servers with client
7675 authentication enabled may be affected if they use custom verification
7678 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7679 independently by Hanno Böck.
7684 * PKCS7 crash with missing EnvelopedContent
7686 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7687 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7688 with missing content and trigger a NULL pointer dereference on parsing.
7690 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7691 structures from untrusted sources are affected. OpenSSL clients and
7692 servers are not affected.
7694 This issue was reported to OpenSSL by Michal Zalewski (Google).
7699 * CMS verify infinite loop with unknown hash function
7701 When verifying a signedData message the CMS code can enter an infinite loop
7702 if presented with an unknown hash function OID. This can be used to perform
7703 denial of service against any system which verifies signedData messages using
7705 This issue was reported to OpenSSL by Johannes Bauer.
7710 * Race condition handling NewSessionTicket
7712 If a NewSessionTicket is received by a multi-threaded client when attempting to
7713 reuse a previous ticket then a race condition can occur potentially leading to
7714 a double free of the ticket data.
7719 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7721 * Segmentation fault in ASN1_TYPE_cmp fix
7723 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7724 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7725 certificate signature algorithm consistency this can be used to crash any
7726 certificate verification operation and exploited in a DoS attack. Any
7727 application which performs certificate verification is vulnerable including
7728 OpenSSL clients and servers which enable client authentication.
7733 * ASN.1 structure reuse memory corruption fix
7735 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7736 memory corruption via an invalid write. Such reuse is and has been
7737 strongly discouraged and is believed to be rare.
7739 Applications that parse structures containing CHOICE or ANY DEFINED BY
7740 components may be affected. Certificate parsing (d2i_X509 and related
7741 functions) are however not affected. OpenSSL clients and servers are
7747 * PKCS7 NULL pointer dereferences fix
7749 The PKCS#7 parsing code does not handle missing outer ContentInfo
7750 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7751 missing content and trigger a NULL pointer dereference on parsing.
7753 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7754 otherwise parse PKCS#7 structures from untrusted sources are
7755 affected. OpenSSL clients and servers are not affected.
7757 This issue was reported to OpenSSL by Michal Zalewski (Google).
7762 * DoS via reachable assert in SSLv2 servers fix
7764 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7765 servers that both support SSLv2 and enable export cipher suites by sending
7766 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7768 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7769 (OpenSSL development team).
7774 * Use After Free following d2i_ECPrivatekey error fix
7776 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7777 could cause a use after free condition. This, in turn, could cause a double
7778 free in several private key parsing functions (such as d2i_PrivateKey
7779 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7780 for applications that receive EC private keys from untrusted
7781 sources. This scenario is considered rare.
7783 This issue was discovered by the BoringSSL project and fixed in their
7789 * X509_to_X509_REQ NULL pointer deref fix
7791 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7792 the certificate key is invalid. This function is rarely used in practice.
7794 This issue was discovered by Brian Carpenter.
7799 * Removed the export ciphers from the DEFAULT ciphers
7803 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7805 * Build fixes for the Windows and OpenVMS platforms
7807 *Matt Caswell and Richard Levitte*
7809 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7811 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7812 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7813 dereference. This could lead to a Denial Of Service attack. Thanks to
7814 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7819 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7820 dtls1_buffer_record function under certain conditions. In particular this
7821 could occur if an attacker sent repeated DTLS records with the same
7822 sequence number but for the next epoch. The memory leak could be exploited
7823 by an attacker in a Denial of Service attack through memory exhaustion.
7824 Thanks to Chris Mueller for reporting this issue.
7829 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7830 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7831 method would be set to NULL which could later result in a NULL pointer
7832 dereference. Thanks to Frank Schmirler for reporting this issue.
7837 * Abort handshake if server key exchange message is omitted for ephemeral
7840 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7841 reporting this issue.
7846 * Remove non-export ephemeral RSA code on client and server. This code
7847 violated the TLS standard by allowing the use of temporary RSA keys in
7848 non-export ciphersuites and could be used by a server to effectively
7849 downgrade the RSA key length used to a value smaller than the server
7850 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7851 INRIA or reporting this issue.
7856 * Fixed issue where DH client certificates are accepted without verification.
7857 An OpenSSL server will accept a DH certificate for client authentication
7858 without the certificate verify message. This effectively allows a client to
7859 authenticate without the use of a private key. This only affects servers
7860 which trust a client certificate authority which issues certificates
7861 containing DH keys: these are extremely rare and hardly ever encountered.
7862 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7868 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7869 results on some platforms, including x86_64. This bug occurs at random
7870 with a very low probability, and is not known to be exploitable in any
7871 way, though its exact impact is difficult to determine. Thanks to Pieter
7872 Wuille (Blockstream) who reported this issue and also suggested an initial
7873 fix. Further analysis was conducted by the OpenSSL development team and
7874 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7875 the OpenSSL core team.
7880 * Fix various certificate fingerprint issues.
7882 By using non-DER or invalid encodings outside the signed portion of a
7883 certificate the fingerprint can be changed without breaking the signature.
7884 Although no details of the signed portion of the certificate can be changed
7885 this can cause problems with some applications: e.g. those using the
7886 certificate fingerprint for blacklists.
7888 1. Reject signatures with non zero unused bits.
7890 If the BIT STRING containing the signature has non zero unused bits reject
7891 the signature. All current signature algorithms require zero unused bits.
7893 2. Check certificate algorithm consistency.
7895 Check the AlgorithmIdentifier inside TBS matches the one in the
7896 certificate signature. NB: this will result in signature failure
7897 errors for some broken certificates.
7899 Thanks to Konrad Kraszewski from Google for reporting this issue.
7901 3. Check DSA/ECDSA signatures use DER.
7903 Reencode DSA/ECDSA signatures and compare with the original received
7904 signature. Return an error if there is a mismatch.
7906 This will reject various cases including garbage after signature
7907 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7908 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7909 (negative or with leading zeroes).
7911 Further analysis was conducted and fixes were developed by Stephen Henson
7912 of the OpenSSL core team.
7918 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7920 * Session Ticket Memory Leak.
7922 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7923 integrity of that ticket is first verified. In the event of a session
7924 ticket integrity check failing, OpenSSL will fail to free memory
7925 causing a memory leak. By sending a large number of invalid session
7926 tickets an attacker could exploit this issue in a Denial Of Service
7932 * Build option no-ssl3 is incomplete.
7934 When OpenSSL is configured with "no-ssl3" as a build option, servers
7935 could accept and complete a SSL 3.0 handshake, and clients could be
7936 configured to send them.
7939 *Akamai and the OpenSSL team*
7941 * Add support for TLS_FALLBACK_SCSV.
7942 Client applications doing fallback retries should call
7943 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7946 *Adam Langley, Bodo Moeller*
7948 * Add additional DigestInfo checks.
7950 Reencode DigestInto in DER and check against the original when
7951 verifying RSA signature: this will reject any improperly encoded
7952 DigestInfo structures.
7954 Note: this is a precautionary measure and no attacks are currently known.
7958 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7960 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7961 to a denial of service attack. A malicious server can crash the client
7962 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7963 ciphersuite and sending carefully crafted handshake messages.
7965 Thanks to Felix Gröbert (Google) for discovering and researching this
7971 * By sending carefully crafted DTLS packets an attacker could cause openssl
7972 to leak memory. This can be exploited through a Denial of Service attack.
7973 Thanks to Adam Langley for discovering and researching this issue.
7978 * An attacker can force openssl to consume large amounts of memory whilst
7979 processing DTLS handshake messages. This can be exploited through a
7980 Denial of Service attack.
7981 Thanks to Adam Langley for discovering and researching this issue.
7986 * An attacker can force an error condition which causes openssl to crash
7987 whilst processing DTLS packets due to memory being freed twice. This
7988 can be exploited through a Denial of Service attack.
7989 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7995 * If a multithreaded client connects to a malicious server using a resumed
7996 session and the server sends an ec point format extension it could write
7997 up to 255 bytes to freed memory.
7999 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8005 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8006 X509_name_oneline, X509_name_print_ex et al. to leak some information
8007 from the stack. Applications may be affected if they echo pretty printing
8008 output to the attacker.
8010 Thanks to Ivan Fratric (Google) for discovering this issue.
8013 *Emilia Käsper, and Steve Henson*
8015 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8016 for corner cases. (Certain input points at infinity could lead to
8017 bogus results, with non-infinity inputs mapped to infinity too.)
8021 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8023 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8024 handshake can force the use of weak keying material in OpenSSL
8025 SSL/TLS clients and servers.
8027 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8028 researching this issue. ([CVE-2014-0224])
8030 *KIKUCHI Masashi, Steve Henson*
8032 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8033 OpenSSL DTLS client the code can be made to recurse eventually crashing
8036 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8039 *Imre Rad, Steve Henson*
8041 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8042 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8043 client or server. This is potentially exploitable to run arbitrary
8044 code on a vulnerable client or server.
8046 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8048 *Jüri Aedla, Steve Henson*
8050 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8051 are subject to a denial of service attack.
8053 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8054 this issue. ([CVE-2014-3470])
8056 *Felix Gröbert, Ivan Fratric, Steve Henson*
8058 * Harmonize version and its documentation. -f flag is used to display
8061 *mancha <mancha1@zoho.com>*
8063 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8064 in i2d_ECPrivateKey.
8066 *mancha <mancha1@zoho.com>*
8068 * Fix some double frees. These are not thought to be exploitable.
8070 *mancha <mancha1@zoho.com>*
8072 * Fix for the attack described in the paper "Recovering OpenSSL
8073 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8074 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8075 <http://eprint.iacr.org/2014/140>
8077 Thanks to Yuval Yarom and Naomi Benger for discovering this
8078 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8080 *Yuval Yarom and Naomi Benger*
8082 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8084 * Keep original DTLS digest and encryption contexts in retransmission
8085 structures so we can use the previous session parameters if they need
8086 to be resent. ([CVE-2013-6450])
8090 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8091 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8092 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8093 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8094 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8095 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8097 *Rob Stradling, Adam Langley*
8099 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8101 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8103 This addresses the flaw in CBC record processing discovered by
8104 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8105 at: <http://www.isg.rhul.ac.uk/tls/>
8107 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8108 Security Group at Royal Holloway, University of London
8109 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8110 Emilia Käsper for the initial patch.
8113 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8115 * Return an error when checking OCSP signatures when key is NULL.
8116 This fixes a DoS attack. ([CVE-2013-0166])
8120 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8121 the right response is stapled. Also change SSL_get_certificate()
8122 so it returns the certificate actually sent.
8123 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8124 (This is a backport)
8126 *Rob Stradling <rob.stradling@comodo.com>*
8128 * Fix possible deadlock when decoding public keys.
8132 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8134 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8137 * Sanity check record length before skipping explicit IV in DTLS
8140 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8141 fuzzing as a service testing platform.
8146 * Initialise tkeylen properly when encrypting CMS messages.
8147 Thanks to Solar Designer of Openwall for reporting this issue.
8151 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8153 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8154 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8155 in CRYPTO_realloc_clean.
8157 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8158 issue and to Adam Langley <agl@chromium.org> for fixing it.
8161 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8163 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8165 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8166 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8167 content decryption and always return the same error. Note: this attack
8168 needs on average 2^20 messages so it only affects automated senders. The
8169 old behaviour can be re-enabled in the CMS code by setting the
8170 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8171 an MMA defence is not necessary.
8172 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8173 this issue. ([CVE-2012-0884])
8177 * Fix CVE-2011-4619: make sure we really are receiving a
8178 client hello before rejecting multiple SGC restarts. Thanks to
8179 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8183 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8185 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8186 Thanks to Antonio Martin, Enterprise Secure Access Research and
8187 Development, Cisco Systems, Inc. for discovering this bug and
8188 preparing a fix. ([CVE-2012-0050])
8192 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8194 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8195 of the Vaudenay padding oracle attack on CBC mode encryption
8196 which enables an efficient plaintext recovery attack against
8197 the OpenSSL implementation of DTLS. Their attack exploits timing
8198 differences arising during decryption processing. A research
8199 paper describing this attack can be found at:
8200 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8201 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8202 Security Group at Royal Holloway, University of London
8203 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8204 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8205 for preparing the fix. ([CVE-2011-4108])
8207 *Robin Seggelmann, Michael Tuexen*
8209 * Clear bytes used for block padding of SSL 3.0 records.
8212 *Adam Langley (Google)*
8214 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8215 Kadianakis <desnacked@gmail.com> for discovering this issue and
8216 Adam Langley for preparing the fix. ([CVE-2011-4619])
8218 *Adam Langley (Google)*
8220 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8222 *Andrey Kulikov <amdeich@gmail.com>*
8224 * Prevent malformed RFC3779 data triggering an assertion failure.
8225 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8226 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8228 *Rob Austein <sra@hactrn.net>*
8230 * Improved PRNG seeding for VOS.
8232 *Paul Green <Paul.Green@stratus.com>*
8234 * Fix ssl_ciph.c set-up race.
8236 *Adam Langley (Google)*
8238 * Fix spurious failures in ecdsatest.c.
8240 *Emilia Käsper (Google)*
8242 * Fix the BIO_f_buffer() implementation (which was mixing different
8243 interpretations of the `..._len` fields).
8245 *Adam Langley (Google)*
8247 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8248 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8249 threads won't reuse the same blinding coefficients.
8251 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8252 lock to call BN_BLINDING_invert_ex, and avoids one use of
8253 BN_BLINDING_update for each BN_BLINDING structure (previously,
8254 the last update always remained unused).
8256 *Emilia Käsper (Google)*
8258 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8260 *Bob Buckholz (Google)*
8262 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8264 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8265 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8267 *Kaspar Brand <ossl@velox.ch>*
8269 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8270 for multi-threaded use of ECDH. ([CVE-2011-3210])
8272 *Adam Langley (Google)*
8274 * Fix x509_name_ex_d2i memory leak on bad inputs.
8278 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8279 signature public key algorithm by using OID xref utilities instead.
8280 Before this you could only use some ECC ciphersuites with SHA1 only.
8284 * Add protection against ECDSA timing attacks as mentioned in the paper
8285 by Billy Bob Brumley and Nicola Tuveri, see:
8286 <http://eprint.iacr.org/2011/232.pdf>
8288 *Billy Bob Brumley and Nicola Tuveri*
8290 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8292 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8294 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8296 * Fix bug in string printing code: if *any* escaping is enabled we must
8297 escape the escape character (backslash) or the resulting string is
8302 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8304 * Disable code workaround for ancient and obsolete Netscape browsers
8305 and servers: an attacker can use it in a ciphersuite downgrade attack.
8306 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8310 * Fixed J-PAKE implementation error, originally discovered by
8311 Sebastien Martini, further info and confirmation from Stefan
8312 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8316 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8318 * Fix extension code to avoid race conditions which can result in a buffer
8319 overrun vulnerability: resumed sessions must not be modified as they can
8320 be shared by multiple threads. CVE-2010-3864
8324 * Fix WIN32 build system to correctly link an ENGINE directory into
8329 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8331 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8334 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8336 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8338 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8339 context. The operation can be customised via the ctrl mechanism in
8340 case ENGINEs want to include additional functionality.
8344 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8348 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8349 output hashes compatible with older versions of OpenSSL.
8351 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8353 * Fix compression algorithm handling: if resuming a session use the
8354 compression algorithm of the resumed session instead of determining
8355 it from client hello again. Don't allow server to change algorithm.
8359 * Add load_crls() function to commands tidying load_certs() too. Add option
8360 to verify utility to allow additional CRLs to be included.
8364 * Update OCSP request code to permit adding custom headers to the request:
8365 some responders need this.
8369 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8372 *Julia Lawall <julia@diku.dk>*
8374 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8375 needlessly dereferenced structures, used obsolete functions and
8376 didn't handle all updated verify codes correctly.
8380 * Disable MD2 in the default configuration.
8384 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8385 indicate the initial BIO being pushed or popped. This makes it possible
8386 to determine whether the BIO is the one explicitly called or as a result
8387 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8388 it handles reference counts correctly and doesn't zero out the I/O bio
8389 when it is not being explicitly popped. WARNING: applications which
8390 included workarounds for the old buggy behaviour will need to be modified
8391 or they could free up already freed BIOs.
8395 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8396 renaming to all platforms (within the 0.9.8 branch, this was
8397 done conditionally on Netware platforms to avoid a name clash).
8399 *Guenter <lists@gknw.net>*
8401 * Add ECDHE and PSK support to DTLS.
8403 *Michael Tuexen <tuexen@fh-muenster.de>*
8405 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8410 * Add "missing" function EVP_MD_flags() (without this the only way to
8411 retrieve a digest flags is by accessing the structure directly. Update
8412 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8413 or cipher is registered as in the "from" argument. Print out all
8414 registered digests in the dgst usage message instead of manually
8415 attempting to work them out.
8419 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8420 this allows the use of compression and extensions. Change default cipher
8421 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8422 by default unless an application cipher string requests it.
8426 * Alter match criteria in PKCS12_parse(). It used to try to use local
8427 key ids to find matching certificates and keys but some PKCS#12 files
8428 don't follow the (somewhat unwritten) rules and this strategy fails.
8429 Now just gather all certificates together and the first private key
8430 then look for the first certificate that matches the key.
8434 * Support use of registered digest and cipher names for dgst and cipher
8435 commands instead of having to add each one as a special case. So now
8442 openssl dgst -sha256 foo
8444 and this works for ENGINE based algorithms too.
8448 * Update Gost ENGINE to support parameter files.
8450 *Victor B. Wagner <vitus@cryptocom.ru>*
8452 * Support GeneralizedTime in ca utility.
8454 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8456 * Enhance the hash format used for certificate directory links. The new
8457 form uses the canonical encoding (meaning equivalent names will work
8458 even if they aren't identical) and uses SHA1 instead of MD5. This form
8459 is incompatible with the older format and as a result c_rehash should
8460 be used to rebuild symbolic links.
8464 * Make PKCS#8 the default write format for private keys, replacing the
8465 traditional format. This form is standardised, more secure and doesn't
8466 include an implicit MD5 dependency.
8470 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8471 committed to OpenSSL should pass this lot as a minimum.
8475 * Add session ticket override functionality for use by EAP-FAST.
8477 *Jouni Malinen <j@w1.fi>*
8479 * Modify HMAC functions to return a value. Since these can be implemented
8480 in an ENGINE errors can occur.
8484 * Type-checked OBJ_bsearch_ex.
8488 * Type-checked OBJ_bsearch. Also some constification necessitated
8489 by type-checking. Still to come: TXT_DB, bsearch(?),
8490 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8495 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8496 seconds to a tm structure directly, instead of going through OS
8497 specific date routines. This avoids any issues with OS routines such
8498 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8499 and X509_time_adj_ex() to cover the extended range. The existing
8500 X509_time_adj() is still usable and will no longer have any date issues.
8504 * Delta CRL support. New use deltas option which will attempt to locate
8505 and search any appropriate delta CRLs available.
8507 This work was sponsored by Google.
8511 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8512 code and add additional score elements. Validate alternate CRL paths
8513 as part of the CRL checking and indicate a new error "CRL path validation
8514 error" in this case. Applications wanting additional details can use
8515 the verify callback and check the new "parent" field. If this is not
8516 NULL CRL path validation is taking place. Existing applications won't
8517 see this because it requires extended CRL support which is off by
8520 This work was sponsored by Google.
8524 * Support for freshest CRL extension.
8526 This work was sponsored by Google.
8530 * Initial indirect CRL support. Currently only supported in the CRLs
8531 passed directly and not via lookup. Process certificate issuer
8532 CRL entry extension and lookup CRL entries by bother issuer name
8533 and serial number. Check and process CRL issuer entry in IDP extension.
8535 This work was sponsored by Google.
8539 * Add support for distinct certificate and CRL paths. The CRL issuer
8540 certificate is validated separately in this case. Only enabled if
8541 an extended CRL support flag is set: this flag will enable additional
8542 CRL functionality in future.
8544 This work was sponsored by Google.
8548 * Add support for policy mappings extension.
8550 This work was sponsored by Google.
8554 * Fixes to pathlength constraint, self issued certificate handling,
8555 policy processing to align with RFC3280 and PKITS tests.
8557 This work was sponsored by Google.
8561 * Support for name constraints certificate extension. DN, email, DNS
8562 and URI types are currently supported.
8564 This work was sponsored by Google.
8568 * To cater for systems that provide a pointer-based thread ID rather
8569 than numeric, deprecate the current numeric thread ID mechanism and
8570 replace it with a structure and associated callback type. This
8571 mechanism allows a numeric "hash" to be extracted from a thread ID in
8572 either case, and on platforms where pointers are larger than 'long',
8573 mixing is done to help ensure the numeric 'hash' is usable even if it
8574 can't be guaranteed unique. The default mechanism is to use "&errno"
8575 as a pointer-based thread ID to distinguish between threads.
8577 Applications that want to provide their own thread IDs should now use
8578 CRYPTO_THREADID_set_callback() to register a callback that will call
8579 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8581 Note that ERR_remove_state() is now deprecated, because it is tied
8582 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8583 to free the current thread's error state should be replaced by
8584 ERR_remove_thread_state(NULL).
8586 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8587 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8588 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8589 application was previously providing a numeric thread callback that
8590 was inappropriate for distinguishing threads, then uniqueness might
8591 have been obtained with &errno that happened immediately in the
8592 intermediate development versions of OpenSSL; this is no longer the
8593 case, the numeric thread callback will now override the automatic use
8596 *Geoff Thorpe, with help from Bodo Moeller*
8598 * Initial support for different CRL issuing certificates. This covers a
8599 simple case where the self issued certificates in the chain exist and
8600 the real CRL issuer is higher in the existing chain.
8602 This work was sponsored by Google.
8606 * Removed effectively defunct crypto/store from the build.
8610 * Revamp of STACK to provide stronger type-checking. Still to come:
8611 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8612 ASN1_STRING, CONF_VALUE.
8616 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8617 RAM on SSL connections. This option can save about 34k per idle SSL.
8621 * Revamp of LHASH to provide stronger type-checking. Still to come:
8622 STACK, TXT_DB, bsearch, qsort.
8626 * Initial support for Cryptographic Message Syntax (aka CMS) based
8627 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8628 support for data, signedData, compressedData, digestedData and
8629 encryptedData, envelopedData types included. Scripts to check against
8630 RFC4134 examples draft and interop and consistency checks of many
8631 content types and variants.
8635 * Add options to enc utility to support use of zlib compression BIO.
8639 * Extend mk1mf to support importing of options and assembly language
8640 files from Configure script, currently only included in VC-WIN32.
8641 The assembly language rules can now optionally generate the source
8642 files from the associated perl scripts.
8646 * Implement remaining functionality needed to support GOST ciphersuites.
8647 Interop testing has been performed using CryptoPro implementations.
8649 *Victor B. Wagner <vitus@cryptocom.ru>*
8651 * s390x assembler pack.
8655 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8660 * Implement Opaque PRF Input TLS extension as specified in
8661 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8662 official specification yet and no extension type assignment by
8663 IANA exists, this extension (for now) will have to be explicitly
8664 enabled when building OpenSSL by providing the extension number
8665 to use. For example, specify an option
8667 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8669 to the "config" or "Configure" script to enable the extension,
8670 assuming extension number 0x9527 (which is a completely arbitrary
8671 and unofficial assignment based on the MD5 hash of the Internet
8672 Draft). Note that by doing so, you potentially lose
8673 interoperability with other TLS implementations since these might
8674 be using the same extension number for other purposes.
8676 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8677 opaque PRF input value to use in the handshake. This will create
8678 an internal copy of the length-'len' string at 'src', and will
8679 return non-zero for success.
8681 To get more control and flexibility, provide a callback function
8684 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8685 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8689 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8692 Callback function 'cb' will be called in handshakes, and is
8693 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8694 Argument 'arg' is for application purposes (the value as given to
8695 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8696 be provided to the callback function). The callback function
8697 has to return non-zero to report success: usually 1 to use opaque
8698 PRF input just if possible, or 2 to enforce use of the opaque PRF
8699 input. In the latter case, the library will abort the handshake
8700 if opaque PRF input is not successfully negotiated.
8702 Arguments 'peerinput' and 'len' given to the callback function
8703 will always be NULL and 0 in the case of a client. A server will
8704 see the client's opaque PRF input through these variables if
8705 available (NULL and 0 otherwise). Note that if the server
8706 provides an opaque PRF input, the length must be the same as the
8707 length of the client's opaque PRF input.
8709 Note that the callback function will only be called when creating
8710 a new session (session resumption can resume whatever was
8711 previously negotiated), and will not be called in SSL 2.0
8712 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8713 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8714 for applications that need to enforce opaque PRF input.
8718 * Update ssl code to support digests other than SHA1+MD5 for handshake
8721 *Victor B. Wagner <vitus@cryptocom.ru>*
8723 * Add RFC4507 support to OpenSSL. This includes the corrections in
8724 RFC4507bis. The encrypted ticket format is an encrypted encoded
8725 SSL_SESSION structure, that way new session features are automatically
8728 If a client application caches session in an SSL_SESSION structure
8729 support is transparent because tickets are now stored in the encoded
8732 The SSL_CTX structure automatically generates keys for ticket
8733 protection in servers so again support should be possible
8734 with no application modification.
8736 If a client or server wishes to disable RFC4507 support then the option
8737 SSL_OP_NO_TICKET can be set.
8739 Add a TLS extension debugging callback to allow the contents of any client
8740 or server extensions to be examined.
8742 This work was sponsored by Google.
8746 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8747 OpenSSL should now compile cleanly on gcc 4.2
8749 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8751 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8752 support including streaming MAC support: this is required for GOST
8753 ciphersuite support.
8755 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8757 * Add option -stream to use PKCS#7 streaming in smime utility. New
8758 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8759 to output in BER and PEM format.
8763 * Experimental support for use of HMAC via EVP_PKEY interface. This
8764 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8765 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8766 ENGINE support for HMAC keys which are unextractable. New -mac and
8767 -macopt options to dgst utility.
8771 * New option -sigopt to dgst utility. Update dgst to use
8772 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8773 alternative signing parameters such as X9.31 or PSS in the dgst
8778 * Change ssl_cipher_apply_rule(), the internal function that does
8779 the work each time a ciphersuite string requests enabling
8780 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8781 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8782 the order of disabled ciphersuites such that those ciphersuites
8783 that most recently went from enabled to disabled not only stay
8784 in order with respect to each other, but also have higher priority
8785 than other disabled ciphersuites the next time ciphersuites are
8788 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8789 the same ciphersuites as with "HIGH" alone, but in a specific
8790 order where the PSK ciphersuites come first (since they are the
8791 most recently disabled ciphersuites when "HIGH" is parsed).
8793 Also, change ssl_create_cipher_list() (using this new
8794 functionality) such that between otherwise identical
8795 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8800 * Change ssl_create_cipher_list() so that it automatically
8801 arranges the ciphersuites in reasonable order before starting
8802 to process the rule string. Thus, the definition for "DEFAULT"
8803 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8804 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8805 This makes it much easier to arrive at a reasonable default order
8806 in applications for which anonymous ciphers are OK (meaning
8807 that you can't actually use DEFAULT).
8809 *Bodo Moeller; suggested by Victor Duchovni*
8811 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8812 processing) into multiple integers instead of setting
8813 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8814 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8815 (These masks as well as the individual bit definitions are hidden
8816 away into the non-exported interface ssl/ssl_locl.h, so this
8817 change to the definition of the SSL_CIPHER structure shouldn't
8818 affect applications.) This give us more bits for each of these
8819 categories, so there is no longer a need to coagulate AES128 and
8820 AES256 into a single algorithm bit, and to coagulate Camellia128
8821 and Camellia256 into a single algorithm bit, which has led to all
8824 Thus, among other things, the kludge introduced in 0.9.7m and
8825 0.9.8e for masking out AES256 independently of AES128 or masking
8826 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8828 With the change, we also introduce new ciphersuite aliases that
8829 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8834 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8835 Use the leftmost N bytes of the signature input if the input is
8836 larger than the prime q (with N being the size in bytes of q).
8840 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8841 it yet and it is largely untested.
8845 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8849 * Initial incomplete changes to avoid need for function casts in OpenSSL
8850 some compilers (gcc 4.2 and later) reject their use. Safestack is
8851 reimplemented. Update ASN1 to avoid use of legacy functions.
8855 * Win32/64 targets are linked with Winsock2.
8859 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8860 to external functions. This can be used to increase CRL handling
8861 efficiency especially when CRLs are very large by (for example) storing
8862 the CRL revoked certificates in a database.
8866 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8867 new CRLs added to a directory can be used. New command line option
8868 -verify_return_error to s_client and s_server. This causes real errors
8869 to be returned by the verify callback instead of carrying on no matter
8870 what. This reflects the way a "real world" verify callback would behave.
8874 * GOST engine, supporting several GOST algorithms and public key formats.
8875 Kindly donated by Cryptocom.
8879 * Partial support for Issuing Distribution Point CRL extension. CRLs
8880 partitioned by DP are handled but no indirect CRL or reason partitioning
8881 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8882 selected via a scoring technique which handles IDP and AKID in CRLs.
8886 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8887 will ultimately be used for all verify operations: this will remove the
8888 X509_STORE dependency on certificate verification and allow alternative
8889 lookup methods. X509_STORE based implementations of these two callbacks.
8893 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8894 Modify get_crl() to find a valid (unexpired) CRL if possible.
8898 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8899 this would be called X509_CRL_cmp() but that name is already used by
8900 a function that just compares CRL issuer names. Cache several CRL
8901 extensions in X509_CRL structure and cache CRLDP in X509.
8905 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8906 this maps equivalent X509_NAME structures into a consistent structure.
8907 Name comparison can then be performed rapidly using memcmp().
8911 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8916 * Allow digests to supply their own micalg string for S/MIME type using
8917 the ctrl EVP_MD_CTRL_MICALG.
8921 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8922 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8923 ctrl. It can then customise the structure before and/or after signing
8928 * New function OBJ_add_sigid() to allow application defined signature OIDs
8929 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8930 to free up any added signature OIDs.
8934 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8935 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8936 digest and cipher tables. New options added to openssl utility:
8937 list-message-digest-algorithms and list-cipher-algorithms.
8941 * Change the array representation of binary polynomials: the list
8942 of degrees of non-zero coefficients is now terminated with -1.
8943 Previously it was terminated with 0, which was also part of the
8944 value; thus, the array representation was not applicable to
8945 polynomials where t^0 has coefficient zero. This change makes
8946 the array representation useful in a more general context.
8950 * Various modifications and fixes to SSL/TLS cipher string
8951 handling. For ECC, the code now distinguishes between fixed ECDH
8952 with RSA certificates on the one hand and with ECDSA certificates
8953 on the other hand, since these are separate ciphersuites. The
8954 unused code for Fortezza ciphersuites has been removed.
8956 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8957 (not "ECDHE"). For consistency with the code for DH
8958 certificates, use of ECDH certificates is now considered ECDH
8959 authentication, not RSA or ECDSA authentication (the latter is
8960 merely the CA's signing algorithm and not actively used in the
8963 The temporary ciphersuite alias "ECCdraft" is no longer
8964 available, and ECC ciphersuites are no longer excluded from "ALL"
8965 and "DEFAULT". The following aliases now exist for RFC 4492
8966 ciphersuites, most of these by analogy with the DH case:
8968 kECDHr - ECDH cert, signed with RSA
8969 kECDHe - ECDH cert, signed with ECDSA
8970 kECDH - ECDH cert (signed with either RSA or ECDSA)
8971 kEECDH - ephemeral ECDH
8972 ECDH - ECDH cert or ephemeral ECDH
8978 AECDH - anonymous ECDH
8979 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8983 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8984 Use correct micalg parameters depending on digest(s) in signed message.
8988 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8989 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8993 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8994 an engine to register a method. Add ENGINE lookups for methods and
8995 functional reference processing.
8999 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9000 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9005 * New -resign option to smime utility. This adds one or more signers
9006 to an existing PKCS#7 signedData structure. Also -md option to use an
9007 alternative message digest algorithm for signing.
9011 * Tidy up PKCS#7 routines and add new functions to make it easier to
9012 create PKCS7 structures containing multiple signers. Update smime
9013 application to support multiple signers.
9017 * New -macalg option to pkcs12 utility to allow setting of an alternative
9022 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9023 Reorganize PBE internals to lookup from a static table using NIDs,
9024 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9025 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9026 PRF which will be automatically used with PBES2.
9030 * Replace the algorithm specific calls to generate keys in "req" with the
9035 * Update PKCS#7 enveloped data routines to use new API. This is now
9036 supported by any public key method supporting the encrypt operation. A
9037 ctrl is added to allow the public key algorithm to examine or modify
9038 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9043 * Add a ctrl to asn1 method to allow a public key algorithm to express
9044 a default digest type to use. In most cases this will be SHA1 but some
9045 algorithms (such as GOST) need to specify an alternative digest. The
9046 return value indicates how strong the preference is 1 means optional and
9047 2 is mandatory (that is it is the only supported type). Modify
9048 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9049 use the default md. Update openssl utilities to use the default digest
9050 type for signing if it is not explicitly indicated.
9054 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9055 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9056 signing method from the key type. This effectively removes the link
9057 between digests and public key types.
9061 * Add an OID cross reference table and utility functions. Its purpose is to
9062 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9063 rsaEncryption. This will allow some of the algorithm specific hackery
9064 needed to use the correct OID to be removed.
9068 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9069 structures for PKCS7_sign(). They are now set up by the relevant public
9074 * Add provisional EC pkey method with support for ECDSA and ECDH.
9078 * Add support for key derivation (agreement) in the API, DH method and
9083 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9084 public and private key formats. As a side effect these add additional
9085 command line functionality not previously available: DSA signatures can be
9086 generated and verified using pkeyutl and DH key support and generation in
9093 *Oliver Tappe <zooey@hirschkaefer.de>*
9095 * New make target "install_html_docs" installs HTML renditions of the
9098 *Oliver Tappe <zooey@hirschkaefer.de>*
9100 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9101 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9102 support key and parameter generation and add initial key generation
9103 functionality for RSA.
9107 * Add functions for main EVP_PKEY_method operations. The undocumented
9108 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9109 `EVP_PKEY_{encrypt,decrypt}_old`.
9113 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9114 key API, doesn't do much yet.
9118 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9119 public key algorithms. New option to openssl utility:
9120 "list-public-key-algorithms" to print out info.
9124 * Implement the Supported Elliptic Curves Extension for
9125 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9129 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9130 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9134 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9135 utilities such as rsa, dsa, dsaparam etc except they process any key
9140 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9141 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9142 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9147 * Initial support for pluggable public key ASN1.
9148 De-spaghettify the public key ASN1 handling. Move public and private
9149 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9150 algorithm specific handling to a single module within the relevant
9151 algorithm directory. Add functions to allow (near) opaque processing
9152 of public and private key structures.
9156 * Implement the Supported Point Formats Extension for
9157 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9161 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9162 for the psk identity [hint] and the psk callback functions to the
9163 SSL_SESSION, SSL and SSL_CTX structure.
9166 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9170 SSL_CTX_use_psk_identity_hint
9171 SSL_get_psk_identity_hint
9172 SSL_get_psk_identity
9173 SSL_use_psk_identity_hint
9175 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9177 * Add RFC 3161 compliant time stamp request creation, response generation
9178 and response verification functionality.
9180 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9182 * Add initial support for TLS extensions, specifically for the server_name
9183 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9184 have new members for a host name. The SSL data structure has an
9185 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9186 stored in that context to allow for session resumption, even after the
9187 SSL has been switched to a new SSL_CTX in reaction to a client's
9188 server_name extension.
9190 New functions (subject to change):
9192 SSL_get_servername()
9193 SSL_get_servername_type()
9196 New CTRL codes and macros (subject to change):
9198 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9199 - SSL_CTX_set_tlsext_servername_callback()
9200 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9201 - SSL_CTX_set_tlsext_servername_arg()
9202 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9204 openssl s_client has a new '-servername ...' option.
9206 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9207 '-key2 ...', '-servername_fatal' (subject to change). This allows
9208 testing the HostName extension for a specific single host name ('-cert'
9209 and '-key' remain fallbacks for handshakes without HostName
9210 negotiation). If the unrecognized_name alert has to be sent, this by
9211 default is a warning; it becomes fatal with the '-servername_fatal'
9214 *Peter Sylvester, Remy Allais, Christophe Renou*
9216 * Whirlpool hash implementation is added.
9220 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9221 bn(64,32). Because of instruction set limitations it doesn't have
9222 any negative impact on performance. This was done mostly in order
9223 to make it possible to share assembler modules, such as bn_mul_mont
9224 implementations, between 32- and 64-bit builds without hassle.
9228 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9229 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9234 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9235 dedicated Montgomery multiplication procedure, is introduced.
9236 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9237 "64-bit" performance on certain 32-bit targets.
9241 * New option SSL_OP_NO_COMP to disable use of compression selectively
9242 in SSL structures. New SSL ctrl to set maximum send fragment size.
9243 Save memory by setting the I/O buffer sizes dynamically instead of
9244 using the maximum available value.
9248 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9249 in addition to the text details.
9253 * Very, very preliminary EXPERIMENTAL support for printing of general
9254 ASN1 structures. This currently produces rather ugly output and doesn't
9255 handle several customised structures at all.
9259 * Integrated support for PVK file format and some related formats such
9260 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9261 these in the 'rsa' and 'dsa' utilities.
9265 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9269 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9270 place for the (very old) "NETSCAPE" format certificates which are now
9271 handled using new ASN1 code equivalents.
9275 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9276 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9277 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9281 * Modify CRL distribution points extension code to print out previously
9282 unsupported fields. Enhance extension setting code to allow setting of
9287 * Add print and set support for Issuing Distribution Point CRL extension.
9291 * Change 'Configure' script to enable Camellia by default.
9298 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9300 * When rejecting SSL/TLS records due to an incorrect version number, never
9301 update s->server with a new major version number. As of
9302 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9303 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9304 the previous behavior could result in a read attempt at NULL when
9305 receiving specific incorrect SSL/TLS records once record payload
9306 protection is active. ([CVE-2010-0740])
9308 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9310 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9311 could be crashed if the relevant tables were not present (e.g. chrooted).
9313 *Tomas Hoger <thoger@redhat.com>*
9315 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9317 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9319 *Martin Olsson, Neel Mehta*
9321 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9322 accommodate for stack sorting, always a write lock!).
9326 * On some versions of WIN32 Heap32Next is very slow. This can cause
9327 excessive delays in the RAND_poll(): over a minute. As a workaround
9328 include a time check in the inner Heap32Next loop too.
9332 * The code that handled flushing of data in SSL/TLS originally used the
9333 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9334 the problem outlined in PR#1949. The fix suggested there however can
9335 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9336 of Apache). So instead simplify the code to flush unconditionally.
9337 This should be fine since flushing with no data to flush is a no op.
9341 * Handle TLS versions 2.0 and later properly and correctly use the
9342 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9343 off ancient servers have a habit of sticking around for a while...
9347 * Modify compression code so it frees up structures without using the
9348 ex_data callbacks. This works around a problem where some applications
9349 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9350 restarting) then use compression (e.g. SSL with compression) later.
9351 This results in significant per-connection memory leaks and
9352 has caused some security issues including CVE-2008-1678 and
9357 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9358 change when encrypting or decrypting.
9362 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9363 connect and renegotiate with servers which do not support RI.
9364 Until RI is more widely deployed this option is enabled by default.
9368 * Add "missing" ssl ctrls to clear options and mode.
9372 * If client attempts to renegotiate and doesn't support RI respond with
9373 a no_renegotiation alert as required by RFC5746. Some renegotiating
9374 TLS clients will continue a connection gracefully when they receive
9375 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9376 waiting for a server hello which it will never receive. Now we treat a
9377 received no_renegotiation alert as a fatal error. This is because
9378 applications requesting a renegotiation might well expect it to succeed
9379 and would have no code in place to handle the server denying it so the
9380 only safe thing to do is to terminate the connection.
9384 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9385 peer supports secure renegotiation and 0 otherwise. Print out peer
9386 renegotiation support in s_client/s_server.
9390 * Replace the highly broken and deprecated SPKAC certification method with
9391 the updated NID creation version. This should correctly handle UTF8.
9395 * Implement RFC5746. Re-enable renegotiation but require the extension
9396 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9397 turns out to be a bad idea. It has been replaced by
9398 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9399 SSL_CTX_set_options(). This is really not recommended unless you
9400 know what you are doing.
9402 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9404 * Fixes to stateless session resumption handling. Use initial_ctx when
9405 issuing and attempting to decrypt tickets in case it has changed during
9406 servername handling. Use a non-zero length session ID when attempting
9407 stateless session resumption: this makes it possible to determine if
9408 a resumption has occurred immediately after receiving server hello
9409 (several places in OpenSSL subtly assume this) instead of later in
9414 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9415 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9416 fixes for a few places where the return code is not checked
9419 *Julia Lawall <julia@diku.dk>*
9421 * Add --strict-warnings option to Configure script to include devteam
9422 warnings in other configurations.
9426 * Add support for --libdir option and LIBDIR variable in makefiles. This
9427 makes it possible to install openssl libraries in locations which
9428 have names other than "lib", for example "/usr/lib64" which some
9431 *Steve Henson, based on patch from Jeremy Utley*
9433 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9434 X690 8.9.12 and can produce some misleading textual output of OIDs.
9436 *Steve Henson, reported by Dan Kaminsky*
9438 * Delete MD2 from algorithm tables. This follows the recommendation in
9439 several standards that it is not used in new applications due to
9440 several cryptographic weaknesses. For binary compatibility reasons
9441 the MD2 API is still compiled in by default.
9445 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9450 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9451 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9454 *Guenter <lists@gknw.net>*
9456 * Fix the server certificate chain building code to use X509_verify_cert(),
9457 it used to have an ad-hoc builder which was unable to cope with anything
9458 other than a simple chain.
9460 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9462 * Don't check self signed certificate signatures in X509_verify_cert()
9463 by default (a flag can override this): it just wastes time without
9464 adding any security. As a useful side effect self signed root CAs
9465 with non-FIPS digests are now usable in FIPS mode.
9469 * In dtls1_process_out_of_seq_message() the check if the current message
9470 is already buffered was missing. For every new message was memory
9471 allocated, allowing an attacker to perform an denial of service attack
9472 with sending out of seq handshake messages until there is no memory
9473 left. Additionally every future message was buffered, even if the
9474 sequence number made no sense and would be part of another handshake.
9475 So only messages with sequence numbers less than 10 in advance will be
9476 buffered. ([CVE-2009-1378])
9478 *Robin Seggelmann, discovered by Daniel Mentz*
9480 * Records are buffered if they arrive with a future epoch to be
9481 processed after finishing the corresponding handshake. There is
9482 currently no limitation to this buffer allowing an attacker to perform
9483 a DOS attack with sending records with future epochs until there is no
9484 memory left. This patch adds the pqueue_size() function to determine
9485 the size of a buffer and limits the record buffer to 100 entries.
9488 *Robin Seggelmann, discovered by Daniel Mentz*
9490 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9491 parent structure is freed. ([CVE-2009-1379])
9495 * Handle non-blocking I/O properly in SSL_shutdown() call.
9497 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9499 * Add `2.5.4.*` OIDs
9501 *Ilya O. <vrghost@gmail.com>*
9503 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9505 * Disable renegotiation completely - this fixes a severe security
9506 problem ([CVE-2009-3555]) at the cost of breaking all
9507 renegotiation. Renegotiation can be re-enabled by setting
9508 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9509 run-time. This is really not recommended unless you know what
9514 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9516 * Don't set val to NULL when freeing up structures, it is freed up by
9517 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9518 zeroing past the valid field. ([CVE-2009-0789])
9520 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9522 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9523 checked correctly. This would allow some invalid signed attributes to
9524 appear to verify correctly. ([CVE-2009-0591])
9526 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9528 * Reject UniversalString and BMPString types with invalid lengths. This
9529 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9530 a legal length. ([CVE-2009-0590])
9534 * Set S/MIME signing as the default purpose rather than setting it
9535 unconditionally. This allows applications to override it at the store
9540 * Permit restricted recursion of ASN1 strings. This is needed in practice
9541 to handle some structures.
9545 * Improve efficiency of mem_gets: don't search whole buffer each time
9548 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9550 * New -hex option for openssl rand.
9554 * Print out UTF8String and NumericString when parsing ASN1.
9558 * Support NumericString type for name components.
9562 * Allow CC in the environment to override the automatically chosen
9563 compiler. Note that nothing is done to ensure flags work with the
9568 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9570 * Properly check EVP_VerifyFinal() and similar return values
9573 *Ben Laurie, Bodo Moeller, Google Security Team*
9575 * Enable TLS extensions by default.
9579 * Allow the CHIL engine to be loaded, whether the application is
9580 multithreaded or not. (This does not release the developer from the
9581 obligation to set up the dynamic locking callbacks.)
9583 *Sander Temme <sander@temme.net>*
9585 * Use correct exit code if there is an error in dgst command.
9587 *Steve Henson; problem pointed out by Roland Dirlewanger*
9589 * Tweak Configure so that you need to say "experimental-jpake" to enable
9590 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9594 * Add experimental JPAKE support, including demo authentication in
9595 s_client and s_server.
9599 * Set the comparison function in v3_addr_canonize().
9601 *Rob Austein <sra@hactrn.net>*
9603 * Add support for XMPP STARTTLS in s_client.
9605 *Philip Paeps <philip@freebsd.org>*
9607 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9608 to ensure that even with this option, only ciphersuites in the
9609 server's preference list will be accepted. (Note that the option
9610 applies only when resuming a session, so the earlier behavior was
9611 just about the algorithm choice for symmetric cryptography.)
9615 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9617 * Fix NULL pointer dereference if a DTLS server received
9618 ChangeCipherSpec as first record ([CVE-2009-1386]).
9622 * Fix a state transition in s3_srvr.c and d1_srvr.c
9623 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9627 * The fix in 0.9.8c that supposedly got rid of unsafe
9628 double-checked locking was incomplete for RSA blinding,
9629 addressing just one layer of what turns out to have been
9630 doubly unsafe triple-checked locking.
9632 So now fix this for real by retiring the MONT_HELPER macro
9633 in crypto/rsa/rsa_eay.c.
9635 *Bodo Moeller; problem pointed out by Marius Schilder*
9637 * Various precautionary measures:
9639 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9641 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9642 (NB: This would require knowledge of the secret session ticket key
9643 to exploit, in which case you'd be SOL either way.)
9645 - Change bn_nist.c so that it will properly handle input BIGNUMs
9646 outside the expected range.
9648 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9651 *Neel Mehta, Bodo Moeller*
9653 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9654 the load fails. Useful for distros.
9656 *Ben Laurie and the FreeBSD team*
9658 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9662 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9666 * Expand ENGINE to support engine supplied SSL client certificate functions.
9668 This work was sponsored by Logica.
9672 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9673 keystores. Support for SSL/TLS client authentication too.
9674 Not compiled unless enable-capieng specified to Configure.
9676 This work was sponsored by Logica.
9680 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9681 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9682 attribute creation routines such as certificate requests and PKCS#12
9687 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9689 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9690 handshake which could lead to a client crash as found using the
9691 Codenomicon TLS test suite ([CVE-2008-1672])
9693 *Steve Henson, Mark Cox*
9695 * Fix double free in TLS server name extensions which could lead to
9696 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9700 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9702 Clear the error queue to ensure that error entries left from
9703 older function calls do not interfere with the correct operation.
9705 *Lutz Jaenicke, Erik de Castro Lopo*
9707 * Remove root CA certificates of commercial CAs:
9709 The OpenSSL project does not recommend any specific CA and does not
9710 have any policy with respect to including or excluding any CA.
9711 Therefore it does not make any sense to ship an arbitrary selection
9712 of root CA certificates with the OpenSSL software.
9716 * RSA OAEP patches to fix two separate invalid memory reads.
9717 The first one involves inputs when 'lzero' is greater than
9718 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9719 before the beginning of from). The second one involves inputs where
9720 the 'db' section contains nothing but zeroes (there is a one-byte
9721 invalid read after the end of 'db').
9723 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9725 * Partial backport from 0.9.9-dev:
9727 Introduce bn_mul_mont (dedicated Montgomery multiplication
9728 procedure) as a candidate for BIGNUM assembler implementation.
9729 While 0.9.9-dev uses assembler for various architectures, only
9730 x86_64 is available by default here in the 0.9.8 branch, and
9731 32-bit x86 is available through a compile-time setting.
9733 To try the 32-bit x86 assembler implementation, use Configure
9734 option "enable-montasm" (which exists only for this backport).
9736 As "enable-montasm" for 32-bit x86 disclaims code stability
9737 anyway, in this constellation we activate additional code
9738 backported from 0.9.9-dev for further performance improvements,
9739 namely BN_from_montgomery_word. (To enable this otherwise,
9740 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9742 *Andy Polyakov (backport partially by Bodo Moeller)*
9744 * Add TLS session ticket callback. This allows an application to set
9745 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9746 values. This is useful for key rollover for example where several key
9747 sets may exist with different names.
9751 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9752 This was broken until now in 0.9.8 releases, such that the only way
9753 a registered ENGINE could be used (assuming it initialises
9754 successfully on the host) was to explicitly set it as the default
9755 for the relevant algorithms. This is in contradiction with 0.9.7
9756 behaviour and the documentation. With this fix, when an ENGINE is
9757 registered into a given algorithm's table of implementations, the
9758 'uptodate' flag is reset so that auto-discovery will be used next
9759 time a new context for that algorithm attempts to select an
9762 *Ian Lister (tweaked by Geoff Thorpe)*
9764 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9765 implementation in the following ways:
9767 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9770 Lack of BER streaming support means one pass streaming processing is
9771 only supported if data is detached: setting the streaming flag is
9772 ignored for embedded content.
9774 CMS support is disabled by default and must be explicitly enabled
9775 with the enable-cms configuration option.
9779 * Update the GMP engine glue to do direct copies between BIGNUM and
9780 mpz_t when openssl and GMP use the same limb size. Otherwise the
9781 existing "conversion via a text string export" trick is still used.
9783 *Paul Sheer <paulsheer@gmail.com>*
9785 * Zlib compression BIO. This is a filter BIO which compressed and
9786 uncompresses any data passed through it.
9790 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9791 RFC3394 compatible AES key wrapping.
9795 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9796 sets string data without copying. X509_ALGOR_set0() and
9797 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9798 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9799 from an X509_ATTRIBUTE structure optionally checking it occurs only
9800 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9805 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9806 to get the expected BN_FLG_CONSTTIME behavior.
9808 *Bodo Moeller (Google)*
9812 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9813 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9814 - added some more tests to do_tests.pl
9815 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9816 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9817 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9818 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9819 - various changes to netware.pl to enable gcc-cross builds on Win32
9821 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9822 - various changes to fix missing prototype warnings
9823 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9824 - added AES, WHIRLPOOL and CPUID assembler code to build files
9825 - added missing AES assembler make rules to mk1mf.pl
9826 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9828 *Guenter Knauf <eflash@gmx.net>*
9830 * Implement certificate status request TLS extension defined in RFC3546.
9831 A client can set the appropriate parameters and receive the encoded
9832 OCSP response via a callback. A server can query the supplied parameters
9833 and set the encoded OCSP response in the callback. Add simplified examples
9834 to s_client and s_server.
9838 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9841 + Binary incompatibility of ssl_ctx_st structure
9842 + DTLS interoperation with non-compliant servers
9843 + Don't call get_session_cb() without proposed session
9844 + Fix ia64 assembler code
9846 *Andy Polyakov, Steve Henson*
9848 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9850 * DTLS Handshake overhaul. There were longstanding issues with
9851 OpenSSL DTLS implementation, which were making it impossible for
9852 RFC 4347 compliant client to communicate with OpenSSL server.
9853 Unfortunately just fixing these incompatibilities would "cut off"
9854 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9855 server keeps tolerating non RFC compliant syntax. The opposite is
9856 not true, 0.9.8f client can not communicate with earlier server.
9857 This update even addresses CVE-2007-4995.
9861 * Changes to avoid need for function casts in OpenSSL: some compilers
9862 (gcc 4.2 and later) reject their use.
9863 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9866 * Add RFC4507 support to OpenSSL. This includes the corrections in
9867 RFC4507bis. The encrypted ticket format is an encrypted encoded
9868 SSL_SESSION structure, that way new session features are automatically
9871 If a client application caches session in an SSL_SESSION structure
9872 support is transparent because tickets are now stored in the encoded
9875 The SSL_CTX structure automatically generates keys for ticket
9876 protection in servers so again support should be possible
9877 with no application modification.
9879 If a client or server wishes to disable RFC4507 support then the option
9880 SSL_OP_NO_TICKET can be set.
9882 Add a TLS extension debugging callback to allow the contents of any client
9883 or server extensions to be examined.
9885 This work was sponsored by Google.
9889 * Add initial support for TLS extensions, specifically for the server_name
9890 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9891 have new members for a host name. The SSL data structure has an
9892 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9893 stored in that context to allow for session resumption, even after the
9894 SSL has been switched to a new SSL_CTX in reaction to a client's
9895 server_name extension.
9897 New functions (subject to change):
9899 SSL_get_servername()
9900 SSL_get_servername_type()
9903 New CTRL codes and macros (subject to change):
9905 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9906 - SSL_CTX_set_tlsext_servername_callback()
9907 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9908 - SSL_CTX_set_tlsext_servername_arg()
9909 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9911 openssl s_client has a new '-servername ...' option.
9913 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9914 '-key2 ...', '-servername_fatal' (subject to change). This allows
9915 testing the HostName extension for a specific single host name ('-cert'
9916 and '-key' remain fallbacks for handshakes without HostName
9917 negotiation). If the unrecognized_name alert has to be sent, this by
9918 default is a warning; it becomes fatal with the '-servername_fatal'
9921 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9923 * Add AES and SSE2 assembly language support to VC++ build.
9927 * Mitigate attack on final subtraction in Montgomery reduction.
9931 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9932 (which previously caused an internal error).
9936 * Squeeze another 10% out of IGE mode when in != out.
9940 * AES IGE mode speedup.
9942 *Dean Gaudet (Google)*
9944 * Add the Korean symmetric 128-bit cipher SEED (see
9945 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9946 add SEED ciphersuites from RFC 4162:
9948 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9949 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9950 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9951 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9953 To minimize changes between patchlevels in the OpenSSL 0.9.8
9954 series, SEED remains excluded from compilation unless OpenSSL
9955 is configured with 'enable-seed'.
9957 *KISA, Bodo Moeller*
9959 * Mitigate branch prediction attacks, which can be practical if a
9960 single processor is shared, allowing a spy process to extract
9961 information. For detailed background information, see
9962 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9963 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9964 and Necessary Software Countermeasures"). The core of the change
9965 are new versions BN_div_no_branch() and
9966 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9967 respectively, which are slower, but avoid the security-relevant
9968 conditional branches. These are automatically called by BN_div()
9969 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9970 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9971 remove a conditional branch.
9973 BN_FLG_CONSTTIME is the new name for the previous
9974 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9975 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9976 in the exponent causes BN_mod_exp_mont() to use the alternative
9977 implementation in BN_mod_exp_mont_consttime().) The old name
9978 remains as a deprecated alias.
9980 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9981 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9982 constant-time implementations for more than just exponentiation.
9983 Here too the old name is kept as a deprecated alias.
9985 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9986 the BN_BLINDING structure gets an independent copy of the
9987 modulus. This means that the previous `BIGNUM *m` argument to
9988 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9989 essentially becomes `const BIGNUM *m`, although we can't actually
9990 change this in the header file before 0.9.9. It allows
9991 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9992 enable BN_FLG_CONSTTIME.
9994 *Matthew D Wood (Intel Corp)*
9996 * In the SSL/TLS server implementation, be strict about session ID
9997 context matching (which matters if an application uses a single
9998 external cache for different purposes). Previously,
9999 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10000 set. This did ensure strict client verification, but meant that,
10001 with applications using a single external cache for quite
10002 different requirements, clients could circumvent ciphersuite
10003 restrictions for a given session ID context by starting a session
10004 in a different context.
10008 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10009 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10010 authentication-only ciphersuites.
10014 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10015 not complete and could lead to a possible single byte overflow
10016 ([CVE-2007-5135]) [Ben Laurie]
10018 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10020 * Since AES128 and AES256 (and similarly Camellia128 and
10021 Camellia256) share a single mask bit in the logic of
10022 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10023 kludge to work properly if AES128 is available and AES256 isn't
10024 (or if Camellia128 is available and Camellia256 isn't).
10028 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10029 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10030 When a point or a seed is encoded in a BIT STRING, we need to
10031 prevent the removal of trailing zero bits to get the proper DER
10032 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10033 of a NamedBitList, for which trailing 0 bits need to be removed.)
10037 * Have SSL/TLS server implementation tolerate "mismatched" record
10038 protocol version while receiving ClientHello even if the
10039 ClientHello is fragmented. (The server can't insist on the
10040 particular protocol version it has chosen before the ServerHello
10041 message has informed the client about his choice.)
10045 * Add RFC 3779 support.
10047 *Rob Austein for ARIN, Ben Laurie*
10049 * Load error codes if they are not already present instead of using a
10050 static variable. This allows them to be cleanly unloaded and reloaded.
10051 Improve header file function name parsing.
10055 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10056 or CAPABILITY handshake as required by RFCs.
10058 *Goetz Babin-Ebell*
10060 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10062 * Introduce limits to prevent malicious keys being able to
10063 cause a denial of service. ([CVE-2006-2940])
10065 *Steve Henson, Bodo Moeller*
10067 * Fix ASN.1 parsing of certain invalid structures that can result
10068 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10070 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10071 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10073 * Fix SSL client code which could crash if connecting to a
10074 malicious SSLv2 server. ([CVE-2006-4343])
10076 *Tavis Ormandy and Will Drewry, Google Security Team*
10078 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10079 match only those. Before that, "AES256-SHA" would be interpreted
10080 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10081 the same strength classification in 0.9.7h) as we currently only
10082 have a single AES bit in the ciphersuite description bitmap.
10083 That change, however, also applied to ciphersuite strings such as
10084 "RC4-MD5" that intentionally matched multiple ciphersuites --
10085 namely, SSL 2.0 ciphersuites in addition to the more common ones
10086 from SSL 3.0/TLS 1.0.
10088 So we change the selection algorithm again: Naming an explicit
10089 ciphersuite selects this one ciphersuite, and any other similar
10090 ciphersuite (same bitmap) from *other* protocol versions.
10091 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10092 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10094 Since SSL 2.0 does not have any ciphersuites for which the
10095 128/256 bit distinction would be relevant, this works for now.
10096 The proper fix will be to use different bits for AES128 and
10097 AES256, which would have avoided the problems from the beginning;
10098 however, bits are scarce, so we can only do this in a new release
10099 (not just a patchlevel) when we can change the SSL_CIPHER
10100 definition to split the single 'unsigned long mask' bitmap into
10101 multiple values to extend the available space.
10105 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10107 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10108 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10110 * Add AES IGE and biIGE modes.
10114 * Change the Unix randomness entropy gathering to use poll() when
10115 possible instead of select(), since the latter has some
10116 undesirable limitations.
10118 *Darryl Miles via Richard Levitte and Bodo Moeller*
10120 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10121 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10122 cannot be implicitly activated as part of, e.g., the "AES" alias.
10123 However, please upgrade to OpenSSL 0.9.9[-dev] for
10124 non-experimental use of the ECC ciphersuites to get TLS extension
10125 support, which is required for curve and point format negotiation
10126 to avoid potential handshake problems.
10130 * Disable rogue ciphersuites:
10132 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10133 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10134 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10136 The latter two were purportedly from
10137 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10140 Also deactivate the remaining ciphersuites from
10141 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10142 unofficial, and the ID has long expired.
10146 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10147 dual-core machines) and other potential thread-safety issues.
10151 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10152 versions), which is now available for royalty-free use
10153 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10154 Also, add Camellia TLS ciphersuites from RFC 4132.
10156 To minimize changes between patchlevels in the OpenSSL 0.9.8
10157 series, Camellia remains excluded from compilation unless OpenSSL
10158 is configured with 'enable-camellia'.
10162 * Disable the padding bug check when compression is in use. The padding
10163 bug check assumes the first packet is of even length, this is not
10164 necessarily true if compression is enabled and can result in false
10165 positives causing handshake failure. The actual bug test is ancient
10166 code so it is hoped that implementations will either have fixed it by
10167 now or any which still have the bug do not support compression.
10171 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10173 * When applying a cipher rule check to see if string match is an explicit
10174 cipher suite and only match that one cipher suite if it is.
10178 * Link in manifests for VC++ if needed.
10180 *Austin Ziegler <halostatue@gmail.com>*
10182 * Update support for ECC-based TLS ciphersuites according to
10183 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10184 TLS extensions, which are supported starting with the 0.9.9
10185 branch, not in the OpenSSL 0.9.8 branch).
10189 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10190 opaque EVP_CIPHER_CTX handling.
10194 * Fixes and enhancements to zlib compression code. We now only use
10195 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10196 to conform with the standards mentioned here:
10197 <http://www.zlib.net/DLL_FAQ.txt>
10198 Static zlib linking now works on Windows and the new --with-zlib-include
10199 --with-zlib-lib options to Configure can be used to supply the location
10200 of the headers and library. Gracefully handle case where zlib library
10205 * Several fixes and enhancements to the OID generation code. The old code
10206 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10207 handle numbers larger than ULONG_MAX, truncated printing and had a
10208 non standard OBJ_obj2txt() behaviour.
10212 * Add support for building of engines under engine/ as shared libraries
10213 under VC++ build system.
10217 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10218 Hopefully, we will not see any false combination of paths any more.
10222 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10224 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10225 (part of SSL_OP_ALL). This option used to disable the
10226 countermeasure against man-in-the-middle protocol-version
10227 rollback in the SSL 2.0 server implementation, which is a bad
10228 idea. ([CVE-2005-2969])
10230 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10231 for Information Security, National Institute of Advanced Industrial
10232 Science and Technology [AIST], Japan)*
10234 * Add two function to clear and return the verify parameter flags.
10238 * Keep cipherlists sorted in the source instead of sorting them at
10239 runtime, thus removing the need for a lock.
10243 * Avoid some small subgroup attacks in Diffie-Hellman.
10245 *Nick Mathewson and Ben Laurie*
10247 * Add functions for well-known primes.
10251 * Extended Windows CE support.
10253 *Satoshi Nakamura and Andy Polyakov*
10255 * Initialize SSL_METHOD structures at compile time instead of during
10256 runtime, thus removing the need for a lock.
10260 * Make PKCS7_decrypt() work even if no certificate is supplied by
10261 attempting to decrypt each encrypted key in turn. Add support to
10266 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10268 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10271 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10275 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10276 key into the same file any more.
10280 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10284 * Add -utf8 command line and config file option to 'ca'.
10286 *Stefan <stf@udoma.org*
10288 * Removed the macro des_crypt(), as it seems to conflict with some
10289 libraries. Use DES_crypt().
10293 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10294 involves renaming the source and generated shared-libs for
10295 both. The engines will accept the corrected or legacy ids
10296 ('ncipher' and '4758_cca' respectively) when binding. NB,
10297 this only applies when building 'shared'.
10299 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10301 * Add attribute functions to EVP_PKEY structure. Modify
10302 PKCS12_create() to recognize a CSP name attribute and
10303 use it. Make -CSP option work again in pkcs12 utility.
10307 * Add new functionality to the bn blinding code:
10308 - automatic re-creation of the BN_BLINDING parameters after
10309 a fixed number of uses (currently 32)
10310 - add new function for parameter creation
10311 - introduce flags to control the update behaviour of the
10312 BN_BLINDING parameters
10313 - hide BN_BLINDING structure
10314 Add a second BN_BLINDING slot to the RSA structure to improve
10315 performance when a single RSA object is shared among several
10320 * Add support for DTLS.
10322 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10324 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10325 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10329 * Remove buggy and incomplete DH cert support from
10330 ssl/ssl_rsa.c and ssl/s3_both.c
10334 * Use SHA-1 instead of MD5 as the default digest algorithm for
10335 the `apps/openssl` commands.
10339 * Compile clean with "-Wall -Wmissing-prototypes
10340 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10341 DEBUG_SAFESTACK must also be set.
10345 * Change ./Configure so that certain algorithms can be disabled by default.
10346 The new counterpiece to "no-xxx" is "enable-xxx".
10348 The patented RC5 and MDC2 algorithms will now be disabled unless
10349 "enable-rc5" and "enable-mdc2", respectively, are specified.
10351 (IDEA remains enabled despite being patented. This is because IDEA
10352 is frequently required for interoperability, and there is no license
10353 fee for non-commercial use. As before, "no-idea" can be used to
10354 avoid this algorithm.)
10358 * Add processing of proxy certificates (see RFC 3820). This work was
10359 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10360 EGEE (Enabling Grids for E-science in Europe).
10364 * RC4 performance overhaul on modern architectures/implementations, such
10365 as Intel P4, IA-64 and AMD64.
10369 * New utility extract-section.pl. This can be used specify an alternative
10370 section number in a pod file instead of having to treat each file as
10371 a separate case in Makefile. This can be done by adding two lines to the
10374 =for comment openssl_section:XXX
10376 The blank line is mandatory.
10380 * New arguments -certform, -keyform and -pass for s_client and s_server
10381 to allow alternative format key and certificate files and passphrase
10386 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10387 update associated structures and add various utility functions.
10389 Add new policy related verify parameters, include policy checking in
10390 standard verify code. Enhance 'smime' application with extra parameters
10391 to support policy checking and print out.
10395 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10396 Nehemiah processors. These extensions support AES encryption in hardware
10397 as well as RNG (though RNG support is currently disabled).
10399 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10401 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10405 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10407 *Andy Polyakov and a number of other people*
10409 * Improved PowerPC platform support. Most notably BIGNUM assembler
10410 implementation contributed by IBM.
10412 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10414 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10415 exponent rather than 'unsigned long'. There is a corresponding change to
10416 the new 'rsa_keygen' element of the RSA_METHOD structure.
10418 *Jelte Jansen, Geoff Thorpe*
10420 * Functionality for creating the initial serial number file is now
10421 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10423 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10424 number file to 1, which is bound to cause problems. To avoid
10425 the problems while respecting compatibility between different 0.9.7
10426 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10427 CA.pl for serial number initialization. With the new release 0.9.8,
10428 we can fix the problem directly in the 'ca' utility.)
10432 * Reduced header interdependencies by declaring more opaque objects in
10433 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10434 give fewer recursive includes, which could break lazy source code - so
10435 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10436 developers should define this symbol when building and using openssl to
10437 ensure they track the recommended behaviour, interfaces, [etc], but
10438 backwards-compatible behaviour prevails when this isn't defined.
10442 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10446 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10447 This will generate a random key of the appropriate length based on the
10448 cipher context. The EVP_CIPHER can provide its own random key generation
10449 routine to support keys of a specific form. This is used in the des and
10450 3des routines to generate a key of the correct parity. Update S/MIME
10451 code to use new functions and hence generate correct parity DES keys.
10452 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10453 valid (weak or incorrect parity).
10457 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10458 as looking them up. This is useful when the verified structure may contain
10459 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10460 present unless the new PKCS7_NO_CRL flag is asserted.
10464 * Extend ASN1 oid configuration module. It now additionally accepts the
10467 shortName = some long name, 1.2.3.4
10471 * Reimplemented the BN_CTX implementation. There is now no more static
10472 limitation on the number of variables it can handle nor the depth of the
10473 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10474 information can now expand as required, and rather than having a single
10475 static array of bignums, BN_CTX now uses a linked-list of such arrays
10476 allowing it to expand on demand whilst maintaining the usefulness of
10477 BN_CTX's "bundling".
10481 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10482 to allow all RSA operations to function using a single BN_CTX.
10486 * Preliminary support for certificate policy evaluation and checking. This
10487 is initially intended to pass the tests outlined in "Conformance Testing
10488 of Relying Party Client Certificate Path Processing Logic" v1.07.
10492 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10493 remained unused and not that useful. A variety of other little bignum
10494 tweaks and fixes have also been made continuing on from the audit (see
10499 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10500 associated ASN1, EVP and SSL functions and old ASN1 macros.
10504 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10505 and this should never fail. So the return value from the use of
10506 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10507 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10511 * BN_CTX_get() should return zero-valued bignums, providing the same
10512 initialised value as BN_new().
10514 *Geoff Thorpe, suggested by Ulf Möller*
10516 * Support for inhibitAnyPolicy certificate extension.
10520 * An audit of the BIGNUM code is underway, for which debugging code is
10521 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10522 is considered valid when processing BIGNUMs, and causes execution to
10523 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10524 further steps are taken to deliberately pollute unused data in BIGNUM
10525 structures to try and expose faulty code further on. For now, openssl will
10526 (in its default mode of operation) continue to tolerate the inconsistent
10527 forms that it has tolerated in the past, but authors and packagers should
10528 consider trying openssl and their own applications when compiled with
10529 these debugging symbols defined. It will help highlight potential bugs in
10530 their own code, and will improve the test coverage for OpenSSL itself. At
10531 some point, these tighter rules will become openssl's default to improve
10532 maintainability, though the assert()s and other overheads will remain only
10533 in debugging configurations. See bn.h for more details.
10535 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10537 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10538 that can only be obtained through BN_CTX_new() (which implicitly
10539 initialises it). The presence of this function only made it possible
10540 to overwrite an existing structure (and cause memory leaks).
10544 * Because of the callback-based approach for implementing LHASH as a
10545 template type, lh_insert() adds opaque objects to hash-tables and
10546 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10547 to clean up those corresponding objects before destroying the hash table
10548 (and losing the object pointers). So some over-zealous constifications in
10549 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10550 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10551 prototyped to have "const" restrictions on the object pointers they are
10552 given (and so aren't required to cast them away any more).
10556 * The tmdiff.h API was so ugly and minimal that our own timing utility
10557 (speed) prefers to use its own implementation. The two implementations
10558 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10559 its object type properly exposed (MS_TM) instead of casting to/from
10560 `char *`. This may still change yet if someone realises MS_TM and
10562 aren't necessarily the greatest nomenclatures - but this is what was used
10563 internally to the implementation so I've used that for now.
10567 * Ensure that deprecated functions do not get compiled when
10568 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10569 the self-tests were still using deprecated key-generation functions so
10570 these have been updated also.
10574 * Reorganise PKCS#7 code to separate the digest location functionality
10575 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10576 New function PKCS7_set_digest() to set the digest type for PKCS#7
10577 digestedData type. Add additional code to correctly generate the
10578 digestedData type and add support for this type in PKCS7 initialization
10583 * New function PKCS7_set0_type_other() this initializes a PKCS7
10584 structure of type "other".
10588 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10589 sure the loop does correctly stop and breaking ("division by zero")
10590 modulus operations are not performed. The (pre-generated) prime
10591 table crypto/bn/bn_prime.h was already correct, but it could not be
10592 re-generated on some platforms because of the "division by zero"
10593 situation in the script.
10595 *Ralf S. Engelschall*
10597 * Update support for ECC-based TLS ciphersuites according to
10598 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10599 SHA-1 now is only used for "small" curves (where the
10600 representation of a field element takes up to 24 bytes); for
10601 larger curves, the field element resulting from ECDH is directly
10602 used as premaster secret.
10604 *Douglas Stebila (Sun Microsystems Laboratories)*
10606 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10607 curve secp160r1 to the tests.
10609 *Douglas Stebila (Sun Microsystems Laboratories)*
10611 * Add the possibility to load symbols globally with DSO.
10613 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10615 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10616 control of the error stack.
10620 * Add support for STORE in ENGINE.
10624 * Add the STORE type. The intention is to provide a common interface
10625 to certificate and key stores, be they simple file-based stores, or
10626 HSM-type store, or LDAP stores, or...
10627 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10631 * Add a generic structure called OPENSSL_ITEM. This can be used to
10632 pass a list of arguments to any function as well as provide a way
10633 for a function to pass data back to the caller.
10637 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10638 works like BUF_strdup() but can be used to duplicate a portion of
10639 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10644 * Add the function sk_find_ex() which works like sk_find(), but will
10645 return an index to an element even if an exact match couldn't be
10646 found. The index is guaranteed to point at the element where the
10647 searched-for key would be inserted to preserve sorting order.
10651 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10652 takes an extra flags argument for optional functionality. Currently,
10653 the following flags are defined:
10655 OBJ_BSEARCH_VALUE_ON_NOMATCH
10656 This one gets OBJ_bsearch_ex() to return a pointer to the first
10657 element where the comparing function returns a negative or zero
10660 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10661 This one gets OBJ_bsearch_ex() to return a pointer to the first
10662 element where the comparing function returns zero. This is useful
10663 if there are more than one element where the comparing function
10668 * Make it possible to create self-signed certificates with 'openssl ca'
10669 in such a way that the self-signed certificate becomes part of the
10670 CA database and uses the same mechanisms for serial number generation
10671 as all other certificate signing. The new flag '-selfsign' enables
10672 this functionality. Adapt CA.sh and CA.pl.in.
10676 * Add functionality to check the public key of a certificate request
10677 against a given private. This is useful to check that a certificate
10678 request can be signed by that key (self-signing).
10682 * Make it possible to have multiple active certificates with the same
10683 subject in the CA index file. This is done only if the keyword
10684 'unique_subject' is set to 'no' in the main CA section (default
10685 if 'CA_default') of the configuration file. The value is saved
10686 with the database itself in a separate index attribute file,
10687 named like the index file with '.attr' appended to the name.
10691 * Generate multi-valued AVAs using '+' notation in config files for
10696 * Support for nameConstraints certificate extension.
10700 * Support for policyConstraints certificate extension.
10704 * Support for policyMappings certificate extension.
10708 * Make sure the default DSA_METHOD implementation only uses its
10709 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10710 and change its own handlers to be NULL so as to remove unnecessary
10711 indirection. This lets alternative implementations fallback to the
10712 default implementation more easily.
10716 * Support for directoryName in GeneralName related extensions
10721 * Make it possible to link applications using Makefile.shared.
10722 Make that possible even when linking against static libraries!
10726 * Support for single pass processing for S/MIME signing. This now
10727 means that S/MIME signing can be done from a pipe, in addition
10728 cleartext signing (multipart/signed type) is effectively streaming
10729 and the signed data does not need to be all held in memory.
10731 This is done with a new flag PKCS7_STREAM. When this flag is set
10732 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10733 is done after the data is output (and digests calculated) in
10734 SMIME_write_PKCS7().
10738 * Add full support for -rpath/-R, both in shared libraries and
10739 applications, at least on the platforms where it's known how
10744 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10745 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10746 will now compute a table of multiples of the generator that
10747 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10748 faster (notably in the case of a single point multiplication,
10749 scalar * generator).
10751 *Nils Larsch, Bodo Moeller*
10753 * IPv6 support for certificate extensions. The various extensions
10754 which use the IP:a.b.c.d can now take IPv6 addresses using the
10755 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10760 * Added an ENGINE that implements RSA by performing private key
10761 exponentiations with the GMP library. The conversions to and from
10762 GMP's mpz_t format aren't optimised nor are any montgomery forms
10763 cached, and on x86 it appears OpenSSL's own performance has caught up.
10764 However there are likely to be other architectures where GMP could
10765 provide a boost. This ENGINE is not built in by default, but it can be
10766 specified at Configure time and should be accompanied by the necessary
10767 linker additions, eg;
10768 ./config -DOPENSSL_USE_GMP -lgmp
10772 * "openssl engine" will not display ENGINE/DSO load failure errors when
10773 testing availability of engines with "-t" - the old behaviour is
10774 produced by increasing the feature's verbosity with "-tt".
10778 * ECDSA routines: under certain error conditions uninitialized BN objects
10779 could be freed. Solution: make sure initialization is performed early
10780 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10785 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10786 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10787 software implementations. For DSA and DH, parameter generation can
10788 also be overridden by providing the appropriate method callbacks.
10792 * Change the "progress" mechanism used in key-generation and
10793 primality testing to functions that take a new BN_GENCB pointer in
10794 place of callback/argument pairs. The new API functions have `_ex`
10795 postfixes and the older functions are reimplemented as wrappers for
10796 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10797 declarations of the old functions to help (graceful) attempts to
10798 migrate to the new functions. Also, the new key-generation API
10799 functions operate on a caller-supplied key-structure and return
10800 success/failure rather than returning a key or NULL - this is to
10801 help make "keygen" another member function of RSA_METHOD etc.
10803 Example for using the new callback interface:
10805 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10806 void *my_arg = ...;
10809 BN_GENCB_set(&my_cb, my_callback, my_arg);
10811 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10812 /* For the meaning of a, b in calls to my_callback(), see the
10813 * documentation of the function that calls the callback.
10814 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10815 * my_callback should return 1 if it wants BN_is_prime_ex()
10816 * to continue, or 0 to stop.
10821 * Change the ZLIB compression method to be stateful, and make it
10822 available to TLS with the number defined in
10823 draft-ietf-tls-compression-04.txt.
10827 * Add the ASN.1 structures and functions for CertificatePair, which
10828 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10830 CertificatePair ::= SEQUENCE {
10831 forward [0] Certificate OPTIONAL,
10832 reverse [1] Certificate OPTIONAL,
10833 -- at least one of the pair shall be present -- }
10835 Also implement the PEM functions to read and write certificate
10836 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10838 This needed to be defined, mostly for the sake of the LDAP
10839 attribute crossCertificatePair, but may prove useful elsewhere as
10844 * Make it possible to inhibit symlinking of shared libraries in
10845 Makefile.shared, for Cygwin's sake.
10849 * Extend the BIGNUM API by creating a function
10850 void BN_set_negative(BIGNUM *a, int neg);
10851 and a macro that behave like
10852 int BN_is_negative(const BIGNUM *a);
10854 to avoid the need to access 'a->neg' directly in applications.
10858 * Implement fast modular reduction for pseudo-Mersenne primes
10859 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10860 EC_GROUP_new_curve_GFp() will now automatically use this
10863 *Nils Larsch <nla@trustcenter.de>*
10865 * Add new lock type (CRYPTO_LOCK_BN).
10869 * Change the ENGINE framework to automatically load engines
10870 dynamically from specific directories unless they could be
10871 found to already be built in or loaded. Move all the
10872 current engines except for the cryptodev one to a new
10873 directory engines/.
10874 The engines in engines/ are built as shared libraries if
10875 the "shared" options was given to ./Configure or ./config.
10876 Otherwise, they are inserted in libcrypto.a.
10877 /usr/local/ssl/engines is the default directory for dynamic
10878 engines, but that can be overridden at configure time through
10879 the usual use of --prefix and/or --openssldir, and at run
10880 time with the environment variable OPENSSL_ENGINES.
10882 *Geoff Thorpe and Richard Levitte*
10884 * Add Makefile.shared, a helper makefile to build shared
10885 libraries. Adapt Makefile.org.
10889 * Add version info to Win32 DLLs.
10891 *Peter 'Luna' Runestig" <peter@runestig.com>*
10893 * Add new 'medium level' PKCS#12 API. Certificates and keys
10894 can be added using this API to created arbitrary PKCS#12
10895 files while avoiding the low-level API.
10897 New options to PKCS12_create(), key or cert can be NULL and
10898 will then be omitted from the output file. The encryption
10899 algorithm NIDs can be set to -1 for no encryption, the mac
10900 iteration count can be set to 0 to omit the mac.
10902 Enhance pkcs12 utility by making the -nokeys and -nocerts
10903 options work when creating a PKCS#12 file. New option -nomac
10904 to omit the mac, NONE can be set for an encryption algorithm.
10905 New code is modified to use the enhanced PKCS12_create()
10906 instead of the low-level API.
10910 * Extend ASN1 encoder to support indefinite length constructed
10911 encoding. This can output sequences tags and octet strings in
10912 this form. Modify pk7_asn1.c to support indefinite length
10913 encoding. This is experimental and needs additional code to
10914 be useful, such as an ASN1 bio and some enhanced streaming
10917 Extend template encode functionality so that tagging is passed
10918 down to the template encoder.
10922 * Let 'openssl req' fail if an argument to '-newkey' is not
10923 recognized instead of using RSA as a default.
10927 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10928 As these are not official, they are not included in "ALL";
10929 the "ECCdraft" ciphersuite group alias can be used to select them.
10931 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10933 * Add ECDH engine support.
10935 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10937 * Add ECDH in new directory crypto/ecdh/.
10939 *Douglas Stebila (Sun Microsystems Laboratories)*
10941 * Let BN_rand_range() abort with an error after 100 iterations
10942 without success (which indicates a broken PRNG).
10946 * Change BN_mod_sqrt() so that it verifies that the input value
10947 is really the square of the return value. (Previously,
10948 BN_mod_sqrt would show GIGO behaviour.)
10952 * Add named elliptic curves over binary fields from X9.62, SECG,
10953 and WAP/WTLS; add OIDs that were still missing.
10955 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10957 * Extend the EC library for elliptic curves over binary fields
10958 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10961 EC_GF2m_simple_method
10965 EC_GROUP_new_curve_GF2m
10966 EC_GROUP_set_curve_GF2m
10967 EC_GROUP_get_curve_GF2m
10968 EC_POINT_set_affine_coordinates_GF2m
10969 EC_POINT_get_affine_coordinates_GF2m
10970 EC_POINT_set_compressed_coordinates_GF2m
10972 Point compression for binary fields is disabled by default for
10973 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10976 As binary polynomials are represented as BIGNUMs, various members
10977 of the EC_GROUP and EC_POINT data structures can be shared
10978 between the implementations for prime fields and binary fields;
10979 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10980 are essentially identical to their `..._GFp` counterparts.
10981 (For simplicity, the `..._GFp` prefix has been dropped from
10982 various internal method names.)
10984 An internal 'field_div' method (similar to 'field_mul' and
10985 'field_sqr') has been added; this is used only for binary fields.
10987 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10989 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10990 through methods ('mul', 'precompute_mult').
10992 The generic implementations (now internally called 'ec_wNAF_mul'
10993 and 'ec_wNAF_precomputed_mult') remain the default if these
10994 methods are undefined.
10996 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10998 * New function EC_GROUP_get_degree, which is defined through
10999 EC_METHOD. For curves over prime fields, this returns the bit
11000 length of the modulus.
11002 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11004 * New functions EC_GROUP_dup, EC_POINT_dup.
11005 (These simply call ..._new and ..._copy).
11007 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11009 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11010 Polynomials are represented as BIGNUMs (where the sign bit is not
11011 used) in the following functions [macros]:
11014 BN_GF2m_sub [= BN_GF2m_add]
11015 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11016 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11017 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11019 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11020 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11021 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11022 BN_GF2m_cmp [= BN_ucmp]
11024 (Note that only the 'mod' functions are actually for fields GF(2^m).
11025 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11027 For some functions, an the irreducible polynomial defining a
11028 field can be given as an 'unsigned int[]' with strictly
11029 decreasing elements giving the indices of those bits that are set;
11030 i.e., p[] represents the polynomial
11031 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11033 p[0] > p[1] > ... > p[k] = 0.
11034 This applies to the following functions:
11037 BN_GF2m_mod_mul_arr
11038 BN_GF2m_mod_sqr_arr
11039 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11040 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11041 BN_GF2m_mod_exp_arr
11042 BN_GF2m_mod_sqrt_arr
11043 BN_GF2m_mod_solve_quad_arr
11047 Conversion can be performed by the following functions:
11052 bntest.c has additional tests for binary polynomial arithmetic.
11054 Two implementations for BN_GF2m_mod_div() are available.
11055 The default algorithm simply uses BN_GF2m_mod_inv() and
11056 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11057 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11058 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11060 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11062 * Add new error code 'ERR_R_DISABLED' that can be used when some
11063 functionality is disabled at compile-time.
11065 *Douglas Stebila <douglas.stebila@sun.com>*
11067 * Change default behaviour of 'openssl asn1parse' so that more
11068 information is visible when viewing, e.g., a certificate:
11070 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11071 mode the content of non-printable OCTET STRINGs is output in a
11072 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11073 avoid the appearance of a printable string.
11075 *Nils Larsch <nla@trustcenter.de>*
11077 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11079 EC_GROUP_set_asn1_flag()
11080 EC_GROUP_get_asn1_flag()
11081 EC_GROUP_set_point_conversion_form()
11082 EC_GROUP_get_point_conversion_form()
11083 These control ASN1 encoding details:
11084 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11085 has been set to OPENSSL_EC_NAMED_CURVE.
11086 - Points are encoded in uncompressed form by default; options for
11087 asn1_for are as for point2oct, namely
11088 POINT_CONVERSION_COMPRESSED
11089 POINT_CONVERSION_UNCOMPRESSED
11090 POINT_CONVERSION_HYBRID
11092 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11094 EC_GROUP_set_seed()
11095 EC_GROUP_get0_seed()
11096 EC_GROUP_get_seed_len()
11097 This is used only for ASN1 purposes (so far).
11099 *Nils Larsch <nla@trustcenter.de>*
11101 * Add 'field_type' member to EC_METHOD, which holds the NID
11102 of the appropriate field type OID. The new function
11103 EC_METHOD_get_field_type() returns this value.
11105 *Nils Larsch <nla@trustcenter.de>*
11108 EC_POINT_point2bn()
11109 EC_POINT_bn2point()
11110 EC_POINT_point2hex()
11111 EC_POINT_hex2point()
11112 providing useful interfaces to EC_POINT_point2oct() and
11113 EC_POINT_oct2point().
11115 *Nils Larsch <nla@trustcenter.de>*
11117 * Change internals of the EC library so that the functions
11118 EC_GROUP_set_generator()
11119 EC_GROUP_get_generator()
11120 EC_GROUP_get_order()
11121 EC_GROUP_get_cofactor()
11122 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11123 to methods, which would lead to unnecessary code duplication when
11124 adding different types of curves.
11126 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11128 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11129 arithmetic, and such that modified wNAFs are generated
11130 (which avoid length expansion in many cases).
11134 * Add a function EC_GROUP_check_discriminant() (defined via
11135 EC_METHOD) that verifies that the curve discriminant is non-zero.
11137 Add a function EC_GROUP_check() that makes some sanity tests
11138 on a EC_GROUP, its generator and order. This includes
11139 EC_GROUP_check_discriminant().
11141 *Nils Larsch <nla@trustcenter.de>*
11143 * Add ECDSA in new directory crypto/ecdsa/.
11145 Add applications 'openssl ecparam' and 'openssl ecdsa'
11146 (these are based on 'openssl dsaparam' and 'openssl dsa').
11148 ECDSA support is also included in various other files across the
11149 library. Most notably,
11150 - 'openssl req' now has a '-newkey ecdsa:file' option;
11151 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11152 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11153 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11154 them suitable for ECDSA where domain parameters must be
11155 extracted before the specific public key;
11156 - ECDSA engine support has been added.
11158 *Nils Larsch <nla@trustcenter.de>*
11160 * Include some named elliptic curves, and add OIDs from X9.62,
11161 SECG, and WAP/WTLS. Each curve can be obtained from the new
11163 EC_GROUP_new_by_curve_name(),
11164 and the list of available named curves can be obtained with
11165 EC_get_builtin_curves().
11166 Also add a 'curve_name' member to EC_GROUP objects, which can be
11168 EC_GROUP_set_curve_name()
11169 EC_GROUP_get_curve_name()
11171 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11173 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11174 was actually never needed) and in BN_mul(). The removal in BN_mul()
11175 required a small change in bn_mul_part_recursive() and the addition
11176 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11177 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11178 bn_sub_words() and bn_add_words() except they take arrays with
11183 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11185 * Cleanse PEM buffers before freeing them since they may contain
11188 *Benjamin Bennett <ben@psc.edu>*
11190 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11191 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11192 authentication-only ciphersuites.
11196 * Since AES128 and AES256 share a single mask bit in the logic of
11197 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11198 kludge to work properly if AES128 is available and AES256 isn't.
11202 * Expand security boundary to match 1.1.1 module.
11206 * Remove redundant features: hash file source, editing of test vectors
11207 modify fipsld to use external fips_premain.c signature.
11211 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11212 run algorithm test programs.
11216 * Make algorithm test programs more tolerant of whitespace.
11220 * Have SSL/TLS server implementation tolerate "mismatched" record
11221 protocol version while receiving ClientHello even if the
11222 ClientHello is fragmented. (The server can't insist on the
11223 particular protocol version it has chosen before the ServerHello
11224 message has informed the client about his choice.)
11228 * Load error codes if they are not already present instead of using a
11229 static variable. This allows them to be cleanly unloaded and reloaded.
11233 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11235 * Introduce limits to prevent malicious keys being able to
11236 cause a denial of service. ([CVE-2006-2940])
11238 *Steve Henson, Bodo Moeller*
11240 * Fix ASN.1 parsing of certain invalid structures that can result
11241 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11243 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11244 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11246 * Fix SSL client code which could crash if connecting to a
11247 malicious SSLv2 server. ([CVE-2006-4343])
11249 *Tavis Ormandy and Will Drewry, Google Security Team*
11251 * Change ciphersuite string processing so that an explicit
11252 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11253 will no longer include "AES128-SHA"), and any other similar
11254 ciphersuite (same bitmap) from *other* protocol versions (so that
11255 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11256 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11257 changes from 0.9.8b and 0.9.8d.
11261 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11263 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11264 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11266 * Change the Unix randomness entropy gathering to use poll() when
11267 possible instead of select(), since the latter has some
11268 undesirable limitations.
11270 *Darryl Miles via Richard Levitte and Bodo Moeller*
11272 * Disable rogue ciphersuites:
11274 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11275 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11276 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11278 The latter two were purportedly from
11279 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11282 Also deactivate the remaining ciphersuites from
11283 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11284 unofficial, and the ID has long expired.
11288 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11289 dual-core machines) and other potential thread-safety issues.
11293 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11295 * Adapt fipsld and the build system to link against the validated FIPS
11296 module in FIPS mode.
11300 * Fixes for VC++ 2005 build under Windows.
11304 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11305 from a Windows bash shell such as MSYS. It is autodetected from the
11306 "config" script when run from a VC++ environment. Modify standard VC++
11307 build to use fipscanister.o from the GNU make build.
11311 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11313 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11314 The value now differs depending on if you build for FIPS or not.
11315 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11316 safely run with a non-FIPSed libcrypto, as it may crash because of
11317 the difference induced by this change.
11321 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11323 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11324 (part of SSL_OP_ALL). This option used to disable the
11325 countermeasure against man-in-the-middle protocol-version
11326 rollback in the SSL 2.0 server implementation, which is a bad
11327 idea. ([CVE-2005-2969])
11329 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11330 for Information Security, National Institute of Advanced Industrial
11331 Science and Technology [AIST, Japan)]*
11333 * Minimal support for X9.31 signatures and PSS padding modes. This is
11334 mainly for FIPS compliance and not fully integrated at this stage.
11338 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11339 the exponentiation using a fixed-length exponent. (Otherwise,
11340 the information leaked through timing could expose the secret key
11341 after many signatures; cf. Bleichenbacher's attack on DSA with
11346 * Make a new fixed-window mod_exp implementation the default for
11347 RSA, DSA, and DH private-key operations so that the sequence of
11348 squares and multiplies and the memory access pattern are
11349 independent of the particular secret key. This will mitigate
11350 cache-timing and potential related attacks.
11352 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11353 and this is automatically used by BN_mod_exp_mont() if the new flag
11354 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11355 will use this BN flag for private exponents unless the flag
11356 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11357 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11359 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11361 * Change the client implementation for SSLv23_method() and
11362 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11363 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11364 (Previously, the SSL 2.0 backwards compatible Client Hello
11365 message format would be used even with SSL_OP_NO_SSLv2.)
11369 * Add support for smime-type MIME parameter in S/MIME messages which some
11374 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11375 a threadsafe manner. Modify rsa code to use new function and add calls
11376 to dsa and dh code (which had race conditions before).
11380 * Include the fixed error library code in the C error file definitions
11381 instead of fixing them up at runtime. This keeps the error code
11382 structures constant.
11386 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11388 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11391 * Fixes for newer kerberos headers. NB: the casts are needed because
11392 the 'length' field is signed on one version and unsigned on another
11393 with no (?) obvious way to tell the difference, without these VC++
11394 complains. Also the "definition" of FAR (blank) is no longer included
11395 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11396 some needed definitions.
11400 * Undo Cygwin change.
11404 * Added support for proxy certificates according to RFC 3820.
11405 Because they may be a security thread to unaware applications,
11406 they must be explicitly allowed in run-time. See
11407 docs/HOWTO/proxy_certificates.txt for further information.
11411 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11413 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11414 server and client random values. Previously
11415 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11416 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11418 This change has negligible security impact because:
11420 1. Server and client random values still have 24 bytes of pseudo random
11423 2. Server and client random values are sent in the clear in the initial
11426 3. The master secret is derived using the premaster secret (48 bytes in
11427 size for static RSA ciphersuites) as well as client server and random
11430 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11433 *Stephen Henson, reported by UK NISCC*
11435 * Use Windows randomness collection on Cygwin.
11439 * Fix hang in EGD/PRNGD query when communication socket is closed
11440 prematurely by EGD/PRNGD.
11442 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11444 * Prompt for pass phrases when appropriate for PKCS12 input format.
11448 * Back-port of selected performance improvements from development
11449 branch, as well as improved support for PowerPC platforms.
11453 * Add lots of checks for memory allocation failure, error codes to indicate
11454 failure and freeing up memory if a failure occurs.
11456 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11458 * Add new -passin argument to dgst.
11462 * Perform some character comparisons of different types in X509_NAME_cmp:
11463 this is needed for some certificates that re-encode DNs into UTF8Strings
11464 (in violation of RFC3280) and can't or won't issue name rollover
11469 * Make an explicit check during certificate validation to see that
11470 the CA setting in each certificate on the chain is correct. As a
11471 side effect always do the following basic checks on extensions,
11472 not just when there's an associated purpose to the check:
11474 - if there is an unhandled critical extension (unless the user
11475 has chosen to ignore this fault)
11476 - if the path length has been exceeded (if one is set at all)
11477 - that certain extensions fit the associated purpose (if one has
11482 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11484 * Avoid a race condition when CRLs are checked in a multi threaded
11485 environment. This would happen due to the reordering of the revoked
11486 entries during signature checking and serial number lookup. Now the
11487 encoding is cached and the serial number sort performed under a lock.
11488 Add new STACK function sk_is_sorted().
11492 * Add Delta CRL to the extension code.
11496 * Various fixes to s3_pkt.c so alerts are sent properly.
11498 *David Holmes <d.holmes@f5.com>*
11500 * Reduce the chances of duplicate issuer name and serial numbers (in
11501 violation of RFC3280) using the OpenSSL certificate creation utilities.
11502 This is done by creating a random 64 bit value for the initial serial
11503 number when a serial number file is created or when a self signed
11504 certificate is created using 'openssl req -x509'. The initial serial
11505 number file is created using 'openssl x509 -next_serial' in CA.pl
11506 rather than being initialized to 1.
11510 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11512 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11513 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11515 *Joe Orton, Steve Henson*
11517 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11520 *Joe Orton, Steve Henson*
11522 * Make it possible to have multiple active certificates with the same
11523 subject in the CA index file. This is done only if the keyword
11524 'unique_subject' is set to 'no' in the main CA section (default
11525 if 'CA_default') of the configuration file. The value is saved
11526 with the database itself in a separate index attribute file,
11527 named like the index file with '.attr' appended to the name.
11531 * X509 verify fixes. Disable broken certificate workarounds when
11532 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11533 keyUsage extension present. Don't accept CRLs with unhandled critical
11534 extensions: since verify currently doesn't process CRL extensions this
11535 rejects a CRL with *any* critical extensions. Add new verify error codes
11540 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11541 A clarification of RFC2560 will require the use of OCTET STRINGs and
11542 some implementations cannot handle the current raw format. Since OpenSSL
11543 copies and compares OCSP nonces as opaque blobs without any attempt at
11544 parsing them this should not create any compatibility issues.
11548 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11549 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11550 this HMAC (and other) operations are several times slower than OpenSSL
11555 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11557 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11559 * Use the correct content when signing type "other".
11563 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11565 * Fix various bugs revealed by running the NISCC test suite:
11567 Stop out of bounds reads in the ASN1 code when presented with
11568 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11570 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11572 If verify callback ignores invalid public key errors don't try to check
11573 certificate signature with the NULL public key.
11577 * New -ignore_err option in ocsp application to stop the server
11578 exiting on the first error in a request.
11582 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11583 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11588 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11589 extra data after the compression methods not only for TLS 1.0
11590 but also for SSL 3.0 (as required by the specification).
11592 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11594 * Change X509_certificate_type() to mark the key as exported/exportable
11595 when it's 512 *bits* long, not 512 bytes.
11599 * Change AES_cbc_encrypt() so it outputs exact multiple of
11600 blocks during encryption.
11604 * Various fixes to base64 BIO and non blocking I/O. On write
11605 flushes were not handled properly if the BIO retried. On read
11606 data was not being buffered properly and had various logic bugs.
11607 This also affects blocking I/O when the data being decoded is a
11612 * Various S/MIME bugfixes and compatibility changes:
11613 output correct application/pkcs7 MIME type if
11614 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11615 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11616 of files as .eml work). Correctly handle very long lines in MIME
11621 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11623 * Countermeasure against the Klima-Pokorny-Rosa extension of
11624 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11625 a protocol version number mismatch like a decryption error
11626 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11630 * Turn on RSA blinding by default in the default implementation
11631 to avoid a timing attack. Applications that don't want it can call
11632 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11633 They would be ill-advised to do so in most cases.
11635 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11637 * Change RSA blinding code so that it works when the PRNG is not
11638 seeded (in this case, the secret RSA exponent is abused as
11639 an unpredictable seed -- if it is not unpredictable, there
11640 is no point in blinding anyway). Make RSA blinding thread-safe
11641 by remembering the creator's thread ID in rsa->blinding and
11642 having all other threads use local one-time blinding factors
11643 (this requires more computation than sharing rsa->blinding, but
11644 avoids excessive locking; and if an RSA object is not shared
11645 between threads, blinding will still be very fast).
11649 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11650 ENGINE as defaults for all supported algorithms irrespective of
11651 the 'flags' parameter. 'flags' is now honoured, so applications
11652 should make sure they are passing it correctly.
11656 * Target "mingw" now allows native Windows code to be generated in
11657 the Cygwin environment as well as with the MinGW compiler.
11661 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11663 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11664 via timing by performing a MAC computation even if incorrect
11665 block cipher padding has been found. This is a countermeasure
11666 against active attacks where the attacker has to distinguish
11667 between bad padding and a MAC verification error. ([CVE-2003-0078])
11669 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11670 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11671 Martin Vuagnoux (EPFL, Ilion)*
11673 * Make the no-err option work as intended. The intention with no-err
11674 is not to have the whole error stack handling routines removed from
11675 libcrypto, it's only intended to remove all the function name and
11676 reason texts, thereby removing some of the footprint that may not
11677 be interesting if those errors aren't displayed anyway.
11679 NOTE: it's still possible for any application or module to have its
11680 own set of error texts inserted. The routines are there, just not
11681 used by default when no-err is given.
11685 * Add support for FreeBSD on IA64.
11687 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11689 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11690 Kerberos function mit_des_cbc_cksum(). Before this change,
11691 the value returned by DES_cbc_cksum() was like the one from
11692 mit_des_cbc_cksum(), except the bytes were swapped.
11694 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11696 * Allow an application to disable the automatic SSL chain building.
11697 Before this a rather primitive chain build was always performed in
11698 ssl3_output_cert_chain(): an application had no way to send the
11699 correct chain if the automatic operation produced an incorrect result.
11701 Now the chain builder is disabled if either:
11703 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11705 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11707 The reasoning behind this is that an application would not want the
11708 auto chain building to take place if extra chain certificates are
11709 present and it might also want a means of sending no additional
11710 certificates (for example the chain has two certificates and the
11715 * Add the possibility to build without the ENGINE framework.
11717 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11719 * Under Win32 gmtime() can return NULL: check return value in
11720 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11724 * DSA routines: under certain error conditions uninitialized BN objects
11725 could be freed. Solution: make sure initialization is performed early
11726 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11727 Nils Larsch <nla@trustcenter.de> via PR#459)
11731 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11732 checked on reconnect on the client side, therefore session resumption
11733 could still fail with a "ssl session id is different" error. This
11734 behaviour is masked when SSL_OP_ALL is used due to
11735 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11736 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11737 followup to PR #377.
11741 * IA-32 assembler support enhancements: unified ELF targets, support
11742 for SCO/Caldera platforms, fix for Cygwin shared build.
11746 * Add support for FreeBSD on sparc64. As a consequence, support for
11747 FreeBSD on non-x86 processors is separate from x86 processors on
11748 the config script, much like the NetBSD support.
11750 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11752 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11754 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11757 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11758 code (06) was taken as the first octet of the session ID and the last
11759 octet was ignored consequently. As a result SSLv2 client side session
11760 caching could not have worked due to the session ID mismatch between
11762 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11767 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11768 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11773 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11774 seems that in spite of existing for more than a year, many application
11775 author have done nothing to provide the necessary callbacks, which
11776 means that this particular engine will not work properly anywhere.
11777 This is a very unfortunate situation which forces us, in the name
11778 of usability, to give the hw_ncipher.c a static lock, which is part
11780 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11781 appear in 0.9.8 or later. We EXPECT application authors to have
11782 dealt properly with this when 0.9.8 is released (unless we actually
11783 make such changes in the libcrypto locking code that changes will
11784 have to be made anyway).
11788 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11789 octets have been read, EOF or an error occurs. Without this change
11790 some truncated ASN1 structures will not produce an error.
11794 * Disable Heimdal support, since it hasn't been fully implemented.
11795 Still give the possibility to force the use of Heimdal, but with
11796 warnings and a request that patches get sent to openssl-dev.
11800 * Add the VC-CE target, introduce the WINCE sysname, and add
11801 INSTALL.WCE and appropriate conditionals to make it build.
11803 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11805 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11806 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11807 edit numbers of the version.
11809 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11811 * Introduce safe string copy and catenation functions
11812 (BUF_strlcpy() and BUF_strlcat()).
11814 *Ben Laurie (CHATS) and Richard Levitte*
11816 * Avoid using fixed-size buffers for one-line DNs.
11818 *Ben Laurie (CHATS)*
11820 * Add BUF_MEM_grow_clean() to avoid information leakage when
11821 resizing buffers containing secrets, and use where appropriate.
11823 *Ben Laurie (CHATS)*
11825 * Avoid using fixed size buffers for configuration file location.
11827 *Ben Laurie (CHATS)*
11829 * Avoid filename truncation for various CA files.
11831 *Ben Laurie (CHATS)*
11833 * Use sizeof in preference to magic numbers.
11835 *Ben Laurie (CHATS)*
11837 * Avoid filename truncation in cert requests.
11839 *Ben Laurie (CHATS)*
11841 * Add assertions to check for (supposedly impossible) buffer
11844 *Ben Laurie (CHATS)*
11846 * Don't cache truncated DNS entries in the local cache (this could
11847 potentially lead to a spoofing attack).
11849 *Ben Laurie (CHATS)*
11851 * Fix various buffers to be large enough for hex/decimal
11852 representations in a platform independent manner.
11854 *Ben Laurie (CHATS)*
11856 * Add CRYPTO_realloc_clean() to avoid information leakage when
11857 resizing buffers containing secrets, and use where appropriate.
11859 *Ben Laurie (CHATS)*
11861 * Add BIO_indent() to avoid much slightly worrying code to do
11864 *Ben Laurie (CHATS)*
11866 * Convert sprintf()/BIO_puts() to BIO_printf().
11868 *Ben Laurie (CHATS)*
11870 * buffer_gets() could terminate with the buffer only half
11873 *Ben Laurie (CHATS)*
11875 * Add assertions to prevent user-supplied crypto functions from
11876 overflowing internal buffers by having large block sizes, etc.
11878 *Ben Laurie (CHATS)*
11880 * New OPENSSL_assert() macro (similar to assert(), but enabled
11883 *Ben Laurie (CHATS)*
11885 * Eliminate unused copy of key in RC4.
11887 *Ben Laurie (CHATS)*
11889 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11891 *Ben Laurie (CHATS)*
11893 * Fix off-by-one error in EGD path.
11895 *Ben Laurie (CHATS)*
11897 * If RANDFILE path is too long, ignore instead of truncating.
11899 *Ben Laurie (CHATS)*
11901 * Eliminate unused and incorrectly sized X.509 structure
11904 *Ben Laurie (CHATS)*
11906 * Eliminate unused and dangerous function knumber().
11908 *Ben Laurie (CHATS)*
11910 * Eliminate unused and dangerous structure, KSSL_ERR.
11912 *Ben Laurie (CHATS)*
11914 * Protect against overlong session ID context length in an encoded
11915 session object. Since these are local, this does not appear to be
11918 *Ben Laurie (CHATS)*
11920 * Change from security patch (see 0.9.6e below) that did not affect
11921 the 0.9.6 release series:
11923 Remote buffer overflow in SSL3 protocol - an attacker could
11924 supply an oversized master key in Kerberos-enabled versions.
11927 *Ben Laurie (CHATS)*
11929 * Change the SSL kerb5 codes to match RFC 2712.
11933 * Make -nameopt work fully for req and add -reqopt switch.
11935 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11937 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11939 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11941 * Make sure tests can be performed even if the corresponding algorithms
11942 have been removed entirely. This was also the last step to make
11943 OpenSSL compilable with DJGPP under all reasonable conditions.
11945 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11947 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11948 to allow version independent disabling of normally unselected ciphers,
11949 which may be activated as a side-effect of selecting a single cipher.
11951 (E.g., cipher list string "RSA" enables ciphersuites that are left
11952 out of "ALL" because they do not provide symmetric encryption.
11953 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11955 *Lutz Jaenicke, Bodo Moeller*
11957 * Add appropriate support for separate platform-dependent build
11958 directories. The recommended way to make a platform-dependent
11959 build directory is the following (tested on Linux), maybe with
11962 # Place yourself outside of the OpenSSL source tree. In
11963 # this example, the environment variable OPENSSL_SOURCE
11964 # is assumed to contain the absolute OpenSSL source directory.
11965 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11966 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11967 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11968 mkdir -p `dirname $F`
11969 ln -s $OPENSSL_SOURCE/$F $F
11972 To be absolutely sure not to disturb the source tree, a "make clean"
11973 is a good thing. If it isn't successful, don't worry about it,
11974 it probably means the source directory is very clean.
11978 * Make sure any ENGINE control commands make local copies of string
11979 pointers passed to them whenever necessary. Otherwise it is possible
11980 the caller may have overwritten (or deallocated) the original string
11981 data when a later ENGINE operation tries to use the stored values.
11983 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11985 * Improve diagnostics in file reading and command-line digests.
11987 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11989 * Add AES modes CFB and OFB to the object database. Correct an
11990 error in AES-CFB decryption.
11994 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11995 allows existing EVP_CIPHER_CTX structures to be reused after
11996 calling `EVP_*Final()`. This behaviour is used by encryption
11997 BIOs and some applications. This has the side effect that
11998 applications must explicitly clean up cipher contexts with
11999 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12003 * Check the values of dna and dnb in bn_mul_recursive before calling
12004 bn_mul_comba (a non zero value means the a or b arrays do not contain
12005 n2 elements) and fallback to bn_mul_normal if either is not zero.
12009 * Fix escaping of non-ASCII characters when using the -subj option
12010 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12014 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12015 form for "surname", serialNumber has no short form.
12016 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12017 therefore remove "mail" short name for "internet 7".
12018 The OID for unique identifiers in X509 certificates is
12019 x500UniqueIdentifier, not uniqueIdentifier.
12020 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12024 * Add an "init" command to the ENGINE config module and auto initialize
12025 ENGINEs. Without any "init" command the ENGINE will be initialized
12026 after all ctrl commands have been executed on it. If init=1 the
12027 ENGINE is initialized at that point (ctrls before that point are run
12028 on the uninitialized ENGINE and after on the initialized one). If
12029 init=0 then the ENGINE will not be initialized at all.
12033 * Fix the 'app_verify_callback' interface so that the user-defined
12034 argument is actually passed to the callback: In the
12035 SSL_CTX_set_cert_verify_callback() prototype, the callback
12036 declaration has been changed from
12039 int (*cb)(X509_STORE_CTX *,void *);
12040 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12041 i=s->ctx->app_verify_callback(&ctx)
12042 has been changed into
12043 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12045 To update applications using SSL_CTX_set_cert_verify_callback(),
12046 a dummy argument can be added to their callback functions.
12048 *D. K. Smetters <smetters@parc.xerox.com>*
12050 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12052 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12054 * Add and OPENSSL_LOAD_CONF define which will cause
12055 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12056 This allows older applications to transparently support certain
12057 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12058 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12059 load the config file and OPENSSL_add_all_algorithms_conf() which will
12060 always load it have also been added.
12064 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12065 Adjust NIDs and EVP layer.
12067 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12069 * Config modules support in openssl utility.
12071 Most commands now load modules from the config file,
12072 though in a few (such as version) this isn't done
12073 because it couldn't be used for anything.
12075 In the case of ca and req the config file used is
12076 the same as the utility itself: that is the -config
12077 command line option can be used to specify an
12082 * Move default behaviour from OPENSSL_config(). If appname is NULL
12083 use "openssl_conf" if filename is NULL use default openssl config file.
12087 * Add an argument to OPENSSL_config() to allow the use of an alternative
12088 config section name. Add a new flag to tolerate a missing config file
12089 and move code to CONF_modules_load_file().
12093 * Support for crypto accelerator cards from Accelerated Encryption
12094 Processing, www.aep.ie. (Use engine 'aep')
12095 The support was copied from 0.9.6c [engine] and adapted/corrected
12096 to work with the new engine framework.
12098 *AEP Inc. and Richard Levitte*
12100 * Support for SureWare crypto accelerator cards from Baltimore
12101 Technologies. (Use engine 'sureware')
12102 The support was copied from 0.9.6c [engine] and adapted
12103 to work with the new engine framework.
12107 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12108 make the newer ENGINE framework commands for the CHIL engine work.
12110 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12112 * Make it possible to produce shared libraries on ReliantUNIX.
12114 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12116 * Add the configuration target debug-linux-ppro.
12117 Make 'openssl rsa' use the general key loading routines
12118 implemented in `apps.c`, and make those routines able to
12119 handle the key format FORMAT_NETSCAPE and the variant
12122 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12124 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12126 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12128 * Add -keyform to rsautl, and document -engine.
12130 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12132 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12133 BIO_R_NO_SUCH_FILE error code rather than the generic
12134 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12138 * Add new functions
12139 ERR_peek_last_error
12140 ERR_peek_last_error_line
12141 ERR_peek_last_error_line_data.
12142 These are similar to
12144 ERR_peek_error_line
12145 ERR_peek_error_line_data,
12146 but report on the latest error recorded rather than the first one
12147 still in the error queue.
12149 *Ben Laurie, Bodo Moeller*
12151 * default_algorithms option in ENGINE config module. This allows things
12153 default_algorithms = ALL
12154 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12158 * Preliminary ENGINE config module.
12162 * New experimental application configuration code.
12166 * Change the AES code to follow the same name structure as all other
12167 symmetric ciphers, and behave the same way. Move everything to
12168 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12170 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12172 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12174 *Ben Laurie and Theo de Raadt*
12176 * Add option to output public keys in req command.
12178 *Massimiliano Pala madwolf@openca.org*
12180 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12181 (up to about 10% better than before for P-192 and P-224).
12185 * New functions/macros
12187 SSL_CTX_set_msg_callback(ctx, cb)
12188 SSL_CTX_set_msg_callback_arg(ctx, arg)
12189 SSL_set_msg_callback(ssl, cb)
12190 SSL_set_msg_callback_arg(ssl, arg)
12192 to request calling a callback function
12194 void cb(int write_p, int version, int content_type,
12195 const void *buf, size_t len, SSL *ssl, void *arg)
12197 whenever a protocol message has been completely received
12198 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12199 protocol version according to which the SSL library interprets
12200 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12201 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12202 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12203 specification (change_cipher_spec(20), alert(21), handshake(22)).
12204 'buf' and 'len' point to the actual message, 'ssl' to the
12205 SSL object, and 'arg' is the application-defined value set by
12206 SSL[_CTX]_set_msg_callback_arg().
12208 'openssl s_client' and 'openssl s_server' have new '-msg' options
12209 to enable a callback that displays all protocol messages.
12213 * Change the shared library support so shared libraries are built as
12214 soon as the corresponding static library is finished, and thereby get
12215 openssl and the test programs linked against the shared library.
12216 This still only happens when the keyword "shard" has been given to
12217 the configuration scripts.
12219 NOTE: shared library support is still an experimental thing, and
12220 backward binary compatibility is still not guaranteed.
12222 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12224 * Add support for Subject Information Access extension.
12226 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12228 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12229 additional bytes when new memory had to be allocated, not just
12230 when reusing an existing buffer.
12234 * New command line and configuration option 'utf8' for the req command.
12235 This allows field values to be specified as UTF8 strings.
12239 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12240 runs for the former and machine-readable output for the latter.
12244 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12245 of the e-mail address in the DN (i.e., it will go into a certificate
12246 extension only). The new configuration file option 'email_in_dn = no'
12247 has the same effect.
12249 *Massimiliano Pala madwolf@openca.org*
12251 * Change all functions with names starting with `des_` to be starting
12252 with `DES_` instead. Add wrappers that are compatible with libdes,
12253 but are named `_ossl_old_des_*`. Finally, add macros that map the
12254 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12255 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12256 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12259 Since we provide two compatibility mappings, the user needs to
12260 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12261 compatibility is desired. The default (i.e., when that macro
12262 isn't defined) is OpenSSL 0.9.6c compatibility.
12264 There are also macros that enable and disable the support of old
12265 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12266 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12267 are defined, the default will apply: to support the old des routines.
12269 In either case, one must include openssl/des.h to get the correct
12270 definitions. Do not try to just include openssl/des_old.h, that
12273 NOTE: This is a major break of an old API into a new one. Software
12274 authors are encouraged to switch to the `DES_` style functions. Some
12275 time in the future, des_old.h and the libdes compatibility functions
12276 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12277 default), and then completely removed.
12281 * Test for certificates which contain unsupported critical extensions.
12282 If such a certificate is found during a verify operation it is
12283 rejected by default: this behaviour can be overridden by either
12284 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12285 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12286 X509_supported_extension() has also been added which returns 1 if a
12287 particular extension is supported.
12291 * Modify the behaviour of EVP cipher functions in similar way to digests
12292 to retain compatibility with existing code.
12296 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12297 compatibility with existing code. In particular the 'ctx' parameter does
12298 not have to be to be initialized before the call to EVP_DigestInit() and
12299 it is tidied up after a call to EVP_DigestFinal(). New function
12300 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12301 EVP_MD_CTX_copy() changed to not require the destination to be
12302 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12303 requires the destination to be valid.
12305 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12306 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12310 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12311 so that complete 'Handshake' protocol structures are kept in memory
12312 instead of overwriting 'msg_type' and 'length' with 'body' data.
12316 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12318 *Massimo Santin via Richard Levitte*
12320 * Major restructuring to the underlying ENGINE code. This includes
12321 reduction of linker bloat, separation of pure "ENGINE" manipulation
12322 (initialisation, etc) from functionality dealing with implementations
12323 of specific crypto interfaces. This change also introduces integrated
12324 support for symmetric ciphers and digest implementations - so ENGINEs
12325 can now accelerate these by providing EVP_CIPHER and EVP_MD
12326 implementations of their own. This is detailed in
12327 [crypto/engine/README.md](crypto/engine/README.md)
12328 as it couldn't be adequately described here. However, there are a few
12329 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12330 were changed in the original introduction of ENGINE code have now
12331 reverted back - the hooking from this code to ENGINE is now a good
12332 deal more passive and at run-time, operations deal directly with
12333 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12334 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12335 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12336 they were not being used by the framework as there is no concept of a
12337 BIGNUM_METHOD and they could not be generalised to the new
12338 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12339 ENGINE_cpy() has been removed as it cannot be consistently defined in
12344 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12348 * Change mkdef.pl to sort symbols that get the same entry number,
12349 and make sure the automatically generated functions `ERR_load_*`
12350 become part of libeay.num as well.
12354 * New function SSL_renegotiate_pending(). This returns true once
12355 renegotiation has been requested (either SSL_renegotiate() call
12356 or HelloRequest/ClientHello received from the peer) and becomes
12357 false once a handshake has been completed.
12358 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12359 sends a HelloRequest, but does not ensure that a handshake takes
12360 place. SSL_renegotiate_pending() is useful for checking if the
12361 client has followed the request.)
12365 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12366 By default, clients may request session resumption even during
12367 renegotiation (if session ID contexts permit); with this option,
12368 session resumption is possible only in the first handshake.
12370 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12371 more bits available for options that should not be part of
12372 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12376 * Add some demos for certificate and certificate request creation.
12380 * Make maximum certificate chain size accepted from the peer application
12381 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12382 "Douglas E. Engert" <deengert@anl.gov>.
12386 * Add support for shared libraries for Unixware-7
12387 (Boyd Lynn Gerber <gerberb@zenez.com>).
12391 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12392 be done prior to destruction. Use this to unload error strings from
12393 ENGINEs that load their own error strings. NB: This adds two new API
12394 functions to "get" and "set" this destroy handler in an ENGINE.
12398 * Alter all existing ENGINE implementations (except "openssl" and
12399 "openbsd") to dynamically instantiate their own error strings. This
12400 makes them more flexible to be built both as statically-linked ENGINEs
12401 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12402 Also, add stub code to each that makes building them as self-contained
12403 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12407 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12408 implementations into applications that are completely implemented in
12409 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12410 commands that can be used to configure what shared-library to load and
12411 to control aspects of the way it is handled. Also, made an update to
12412 the [README-Engine.md](README-Engine.md) file
12413 that brings its information up-to-date and
12414 provides some information and instructions on the "dynamic" ENGINE
12415 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12419 * Make it possible to unload ranges of ERR strings with a new
12420 "ERR_unload_strings" function.
12424 * Add a copy() function to EVP_MD.
12428 * Make EVP_MD routines take a context pointer instead of just the
12429 md_data void pointer.
12433 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12434 that the digest can only process a single chunk of data
12435 (typically because it is provided by a piece of
12436 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12437 is only going to provide a single chunk of data, and hence the
12438 framework needn't accumulate the data for oneshot drivers.
12442 * As with "ERR", make it possible to replace the underlying "ex_data"
12443 functions. This change also alters the storage and management of global
12444 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12445 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12446 index counters. The API functions that use this state have been changed
12447 to take a "class_index" rather than pointers to the class's local STACK
12448 and counter, and there is now an API function to dynamically create new
12449 classes. This centralisation allows us to (a) plug a lot of the
12450 thread-safety problems that existed, and (b) makes it possible to clean
12451 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12452 such data would previously have always leaked in application code and
12453 workarounds were in place to make the memory debugging turn a blind eye
12454 to it. Application code that doesn't use this new function will still
12455 leak as before, but their memory debugging output will announce it now
12456 rather than letting it slide.
12458 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12459 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12460 has a return value to indicate success or failure.
12464 * Make it possible to replace the underlying "ERR" functions such that the
12465 global state (2 LHASH tables and 2 locks) is only used by the "default"
12466 implementation. This change also adds two functions to "get" and "set"
12467 the implementation prior to it being automatically set the first time
12468 any other ERR function takes place. Ie. an application can call "get",
12469 pass the return value to a module it has just loaded, and that module
12470 can call its own "set" function using that value. This means the
12471 module's "ERR" operations will use (and modify) the error state in the
12472 application and not in its own statically linked copy of OpenSSL code.
12476 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12477 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12478 the operation, and provides a more encapsulated way for external code
12479 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12480 to use these functions rather than manually incrementing the counts.
12482 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12486 * Add EVP test program.
12490 * Add symmetric cipher support to ENGINE. Expect the API to change!
12494 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12495 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12496 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12497 These allow a CRL to be built without having to access X509_CRL fields
12498 directly. Modify 'ca' application to use new functions.
12502 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12503 bug workarounds. Rollback attack detection is a security feature.
12504 The problem will only arise on OpenSSL servers when TLSv1 is not
12505 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12506 Software authors not wanting to support TLSv1 will have special reasons
12507 for their choice and can explicitly enable this option.
12509 *Bodo Moeller, Lutz Jaenicke*
12511 * Rationalise EVP so it can be extended: don't include a union of
12512 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12513 (similar to those existing for EVP_CIPHER_CTX).
12518 EVP_MD_CTX_init(&md); /* new function call */
12519 EVP_DigestInit(&md, EVP_sha1());
12520 EVP_DigestUpdate(&md, in, len);
12521 EVP_DigestFinal(&md, out, NULL);
12522 EVP_MD_CTX_cleanup(&md); /* new function call */
12526 * Make DES key schedule conform to the usual scheme, as well as
12527 correcting its structure. This means that calls to DES functions
12528 now have to pass a pointer to a des_key_schedule instead of a
12529 plain des_key_schedule (which was actually always a pointer
12532 des_key_schedule ks;
12534 des_set_key_checked(..., &ks);
12535 des_ncbc_encrypt(..., &ks, ...);
12537 (Note that a later change renames 'des_...' into 'DES_...'.)
12541 * Initial reduction of linker bloat: the use of some functions, such as
12542 PEM causes large amounts of unused functions to be linked in due to
12543 poor organisation. For example pem_all.c contains every PEM function
12544 which has a knock on effect of linking in large amounts of (unused)
12545 ASN1 code. Grouping together similar functions and splitting unrelated
12546 functions prevents this.
12550 * Cleanup of EVP macros.
12554 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12555 correct `_ecb suffix`.
12559 * Add initial OCSP responder support to ocsp application. The
12560 revocation information is handled using the text based index
12561 use by the ca application. The responder can either handle
12562 requests generated internally, supplied in files (for example
12563 via a CGI script) or using an internal minimal server.
12567 * Add configuration choices to get zlib compression for TLS.
12571 * Changes to Kerberos SSL for RFC 2712 compliance:
12572 1. Implemented real KerberosWrapper, instead of just using
12573 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12574 2. Implemented optional authenticator field of KerberosWrapper.
12576 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12577 and authenticator structs; see crypto/krb5/.
12579 Generalized Kerberos calls to support multiple Kerberos libraries.
12580 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12581 via Richard Levitte*
12583 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12584 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12585 values for each of the key sizes rather than having just
12586 parameters (and 'speed' generating keys each time).
12590 * Speed up EVP routines.
12593 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12594 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12595 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12596 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12598 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12599 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12600 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12603 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12605 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12609 * Added the OS2-EMX target.
12611 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12613 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12614 New functions to support `NCONF` routines in extension code.
12615 New function `CONF_set_nconf()`
12616 to allow functions which take an `NCONF` to also handle the old `LHASH`
12617 structure: this means that the old `CONF` compatible routines can be
12618 retained (in particular w.rt. extensions) without having to duplicate the
12619 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12623 * Enhance the general user interface with mechanisms for inner control
12624 and with possibilities to have yes/no kind of prompts.
12628 * Change all calls to low-level digest routines in the library and
12629 applications to use EVP. Add missing calls to HMAC_cleanup() and
12630 don't assume HMAC_CTX can be copied using memcpy().
12632 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12634 * Add the possibility to control engines through control names but with
12635 arbitrary arguments instead of just a string.
12636 Change the key loaders to take a UI_METHOD instead of a callback
12637 function pointer. NOTE: this breaks binary compatibility with earlier
12638 versions of OpenSSL [engine].
12639 Adapt the nCipher code for these new conditions and add a card insertion
12644 * Enhance the general user interface with mechanisms to better support
12645 dialog box interfaces, application-defined prompts, the possibility
12646 to use defaults (for example default passwords from somewhere else)
12647 and interrupts/cancellations.
12651 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12652 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12656 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12657 tidy up some unnecessarily weird code in 'sk_new()').
12659 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12661 * Change the key loading routines for ENGINEs to use the same kind
12662 callback (pem_password_cb) as all other routines that need this
12667 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12668 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12669 than this minimum value is recommended.
12673 * New random seeder for OpenVMS, using the system process statistics
12674 that are easily reachable.
12678 * Windows apparently can't transparently handle global
12679 variables defined in DLLs. Initialisations such as:
12681 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12683 won't compile. This is used by the any applications that need to
12684 declare their own ASN1 modules. This was fixed by adding the option
12685 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12686 needed for static libraries under Win32.
12690 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12691 setting of purpose and trust fields. New X509_STORE trust and
12692 purpose functions and tidy up setting in other SSL functions.
12696 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12697 structure. These are inherited by X509_STORE_CTX when it is
12698 initialised. This allows various defaults to be set in the
12699 X509_STORE structure (such as flags for CRL checking and custom
12700 purpose or trust settings) for functions which only use X509_STORE_CTX
12701 internally such as S/MIME.
12703 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12704 trust settings if they are not set in X509_STORE. This allows X509_STORE
12705 purposes and trust (in S/MIME for example) to override any set by default.
12707 Add command line options for CRL checking to smime, s_client and s_server
12712 * Initial CRL based revocation checking. If the CRL checking flag(s)
12713 are set then the CRL is looked up in the X509_STORE structure and
12714 its validity and signature checked, then if the certificate is found
12715 in the CRL the verify fails with a revoked error.
12717 Various new CRL related callbacks added to X509_STORE_CTX structure.
12719 Command line options added to 'verify' application to support this.
12721 This needs some additional work, such as being able to handle multiple
12722 CRLs with different times, extension based lookup (rather than just
12723 by subject name) and ultimately more complete V2 CRL extension
12728 * Add a general user interface API (crypto/ui/). This is designed
12729 to replace things like des_read_password and friends (backward
12730 compatibility functions using this new API are provided).
12731 The purpose is to remove prompting functions from the DES code
12732 section as well as provide for prompting through dialog boxes in
12733 a window system and the like.
12737 * Add "ex_data" support to ENGINE so implementations can add state at a
12738 per-structure level rather than having to store it globally.
12742 * Make it possible for ENGINE structures to be copied when retrieved by
12743 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12744 This causes the "original" ENGINE structure to act like a template,
12745 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12746 operational state can be localised to each ENGINE structure, despite the
12747 fact they all share the same "methods". New ENGINE structures returned in
12748 this case have no functional references and the return value is the single
12749 structural reference. This matches the single structural reference returned
12750 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12755 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12756 needs to match any other type at all we need to manually clear the
12761 * Changes to the "openssl engine" utility to include;
12762 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12763 about an ENGINE's available control commands.
12764 - executing control commands from command line arguments using the
12765 '-pre' and '-post' switches. '-post' is only used if '-t' is
12766 specified and the ENGINE is successfully initialised. The syntax for
12767 the individual commands are colon-separated, for example;
12768 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12772 * New dynamic control command support for ENGINEs. ENGINEs can now
12773 declare their own commands (numbers), names (strings), descriptions,
12774 and input types for run-time discovery by calling applications. A
12775 subset of these commands are implicitly classed as "executable"
12776 depending on their input type, and only these can be invoked through
12777 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12778 can be based on user input, config files, etc). The distinction is
12779 that "executable" commands cannot return anything other than a boolean
12780 result and can only support numeric or string input, whereas some
12781 discoverable commands may only be for direct use through
12782 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12783 pointers, or other custom uses. The "executable" commands are to
12784 support parameterisations of ENGINE behaviour that can be
12785 unambiguously defined by ENGINEs and used consistently across any
12786 OpenSSL-based application. Commands have been added to all the
12787 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12788 control over shared-library paths without source code alterations.
12792 * Changed all ENGINE implementations to dynamically allocate their
12793 ENGINEs rather than declaring them statically. Apart from this being
12794 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12795 this also allows the implementations to compile without using the
12796 internal engine_int.h header.
12800 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12801 'const' value. Any code that should be able to modify a RAND_METHOD
12802 should already have non-const pointers to it (ie. they should only
12803 modify their own ones).
12807 * Made a variety of little tweaks to the ENGINE code.
12808 - "atalla" and "ubsec" string definitions were moved from header files
12809 to C code. "nuron" string definitions were placed in variables
12810 rather than hard-coded - allowing parameterisation of these values
12811 later on via ctrl() commands.
12812 - Removed unused "#if 0"'d code.
12813 - Fixed engine list iteration code so it uses ENGINE_free() to release
12814 structural references.
12815 - Constified the RAND_METHOD element of ENGINE structures.
12816 - Constified various get/set functions as appropriate and added
12817 missing functions (including a catch-all ENGINE_cpy that duplicates
12818 all ENGINE values onto a new ENGINE except reference counts/state).
12819 - Removed NULL parameter checks in get/set functions. Setting a method
12820 or function to NULL is a way of cancelling out a previously set
12821 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12822 and doesn't justify the extra error symbols and code.
12823 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12824 flags from engine_int.h to engine.h.
12825 - Changed prototypes for ENGINE handler functions (init(), finish(),
12826 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12830 * Implement binary inversion algorithm for BN_mod_inverse in addition
12831 to the algorithm using long division. The binary algorithm can be
12832 used only if the modulus is odd. On 32-bit systems, it is faster
12833 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12834 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12835 up to 450 bits. In 64-bit environments, the binary algorithm
12836 appears to be advantageous for much longer moduli; here we use it
12837 for moduli up to 2048 bits.
12841 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12842 could not support the combine flag in choice fields.
12846 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12847 extensions from a certificate request to the certificate.
12851 * Allow multiple 'certopt' and 'nameopt' options to be separated
12852 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12853 file: this allows the display of the certificate about to be
12854 signed to be customised, to allow certain fields to be included
12855 or excluded and extension details. The old system didn't display
12856 multicharacter strings properly, omitted fields not in the policy
12857 and couldn't display additional details such as extensions.
12861 * Function EC_POINTs_mul for multiple scalar multiplication
12862 of an arbitrary number of elliptic curve points
12863 \sum scalars[i]*points[i],
12864 optionally including the generator defined for the EC_GROUP:
12865 scalar*generator + \sum scalars[i]*points[i].
12867 EC_POINT_mul is a simple wrapper function for the typical case
12868 that the point list has just one item (besides the optional
12873 * First EC_METHODs for curves over GF(p):
12875 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12876 operations and provides various method functions that can also
12877 operate with faster implementations of modular arithmetic.
12879 EC_GFp_mont_method() reuses most functions that are part of
12880 EC_GFp_simple_method, but uses Montgomery arithmetic.
12882 *Bodo Moeller; point addition and point doubling
12883 implementation directly derived from source code provided by
12884 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12886 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12887 crypto/ec/ec_lib.c):
12889 Curves are EC_GROUP objects (with an optional group generator)
12890 based on EC_METHODs that are built into the library.
12892 Points are EC_POINT objects based on EC_GROUP objects.
12894 Most of the framework would be able to handle curves over arbitrary
12895 finite fields, but as there are no obvious types for fields other
12896 than GF(p), some functions are limited to that for now.
12900 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12901 that the file contains a complete HTTP response.
12905 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12906 change the def and num file printf format specifier from "%-40sXXX"
12907 to "%-39s XXX". The latter will always guarantee a space after the
12908 field while the former will cause them to run together if the field
12909 is 40 of more characters long.
12913 * Constify the cipher and digest 'method' functions and structures
12914 and modify related functions to take constant EVP_MD and EVP_CIPHER
12919 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12920 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12924 * Modify `EVP_Digest*()` routines so they now return values. Although the
12925 internal software routines can never fail additional hardware versions
12930 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12932 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12933 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12936 ERR_R_NESTED_ASN1_ERROR
12938 ERR_R_MISSING_ASN1_EOS
12939 were 4 .. 9, conflicting with
12940 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12942 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12943 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12945 Add new error code 'ERR_R_INTERNAL_ERROR'.
12949 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12954 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12955 sets the subject name for a new request or supersedes the
12956 subject name in a given request. Formats that can be parsed are
12957 'CN=Some Name, OU=myOU, C=IT'
12959 'CN=Some Name/OU=myOU/C=IT'.
12961 Add options '-batch' and '-verbose' to 'openssl req'.
12963 *Massimiliano Pala <madwolf@hackmasters.net>*
12965 * Introduce the possibility to access global variables through
12966 functions on platform were that's the best way to handle exporting
12967 global variables in shared libraries. To enable this functionality,
12968 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12969 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12970 is normally done by Configure or something similar).
12972 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12973 in the source file (foo.c) like this:
12975 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12976 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12978 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12979 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12981 OPENSSL_DECLARE_GLOBAL(int,foo);
12982 #define foo OPENSSL_GLOBAL_REF(foo)
12983 OPENSSL_DECLARE_GLOBAL(double,bar);
12984 #define bar OPENSSL_GLOBAL_REF(bar)
12986 The #defines are very important, and therefore so is including the
12987 header file everywhere where the defined globals are used.
12989 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12990 of ASN.1 items, but that structure is a bit different.
12992 The largest change is in util/mkdef.pl which has been enhanced with
12993 better and easier to understand logic to choose which symbols should
12994 go into the Windows .def files as well as a number of fixes and code
12995 cleanup (among others, algorithm keywords are now sorted
12996 lexicographically to avoid constant rewrites).
13000 * In BN_div() keep a copy of the sign of 'num' before writing the
13001 result to 'rm' because if rm==num the value will be overwritten
13002 and produce the wrong result if 'num' is negative: this caused
13003 problems with BN_mod() and BN_nnmod().
13007 * Function OCSP_request_verify(). This checks the signature on an
13008 OCSP request and verifies the signer certificate. The signer
13009 certificate is just checked for a generic purpose and OCSP request
13014 * Add OCSP_check_validity() function to check the validity of OCSP
13015 responses. OCSP responses are prepared in real time and may only
13016 be a few seconds old. Simply checking that the current time lies
13017 between thisUpdate and nextUpdate max reject otherwise valid responses
13018 caused by either OCSP responder or client clock inaccuracy. Instead
13019 we allow thisUpdate and nextUpdate to fall within a certain period of
13020 the current time. The age of the response can also optionally be
13021 checked. Two new options -validity_period and -status_age added to
13026 * If signature or public key algorithm is unrecognized print out its
13027 OID rather that just UNKNOWN.
13031 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13032 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13033 ID to be generated from the issuer certificate alone which can then be
13034 passed to OCSP_id_issuer_cmp().
13038 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13039 ASN1 modules to export functions returning ASN1_ITEM pointers
13040 instead of the ASN1_ITEM structures themselves. This adds several
13041 new macros which allow the underlying ASN1 function/structure to
13042 be accessed transparently. As a result code should not use ASN1_ITEM
13043 references directly (such as &X509_it) but instead use the relevant
13044 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13045 use of the new ASN1 code on platforms where exporting structures
13046 is problematical (for example in shared libraries) but exporting
13047 functions returning pointers to structures is not.
13051 * Add support for overriding the generation of SSL/TLS session IDs.
13052 These callbacks can be registered either in an SSL_CTX or per SSL.
13053 The purpose of this is to allow applications to control, if they wish,
13054 the arbitrary values chosen for use as session IDs, particularly as it
13055 can be useful for session caching in multiple-server environments. A
13056 command-line switch for testing this (and any client code that wishes
13057 to use such a feature) has been added to "s_server".
13059 *Geoff Thorpe, Lutz Jaenicke*
13061 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13062 of the form `#if defined(...) || defined(...) || ...` and
13063 `#if !defined(...) && !defined(...) && ...`. This also avoids
13064 the growing number of special cases it was previously handling.
13068 * Make all configuration macros available for application by making
13069 sure they are available in opensslconf.h, by giving them names starting
13070 with `OPENSSL_` to avoid conflicts with other packages and by making
13071 sure e_os2.h will cover all platform-specific cases together with
13073 Additionally, it is now possible to define configuration/platform-
13074 specific names (called "system identities"). In the C code, these
13075 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13076 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13077 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13082 * New option -set_serial to 'req' and 'x509' this allows the serial
13083 number to use to be specified on the command line. Previously self
13084 signed certificates were hard coded with serial number 0 and the
13085 CA options of 'x509' had to use a serial number in a file which was
13090 * New options to 'ca' utility to support V2 CRL entry extensions.
13091 Currently CRL reason, invalidity date and hold instruction are
13092 supported. Add new CRL extensions to V3 code and some new objects.
13096 * New function EVP_CIPHER_CTX_set_padding() this is used to
13097 disable standard block padding (aka PKCS#5 padding) in the EVP
13098 API, which was previously mandatory. This means that the data is
13099 not padded in any way and so the total length much be a multiple
13100 of the block size, otherwise an error occurs.
13104 * Initial (incomplete) OCSP SSL support.
13108 * New function OCSP_parse_url(). This splits up a URL into its host,
13109 port and path components: primarily to parse OCSP URLs. New -url
13110 option to ocsp utility.
13114 * New nonce behavior. The return value of OCSP_check_nonce() now
13115 reflects the various checks performed. Applications can decide
13116 whether to tolerate certain situations such as an absent nonce
13117 in a response when one was present in a request: the ocsp application
13118 just prints out a warning. New function OCSP_add1_basic_nonce()
13119 this is to allow responders to include a nonce in a response even if
13120 the request is nonce-less.
13124 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13125 skipped when using openssl x509 multiple times on a single input file,
13126 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13130 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13131 set string type: to handle setting ASN1_TIME structures. Fix ca
13132 utility to correctly initialize revocation date of CRLs.
13136 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13137 the clients preferred ciphersuites and rather use its own preferences.
13138 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13139 Internet Explorer by ensuring unchanged hash method during stepup.
13140 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13144 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13145 to aes and add a new 'exist' option to print out symbols that don't
13150 * Additional options to ocsp utility to allow flags to be set and
13151 additional certificates supplied.
13155 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13156 OCSP client a number of certificate to only verify the response
13161 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13162 handle the new API. Currently only ECB, CBC modes supported. Add new
13165 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13166 Encryption Standard (AES) Ciphersuites for Transport Layer
13167 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13168 not enabled by default and were not part of the "ALL" ciphersuite
13169 alias because they were not yet official; they could be
13170 explicitly requested by specifying the "AESdraft" ciphersuite
13171 group alias. In the final release of OpenSSL 0.9.7, the group
13172 alias is called "AES" and is part of "ALL".)
13174 *Ben Laurie, Steve Henson, Bodo Moeller*
13176 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13177 request to response.
13181 * Functions for OCSP responders. OCSP_request_onereq_count(),
13182 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13183 extract information from a certificate request. OCSP_response_create()
13184 creates a response and optionally adds a basic response structure.
13185 OCSP_basic_add1_status() adds a complete single response to a basic
13186 response and returns the OCSP_SINGLERESP structure just added (to allow
13187 extensions to be included for example). OCSP_basic_add1_cert() adds a
13188 certificate to a basic response and OCSP_basic_sign() signs a basic
13189 response with various flags. New helper functions ASN1_TIME_check()
13190 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13191 (converts ASN1_TIME to GeneralizedTime).
13195 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13196 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13197 structure from a certificate. X509_pubkey_digest() digests the public_key
13198 contents: this is used in various key identifiers.
13202 * Make sk_sort() tolerate a NULL argument.
13204 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13206 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13207 passed by the function are trusted implicitly. If any of them signed the
13208 response then it is assumed to be valid and is not verified.
13212 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13213 to data. This was previously part of the PKCS7 ASN1 code. This
13214 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13215 *Steve Henson, reported by Kenneth R. Robinette
13216 <support@securenetterm.com>*
13218 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13219 routines: without these tracing memory leaks is very painful.
13220 Fix leaks in PKCS12 and PKCS7 routines.
13224 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13225 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13226 effectively meant GeneralizedTime would never be used. Now it
13227 is initialised to -1 but X509_time_adj() now has to check the value
13228 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13229 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13230 *Steve Henson, reported by Kenneth R. Robinette
13231 <support@securenetterm.com>*
13233 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13234 result in a zero length in the ASN1_INTEGER structure which was
13235 not consistent with the structure when d2i_ASN1_INTEGER() was used
13236 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13237 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13238 where it did not print out a minus for negative ASN1_INTEGER.
13242 * Add summary printout to ocsp utility. The various functions which
13243 convert status values to strings have been renamed to:
13244 OCSP_response_status_str(), OCSP_cert_status_str() and
13245 OCSP_crl_reason_str() and are no longer static. New options
13246 to verify nonce values and to disable verification. OCSP response
13247 printout format cleaned up.
13251 * Add additional OCSP certificate checks. These are those specified
13252 in RFC2560. This consists of two separate checks: the CA of the
13253 certificate being checked must either be the OCSP signer certificate
13254 or the issuer of the OCSP signer certificate. In the latter case the
13255 OCSP signer certificate must contain the OCSP signing extended key
13256 usage. This check is performed by attempting to match the OCSP
13257 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13258 in the OCSP_CERTID structures of the response.
13262 * Initial OCSP certificate verification added to OCSP_basic_verify()
13263 and related routines. This uses the standard OpenSSL certificate
13264 verify routines to perform initial checks (just CA validity) and
13265 to obtain the certificate chain. Then additional checks will be
13266 performed on the chain. Currently the root CA is checked to see
13267 if it is explicitly trusted for OCSP signing. This is used to set
13268 a root CA as a global signing root: that is any certificate that
13269 chains to that CA is an acceptable OCSP signing certificate.
13273 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13274 extensions from a separate configuration file.
13275 As when reading extensions from the main configuration file,
13276 the '-extensions ...' option may be used for specifying the
13279 *Massimiliano Pala <madwolf@comune.modena.it>*
13281 * New OCSP utility. Allows OCSP requests to be generated or
13282 read. The request can be sent to a responder and the output
13283 parsed, outputted or printed in text form. Not complete yet:
13284 still needs to check the OCSP response validity.
13288 * New subcommands for 'openssl ca':
13289 `openssl ca -status <serial>` prints the status of the cert with
13290 the given serial number (according to the index file).
13291 `openssl ca -updatedb` updates the expiry status of certificates
13294 *Massimiliano Pala <madwolf@comune.modena.it>*
13296 * New '-newreq-nodes' command option to CA.pl. This is like
13297 '-newreq', but calls 'openssl req' with the '-nodes' option
13298 so that the resulting key is not encrypted.
13300 *Damien Miller <djm@mindrot.org>*
13302 * New configuration for the GNU Hurd.
13304 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13306 * Initial code to implement OCSP basic response verify. This
13307 is currently incomplete. Currently just finds the signer's
13308 certificate and verifies the signature on the response.
13312 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13313 value of OPENSSLDIR. This is available via the new '-d' option
13314 to 'openssl version', and is also included in 'openssl version -a'.
13318 * Allowing defining memory allocation callbacks that will be given
13319 file name and line number information in additional arguments
13320 (a `const char*` and an int). The basic functionality remains, as
13321 well as the original possibility to just replace malloc(),
13322 realloc() and free() by functions that do not know about these
13323 additional arguments. To register and find out the current
13324 settings for extended allocation functions, the following
13325 functions are provided:
13327 CRYPTO_set_mem_ex_functions
13328 CRYPTO_set_locked_mem_ex_functions
13329 CRYPTO_get_mem_ex_functions
13330 CRYPTO_get_locked_mem_ex_functions
13332 These work the same way as CRYPTO_set_mem_functions and friends.
13333 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13334 extended allocation function is enabled.
13335 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13336 a conventional allocation function is enabled.
13338 *Richard Levitte, Bodo Moeller*
13340 * Finish off removing the remaining LHASH function pointer casts.
13341 There should no longer be any prototype-casting required when using
13342 the LHASH abstraction, and any casts that remain are "bugs". See
13343 the callback types and macros at the head of lhash.h for details
13344 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13348 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13349 If /dev/[u]random devices are not available or do not return enough
13350 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13352 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13353 /etc/entropy will be queried once each in this sequence, querying stops
13354 when enough entropy was collected without querying more sockets.
13358 * Change the Unix RAND_poll() variant to be able to poll several
13359 random devices, as specified by DEVRANDOM, until a sufficient amount
13360 of data has been collected. We spend at most 10 ms on each file
13361 (select timeout) and read in non-blocking mode. DEVRANDOM now
13362 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13363 (previously it was just the string "/dev/urandom"), so on typical
13364 platforms the 10 ms delay will never occur.
13365 Also separate out the Unix variant to its own file, rand_unix.c.
13366 For VMS, there's a currently-empty rand_vms.c.
13370 * Move OCSP client related routines to ocsp_cl.c. These
13371 provide utility functions which an application needing
13372 to issue a request to an OCSP responder and analyse the
13373 response will typically need: as opposed to those which an
13374 OCSP responder itself would need which will be added later.
13376 OCSP_request_sign() signs an OCSP request with an API similar
13377 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13378 response. OCSP_response_get1_basic() extracts basic response
13379 from response. OCSP_resp_find_status(): finds and extracts status
13380 information from an OCSP_CERTID structure (which will be created
13381 when the request structure is built). These are built from lower
13382 level functions which work on OCSP_SINGLERESP structures but
13383 won't normally be used unless the application wishes to examine
13384 extensions in the OCSP response for example.
13386 Replace nonce routines with a pair of functions.
13387 OCSP_request_add1_nonce() adds a nonce value and optionally
13388 generates a random value. OCSP_check_nonce() checks the
13389 validity of the nonce in an OCSP response.
13393 * Change function OCSP_request_add() to OCSP_request_add0_id().
13394 This doesn't copy the supplied OCSP_CERTID and avoids the
13395 need to free up the newly created id. Change return type
13396 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13397 This can then be used to add extensions to the request.
13398 Deleted OCSP_request_new(), since most of its functionality
13399 is now in OCSP_REQUEST_new() (and the case insensitive name
13400 clash) apart from the ability to set the request name which
13401 will be added elsewhere.
13405 * Update OCSP API. Remove obsolete extensions argument from
13406 various functions. Extensions are now handled using the new
13407 OCSP extension code. New simple OCSP HTTP function which
13408 can be used to send requests and parse the response.
13412 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13413 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13414 uses the special reorder version of SET OF to sort the attributes
13415 and reorder them to match the encoded order. This resolves a long
13416 standing problem: a verify on a PKCS7 structure just after signing
13417 it used to fail because the attribute order did not match the
13418 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13419 it uses the received order. This is necessary to tolerate some broken
13420 software that does not order SET OF. This is handled by encoding
13421 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13422 to produce the required SET OF.
13426 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13427 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13428 files to get correct declarations of the ASN.1 item variables.
13432 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13433 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13434 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13435 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13436 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13437 ASN1_ITEM and no wrapper functions.
13441 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13442 replace the old function pointer based I/O routines. Change most of
13443 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13447 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13448 lines, recognize more "algorithms" that can be deselected, and make
13449 it complain about algorithm deselection that isn't recognised.
13453 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13454 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13455 to use new functions. Add NO_ASN1_OLD which can be set to remove
13456 some old style ASN1 functions: this can be used to determine if old
13457 code will still work when these eventually go away.
13461 * New extension functions for OCSP structures, these follow the
13462 same conventions as certificates and CRLs.
13466 * New function X509V3_add1_i2d(). This automatically encodes and
13467 adds an extension. Its behaviour can be customised with various
13468 flags to append, replace or delete. Various wrappers added for
13469 certificates and CRLs.
13473 * Fix to avoid calling the underlying ASN1 print routine when
13474 an extension cannot be parsed. Correct a typo in the
13475 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13479 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13480 entries for variables.
13484 * Add functionality to `apps/openssl.c` for detecting locking
13485 problems: As the program is single-threaded, all we have
13486 to do is register a locking callback using an array for
13487 storing which locks are currently held by the program.
13491 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13492 SSL_get_ex_data_X509_STORE_idx(), which is used in
13493 ssl_verify_cert_chain() and thus can be called at any time
13494 during TLS/SSL handshakes so that thread-safety is essential.
13495 Unfortunately, the ex_data design is not at all suited
13496 for multi-threaded use, so it probably should be abolished.
13500 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13502 *Broadcom, tweaked and integrated by Geoff Thorpe*
13504 * Move common extension printing code to new function
13505 X509V3_print_extensions(). Reorganise OCSP print routines and
13506 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13510 * New function X509_signature_print() to remove duplication in some
13515 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13516 set (this was treated exactly the same as SET OF previously). This
13517 is used to reorder the STACK representing the structure to match the
13518 encoding. This will be used to get round a problem where a PKCS7
13519 structure which was signed could not be verified because the STACK
13520 order did not reflect the encoded order.
13524 * Reimplement the OCSP ASN1 module using the new code.
13528 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13529 for its ASN1 operations. The old style function pointers still exist
13530 for now but they will eventually go away.
13534 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13535 completely replaces the old ASN1 functionality with a table driven
13536 encoder and decoder which interprets an ASN1_ITEM structure describing
13537 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13538 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13539 has also been converted to the new form.
13543 * Change BN_mod_exp_recp so that negative moduli are tolerated
13544 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13545 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13546 for negative moduli.
13550 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13551 of not touching the result's sign bit.
13555 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13560 * Changed the LHASH code to use prototypes for callbacks, and created
13561 macros to declare and implement thin (optionally static) functions
13562 that provide type-safety and avoid function pointer casting for the
13563 type-specific callbacks.
13567 * Added Kerberos Cipher Suites to be used with TLS, as written in
13569 *Veers Staats <staatsvr@asc.hpc.mil>,
13570 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13572 * Reformat the FAQ so the different questions and answers can be divided
13573 in sections depending on the subject.
13577 * Have the zlib compression code load ZLIB.DLL dynamically under
13582 * New function BN_mod_sqrt for computing square roots modulo a prime
13583 (using the probabilistic Tonelli-Shanks algorithm unless
13584 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13585 be handled deterministically).
13587 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13589 * Make BN_mod_inverse faster by explicitly handling small quotients
13590 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13591 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13595 * New function BN_kronecker.
13599 * Fix BN_gcd so that it works on negative inputs; the result is
13600 positive unless both parameters are zero.
13601 Previously something reasonably close to an infinite loop was
13602 possible because numbers could be growing instead of shrinking
13603 in the implementation of Euclid's algorithm.
13607 * Fix BN_is_word() and BN_is_one() macros to take into account the
13608 sign of the number in question.
13610 Fix BN_is_word(a,w) to work correctly for w == 0.
13612 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13613 because its test if the absolute value of 'a' equals 'w'.
13614 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13615 it exists mostly for use in the implementations of BN_is_zero(),
13616 BN_is_one(), and BN_is_word().
13620 * New function BN_swap.
13624 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13625 the exponentiation functions are more likely to produce reasonable
13626 results on negative inputs.
13630 * Change BN_mod_mul so that the result is always non-negative.
13631 Previously, it could be negative if one of the factors was negative;
13632 I don't think anyone really wanted that behaviour.
13636 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13637 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13638 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13639 and add new functions:
13648 BN_mod_lshift1_quick
13650 BN_mod_lshift_quick
13652 These functions always generate non-negative results.
13654 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13655 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13657 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13658 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13659 be reduced modulo `m`.
13661 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13664 The following entry accidentally appeared in the CHANGES file
13665 distributed with OpenSSL 0.9.7. The modifications described in
13666 it do *not* apply to OpenSSL 0.9.7.
13668 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13669 was actually never needed) and in BN_mul(). The removal in BN_mul()
13670 required a small change in bn_mul_part_recursive() and the addition
13671 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13672 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13673 bn_sub_words() and bn_add_words() except they take arrays with
13679 * In 'openssl passwd', verify passwords read from the terminal
13680 unless the '-salt' option is used (which usually means that
13681 verification would just waste user's time since the resulting
13682 hash is going to be compared with some given password hash)
13683 or the new '-noverify' option is used.
13685 This is an incompatible change, but it does not affect
13686 non-interactive use of 'openssl passwd' (passwords on the command
13687 line, '-stdin' option, '-in ...' option) and thus should not
13688 cause any problems.
13692 * Remove all references to RSAref, since there's no more need for it.
13696 * Make DSO load along a path given through an environment variable
13697 (SHLIB_PATH) with shl_load().
13701 * Constify the ENGINE code as a result of BIGNUM constification.
13702 Also constify the RSA code and most things related to it. In a
13703 few places, most notable in the depth of the ASN.1 code, ugly
13704 casts back to non-const were required (to be solved at a later
13709 * Make it so the openssl application has all engines loaded by default.
13713 * Constify the BIGNUM routines a little more.
13717 * Add the following functions:
13719 ENGINE_load_cswift()
13721 ENGINE_load_atalla()
13722 ENGINE_load_nuron()
13723 ENGINE_load_builtin_engines()
13725 That way, an application can itself choose if external engines that
13726 are built-in in OpenSSL shall ever be used or not. The benefit is
13727 that applications won't have to be linked with libdl or other dso
13728 libraries unless it's really needed.
13730 Changed 'openssl engine' to load all engines on demand.
13731 Changed the engine header files to avoid the duplication of some
13732 declarations (they differed!).
13736 * 'openssl engine' can now list capabilities.
13740 * Better error reporting in 'openssl engine'.
13744 * Never call load_dh_param(NULL) in s_server.
13748 * Add engine application. It can currently list engines by name and
13749 identity, and test if they are actually available.
13753 * Improve RPM specification file by forcing symbolic linking and making
13754 sure the installed documentation is also owned by root.root.
13756 *Damien Miller <djm@mindrot.org>*
13758 * Give the OpenSSL applications more possibilities to make use of
13759 keys (public as well as private) handled by engines.
13763 * Add OCSP code that comes from CertCo.
13767 * Add VMS support for the Rijndael code.
13771 * Added untested support for Nuron crypto accelerator.
13775 * Add support for external cryptographic devices. This code was
13776 previously distributed separately as the "engine" branch.
13778 *Geoff Thorpe, Richard Levitte*
13780 * Rework the filename-translation in the DSO code. It is now possible to
13781 have far greater control over how a "name" is turned into a filename
13782 depending on the operating environment and any oddities about the
13783 different shared library filenames on each system.
13787 * Support threads on FreeBSD-elf in Configure.
13791 * Fix for SHA1 assembly problem with MASM: it produces
13792 warnings about corrupt line number information when assembling
13793 with debugging information. This is caused by the overlapping
13796 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13799 NCONF_get_number() has no error checking at all. As a replacement,
13800 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13801 promoted strongly. The old NCONF_get_number is kept around for
13802 binary backward compatibility.
13803 Make it possible for methods to load from something other than a BIO,
13804 by providing a function pointer that is given a name instead of a BIO.
13805 For example, this could be used to load configuration data from an
13810 * Fix for non blocking accept BIOs. Added new I/O special reason
13811 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13812 with non blocking I/O was not possible because no retry code was
13813 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13818 * Added the beginnings of Rijndael support.
13822 * Fix for bug in DirectoryString mask setting. Add support for
13823 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13824 to allow certificate printing to more controllable, additional
13825 'certopt' option to 'x509' to allow new printing options to be
13830 * Clean old EAY MD5 hack from e_os.h.
13834 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13836 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13837 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13839 *Joe Orton, Steve Henson*
13841 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13843 * Fix additional bug revealed by the NISCC test suite:
13845 Stop bug triggering large recursion when presented with
13846 certain ASN.1 tags ([CVE-2003-0851])
13850 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13852 * Fix various bugs revealed by running the NISCC test suite:
13854 Stop out of bounds reads in the ASN1 code when presented with
13855 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13857 If verify callback ignores invalid public key errors don't try to check
13858 certificate signature with the NULL public key.
13862 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13863 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13868 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13869 extra data after the compression methods not only for TLS 1.0
13870 but also for SSL 3.0 (as required by the specification).
13872 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13874 * Change X509_certificate_type() to mark the key as exported/exportable
13875 when it's 512 *bits* long, not 512 bytes.
13879 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13881 * Countermeasure against the Klima-Pokorny-Rosa extension of
13882 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13883 a protocol version number mismatch like a decryption error
13884 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13888 * Turn on RSA blinding by default in the default implementation
13889 to avoid a timing attack. Applications that don't want it can call
13890 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13891 They would be ill-advised to do so in most cases.
13893 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13895 * Change RSA blinding code so that it works when the PRNG is not
13896 seeded (in this case, the secret RSA exponent is abused as
13897 an unpredictable seed -- if it is not unpredictable, there
13898 is no point in blinding anyway). Make RSA blinding thread-safe
13899 by remembering the creator's thread ID in rsa->blinding and
13900 having all other threads use local one-time blinding factors
13901 (this requires more computation than sharing rsa->blinding, but
13902 avoids excessive locking; and if an RSA object is not shared
13903 between threads, blinding will still be very fast).
13907 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13909 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13910 via timing by performing a MAC computation even if incorrect
13911 block cipher padding has been found. This is a countermeasure
13912 against active attacks where the attacker has to distinguish
13913 between bad padding and a MAC verification error. ([CVE-2003-0078])
13915 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13916 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13917 Martin Vuagnoux (EPFL, Ilion)*
13919 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13921 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13922 memory from its contents. This is done with a counter that will
13923 place alternating values in each byte. This can be used to solve
13924 two issues: 1) the removal of calls to memset() by highly optimizing
13925 compilers, and 2) cleansing with other values than 0, since those can
13926 be read through on certain media, for example a swap space on disk.
13930 * Bugfix: client side session caching did not work with external caching,
13931 because the session->cipher setting was not restored when reloading
13932 from the external cache. This problem was masked, when
13933 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13934 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13938 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13939 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13941 *Zeev Lieber <zeev-l@yahoo.com>*
13943 * Undo an undocumented change introduced in 0.9.6e which caused
13944 repeated calls to OpenSSL_add_all_ciphers() and
13945 OpenSSL_add_all_digests() to be ignored, even after calling
13950 * Change the default configuration reader to deal with last line not
13951 being properly terminated.
13955 * Change X509_NAME_cmp() so it applies the special rules on handling
13956 DN values that are of type PrintableString, as well as RDNs of type
13957 emailAddress where the value has the type ia5String.
13959 *stefank@valicert.com via Richard Levitte*
13961 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13962 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13963 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13964 the bitwise-OR of the two for use by the majority of applications
13965 wanting this behaviour, and update the docs. The documented
13966 behaviour and actual behaviour were inconsistent and had been
13967 changing anyway, so this is more a bug-fix than a behavioural
13970 *Geoff Thorpe, diagnosed by Nadav Har'El*
13972 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13973 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13977 * Fix initialization code race conditions in
13978 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13979 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13980 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13981 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13982 ssl2_get_cipher_by_char(),
13983 ssl3_get_cipher_by_char().
13985 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13987 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13988 the cached sessions are flushed, as the remove_cb() might use ex_data
13989 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13990 (see [openssl.org #212]).
13992 *Geoff Thorpe, Lutz Jaenicke*
13994 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13995 length, instead of the encoding length to d2i_ASN1_OBJECT.
13999 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14001 * [In 0.9.6g-engine release:]
14002 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14004 *Lynn Gazis <lgazis@rainbow.com>*
14006 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14008 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14009 and get fix the header length calculation.
14010 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14011 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14013 * Use proper error handling instead of 'assertions' in buffer
14014 overflow checks added in 0.9.6e. This prevents DoS (the
14015 assertions could call abort()).
14017 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14019 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14021 * Add various sanity checks to asn1_get_length() to reject
14022 the ASN1 length bytes if they exceed sizeof(long), will appear
14023 negative or the content length exceeds the length of the
14026 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14028 * Fix cipher selection routines: ciphers without encryption had no flags
14029 for the cipher strength set and where therefore not handled correctly
14030 by the selection routines (PR #130).
14034 * Fix EVP_dsa_sha macro.
14039 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14040 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14041 that was added in OpenSSL 0.9.6d.
14043 As the countermeasure turned out to be incompatible with some
14044 broken SSL implementations, the new option is part of SSL_OP_ALL.
14045 SSL_OP_ALL is usually employed when compatibility with weird SSL
14046 implementations is desired (e.g. '-bugs' option to 's_client' and
14047 's_server'), so the new option is automatically set in many
14052 * Changes in security patch:
14054 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14055 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14056 Air Force Materiel Command, USAF, under agreement number
14059 * Add various sanity checks to asn1_get_length() to reject
14060 the ASN1 length bytes if they exceed sizeof(long), will appear
14061 negative or the content length exceeds the length of the
14062 supplied buffer. ([CVE-2002-0659])
14064 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14066 * Assertions for various potential buffer overflows, not known to
14067 happen in practice.
14069 *Ben Laurie (CHATS)*
14071 * Various temporary buffers to hold ASCII versions of integers were
14072 too small for 64 bit platforms. ([CVE-2002-0655])
14073 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14075 * Remote buffer overflow in SSL3 protocol - an attacker could
14076 supply an oversized session ID to a client. ([CVE-2002-0656])
14078 *Ben Laurie (CHATS)*
14080 * Remote buffer overflow in SSL2 protocol - an attacker could
14081 supply an oversized client master key. ([CVE-2002-0656])
14083 *Ben Laurie (CHATS)*
14085 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14087 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14088 encoded as NULL) with id-dsa-with-sha1.
14090 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14092 * Check various `X509_...()` return values in `apps/req.c`.
14094 *Nils Larsch <nla@trustcenter.de>*
14096 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14097 an end-of-file condition would erroneously be flagged, when the CRLF
14098 was just at the end of a processed block. The bug was discovered when
14099 processing data through a buffering memory BIO handing the data to a
14100 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14101 <ptsekov@syntrex.com> and Nedelcho Stanev.
14105 * Implement a countermeasure against a vulnerability recently found
14106 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14107 before application data chunks to avoid the use of known IVs
14108 with data potentially chosen by the attacker.
14112 * Fix length checks in ssl3_get_client_hello().
14116 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14117 to prevent ssl3_read_internal() from incorrectly assuming that
14118 ssl3_read_bytes() found application data while handshake
14119 processing was enabled when in fact s->s3->in_read_app_data was
14120 merely automatically cleared during the initial handshake.
14122 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14124 * Fix object definitions for Private and Enterprise: they were not
14125 recognized in their shortname (=lowercase) representation. Extend
14126 obj_dat.pl to issue an error when using undefined keywords instead
14127 of silently ignoring the problem (Svenning Sorensen
14128 <sss@sss.dnsalias.net>).
14132 * Fix DH_generate_parameters() so that it works for 'non-standard'
14133 generators, i.e. generators other than 2 and 5. (Previously, the
14134 code did not properly initialise the 'add' and 'rem' values to
14135 BN_generate_prime().)
14137 In the new general case, we do not insist that 'generator' is
14138 actually a primitive root: This requirement is rather pointless;
14139 a generator of the order-q subgroup is just as good, if not
14144 * Map new X509 verification errors to alerts. Discovered and submitted by
14145 Tom Wu <tom@arcot.com>.
14149 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14150 returning non-zero before the data has been completely received
14151 when using non-blocking I/O.
14153 *Bodo Moeller; problem pointed out by John Hughes*
14155 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14157 *Ben Laurie, Lutz Jaenicke*
14159 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14160 Yoram Zahavi <YoramZ@gilian.com>).
14164 * Add information about CygWin 1.3 and on, and preserve proper
14165 configuration for the versions before that.
14167 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14169 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14170 check whether we deal with a copy of a session and do not delete from
14171 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14172 <izhar@checkpoint.com>.
14176 * Do not store session data into the internal session cache, if it
14177 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14178 flag is set). Proposed by Aslam <aslam@funk.com>.
14182 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14187 * [In 0.9.6d-engine release:]
14188 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14190 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14192 * Add the configuration target linux-s390x.
14194 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14196 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14197 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14198 variable as an indication that a ClientHello message has been
14199 received. As the flag value will be lost between multiple
14200 invocations of ssl3_accept when using non-blocking I/O, the
14201 function may not be aware that a handshake has actually taken
14202 place, thus preventing a new session from being added to the
14205 To avoid this problem, we now set s->new_session to 2 instead of
14206 using a local variable.
14208 *Lutz Jaenicke, Bodo Moeller*
14210 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14211 if the SSL_R_LENGTH_MISMATCH error is detected.
14213 *Geoff Thorpe, Bodo Moeller*
14215 * New 'shared_ldflag' column in Configure platform table.
14219 * Fix EVP_CIPHER_mode macro.
14221 *"Dan S. Camper" <dan@bti.net>*
14223 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14224 type, we must throw them away by setting rr->length to 0.
14226 *D P Chang <dpc@qualys.com>*
14228 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14230 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14231 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14232 worked incorrectly for those cases where range = `10..._2` and
14233 `3*range` is two bits longer than range.)
14237 * Only add signing time to PKCS7 structures if it is not already
14242 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14243 OBJ_ld_ce should be OBJ_id_ce.
14244 Also some ip-pda OIDs in crypto/objects/objects.txt were
14245 incorrect (cf. RFC 3039).
14247 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14249 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14250 returns early because it has nothing to do.
14252 *Andy Schneider <andy.schneider@bjss.co.uk>*
14254 * [In 0.9.6c-engine release:]
14255 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14257 *Andy Schneider <andy.schneider@bjss.co.uk>*
14259 * [In 0.9.6c-engine release:]
14260 Add support for Cryptographic Appliance's keyserver technology.
14261 (Use engine 'keyclient')
14263 *Cryptographic Appliances and Geoff Thorpe*
14265 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14266 is called via tools/c89.sh because arguments have to be
14267 rearranged (all '-L' options must appear before the first object
14270 *Richard Shapiro <rshapiro@abinitio.com>*
14272 * [In 0.9.6c-engine release:]
14273 Add support for Broadcom crypto accelerator cards, backported
14276 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14278 * [In 0.9.6c-engine release:]
14279 Add support for SureWare crypto accelerator cards from
14280 Baltimore Technologies. (Use engine 'sureware')
14282 *Baltimore Technologies and Mark Cox*
14284 * [In 0.9.6c-engine release:]
14285 Add support for crypto accelerator cards from Accelerated
14286 Encryption Processing, www.aep.ie. (Use engine 'aep')
14288 *AEP Inc. and Mark Cox*
14290 * Add a configuration entry for gcc on UnixWare.
14292 *Gary Benson <gbenson@redhat.com>*
14294 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14295 messages are stored in a single piece (fixed-length part and
14296 variable-length part combined) and fix various bugs found on the way.
14300 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14301 instead. BIO_gethostbyname() does not know what timeouts are
14302 appropriate, so entries would stay in cache even when they have
14304 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14306 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14307 faced with a pathologically small ClientHello fragment that does
14308 not contain client_version: Instead of aborting with an error,
14309 simply choose the highest available protocol version (i.e.,
14310 TLS 1.0 unless it is disabled). In practice, ClientHello
14311 messages are never sent like this, but this change gives us
14312 strictly correct behaviour at least for TLS.
14316 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14317 never resets s->method to s->ctx->method when called from within
14318 one of the SSL handshake functions.
14320 *Bodo Moeller; problem pointed out by Niko Baric*
14322 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14323 (sent using the client's version number) if client_version is
14324 smaller than the protocol version in use. Also change
14325 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14326 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14327 the client will at least see that alert.
14331 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14336 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14337 client receives HelloRequest while in a handshake.
14339 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14341 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14342 should end in 'break', not 'goto end' which circumvents various
14343 cleanups done in state SSL_ST_OK. But session related stuff
14344 must be disabled for SSL_ST_OK in the case that we just sent a
14347 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14348 before just sending a HelloRequest.
14350 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14352 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14353 reveal whether illegal block cipher padding was found or a MAC
14354 verification error occurred. (Neither SSLerr() codes nor alerts
14355 are directly visible to potential attackers, but the information
14356 may leak via logfiles.)
14358 Similar changes are not required for the SSL 2.0 implementation
14359 because the number of padding bytes is sent in clear for SSL 2.0,
14360 and the extra bytes are just ignored. However ssl/s2_pkt.c
14361 failed to verify that the purported number of padding bytes is in
14366 * Add OpenUNIX-8 support including shared libraries
14367 (Boyd Lynn Gerber <gerberb@zenez.com>).
14371 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14372 'wristwatch attack' using huge encoding parameters (cf.
14373 James H. Manger's CRYPTO 2001 paper). Note that the
14374 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14375 encoding parameters and hence was not vulnerable.
14379 * BN_sqr() bug fix.
14381 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14383 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14384 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14385 followed by modular reduction.
14387 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14389 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14390 equivalent based on BN_pseudo_rand() instead of BN_rand().
14394 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14395 This function was broken, as the check for a new client hello message
14396 to handle SGC did not allow these large messages.
14397 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14401 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14405 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14406 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14410 * Rework the configuration and shared library support for Tru64 Unix.
14411 The configuration part makes use of modern compiler features and
14412 still retains old compiler behavior for those that run older versions
14413 of the OS. The shared library support part includes a variant that
14414 uses the RPATH feature, and is available through the special
14415 configuration target "alpha-cc-rpath", which will never be selected
14418 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14420 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14421 with the same message size as in ssl3_get_certificate_request().
14422 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14423 messages might inadvertently be reject as too long.
14425 *Petr Lampa <lampa@fee.vutbr.cz>*
14427 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14431 * Modified SSL library such that the verify_callback that has been set
14432 specifically for an SSL object with SSL_set_verify() is actually being
14433 used. Before the change, a verify_callback set with this function was
14434 ignored and the verify_callback() set in the SSL_CTX at the time of
14435 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14436 to allow the necessary settings.
14440 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14441 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14442 done automatically (in contradiction to the requirements of the C
14443 standard). This made problems when used from OpenSSH.
14447 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14448 dh->length and always used
14450 BN_rand_range(priv_key, dh->p).
14452 BN_rand_range() is not necessary for Diffie-Hellman, and this
14453 specific range makes Diffie-Hellman unnecessarily inefficient if
14454 dh->length (recommended exponent length) is much smaller than the
14455 length of dh->p. We could use BN_rand_range() if the order of
14456 the subgroup was stored in the DH structure, but we only have
14461 BN_rand(priv_key, l, ...)
14463 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14470 RSA_eay_public_encrypt
14471 RSA_eay_private_decrypt
14472 RSA_eay_private_encrypt (signing)
14473 RSA_eay_public_decrypt (signature verification)
14475 (default implementations for RSA_public_encrypt,
14476 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14477 always reject numbers >= n.
14481 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14482 to synchronize access to 'locking_thread'. This is necessary on
14483 systems where access to 'locking_thread' (an 'unsigned long'
14484 variable) is not atomic.
14488 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14489 *before* setting the 'crypto_lock_rand' flag. The previous code had
14490 a race condition if 0 is a valid thread ID.
14492 *Travis Vitek <vitek@roguewave.com>*
14494 * Add support for shared libraries under Irix.
14496 *Albert Chin-A-Young <china@thewrittenword.com>*
14498 * Add configuration option to build on Linux on both big-endian and
14499 little-endian MIPS.
14501 *Ralf Baechle <ralf@uni-koblenz.de>*
14503 * Add the possibility to create shared libraries on HP-UX.
14507 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14509 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14510 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14511 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14512 PRNG state recovery was possible based on the output of
14513 one PRNG request appropriately sized to gain knowledge on
14514 'md' followed by enough consecutive 1-byte PRNG requests
14515 to traverse all of 'state'.
14517 1. When updating 'md_local' (the current thread's copy of 'md')
14518 during PRNG output generation, hash all of the previous
14519 'md_local' value, not just the half used for PRNG output.
14521 2. Make the number of bytes from 'state' included into the hash
14522 independent from the number of PRNG bytes requested.
14524 The first measure alone would be sufficient to avoid
14525 Markku-Juhani's attack. (Actually it had never occurred
14526 to me that the half of 'md_local' used for chaining was the
14527 half from which PRNG output bytes were taken -- I had always
14528 assumed that the secret half would be used.) The second
14529 measure makes sure that additional data from 'state' is never
14530 mixed into 'md_local' in small portions; this heuristically
14531 further strengthens the PRNG.
14535 * Fix crypto/bn/asm/mips3.s.
14539 * When only the key is given to "enc", the IV is undefined. Print out
14540 an error message in this case.
14544 * Handle special case when X509_NAME is empty in X509 printing routines.
14548 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14549 positive and less than q.
14553 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14554 used: it isn't thread safe and the add_lock_callback should handle
14557 *Paul Rose <Paul.Rose@bridge.com>*
14559 * Verify that incoming data obeys the block size in
14560 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14566 *Ulf Möller, Bodo Möller*
14568 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14569 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14570 when fixing the server behaviour for backwards-compatible 'client
14571 hello' messages. (Note that the attack is impractical against
14572 SSL 3.0 and TLS 1.0 anyway because length and version checking
14573 means that the probability of guessing a valid ciphertext is
14574 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14577 Before 0.9.5, the countermeasure (hide the error by generating a
14578 random 'decryption result') did not work properly because
14579 ERR_clear_error() was missing, meaning that SSL_get_error() would
14580 detect the supposedly ignored error.
14582 Both problems are now fixed.
14586 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14587 (previously it was 1024).
14591 * Fix for compatibility mode trust settings: ignore trust settings
14592 unless some valid trust or reject settings are present.
14596 * Fix for blowfish EVP: its a variable length cipher.
14600 * Fix various bugs related to DSA S/MIME verification. Handle missing
14601 parameters in DSA public key structures and return an error in the
14602 DSA routines if parameters are absent.
14606 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14607 in the current directory if neither $RANDFILE nor $HOME was set.
14608 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14609 caused some confusion to Windows users who haven't defined $HOME.
14610 Thus RAND_file_name() is changed again: e_os.h can define a
14611 DEFAULT_HOME, which will be used if $HOME is not set.
14612 For Windows, we use "C:"; on other platforms, we still require
14613 environment variables.
14615 * Move 'if (!initialized) RAND_poll()' into regions protected by
14616 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14617 having multiple threads call RAND_poll() concurrently.
14621 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14622 combination of a flag and a thread ID variable.
14623 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14624 flag), *other* threads can enter ssleay_add_bytes without obeying
14625 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14626 that they do not hold after the first thread unsets add_do_not_lock).
14630 * Change bctest again: '-x' expressions are not available in all
14631 versions of 'test'.
14635 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14637 * Fix a couple of memory leaks in PKCS7_dataDecode()
14639 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14641 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14642 the default extension for executables, if any. Also, make the perl
14643 scripts that use symlink() to test if it really exists and use "cp"
14644 if it doesn't. All this made OpenSSL compilable and installable in
14649 * Fix for asn1_GetSequence() for indefinite length constructed data.
14650 If SEQUENCE is length is indefinite just set c->slen to the total
14651 amount of data available.
14653 *Steve Henson, reported by shige@FreeBSD.org*
14655 *This change does not apply to 0.9.7.*
14657 * Change bctest to avoid here-documents inside command substitution
14658 (workaround for FreeBSD /bin/sh bug).
14659 For compatibility with Ultrix, avoid shell functions (introduced
14660 in the bctest version that searches along $PATH).
14664 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14665 with des_encrypt() defined on some operating systems, like Solaris
14670 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14671 On the Importance of Eliminating Errors in Cryptographic
14672 Computations, J. Cryptology 14 (2001) 2, 101-119,
14673 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14677 * MIPS assembler BIGNUM division bug fix.
14681 * Disabled incorrect Alpha assembler code.
14685 * Fix PKCS#7 decode routines so they correctly update the length
14686 after reading an EOC for the EXPLICIT tag.
14690 *This change does not apply to 0.9.7.*
14692 * Fix bug in PKCS#12 key generation routines. This was triggered
14693 if a 3DES key was generated with a 0 initial byte. Include
14694 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14695 (but broken) behaviour.
14699 * Enhance bctest to search for a working bc along $PATH and print
14702 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14704 * Fix memory leaks in err.c: free err_data string if necessary;
14705 don't write to the wrong index in ERR_set_error_data.
14709 * Implement ssl23_peek (analogous to ssl23_read), which previously
14714 * Replace rdtsc with `_emit` statements for VC++ version 5.
14716 *Jeremy Cooper <jeremy@baymoo.org>*
14718 * Make it possible to reuse SSLv2 sessions.
14722 * In copy_email() check for >= 0 as a return value for
14723 X509_NAME_get_index_by_NID() since 0 is a valid index.
14725 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14727 * Avoid coredump with unsupported or invalid public keys by checking if
14728 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14729 PKCS7_verify() fails with non detached data.
14733 * Don't use getenv in library functions when run as setuid/setgid.
14734 New function OPENSSL_issetugid().
14738 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14739 due to incorrect handling of multi-threading:
14741 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14743 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14745 3. Count how many times MemCheck_off() has been called so that
14746 nested use can be treated correctly. This also avoids
14747 inband-signalling in the previous code (which relied on the
14748 assumption that thread ID 0 is impossible).
14752 * Add "-rand" option also to s_client and s_server.
14756 * Fix CPU detection on Irix 6.x.
14757 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14758 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14760 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14765 *This change does not apply to 0.9.7.*
14767 * Use the cached encoding of an X509_NAME structure rather than
14768 copying it. This is apparently the reason for the libsafe "errors"
14769 but the code is actually correct.
14773 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14774 Bleichenbacher's DSA attack.
14775 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14776 to be set and top=0 forces the highest bit to be set; top=-1 is new
14777 and leaves the highest bit random.
14779 *Ulf Moeller, Bodo Moeller*
14781 * In the `NCONF_...`-based implementations for `CONF_...` queries
14782 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14783 a temporary CONF structure with the data component set to NULL
14784 (which gives segmentation faults in lh_retrieve).
14785 Instead, use NULL for the CONF pointer in CONF_get_string and
14786 CONF_get_number (which may use environment variables) and directly
14787 return NULL from CONF_get_section.
14791 * Fix potential buffer overrun for EBCDIC.
14795 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14796 keyUsage if basicConstraints absent for a CA.
14800 * Make SMIME_write_PKCS7() write mail header values with a format that
14801 is more generally accepted (no spaces before the semicolon), since
14802 some programs can't parse those values properly otherwise. Also make
14803 sure BIO's that break lines after each write do not create invalid
14808 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14809 macros previously used would not encode an empty SEQUENCE OF
14810 and break the signature.
14814 *This change does not apply to 0.9.7.*
14816 * Zero the premaster secret after deriving the master secret in
14821 * Add some EVP_add_digest_alias registrations (as found in
14822 OpenSSL_add_all_digests()) to SSL_library_init()
14823 aka OpenSSL_add_ssl_algorithms(). This provides improved
14824 compatibility with peers using X.509 certificates
14825 with unconventional AlgorithmIdentifier OIDs.
14829 * Fix for Irix with NO_ASM.
14831 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14833 * ./config script fixes.
14835 *Ulf Moeller, Richard Levitte*
14837 * Fix 'openssl passwd -1'.
14841 * Change PKCS12_key_gen_asc() so it can cope with non null
14842 terminated strings whose length is passed in the passlen
14843 parameter, for example from PEM callbacks. This was done
14844 by adding an extra length parameter to asc2uni().
14846 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14848 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14849 call failed, free the DSA structure.
14853 * Fix to uni2asc() to cope with zero length Unicode strings.
14854 These are present in some PKCS#12 files.
14858 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14859 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14860 when writing a 32767 byte record.
14862 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14864 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14865 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14867 (RSA objects have a reference count access to which is protected
14868 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14869 so they are meant to be shared between threads.)
14870 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14871 "Reddie, Steven" <Steven.Reddie@ca.com>*
14873 * Fix a deadlock in CRYPTO_mem_leaks().
14877 * Use better test patterns in bntest.
14881 * rand_win.c fix for Borland C.
14885 * BN_rshift bugfix for n == 0.
14889 * Add a 'bctest' script that checks for some known 'bc' bugs
14890 so that 'make test' does not abort just because 'bc' is broken.
14894 * Store verify_result within SSL_SESSION also for client side to
14895 avoid potential security hole. (Re-used sessions on the client side
14896 always resulted in verify_result==X509_V_OK, not using the original
14897 result of the server certificate verification.)
14901 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14902 SSL3_RT_APPLICATION_DATA, return 0.
14903 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14908 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14909 releases, have been re-implemented by renaming the previous
14910 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14911 and ssl3_read_internal, respectively, and adding 'peek' parameters
14912 to them. The new ssl[23]_{read,peek} functions are calls to
14913 ssl[23]_read_internal with the 'peek' flag set appropriately.
14914 A 'peek' parameter has also been added to ssl3_read_bytes, which
14915 does the actual work for ssl3_read_internal.
14919 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14920 the method-specific "init()" handler. Also clean up ex_data after
14921 calling the method-specific "finish()" handler. Previously, this was
14922 happening the other way round.
14926 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14927 The previous value, 12, was not always sufficient for BN_mod_exp().
14931 * Make sure that shared libraries get the internal name engine with
14932 the full version number and not just 0. This should mark the
14933 shared libraries as not backward compatible. Of course, this should
14934 be changed again when we can guarantee backward binary compatibility.
14938 * Fix typo in get_cert_by_subject() in by_dir.c
14940 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14942 * Rework the system to generate shared libraries:
14944 - Make note of the expected extension for the shared libraries and
14945 if there is a need for symbolic links from for example libcrypto.so.0
14946 to libcrypto.so.0.9.7. There is extended info in Configure for
14949 - Make as few rebuilds of the shared libraries as possible.
14951 - Still avoid linking the OpenSSL programs with the shared libraries.
14953 - When installing, install the shared libraries separately from the
14958 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14960 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14961 and not in SSL_clear because the latter is also used by the
14962 accept/connect functions; previously, the settings made by
14963 SSL_set_read_ahead would be lost during the handshake.
14965 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14967 * Correct util/mkdef.pl to be selective about disabled algorithms.
14968 Previously, it would create entries for disabled algorithms no
14973 * Added several new manual pages for SSL_* function.
14977 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14979 * In ssl23_get_client_hello, generate an error message when faced
14980 with an initial SSL 3.0/TLS record that is too small to contain the
14981 first two bytes of the ClientHello message, i.e. client_version.
14982 (Note that this is a pathologic case that probably has never happened
14983 in real life.) The previous approach was to use the version number
14984 from the record header as a substitute; but our protocol choice
14985 should not depend on that one because it is not authenticated
14986 by the Finished messages.
14990 * More robust randomness gathering functions for Windows.
14992 *Jeffrey Altman <jaltman@columbia.edu>*
14994 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14995 not set then we don't setup the error code for issuer check errors
14996 to avoid possibly overwriting other errors which the callback does
14997 handle. If an application does set the flag then we assume it knows
14998 what it is doing and can handle the new informational codes
15003 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15004 a general "ANY" type, as such it should be able to decode anything
15005 including tagged types. However it didn't check the class so it would
15006 wrongly interpret tagged types in the same way as their universal
15007 counterpart and unknown types were just rejected. Changed so that the
15008 tagged and unknown types are handled in the same way as a SEQUENCE:
15009 that is the encoding is stored intact. There is also a new type
15010 "V_ASN1_OTHER" which is used when the class is not universal, in this
15011 case we have no idea what the actual type is so we just lump them all
15016 * On VMS, stdout may very well lead to a file that is written to
15017 in a record-oriented fashion. That means that every write() will
15018 write a separate record, which will be read separately by the
15019 programs trying to read from it. This can be very confusing.
15021 The solution is to put a BIO filter in the way that will buffer
15022 text until a linefeed is reached, and then write everything a
15023 line at a time, so every record written will be an actual line,
15024 not chunks of lines and not (usually doesn't happen, but I've
15025 seen it once) several lines in one record. BIO_f_linebuffer() is
15028 Currently, it's a VMS-only method, because that's where it has
15029 been tested well enough.
15033 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15034 it can return incorrect results.
15035 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15036 but it was in 0.9.6-beta[12].)
15040 * Disable the check for content being present when verifying detached
15041 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15042 include zero length content when signing messages.
15046 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15047 BIO_ctrl (for BIO pairs).
15051 * Add DSO method for VMS.
15055 * Bug fix: Montgomery multiplication could produce results with the
15060 * Add RPM specification openssl.spec and modify it to build three
15061 packages. The default package contains applications, application
15062 documentation and run-time libraries. The devel package contains
15063 include files, static libraries and function documentation. The
15064 doc package contains the contents of the doc directory. The original
15065 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15069 * Add a large number of documentation files for many SSL routines.
15071 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15073 * Add a configuration entry for Sony News 4.
15075 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15077 * Don't set the two most significant bits to one when generating a
15078 random number < q in the DSA library.
15082 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15083 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15084 the underlying transport is blocking) if a handshake took place.
15085 (The default behaviour is needed by applications such as s_client
15086 and s_server that use select() to determine when to use SSL_read;
15087 but for applications that know in advance when to expect data, it
15088 just makes things more complicated.)
15092 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15097 * Add a few more EBCDIC conditionals that make `req` and `x509`
15098 work better on such systems.
15100 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15102 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15103 Update PKCS12_parse() so it copies the friendlyName and the
15104 keyid to the certificates aux info.
15108 * Fix bug in PKCS7_verify() which caused an infinite loop
15109 if there was more than one signature.
15111 *Sven Uszpelkat <su@celocom.de>*
15113 * Major change in util/mkdef.pl to include extra information
15114 about each symbol, as well as presenting variables as well
15115 as functions. This change means that there's n more need
15116 to rebuild the .num files when some algorithms are excluded.
15120 * Allow the verify time to be set by an application,
15121 rather than always using the current time.
15125 * Phase 2 verify code reorganisation. The certificate
15126 verify code now looks up an issuer certificate by a
15127 number of criteria: subject name, authority key id
15128 and key usage. It also verifies self signed certificates
15129 by the same criteria. The main comparison function is
15130 X509_check_issued() which performs these checks.
15132 Lot of changes were necessary in order to support this
15133 without completely rewriting the lookup code.
15135 Authority and subject key identifier are now cached.
15137 The LHASH 'certs' is X509_STORE has now been replaced
15138 by a STACK_OF(X509_OBJECT). This is mainly because an
15139 LHASH can't store or retrieve multiple objects with
15140 the same hash value.
15142 As a result various functions (which were all internal
15143 use only) have changed to handle the new X509_STORE
15144 structure. This will break anything that messed round
15145 with X509_STORE internally.
15147 The functions X509_STORE_add_cert() now checks for an
15148 exact match, rather than just subject name.
15150 The X509_STORE API doesn't directly support the retrieval
15151 of multiple certificates matching a given criteria, however
15152 this can be worked round by performing a lookup first
15153 (which will fill the cache with candidate certificates)
15154 and then examining the cache for matches. This is probably
15155 the best we can do without throwing out X509_LOOKUP
15156 entirely (maybe later...).
15158 The X509_VERIFY_CTX structure has been enhanced considerably.
15160 All certificate lookup operations now go via a get_issuer()
15161 callback. Although this currently uses an X509_STORE it
15162 can be replaced by custom lookups. This is a simple way
15163 to bypass the X509_STORE hackery necessary to make this
15164 work and makes it possible to use more efficient techniques
15165 in future. A very simple version which uses a simple
15166 STACK for its trusted certificate store is also provided
15167 using X509_STORE_CTX_trusted_stack().
15169 The verify_cb() and verify() callbacks now have equivalents
15170 in the X509_STORE_CTX structure.
15172 X509_STORE_CTX also has a 'flags' field which can be used
15173 to customise the verify behaviour.
15177 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15178 excludes S/MIME capabilities.
15182 * When a certificate request is read in keep a copy of the
15183 original encoding of the signed data and use it when outputting
15184 again. Signatures then use the original encoding rather than
15185 a decoded, encoded version which may cause problems if the
15186 request is improperly encoded.
15190 * For consistency with other BIO_puts implementations, call
15191 buffer_write(b, ...) directly in buffer_puts instead of calling
15194 In BIO_puts, increment b->num_write as in BIO_write.
15196 *Peter.Sylvester@EdelWeb.fr*
15198 * Fix BN_mul_word for the case where the word is 0. (We have to use
15199 BN_zero, we may not return a BIGNUM with an array consisting of
15200 words set to zero.)
15204 * Avoid calling abort() from within the library when problems are
15205 detected, except if preprocessor symbols have been defined
15206 (such as REF_CHECK, BN_DEBUG etc.).
15210 * New openssl application 'rsautl'. This utility can be
15211 used for low-level RSA operations. DER public key
15212 BIO/fp routines also added.
15216 * New Configure entry and patches for compiling on QNX 4.
15218 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15220 * A demo state-machine implementation was sponsored by
15221 Nuron (<http://www.nuron.com/>) and is now available in
15222 demos/state_machine.
15226 * New options added to the 'dgst' utility for signature
15227 generation and verification.
15231 * Unrecognized PKCS#7 content types are now handled via a
15232 catch all ASN1_TYPE structure. This allows unsupported
15233 types to be stored as a "blob" and an application can
15234 encode and decode it manually.
15238 * Fix various signed/unsigned issues to make a_strex.c
15239 compile under VC++.
15241 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15243 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15244 length if passed a buffer. ASN1_INTEGER_to_BN failed
15245 if passed a NULL BN and its argument was negative.
15247 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15249 * Modification to PKCS#7 encoding routines to output definite
15250 length encoding. Since currently the whole structures are in
15251 memory there's not real point in using indefinite length
15252 constructed encoding. However if OpenSSL is compiled with
15253 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15257 * Added BIO_vprintf() and BIO_vsnprintf().
15261 * Added more prefixes to parse for in the strings written
15262 through a logging bio, to cover all the levels that are available
15263 through syslog. The prefixes are now:
15265 PANIC, EMERG, EMR => LOG_EMERG
15266 ALERT, ALR => LOG_ALERT
15267 CRIT, CRI => LOG_CRIT
15268 ERROR, ERR => LOG_ERR
15269 WARNING, WARN, WAR => LOG_WARNING
15270 NOTICE, NOTE, NOT => LOG_NOTICE
15271 INFO, INF => LOG_INFO
15272 DEBUG, DBG => LOG_DEBUG
15274 and as before, if none of those prefixes are present at the
15275 beginning of the string, LOG_ERR is chosen.
15277 On Win32, the `LOG_*` levels are mapped according to this:
15279 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15280 LOG_WARNING => EVENTLOG_WARNING_TYPE
15281 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15285 * Made it possible to reconfigure with just the configuration
15286 argument "reconf" or "reconfigure". The command line arguments
15287 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15288 and are retrieved from there when reconfiguring.
15294 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15296 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15300 * The obj_dat.pl script was messing up the sorting of object
15301 names. The reason was that it compared the quoted version
15302 of strings as a result "OCSP" > "OCSP Signing" because
15303 " > SPACE. Changed script to store unquoted versions of
15304 names and add quotes on output. It was also omitting some
15305 names from the lookup table if they were given a default
15306 value (that is if SN is missing it is given the same
15307 value as LN and vice versa), these are now added on the
15308 grounds that if an object has a name we should be able to
15309 look it up. Finally added warning output when duplicate
15310 short or long names are found.
15314 * Changes needed for Tandem NSK.
15316 *Scott Uroff <scott@xypro.com>*
15318 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15319 RSA_padding_check_SSLv23(), special padding was never detected
15320 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15321 version rollback attacks was not effective.
15323 In s23_clnt.c, don't use special rollback-attack detection padding
15324 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15325 client; similarly, in s23_srvr.c, don't do the rollback check if
15326 SSL 2.0 is the only protocol enabled in the server.
15330 * Make it possible to get hexdumps of unprintable data with 'openssl
15331 asn1parse'. By implication, the functions ASN1_parse_dump() and
15332 BIO_dump_indent() are added.
15336 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15337 these print out strings and name structures based on various
15338 flags including RFC2253 support and proper handling of
15339 multibyte characters. Added options to the 'x509' utility
15340 to allow the various flags to be set.
15344 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15345 Also change the functions X509_cmp_current_time() and
15346 X509_gmtime_adj() work with an ASN1_TIME structure,
15347 this will enable certificates using GeneralizedTime in validity
15348 dates to be checked.
15352 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15353 negative public key encodings) on by default,
15354 NO_NEG_PUBKEY_BUG can be set to disable it.
15358 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15359 content octets. An i2c_ASN1_OBJECT is unnecessary because
15360 the encoding can be trivially obtained from the structure.
15364 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15365 not read locks (`CRYPTO_r_[un]lock`).
15369 * A first attempt at creating official support for shared
15370 libraries through configuration. I've kept it so the
15371 default is static libraries only, and the OpenSSL programs
15372 are always statically linked for now, but there are
15373 preparations for dynamic linking in place.
15374 This has been tested on Linux and Tru64.
15378 * Randomness polling function for Win9x, as described in:
15379 Peter Gutmann, Software Generation of Practically Strong
15384 * Fix so PRNG is seeded in req if using an already existing
15389 * New options to smime application. -inform and -outform
15390 allow alternative formats for the S/MIME message including
15391 PEM and DER. The -content option allows the content to be
15392 specified separately. This should allow things like Netscape
15393 form signing output easier to verify.
15397 * Fix the ASN1 encoding of tags using the 'long form'.
15401 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15402 STRING types. These convert content octets to and from the
15403 underlying type. The actual tag and length octets are
15404 already assumed to have been read in and checked. These
15405 are needed because all other string types have virtually
15406 identical handling apart from the tag. By having versions
15407 of the ASN1 functions that just operate on content octets
15408 IMPLICIT tagging can be handled properly. It also allows
15409 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15410 and ASN1_INTEGER are identical apart from the tag.
15414 * Change the handling of OID objects as follows:
15416 - New object identifiers are inserted in objects.txt, following
15417 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15418 - objects.pl is used to process obj_mac.num and create a new
15420 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15423 This is currently kind of a hack, and the perl code in objects.pl
15424 isn't very elegant, but it works as I intended. The simplest way
15425 to check that it worked correctly is to look in obj_dat.h and
15426 check the array nid_objs and make sure the objects haven't moved
15427 around (this is important!). Additions are OK, as well as
15428 consistent name changes.
15432 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15436 * Addition of the command line parameter '-rand file' to 'openssl req'.
15437 The given file adds to whatever has already been seeded into the
15438 random pool through the RANDFILE configuration file option or
15439 environment variable, or the default random state file.
15443 * mkstack.pl now sorts each macro group into lexical order.
15444 Previously the output order depended on the order the files
15445 appeared in the directory, resulting in needless rewriting
15450 * Patches to make OpenSSL compile under Win32 again. Mostly
15451 work arounds for the VC++ problem that it treats func() as
15452 func(void). Also stripped out the parts of mkdef.pl that
15453 added extra typesafe functions: these no longer exist.
15457 * Reorganisation of the stack code. The macros are now all
15458 collected in safestack.h . Each macro is defined in terms of
15459 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15460 DEBUG_SAFESTACK is now handled in terms of function casts,
15461 this has the advantage of retaining type safety without the
15462 use of additional functions. If DEBUG_SAFESTACK is not defined
15463 then the non typesafe macros are used instead. Also modified the
15464 mkstack.pl script to handle the new form. Needs testing to see
15465 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15466 the default if no major problems. Similar behaviour for ASN1_SET_OF
15467 and PKCS12_STACK_OF.
15471 * When some versions of IIS use the 'NET' form of private key the
15472 key derivation algorithm is different. Normally MD5(password) is
15473 used as a 128 bit RC4 key. In the modified case
15474 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15475 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15476 as the old Netscape_RSA functions except they have an additional
15477 'sgckey' parameter which uses the modified algorithm. Also added
15478 an -sgckey command line option to the rsa utility. Thanks to
15479 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15480 algorithm to openssl-dev.
15484 * The evp_local.h macros were using 'c.##kname' which resulted in
15485 invalid expansion on some systems (SCO 5.0.5 for example).
15486 Corrected to 'c.kname'.
15488 *Phillip Porch <root@theporch.com>*
15490 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15491 a STACK of email addresses from a certificate or request, these look
15492 in the subject name and the subject alternative name extensions and
15493 omit any duplicate addresses.
15497 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15498 This makes DSA verification about 2 % faster.
15502 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15503 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15504 plus overhead for 1024 bit moduli).
15505 This makes exponentiations about 0.5 % faster for 1024 bit
15506 exponents (as measured by "openssl speed rsa2048").
15510 * Rename memory handling macros to avoid conflicts with other
15512 Malloc => OPENSSL_malloc
15513 Malloc_locked => OPENSSL_malloc_locked
15514 Realloc => OPENSSL_realloc
15515 Free => OPENSSL_free
15519 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15520 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15524 * CygWin32 support.
15526 *John Jarvie <jjarvie@newsguy.com>*
15528 * The type-safe stack code has been rejigged. It is now only compiled
15529 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15530 by default all type-specific stack functions are "#define"d back to
15531 standard stack functions. This results in more streamlined output
15532 but retains the type-safety checking possibilities of the original
15537 * The STACK code has been cleaned up, and certain type declarations
15538 that didn't make a lot of sense have been brought in line. This has
15539 also involved a cleanup of sorts in safestack.h to more correctly
15540 map type-safe stack functions onto their plain stack counterparts.
15541 This work has also resulted in a variety of "const"ifications of
15542 lots of the code, especially `_cmp` operations which should normally
15543 be prototyped with "const" parameters anyway.
15547 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15548 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15549 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15550 where all of 'md' is used each time the PRNG is used, but 'state'
15551 is used only indexed by a cyclic counter. As entropy may not be
15552 well distributed from the beginning, 'md' is important as a
15553 chaining variable. However, the output function chains only half
15554 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15555 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15556 in all of 'state' being rewritten, with the new values depending
15557 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15561 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15562 the handshake is continued after ssl_verify_cert_chain();
15563 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15564 can lead to 'unexplainable' connection aborts later.
15566 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15568 * Major EVP API cipher revision.
15569 Add hooks for extra EVP features. This allows various cipher
15570 parameters to be set in the EVP interface. Support added for variable
15571 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15572 setting of RC2 and RC5 parameters.
15574 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15577 Remove lots of duplicated code from the EVP library. For example *every*
15578 cipher init() function handles the 'iv' in the same way according to the
15579 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15580 for CFB and OFB modes they zero ctx->num.
15582 New functionality allows removal of S/MIME code RC2 hack.
15584 Most of the routines have the same form and so can be declared in terms
15587 By shifting this to the top level EVP_CipherInit() it can be removed from
15588 all individual ciphers. If the cipher wants to handle IVs or keys
15589 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15592 Change lots of functions like EVP_EncryptUpdate() to now return a
15593 value: although software versions of the algorithms cannot fail
15594 any installed hardware versions can.
15598 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15599 this option is set, tolerate broken clients that send the negotiated
15600 protocol version number instead of the requested protocol version
15605 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15606 i.e. non-zero for export ciphersuites, zero otherwise.
15607 Previous versions had this flag inverted, inconsistent with
15608 rsa_tmp_cb (..._TMP_RSA_CB).
15610 *Bodo Moeller; problem reported by Amit Chopra*
15612 * Add missing DSA library text string. Work around for some IIS
15613 key files with invalid SEQUENCE encoding.
15617 * Add a document (doc/standards.txt) that list all kinds of standards
15618 and so on that are implemented in OpenSSL.
15622 * Enhance c_rehash script. Old version would mishandle certificates
15623 with the same subject name hash and wouldn't handle CRLs at all.
15624 Added -fingerprint option to crl utility, to support new c_rehash
15629 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15633 * Fix for SSL server purpose checking. Server checking was
15634 rejecting certificates which had extended key usage present
15635 but no ssl client purpose.
15637 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15639 * Make PKCS#12 code work with no password. The PKCS#12 spec
15640 is a little unclear about how a blank password is handled.
15641 Since the password in encoded as a BMPString with terminating
15642 double NULL a zero length password would end up as just the
15643 double NULL. However no password at all is different and is
15644 handled differently in the PKCS#12 key generation code. NS
15645 treats a blank password as zero length. MSIE treats it as no
15646 password on export: but it will try both on import. We now do
15647 the same: PKCS12_parse() tries zero length and no password if
15648 the password is set to "" or NULL (NULL is now a valid password:
15649 it wasn't before) as does the pkcs12 application.
15653 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15654 perror when PEM_read_bio_X509_REQ fails, the error message must
15655 be obtained from the error queue.
15659 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15660 it in ERR_remove_state if appropriate, and change ERR_get_state
15661 accordingly to avoid race conditions (this is necessary because
15662 thread_hash is no longer constant once set).
15666 * Bugfix for linux-elf makefile.one.
15670 * RSA_get_default_method() will now cause a default
15671 RSA_METHOD to be chosen if one doesn't exist already.
15672 Previously this was only set during a call to RSA_new()
15673 or RSA_new_method(NULL) meaning it was possible for
15674 RSA_get_default_method() to return NULL.
15678 * Added native name translation to the existing DSO code
15679 that will convert (if the flag to do so is set) filenames
15680 that are sufficiently small and have no path information
15681 into a canonical native form. Eg. "blah" converted to
15682 "libblah.so" or "blah.dll" etc.
15686 * New function ERR_error_string_n(e, buf, len) which is like
15687 ERR_error_string(e, buf), but writes at most 'len' bytes
15688 including the 0 terminator. For ERR_error_string_n, 'buf'
15691 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15693 * CONF library reworked to become more general. A new CONF
15694 configuration file reader "class" is implemented as well as a
15695 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15696 old `CONF_*` functions are still there, but are reimplemented to
15697 work in terms of the new functions. Also, a set of functions
15698 to handle the internal storage of the configuration data is
15699 provided to make it easier to write new configuration file
15700 reader "classes" (I can definitely see something reading a
15701 configuration file in XML format, for example), called `_CONF_*`,
15702 or "the configuration storage API"...
15704 The new configuration file reading functions are:
15706 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15707 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15709 NCONF_default, NCONF_WIN32
15711 NCONF_dump_fp, NCONF_dump_bio
15713 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15714 NCONF_new creates a new CONF object. This works in the same way
15715 as other interfaces in OpenSSL, like the BIO interface.
15716 `NCONF_dump_*` dump the internal storage of the configuration file,
15717 which is useful for debugging. All other functions take the same
15718 arguments as the old `CONF_*` functions with the exception of the
15719 first that must be a `CONF *` instead of a `LHASH *`.
15721 To make it easier to use the new classes with the old `CONF_*` functions,
15722 the function CONF_set_default_method is provided.
15726 * Add '-tls1' option to 'openssl ciphers', which was already
15727 mentioned in the documentation but had not been implemented.
15728 (This option is not yet really useful because even the additional
15729 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15733 * Initial DSO code added into libcrypto for letting OpenSSL (and
15734 OpenSSL-based applications) load shared libraries and bind to
15735 them in a portable way.
15737 *Geoff Thorpe, with contributions from Richard Levitte*
15739 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15741 * Make sure _lrotl and _lrotr are only used with MSVC.
15743 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15744 (the default implementation of RAND_status).
15746 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15747 to '-clrext' (= clear extensions), as intended and documented.
15748 *Bodo Moeller; inconsistency pointed out by Michael Attili
15749 <attili@amaxo.com>*
15751 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15752 was larger than the MD block size.
15754 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15756 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15757 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15758 using the passed key: if the passed key was a private key the result
15759 of X509_print(), for example, would be to print out all the private key
15764 * des_quad_cksum() byte order bug fix.
15765 *Ulf Möller, using the problem description in krb4-0.9.7, where
15766 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15768 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15771 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15773 * For easily testing in shell scripts whether some command
15774 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15775 returns with exit code 0 iff no command of the given name is available.
15776 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15777 the output goes to stdout and nothing is printed to stderr.
15778 Additional arguments are always ignored.
15780 Since for each cipher there is a command of the same name,
15781 the 'no-cipher' compilation switches can be tested this way.
15783 ('openssl no-XXX' is not able to detect pseudo-commands such
15784 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15788 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15792 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15793 is set; it will be thrown away anyway because each handshake creates
15795 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15796 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15797 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15798 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15802 * New s_client option -ign_eof: EOF at stdin is ignored, and
15803 'Q' and 'R' lose their special meanings (quit/renegotiate).
15804 This is part of what -quiet does; unlike -quiet, -ign_eof
15805 does not suppress any output.
15809 * Add compatibility options to the purpose and trust code. The
15810 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15811 accepts a certificate or CA, this was the previous behaviour,
15812 with all the associated security issues.
15814 X509_TRUST_COMPAT is the old trust behaviour: only and
15815 automatically trust self signed roots in certificate store. A
15816 new trust setting X509_TRUST_DEFAULT is used to specify that
15817 a purpose has no associated trust setting and it should instead
15818 use the value in the default purpose.
15822 * Fix the PKCS#8 DSA private key code so it decodes keys again
15823 and fix a memory leak.
15827 * In util/mkerr.pl (which implements 'make errors'), preserve
15828 reason strings from the previous version of the .c file, as
15829 the default to have only downcase letters (and digits) in
15830 automatically generated reasons codes is not always appropriate.
15834 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15835 using strerror. Previously, ERR_reason_error_string() returned
15836 library names as reason strings for SYSerr; but SYSerr is a special
15837 case where small numbers are errno values, not library numbers.
15841 * Add '-dsaparam' option to 'openssl dhparam' application. This
15842 converts DSA parameters into DH parameters. (When creating parameters,
15843 DSA_generate_parameters is used.)
15847 * Include 'length' (recommended exponent length) in C code generated
15848 by 'openssl dhparam -C'.
15852 * The second argument to set_label in perlasm was already being used
15853 so couldn't be used as a "file scope" flag. Moved to third argument
15858 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15859 instead of RAND_bytes for encryption IVs and salts.
15863 * Include RAND_status() into RAND_METHOD instead of implementing
15864 it only for md_rand.c Otherwise replacing the PRNG by calling
15865 RAND_set_rand_method would be impossible.
15869 * Don't let DSA_generate_key() enter an infinite loop if the random
15870 number generation fails.
15874 * New 'rand' application for creating pseudo-random output.
15878 * Added configuration support for Linux/IA64
15880 *Rolf Haberrecker <rolf@suse.de>*
15882 * Assembler module support for Mingw32.
15886 * Shared library support for HPUX (in shlib/).
15888 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15890 * Shared library support for Solaris gcc.
15892 *Lutz Behnke <behnke@trustcenter.de>*
15894 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15896 * PKCS7_encrypt() was adding text MIME headers twice because they
15897 were added manually and by SMIME_crlf_copy().
15901 * In bntest.c don't call BN_rand with zero bits argument.
15903 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15905 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15906 case was implemented. This caused BN_div_recp() to fail occasionally.
15910 * Add an optional second argument to the set_label() in the perl
15911 assembly language builder. If this argument exists and is set
15912 to 1 it signals that the assembler should use a symbol whose
15913 scope is the entire file, not just the current function. This
15914 is needed with MASM which uses the format label:: for this scope.
15916 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15918 * Change the ASN1 types so they are typedefs by default. Before
15919 almost all types were #define'd to ASN1_STRING which was causing
15920 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15925 * Change names of new functions to the new get1/get0 naming
15926 convention: After 'get1', the caller owns a reference count
15927 and has to call `..._free`; 'get0' returns a pointer to some
15928 data structure without incrementing reference counters.
15929 (Some of the existing 'get' functions increment a reference
15930 counter, some don't.)
15931 Similarly, 'set1' and 'add1' functions increase reference
15932 counters or duplicate objects.
15936 * Allow for the possibility of temp RSA key generation failure:
15937 the code used to assume it always worked and crashed on failure.
15941 * Fix potential buffer overrun problem in BIO_printf().
15942 *Ulf Möller, using public domain code by Patrick Powell; problem
15943 pointed out by David Sacerdote <das33@cornell.edu>*
15945 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15946 RAND_egd() and RAND_status(). In the command line application,
15947 the EGD socket can be specified like a seed file using RANDFILE
15952 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15953 Some CAs (e.g. Verisign) distribute certificates in this form.
15957 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15958 list to exclude them. This means that no special compilation option
15959 is needed to use anonymous DH: it just needs to be included in the
15964 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15965 EVP_MD_type. The old functionality is available in a new macro called
15966 EVP_MD_md(). Change code that uses it and update docs.
15970 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15971 where the `void *` argument is replaced by a function pointer argument.
15972 Previously `void *` was abused to point to functions, which works on
15973 many platforms, but is not correct. As these functions are usually
15974 called by macros defined in OpenSSL header files, most source code
15975 should work without changes.
15979 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15980 sections with information on -D... compiler switches used for
15981 compiling the library so that applications can see them. To enable
15982 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15983 must be defined. E.g.,
15984 #define OPENSSL_ALGORITHM_DEFINES
15985 #include <openssl/opensslconf.h>
15986 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15988 *Richard Levitte, Ulf and Bodo Möller*
15990 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15995 * Change the 'other' type in certificate aux info to a STACK_OF
15996 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15997 the required ASN1 format: arbitrary types determined by an OID.
16001 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16002 argument to 'req'. This is not because the function is newer or
16003 better than others it just uses the work 'NEW' in the certificate
16004 request header lines. Some software needs this.
16008 * Reorganise password command line arguments: now passwords can be
16009 obtained from various sources. Delete the PEM_cb function and make
16010 it the default behaviour: i.e. if the callback is NULL and the
16011 usrdata argument is not NULL interpret it as a null terminated pass
16012 phrase. If usrdata and the callback are NULL then the pass phrase
16013 is prompted for as usual.
16017 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16018 the support is automatically enabled. The resulting binaries will
16019 autodetect the card and use it if present.
16021 *Ben Laurie and Compaq Inc.*
16023 * Work around for Netscape hang bug. This sends certificate request
16024 and server done in one record. Since this is perfectly legal in the
16025 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16026 the bugs/SSLv3 entry for more info.
16030 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16034 * Add -rand argument to smime and pkcs12 applications and read/write
16039 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16043 * Add command line password options to the remaining applications.
16047 * Bug fix for BN_div_recp() for numerators with an even number of
16052 * More tests in bntest.c, and changed test_bn output.
16056 * ./config recognizes MacOS X now.
16060 * Bug fix for BN_div() when the first words of num and divisor are
16061 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16065 * Add support for various broken PKCS#8 formats, and command line
16066 options to produce them.
16070 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16071 get temporary BIGNUMs from a BN_CTX.
16075 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16080 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16081 include a #define from the old name to the new. The original intent
16082 was that statically linked binaries could for example just call
16083 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16084 link with digests. This never worked because SSLeay_add_all_digests()
16085 and SSLeay_add_all_ciphers() were in the same source file so calling
16086 one would link with the other. They are now in separate source files.
16090 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16094 * Use a less unusual form of the Miller-Rabin primality test (it used
16095 a binary algorithm for exponentiation integrated into the Miller-Rabin
16096 loop, our standard modexp algorithms are faster).
16100 * Support for the EBCDIC character set completed.
16102 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16104 * Source code cleanups: use const where appropriate, eliminate casts,
16105 use `void *` instead of `char *` in lhash.
16109 * Bugfix: ssl3_send_server_key_exchange was not restartable
16110 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16111 this the server could overwrite ephemeral keys that the client
16116 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16117 using 50 iterations of the Rabin-Miller test.
16119 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16120 iterations of the Rabin-Miller test as required by the appendix
16121 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16122 As BN_is_prime_fasttest includes trial division, DSA parameter
16123 generation becomes much faster.
16125 This implies a change for the callback functions in DSA_is_prime
16126 and DSA_generate_parameters: The callback function is called once
16127 for each positive witness in the Rabin-Miller test, not just
16128 occasionally in the inner loop; and the parameters to the
16129 callback function now provide an iteration count for the outer
16130 loop rather than for the current invocation of the inner loop.
16131 DSA_generate_parameters additionally can call the callback
16132 function with an 'iteration count' of -1, meaning that a
16133 candidate has passed the trial division test (when q is generated
16134 from an application-provided seed, trial division is skipped).
16138 * New function BN_is_prime_fasttest that optionally does trial
16139 division before starting the Rabin-Miller test and has
16140 an additional BN_CTX * argument (whereas BN_is_prime always
16141 has to allocate at least one BN_CTX).
16142 'callback(1, -1, cb_arg)' is called when a number has passed the
16143 trial division stage.
16147 * Fix for bug in CRL encoding. The validity dates weren't being handled
16152 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16156 * New function BN_pseudo_rand().
16160 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16161 bignum version of BN_from_montgomery() with the working code from
16162 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16167 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16168 made it impossible to use the same SSL_SESSION data structure in
16169 SSL2 clients in multiple threads.
16173 * The return value of RAND_load_file() no longer counts bytes obtained
16174 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16175 to seed the PRNG (previously an explicit byte count was required).
16177 *Ulf Möller, Bodo Möller*
16179 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16180 used `char *` instead of `void *` and had casts all over the place.
16184 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16188 * Retain source code compatibility for BN_prime_checks macro:
16189 BN_is_prime(..., BN_prime_checks, ...) now uses
16190 BN_prime_checks_for_size to determine the appropriate number of
16191 Rabin-Miller iterations.
16195 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16196 DH_CHECK_P_NOT_SAFE_PRIME.
16197 (Check if this is true? OpenPGP calls them "strong".)
16201 * Merge the functionality of "dh" and "gendh" programs into a new program
16202 "dhparam". The old programs are retained for now but will handle DH keys
16203 (instead of parameters) in future.
16207 * Make the ciphers, s_server and s_client programs check the return values
16208 when a new cipher list is set.
16212 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16213 ciphers. Before when the 56bit ciphers were enabled the sorting was
16216 The syntax for the cipher sorting has been extended to support sorting by
16217 cipher-strength (using the strength_bits hard coded in the tables).
16218 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16220 Fix a bug in the cipher-command parser: when supplying a cipher command
16221 string with an "undefined" symbol (neither command nor alphanumeric
16222 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16223 an error is flagged.
16225 Due to the strength-sorting extension, the code of the
16226 ssl_create_cipher_list() function was completely rearranged. I hope that
16227 the readability was also increased :-)
16229 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16231 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16232 for the first serial number and places 2 in the serial number file. This
16233 avoids problems when the root CA is created with serial number zero and
16234 the first user certificate has the same issuer name and serial number
16239 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16240 the new code. Add documentation for this stuff.
16244 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16245 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16246 structures and behave in an analogous way to the X509v3 functions:
16247 they shouldn't be called directly but wrapper functions should be used
16250 So we also now have some wrapper functions that call the X509at functions
16251 when passed certificate requests. (TO DO: similar things can be done with
16252 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16253 things. Some of these need some d2i or i2d and print functionality
16254 because they handle more complex structures.)
16258 * Add missing #ifndefs that caused missing symbols when building libssl
16259 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16260 NO_RSA in `ssl/s2*.c`.
16262 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16264 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16265 has a return value which indicates the quality of the random data
16266 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16267 error queue. New function RAND_pseudo_bytes() generates output that is
16268 guaranteed to be unique but not unpredictable. RAND_add is like
16269 RAND_seed, but takes an extra argument for an entropy estimate
16270 (RAND_seed always assumes full entropy).
16274 * Do more iterations of Rabin-Miller probable prime test (specifically,
16275 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16276 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16277 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16278 false-positive rate of at most 2^-80 for random input.
16282 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16286 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16287 in the 0.9.5 release), this returns the chain
16288 from an X509_CTX structure with a dup of the stack and all
16289 the X509 reference counts upped: so the stack will exist
16290 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16293 Also make SSL_SESSION_print() print out the verify return
16298 * Add manpage for the pkcs12 command. Also change the default
16299 behaviour so MAC iteration counts are used unless the new
16300 -nomaciter option is used. This improves file security and
16301 only older versions of MSIE (4.0 for example) need it.
16305 * Honor the no-xxx Configure options when creating .DEF files.
16309 * Add PKCS#10 attributes to field table: challengePassword,
16310 unstructuredName and unstructuredAddress. These are taken from
16311 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16312 international characters are used.
16314 More changes to X509_ATTRIBUTE code: allow the setting of types
16315 based on strings. Remove the 'loc' parameter when adding
16316 attributes because these will be a SET OF encoding which is sorted
16321 * Initial changes to the 'req' utility to allow request generation
16322 automation. This will allow an application to just generate a template
16323 file containing all the field values and have req construct the
16326 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16327 used all over the place including certificate requests and PKCS#7
16328 structures. They are currently handled manually where necessary with
16329 some primitive wrappers for PKCS#7. The new functions behave in a
16330 manner analogous to the X509 extension functions: they allow
16331 attributes to be looked up by NID and added.
16333 Later something similar to the X509V3 code would be desirable to
16334 automatically handle the encoding, decoding and printing of the
16335 more complex types. The string types like challengePassword can
16336 be handled by the string table functions.
16338 Also modified the multi byte string table handling. Now there is
16339 a 'global mask' which masks out certain types. The table itself
16340 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16341 is useful when for example there is only one permissible type
16342 (as in countryName) and using the mask might result in no valid
16347 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16348 SSL_get_peer_finished to allow applications to obtain the latest
16349 Finished messages sent to the peer or expected from the peer,
16350 respectively. (SSL_get_peer_finished is usually the Finished message
16351 actually received from the peer, otherwise the protocol will be aborted.)
16353 As the Finished message are message digests of the complete handshake
16354 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16355 be used for external authentication procedures when the authentication
16356 provided by SSL/TLS is not desired or is not enough.
16360 * Enhanced support for Alpha Linux is added. Now ./config checks if
16361 the host supports BWX extension and if Compaq C is present on the
16362 $PATH. Just exploiting of the BWX extension results in 20-30%
16363 performance kick for some algorithms, e.g. DES and RC4 to mention
16364 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16369 * Add support for MS "fast SGC". This is arguably a violation of the
16370 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16371 weak crypto and after checking the certificate is SGC a second one
16372 with strong crypto. MS SGC stops the first handshake after receiving
16373 the server certificate message and sends a second client hello. Since
16374 a server will typically do all the time consuming operations before
16375 expecting any further messages from the client (server key exchange
16376 is the most expensive) there is little difference between the two.
16378 To get OpenSSL to support MS SGC we have to permit a second client
16379 hello message after we have sent server done. In addition we have to
16380 reset the MAC if we do get this second client hello.
16384 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16385 if a DER encoded private key is RSA or DSA traditional format. Changed
16386 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16387 format DER encoded private key. Newer code should use PKCS#8 format which
16388 has the key type encoded in the ASN1 structure. Added DER private key
16389 support to pkcs8 application.
16393 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16394 ciphersuites has been selected (as required by the SSL 3/TLS 1
16395 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16396 is set, we interpret this as a request to violate the specification
16397 (the worst that can happen is a handshake failure, and 'correct'
16398 behaviour would result in a handshake failure anyway).
16402 * In SSL_CTX_add_session, take into account that there might be multiple
16403 SSL_SESSION structures with the same session ID (e.g. when two threads
16404 concurrently obtain them from an external cache).
16405 The internal cache can handle only one SSL_SESSION with a given ID,
16406 so if there's a conflict, we now throw out the old one to achieve
16411 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16412 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16413 some routines that use cipher OIDs: some ciphers do not have OIDs
16414 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16419 * Simplify the trust setting structure and code. Now we just have
16420 two sequences of OIDs for trusted and rejected settings. These will
16421 typically have values the same as the extended key usage extension
16422 and any application specific purposes.
16424 The trust checking code now has a default behaviour: it will just
16425 check for an object with the same NID as the passed id. Functions can
16426 be provided to override either the default behaviour or the behaviour
16427 for a given id. SSL client, server and email already have functions
16428 in place for compatibility: they check the NID and also return "trusted"
16429 if the certificate is self signed.
16433 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16434 traditional format into an EVP_PKEY structure.
16438 * Add a password callback function PEM_cb() which either prompts for
16439 a password if usr_data is NULL or otherwise assumes it is a null
16440 terminated password. Allow passwords to be passed on command line
16441 environment or config files in a few more utilities.
16445 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16446 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16447 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16448 Update documentation.
16452 * Support for ASN1 "NULL" type. This could be handled before by using
16453 ASN1_TYPE but there wasn't any function that would try to read a NULL
16454 and produce an error if it couldn't. For compatibility we also have
16455 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16456 don't allocate anything because they don't need to.
16460 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16463 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16465 * Rebuild of the memory allocation routines used by OpenSSL code and
16466 possibly others as well. The purpose is to make an interface that
16467 provide hooks so anyone can build a separate set of allocation and
16468 deallocation routines to be used by OpenSSL, for example memory
16469 pool implementations, or something else, which was previously hard
16470 since Malloc(), Realloc() and Free() were defined as macros having
16471 the values malloc, realloc and free, respectively (except for Win32
16472 compilations). The same is provided for memory debugging code.
16473 OpenSSL already comes with functionality to find memory leaks, but
16474 this gives people a chance to debug other memory problems.
16476 With these changes, a new set of functions and macros have appeared:
16478 CRYPTO_set_mem_debug_functions() [F]
16479 CRYPTO_get_mem_debug_functions() [F]
16480 CRYPTO_dbg_set_options() [F]
16481 CRYPTO_dbg_get_options() [F]
16482 CRYPTO_malloc_debug_init() [M]
16484 The memory debug functions are NULL by default, unless the library
16485 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16486 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16487 gives the standard debugging functions that come with OpenSSL) or
16488 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16489 provided by the library user) must be used. When the standard
16490 debugging functions are used, CRYPTO_dbg_set_options can be used to
16491 request additional information:
16492 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16493 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16495 Also, things like CRYPTO_set_mem_functions will always give the
16496 expected result (the new set of functions is used for allocation
16497 and deallocation) at all times, regardless of platform and compiler
16500 To finish it up, some functions that were never use in any other
16501 way than through macros have a new API and new semantic:
16503 CRYPTO_dbg_malloc()
16504 CRYPTO_dbg_realloc()
16507 All macros of value have retained their old syntax.
16509 *Richard Levitte and Bodo Moeller*
16511 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16512 ordering of SMIMECapabilities wasn't in "strength order" and there
16513 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16518 * Some ASN1 types with illegal zero length encoding (INTEGER,
16519 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16521 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16523 * Merge in my S/MIME library for OpenSSL. This provides a simple
16524 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16525 functionality to handle multipart/signed properly) and a utility
16526 called 'smime' to call all this stuff. This is based on code I
16527 originally wrote for Celo who have kindly allowed it to be
16528 included in OpenSSL.
16532 * Add variants des_set_key_checked and des_set_key_unchecked of
16533 des_set_key (aka des_key_sched). Global variable des_check_key
16534 decides which of these is called by des_set_key; this way
16535 des_check_key behaves as it always did, but applications and
16536 the library itself, which was buggy for des_check_key == 1,
16537 have a cleaner way to pick the version they need.
16541 * New function PKCS12_newpass() which changes the password of a
16546 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16547 dynamic mix. In both cases the ids can be used as an index into the
16548 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16549 functions so they accept a list of the field values and the
16550 application doesn't need to directly manipulate the X509_TRUST
16555 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16560 * Modify the way the V3 extension code looks up extensions. This now
16561 works in a similar way to the object code: we have some "standard"
16562 extensions in a static table which is searched with OBJ_bsearch()
16563 and the application can add dynamic ones if needed. The file
16564 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16565 updated whenever a new extension is added to the core code and kept
16566 in ext_nid order. There is a simple program 'tabtest.c' which checks
16567 this. New extensions are not added too often so this file can readily
16568 be maintained manually.
16570 There are two big advantages in doing things this way. The extensions
16571 can be looked up immediately and no longer need to be "added" using
16572 X509V3_add_standard_extensions(): this function now does nothing.
16573 Side note: I get *lots* of email saying the extension code doesn't
16574 work because people forget to call this function.
16575 Also no dynamic allocation is done unless new extensions are added:
16576 so if we don't add custom extensions there is no need to call
16577 X509V3_EXT_cleanup().
16581 * Modify enc utility's salting as follows: make salting the default. Add a
16582 magic header, so unsalted files fail gracefully instead of just decrypting
16583 to garbage. This is because not salting is a big security hole, so people
16584 should be discouraged from doing it.
16588 * Fixes and enhancements to the 'x509' utility. It allowed a message
16589 digest to be passed on the command line but it only used this
16590 parameter when signing a certificate. Modified so all relevant
16591 operations are affected by the digest parameter including the
16592 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16593 DSA key was used because it didn't fix the digest.
16597 * Initial certificate chain verify code. Currently tests the untrusted
16598 certificates for consistency with the verify purpose (which is set
16599 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16601 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16602 this is because it will reject chains with invalid extensions whereas
16603 every previous version of OpenSSL and SSLeay made no checks at all.
16605 Trust code: checks the root CA for the relevant trust settings. Trust
16606 settings have an initial value consistent with the verify purpose: e.g.
16607 if the verify purpose is for SSL client use it expects the CA to be
16608 trusted for SSL client use. However the default value can be changed to
16609 permit custom trust settings: one example of this would be to only trust
16610 certificates from a specific "secure" set of CAs.
16612 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16613 which should be used for version portability: especially since the
16614 verify structure is likely to change more often now.
16616 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16617 to set them. If not set then assume SSL clients will verify SSL servers
16620 Two new options to the verify program: -untrusted allows a set of
16621 untrusted certificates to be passed in and -purpose which sets the
16622 intended purpose of the certificate. If a purpose is set then the
16623 new chain verify code is used to check extension consistency.
16627 * Support for the authority information access extension.
16631 * Modify RSA and DSA PEM read routines to transparently handle
16632 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16633 public keys in a format compatible with certificate
16634 SubjectPublicKeyInfo structures. Unfortunately there were already
16635 functions called *_PublicKey_* which used various odd formats so
16636 these are retained for compatibility: however the DSA variants were
16637 never in a public release so they have been deleted. Changed dsa/rsa
16638 utilities to handle the new format: note no releases ever handled public
16639 keys so we should be OK.
16641 The primary motivation for this change is to avoid the same fiasco
16642 that dogs private keys: there are several incompatible private key
16643 formats some of which are standard and some OpenSSL specific and
16644 require various evil hacks to allow partial transparent handling and
16645 even then it doesn't work with DER formats. Given the option anything
16646 other than PKCS#8 should be dumped: but the other formats have to
16647 stay in the name of compatibility.
16649 With public keys and the benefit of hindsight one standard format
16650 is used which works with EVP_PKEY, RSA or DSA structures: though
16651 it clearly returns an error if you try to read the wrong kind of key.
16653 Added a -pubkey option to the 'x509' utility to output the public key.
16654 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16655 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16656 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16657 that do the same as the `EVP_PKEY_assign_*()` except they up the
16658 reference count of the added key (they don't "swallow" the
16663 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16664 CRLs would fail if the file contained no certificates or no CRLs:
16665 added a new function to read in both types and return the number
16666 read: this means that if none are read it will be an error. The
16667 DER versions of the certificate and CRL reader would always fail
16668 because it isn't possible to mix certificates and CRLs in DER format
16669 without choking one or the other routine. Changed this to just read
16670 a certificate: this is the best we can do. Also modified the code
16671 in `apps/verify.c` to take notice of return codes: it was previously
16672 attempting to read in certificates from NULL pointers and ignoring
16673 any errors: this is one reason why the cert and CRL reader seemed
16674 to work. It doesn't check return codes from the default certificate
16675 routines: these may well fail if the certificates aren't installed.
16679 * Code to support otherName option in GeneralName.
16683 * First update to verify code. Change the verify utility
16684 so it warns if it is passed a self signed certificate:
16685 for consistency with the normal behaviour. X509_verify
16686 has been modified to it will now verify a self signed
16687 certificate if *exactly* the same certificate appears
16688 in the store: it was previously impossible to trust a
16689 single self signed certificate. This means that:
16690 openssl verify ss.pem
16691 now gives a warning about a self signed certificate but
16692 openssl verify -CAfile ss.pem ss.pem
16697 * For servers, store verify_result in SSL_SESSION data structure
16698 (and add it to external session representation).
16699 This is needed when client certificate verifications fails,
16700 but an application-provided verification callback (set by
16701 SSL_CTX_set_cert_verify_callback) allows accepting the session
16702 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16703 but returns 1): When the session is reused, we have to set
16704 ssl->verify_result to the appropriate error code to avoid
16707 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16709 * Fix a bug in the new PKCS#7 code: it didn't consider the
16710 case in PKCS7_dataInit() where the signed PKCS7 structure
16711 didn't contain any existing data because it was being created.
16713 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16715 * Add a salt to the key derivation routines in enc.c. This
16716 forms the first 8 bytes of the encrypted file. Also add a
16717 -S option to allow a salt to be input on the command line.
16721 * New function X509_cmp(). Oddly enough there wasn't a function
16722 to compare two certificates. We do this by working out the SHA1
16723 hash and comparing that. X509_cmp() will be needed by the trust
16728 * SSL_get1_session() is like SSL_get_session(), but increments
16729 the reference count in the SSL_SESSION returned.
16731 *Geoff Thorpe <geoff@eu.c2.net>*
16733 * Fix for 'req': it was adding a null to request attributes.
16734 Also change the X509_LOOKUP and X509_INFO code to handle
16735 certificate auxiliary information.
16739 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16744 * Add the possibility to add extra information to the memory leak
16745 detecting output, to form tracebacks, showing from where each
16746 allocation was originated: CRYPTO_push_info("constant string") adds
16747 the string plus current file name and line number to a per-thread
16748 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16749 is like calling CYRPTO_pop_info() until the stack is empty.
16750 Also updated memory leak detection code to be multi-thread-safe.
16754 * Add options -text and -noout to pkcs7 utility and delete the
16755 encryption options which never did anything. Update docs.
16759 * Add options to some of the utilities to allow the pass phrase
16760 to be included on either the command line (not recommended on
16761 OSes like Unix) or read from the environment. Update the
16762 manpages and fix a few bugs.
16766 * Add a few manpages for some of the openssl commands.
16770 * Fix the -revoke option in ca. It was freeing up memory twice,
16771 leaking and not finding already revoked certificates.
16775 * Extensive changes to support certificate auxiliary information.
16776 This involves the use of X509_CERT_AUX structure and X509_AUX
16777 functions. An X509_AUX function such as PEM_read_X509_AUX()
16778 can still read in a certificate file in the usual way but it
16779 will also read in any additional "auxiliary information". By
16780 doing things this way a fair degree of compatibility can be
16781 retained: existing certificates can have this information added
16782 using the new 'x509' options.
16784 Current auxiliary information includes an "alias" and some trust
16785 settings. The trust settings will ultimately be used in enhanced
16786 certificate chain verification routines: currently a certificate
16787 can only be trusted if it is self signed and then it is trusted
16792 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16793 The problem was that one of the replacement routines had not been working
16794 since SSLeay releases. For now the offending routine has been replaced
16795 with non-optimised assembler. Even so, this now gives around 95%
16796 performance improvement for 1024 bit RSA signs.
16800 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16801 handling. Most clients have the effective key size in bits equal to
16802 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16803 A few however don't do this and instead use the size of the decrypted key
16804 to determine the RC2 key length and the AlgorithmIdentifier to determine
16805 the effective key length. In this case the effective key length can still
16806 be 40 bits but the key length can be 168 bits for example. This is fixed
16807 by manually forcing an RC2 key into the EVP_PKEY structure because the
16808 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16809 the key length and effective key length are equal.
16813 * Add a bunch of functions that should simplify the creation of
16814 X509_NAME structures. Now you should be able to do:
16815 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16816 and have it automatically work out the correct field type and fill in
16817 the structures. The more adventurous can try:
16818 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16819 and it will (hopefully) work out the correct multibyte encoding.
16823 * Change the 'req' utility to use the new field handling and multibyte
16824 copy routines. Before the DN field creation was handled in an ad hoc
16825 way in req, ca, and x509 which was rather broken and didn't support
16826 BMPStrings or UTF8Strings. Since some software doesn't implement
16827 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16828 using the dirstring_type option. See the new comment in the default
16829 openssl.cnf for more info.
16833 * Make crypto/rand/md_rand.c more robust:
16834 - Assure unique random numbers after fork().
16835 - Make sure that concurrent threads access the global counter and
16836 md serializably so that we never lose entropy in them
16837 or use exactly the same state in multiple threads.
16838 Access to the large state is not always serializable because
16839 the additional locking could be a performance killer, and
16840 md should be large enough anyway.
16844 * New file `apps/app_rand.c` with commonly needed functionality
16845 for handling the random seed file.
16847 Use the random seed file in some applications that previously did not:
16849 dsaparam -genkey (which also ignored its '-rand' option),
16852 x509 (when signing).
16853 Except on systems with /dev/urandom, it is crucial to have a random
16854 seed file at least for key creation, DSA signing, and for DH exchanges;
16855 for RSA signatures we could do without one.
16857 gendh and gendsa (unlike genrsa) used to read only the first byte
16858 of each file listed in the '-rand' option. The function as previously
16859 found in genrsa is now in app_rand.c and is used by all programs
16860 that support '-rand'.
16864 * In RAND_write_file, use mode 0600 for creating files;
16865 don't just chmod when it may be too late.
16869 * Report an error from X509_STORE_load_locations
16870 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16874 * New function ASN1_mbstring_copy() this copies a string in either
16875 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16876 into an ASN1_STRING type. A mask of permissible types is passed
16877 and it chooses the "minimal" type to use or an error if not type
16882 * Add function equivalents to the various macros in asn1.h. The old
16883 macros are retained with an `M_` prefix. Code inside the library can
16884 use the `M_` macros. External code (including the openssl utility)
16885 should *NOT* in order to be "shared library friendly".
16889 * Add various functions that can check a certificate's extensions
16890 to see if it usable for various purposes such as SSL client,
16891 server or S/MIME and CAs of these types. This is currently
16892 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16893 verification. Also added a -purpose flag to x509 utility to
16894 print out all the purposes.
16898 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16903 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16904 for, obtain and decode and extension and obtain its critical flag.
16905 This allows all the necessary extension code to be handled in a
16906 single function call.
16910 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16911 platforms. See crypto/rc4/rc4_enc.c for further details.
16915 * New -noout option to asn1parse. This causes no output to be produced
16916 its main use is when combined with -strparse and -out to extract data
16917 from a file (which may not be in ASN.1 format).
16921 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16922 when producing the local key id.
16924 *Richard Levitte <levitte@stacken.kth.se>*
16926 * New option -dhparam in s_server. This allows a DH parameter file to be
16927 stated explicitly. If it is not stated then it tries the first server
16928 certificate file. The previous behaviour hard coded the filename
16933 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16934 a public key to be input or output. For example:
16935 openssl rsa -in key.pem -pubout -out pubkey.pem
16936 Also added necessary DSA public key functions to handle this.
16940 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16941 in the message. This was handled by allowing
16942 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16944 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16946 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16947 to the end of the strings whereas this didn't. This would cause problems
16948 if strings read with d2i_ASN1_bytes() were later modified.
16950 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16952 * Fix for base64 decode bug. When a base64 bio reads only one line of
16953 data and it contains EOF it will end up returning an error. This is
16954 caused by input 46 bytes long. The cause is due to the way base64
16955 BIOs find the start of base64 encoded data. They do this by trying a
16956 trial decode on each line until they find one that works. When they
16957 do a flag is set and it starts again knowing it can pass all the
16958 data directly through the decoder. Unfortunately it doesn't reset
16959 the context it uses. This means that if EOF is reached an attempt
16960 is made to pass two EOFs through the context and this causes the
16961 resulting error. This can also cause other problems as well. As is
16962 usual with these problems it takes *ages* to find and the fix is
16963 trivial: move one line.
16965 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16967 * Ugly workaround to get s_client and s_server working under Windows. The
16968 old code wouldn't work because it needed to select() on sockets and the
16969 tty (for keypresses and to see if data could be written). Win32 only
16970 supports select() on sockets so we select() with a 1s timeout on the
16971 sockets and then see if any characters are waiting to be read, if none
16972 are present then we retry, we also assume we can always write data to
16973 the tty. This isn't nice because the code then blocks until we've
16974 received a complete line of data and it is effectively polling the
16975 keyboard at 1s intervals: however it's quite a bit better than not
16976 working at all :-) A dedicated Windows application might handle this
16977 with an event loop for example.
16981 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16982 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16983 will be called when RSA_sign() and RSA_verify() are used. This is useful
16984 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16985 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16986 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16987 This necessitated the support of an extra signature type NID_md5_sha1
16988 for SSL signatures and modifications to the SSL library to use it instead
16989 of calling RSA_public_decrypt() and RSA_private_encrypt().
16993 * Add new -verify -CAfile and -CApath options to the crl program, these
16994 will lookup a CRL issuers certificate and verify the signature in a
16995 similar way to the verify program. Tidy up the crl program so it
16996 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16997 less strict. It will now permit CRL extensions even if it is not
16998 a V2 CRL: this will allow it to tolerate some broken CRLs.
17002 * Initialize all non-automatic variables each time one of the openssl
17003 sub-programs is started (this is necessary as they may be started
17004 multiple times from the "OpenSSL>" prompt).
17006 *Lennart Bang, Bodo Moeller*
17008 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17009 removing all other RSA functionality (this is what NO_RSA does). This
17010 is so (for example) those in the US can disable those operations covered
17011 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17016 * Non-copying interface to BIO pairs.
17017 (still largely untested)
17021 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17022 ASCII string. This was handled independently in various places before.
17026 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17027 UTF8 strings a character at a time.
17031 * Use client_version from client hello to select the protocol
17032 (s23_srvr.c) and for RSA client key exchange verification
17033 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17037 * Add various utility functions to handle SPKACs, these were previously
17038 handled by poking round in the structure internals. Added new function
17039 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17040 print, verify and generate SPKACs. Based on an original idea from
17041 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17045 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17049 * Allow the config file extension section to be overwritten on the
17050 command line. Based on an original idea from Massimiliano Pala
17051 <madwolf@comune.modena.it>. The new option is called -extensions
17052 and can be applied to ca, req and x509. Also -reqexts to override
17053 the request extensions in req and -crlexts to override the crl extensions
17058 * Add new feature to the SPKAC handling in ca. Now you can include
17059 the same field multiple times by preceding it by "XXXX." for example:
17062 this is the same syntax as used in the req config file.
17066 * Allow certificate extensions to be added to certificate requests. These
17067 are specified in a 'req_extensions' option of the req section of the
17068 config file. They can be printed out with the -text option to req but
17069 are otherwise ignored at present.
17073 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17074 data read consists of only the final block it would not decrypted because
17075 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17076 A misplaced 'break' also meant the decrypted final block might not be
17077 copied until the next read.
17081 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17082 a few extra parameters to the DH structure: these will be useful if
17083 for example we want the value of 'q' or implement X9.42 DH.
17087 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17088 provides hooks that allow the default DSA functions or functions on a
17089 "per key" basis to be replaced. This allows hardware acceleration and
17090 hardware key storage to be handled without major modification to the
17091 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17092 associated functions.
17096 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17097 as "read only": it can't be written to and the buffer it points to will
17098 not be freed. Reading from a read only BIO is much more efficient than
17099 a normal memory BIO. This was added because there are several times when
17100 an area of memory needs to be read from a BIO. The previous method was
17101 to create a memory BIO and write the data to it, this results in two
17102 copies of the data and an O(n^2) reading algorithm. There is a new
17103 function BIO_new_mem_buf() which creates a read only memory BIO from
17104 an area of memory. Also modified the PKCS#7 routines to use read only
17109 * Bugfix: ssl23_get_client_hello did not work properly when called in
17110 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17111 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17112 but a retry condition occurred while trying to read the rest.
17116 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17117 NID_pkcs7_encrypted by default: this was wrong since this should almost
17118 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17119 the encrypted data type: this is a more sensible place to put it and it
17120 allows the PKCS#12 code to be tidied up that duplicated this
17125 * Changed obj_dat.pl script so it takes its input and output files on
17126 the command line. This should avoid shell escape redirection problems
17131 * Initial support for certificate extension requests, these are included
17132 in things like Xenroll certificate requests. Included functions to allow
17133 extensions to be obtained and added.
17137 * -crlf option to s_client and s_server for sending newlines as
17138 CRLF (as required by many protocols).
17142 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17144 * Install libRSAglue.a when OpenSSL is built with RSAref.
17146 *Ralf S. Engelschall*
17148 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17150 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17152 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17157 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17158 DH parameters/keys (q is lost during that conversion, but the resulting
17159 DH parameters contain its length).
17161 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17162 much faster than DH_generate_parameters (which creates parameters
17163 where `p = 2*q + 1`), and also the smaller q makes DH computations
17164 much more efficient (160-bit exponentiation instead of 1024-bit
17165 exponentiation); so this provides a convenient way to support DHE
17166 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17167 utter importance to use
17168 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17170 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17171 when such DH parameters are used, because otherwise small subgroup
17172 attacks may become possible!
17176 * Avoid memory leak in i2d_DHparams.
17180 * Allow the -k option to be used more than once in the enc program:
17181 this allows the same encrypted message to be read by multiple recipients.
17185 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17186 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17187 it will always use the numerical form of the OID, even if it has a short
17192 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17193 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17194 otherwise bn_mod_exp was called. In the case of hardware keys for example
17195 no private key components need be present and it might store extra data
17196 in the RSA structure, which cannot be accessed from bn_mod_exp.
17197 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17198 private key operations.
17202 * Added support for SPARC Linux.
17206 * pem_password_cb function type incompatibly changed from
17207 typedef int pem_password_cb(char *buf, int size, int rwflag);
17209 ....(char *buf, int size, int rwflag, void *userdata);
17210 so that applications can pass data to their callbacks:
17211 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17212 additional void * argument, which is just handed through whenever
17213 the password callback is called.
17215 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17217 New function SSL_CTX_set_default_passwd_cb_userdata.
17219 Compatibility note: As many C implementations push function arguments
17220 onto the stack in reverse order, the new library version is likely to
17221 interoperate with programs that have been compiled with the old
17222 pem_password_cb definition (PEM_whatever takes some data that
17223 happens to be on the stack as its last argument, and the callback
17224 just ignores this garbage); but there is no guarantee whatsoever that
17227 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17228 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17229 problems not only on Windows, but also on some Unix platforms.
17230 To avoid problematic command lines, these definitions are now in an
17231 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17232 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17236 * MIPS III/IV assembler module is reimplemented.
17240 * More DES library cleanups: remove references to srand/rand and
17241 delete an unused file.
17245 * Add support for the free Netwide assembler (NASM) under Win32,
17246 since not many people have MASM (ml) and it can be hard to obtain.
17247 This is currently experimental but it seems to work OK and pass all
17248 the tests. Check out INSTALL.W32 for info.
17252 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17253 without temporary keys kept an extra copy of the server key,
17254 and connections with temporary keys did not free everything in case
17259 * New function RSA_check_key and new openssl rsa option -check
17260 for verifying the consistency of RSA keys.
17262 *Ulf Moeller, Bodo Moeller*
17264 * Various changes to make Win32 compile work:
17265 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17266 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17267 comparison" warnings.
17268 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17272 * Add a debugging option to PKCS#5 v2 key generation function: when
17273 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17274 derived keys are printed to stderr.
17278 * Copy the flags in ASN1_STRING_dup().
17280 *Roman E. Pavlov <pre@mo.msk.ru>*
17282 * The x509 application mishandled signing requests containing DSA
17283 keys when the signing key was also DSA and the parameters didn't match.
17285 It was supposed to omit the parameters when they matched the signing key:
17286 the verifying software was then supposed to automatically use the CA's
17287 parameters if they were absent from the end user certificate.
17289 Omitting parameters is no longer recommended. The test was also
17290 the wrong way round! This was probably due to unusual behaviour in
17291 EVP_cmp_parameters() which returns 1 if the parameters match.
17292 This meant that parameters were omitted when they *didn't* match and
17293 the certificate was useless. Certificates signed with 'ca' didn't have
17296 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17298 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17299 The interface is as follows:
17300 Applications can use
17301 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17302 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17303 "off" is now the default.
17304 The library internally uses
17305 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17306 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17307 to disable memory-checking temporarily.
17309 Some inconsistent states that previously were possible (and were
17310 even the default) are now avoided.
17312 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17313 with each memory chunk allocated; this is occasionally more helpful
17314 than just having a counter.
17316 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17318 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17323 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17324 which largely parallels "options", but is for changing API behaviour,
17325 whereas "options" are about protocol behaviour.
17326 Initial "mode" flags are:
17328 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17329 a single record has been written.
17330 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17331 retries use the same buffer location.
17332 (But all of the contents must be
17337 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17340 * Fix problems with no-hmac etc.
17342 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17344 * New functions RSA_get_default_method(), RSA_set_method() and
17345 RSA_get_method(). These allows replacement of RSA_METHODs without having
17346 to mess around with the internals of an RSA structure.
17350 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17351 Also really enable memory leak checks in openssl.c and in some
17354 *Chad C. Mulligan, Bodo Moeller*
17356 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17357 up the length of negative integers. This has now been simplified to just
17358 store the length when it is first determined and use it later, rather
17359 than trying to keep track of where data is copied and updating it to
17361 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17363 * Add a new function PKCS7_signatureVerify. This allows the verification
17364 of a PKCS#7 signature but with the signing certificate passed to the
17365 function itself. This contrasts with PKCS7_dataVerify which assumes the
17366 certificate is present in the PKCS#7 structure. This isn't always the
17367 case: certificates can be omitted from a PKCS#7 structure and be
17368 distributed by "out of band" means (such as a certificate database).
17372 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17373 function prototypes in pem.h, also change util/mkdef.pl to add the
17374 necessary function names.
17378 * mk1mf.pl (used by Windows builds) did not properly read the
17379 options set by Configure in the top level Makefile, and Configure
17380 was not even able to write more than one option correctly.
17381 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17385 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17386 file to be loaded from a BIO or FILE pointer. The BIO version will
17387 for example allow memory BIOs to contain config info.
17391 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17392 Whoever hopes to achieve shared-library compatibility across versions
17393 must use this, not the compile-time macro.
17394 (Exercise 0.9.4: Which is the minimum library version required by
17396 Note: All this applies only to multi-threaded programs, others don't
17401 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17402 through a BIO pair triggered the default case, i.e.
17403 SSLerr(...,SSL_R_UNKNOWN_STATE).
17407 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17408 can use the SSL library even if none of the specific BIOs is
17413 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17414 for the encoded length.
17416 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17418 * Add initial documentation of the X509V3 functions.
17422 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17423 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17424 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17425 secure PKCS#8 private key format with a high iteration count.
17429 * Fix determination of Perl interpreter: A perl or perl5
17430 *directory* in $PATH was also accepted as the interpreter.
17432 *Ralf S. Engelschall*
17434 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17435 wrong with it but it was very old and did things like calling
17436 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17437 unusual formatting.
17441 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17442 to use the new extension code.
17446 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17447 with macros. This should make it easier to change their form, add extra
17448 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17453 * Add to configuration table a new entry that can specify an alternative
17454 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17455 according to Mark Crispin <MRC@Panda.COM>.
17459 * DES CBC did not update the IV. Weird.
17463 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17464 Changing the behaviour of the former might break existing programs --
17465 where IV updating is needed, des_ncbc_encrypt can be used.
17468 * When bntest is run from "make test" it drives bc to check its
17469 calculations, as well as internally checking them. If an internal check
17470 fails, it needs to cause bc to give a non-zero result or make test carries
17471 on without noticing the failure. Fixed.
17475 * DES library cleanups.
17479 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17480 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17481 ciphers. NOTE: although the key derivation function has been verified
17482 against some published test vectors it has not been extensively tested
17483 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17488 * Instead of "mkdir -p", which is not fully portable, use new
17489 Perl script "util/mkdir-p.pl".
17493 * Rewrite the way password based encryption (PBE) is handled. It used to
17494 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17495 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17496 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17497 the 'parameter' field of the AlgorithmIdentifier is passed to the
17498 underlying key generation function so it must do its own ASN1 parsing.
17499 This has also changed the EVP_PBE_CipherInit() function which now has a
17500 'parameter' argument instead of literal salt and iteration count values
17501 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17505 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17506 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17507 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17508 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17509 value was just used as a "magic string" and not used directly its
17510 value doesn't matter.
17514 * Introduce some semblance of const correctness to BN. Shame C doesn't
17519 * "linux-sparc64" configuration (ultrapenguin).
17521 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17522 "linux-sparc" configuration.
17524 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17526 * config now generates no-xxx options for missing ciphers.
17530 * Support the EBCDIC character set (work in progress).
17531 File ebcdic.c not yet included because it has a different license.
17533 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17535 * Support BS2000/OSD-POSIX.
17537 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17539 * Make callbacks for key generation use `void *` instead of `char *`.
17543 * Make S/MIME samples compile (not yet tested).
17547 * Additional typesafe stacks.
17551 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17555 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17557 * New configuration variant "sco5-gcc".
17559 * Updated some demos.
17561 *Sean O Riordain, Wade Scholine*
17563 * Add missing BIO_free at exit of pkcs12 application.
17567 * Fix memory leak in conf.c.
17571 * Updates for Win32 to assembler version of MD5.
17575 * Set #! path to perl in `apps/der_chop` to where we found it
17576 instead of using a fixed path.
17580 * SHA library changes for irix64-mips4-cc.
17584 * Improvements for VMS support.
17588 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17590 * Bignum library bug fix. IRIX 6 passes "make test" now!
17591 This also avoids the problems with SC4.2 and unpatched SC5.
17593 *Andy Polyakov <appro@fy.chalmers.se>*
17595 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17596 These are required because of the typesafe stack would otherwise break
17597 existing code. If old code used a structure member which used to be STACK
17598 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17599 sk_num or sk_value it would produce an error because the num, data members
17600 are not present in STACK_OF. Now it just produces a warning. sk_set
17601 replaces the old method of assigning a value to sk_value
17602 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17603 that does this will no longer work (and should use sk_set instead) but
17604 this could be regarded as a "questionable" behaviour anyway.
17608 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17609 correctly handle encrypted S/MIME data.
17613 * Change type of various DES function arguments from des_cblock
17614 (which means, in function argument declarations, pointer to char)
17615 to des_cblock * (meaning pointer to array with 8 char elements),
17616 which allows the compiler to do more typechecking; it was like
17617 that back in SSLeay, but with lots of ugly casts.
17619 Introduce new type const_des_cblock.
17623 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17624 problems: find RecipientInfo structure that matches recipient certificate
17625 and initialise the ASN1 structures properly based on passed cipher.
17629 * Belatedly make the BN tests actually check the results.
17633 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17634 to and from BNs: it was completely broken. New compilation option
17635 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17636 key elements as negative integers.
17640 * Reorganize and speed up MD5.
17642 *Andy Polyakov <appro@fy.chalmers.se>*
17646 *Richard Levitte <richard@levitte.org>*
17648 * New option -out to asn1parse to allow the parsed structure to be
17649 output to a file. This is most useful when combined with the -strparse
17650 option to examine the output of things like OCTET STRINGS.
17654 * Make SSL library a little more fool-proof by not requiring any longer
17655 that `SSL_set_{accept,connect}_state` be called before
17656 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17657 in many applications because usually everything *appeared* to work as
17658 intended anyway -- now it really works as intended).
17662 * Move openssl.cnf out of lib/.
17666 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17667 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17668 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17670 *Ralf S. Engelschall*
17672 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17673 handle PKCS#7 enveloped data properly.
17675 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17677 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17678 copying pointers. The cert_st handling is changed by this in
17679 various ways (and thus what used to be known as ctx->default_cert
17680 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17681 any longer when s->cert does not give us what we need).
17682 ssl_cert_instantiate becomes obsolete by this change.
17683 As soon as we've got the new code right (possibly it already is?),
17684 we have solved a couple of bugs of the earlier code where s->cert
17685 was used as if it could not have been shared with other SSL structures.
17687 Note that using the SSL API in certain dirty ways now will result
17688 in different behaviour than observed with earlier library versions:
17689 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17690 does not influence s as it used to.
17692 In order to clean up things more thoroughly, inside SSL_SESSION
17693 we don't use CERT any longer, but a new structure SESS_CERT
17694 that holds per-session data (if available); currently, this is
17695 the peer's certificate chain and, for clients, the server's certificate
17696 and temporary key. CERT holds only those values that can have
17697 meaningful defaults in an SSL_CTX.
17701 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17702 from the internal representation. Various PKCS#7 fixes: remove some
17703 evil casts and set the enc_dig_alg field properly based on the signing
17708 * Allow PKCS#12 password to be set from the command line or the
17709 environment. Let 'ca' get its config file name from the environment
17710 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17715 * Allow certificate policies extension to use an IA5STRING for the
17716 organization field. This is contrary to the PKIX definition but
17717 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17722 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17723 without disallowing inline assembler and the like for non-pedantic builds.
17727 * Support Borland C++ builder.
17729 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17735 * SHA-1 cleanups and performance enhancements.
17737 *Andy Polyakov <appro@fy.chalmers.se>*
17739 * Sparc v8plus assembler for the bignum library.
17741 *Andy Polyakov <appro@fy.chalmers.se>*
17743 * Accept any -xxx and +xxx compiler options in Configure.
17747 * Update HPUX configuration.
17751 * Add missing `sk_<type>_unshift()` function to safestack.h
17753 *Ralf S. Engelschall*
17755 * New function SSL_CTX_use_certificate_chain_file that sets the
17756 "extra_cert"s in addition to the certificate. (This makes sense
17757 only for "PEM" format files, as chains as a whole are not
17762 * Support verify_depth from the SSL API.
17763 x509_vfy.c had what can be considered an off-by-one-error:
17764 Its depth (which was not part of the external interface)
17765 was actually counting the number of certificates in a chain;
17766 now it really counts the depth.
17770 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17771 instead of X509err, which often resulted in confusing error
17772 messages since the error codes are not globally unique
17773 (e.g. an alleged error in ssl3_accept when a certificate
17774 didn't match the private key).
17776 * New function SSL_CTX_set_session_id_context that allows to set a default
17777 value (so that you don't need SSL_set_session_id_context for each
17778 connection using the SSL_CTX).
17782 * OAEP decoding bug fix.
17786 * Support INSTALL_PREFIX for package builders, as proposed by
17791 * New Configure options "threads" and "no-threads". For systems
17792 where the proper compiler options are known (currently Solaris
17793 and Linux), "threads" is the default.
17797 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17801 * Install various scripts to $(OPENSSLDIR)/misc, not to
17802 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17803 such as /usr/local/bin.
17807 * "make linux-shared" to build shared libraries.
17809 *Niels Poppe <niels@netbox.org>*
17811 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17815 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17816 extension adding in x509 utility.
17820 * Remove NOPROTO sections and error code comments.
17824 * Partial rewrite of the DEF file generator to now parse the ANSI
17829 * New Configure options --prefix=DIR and --openssldir=DIR.
17833 * Complete rewrite of the error code script(s). It is all now handled
17834 by one script at the top level which handles error code gathering,
17835 header rewriting and C source file generation. It should be much better
17836 than the old method: it now uses a modified version of Ulf's parser to
17837 read the ANSI prototypes in all header files (thus the old K&R definitions
17838 aren't needed for error creation any more) and do a better job of
17839 translating function codes into names. The old 'ASN1 error code embedded
17840 in a comment' is no longer necessary and it doesn't use .err files which
17841 have now been deleted. Also the error code call doesn't have to appear all
17842 on one line (which resulted in some large lines...).
17846 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17850 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17851 0 (which usually indicates a closed connection), but continue reading.
17855 * Fix some race conditions.
17859 * Add support for CRL distribution points extension. Add Certificate
17860 Policies and CRL distribution points documentation.
17864 * Move the autogenerated header file parts to crypto/opensslconf.h.
17868 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17869 8 of keying material. Merlin has also confirmed interop with this fix
17870 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17872 *Merlin Hughes <merlin@baltimore.ie>*
17874 * Fix lots of warnings.
17876 *Richard Levitte <levitte@stacken.kth.se>*
17878 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17879 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17881 *Richard Levitte <levitte@stacken.kth.se>*
17883 * Fix problems with sizeof(long) == 8.
17885 *Andy Polyakov <appro@fy.chalmers.se>*
17887 * Change functions to ANSI C.
17891 * Fix typos in error codes.
17893 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17895 * Remove defunct assembler files from Configure.
17899 * SPARC v8 assembler BIGNUM implementation.
17901 *Andy Polyakov <appro@fy.chalmers.se>*
17903 * Support for Certificate Policies extension: both print and set.
17904 Various additions to support the r2i method this uses.
17908 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17909 return a const string when you are expecting an allocated buffer.
17913 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17914 types DirectoryString and DisplayText.
17918 * Add code to allow r2i extensions to access the configuration database,
17919 add an LHASH database driver and add several ctx helper functions.
17923 * Fix an evil bug in bn_expand2() which caused various BN functions to
17924 fail when they extended the size of a BIGNUM.
17928 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17929 support typesafe stack.
17933 * Fix typo in SSL_[gs]et_options().
17935 *Nils Frostberg <nils@medcom.se>*
17937 * Delete various functions and files that belonged to the (now obsolete)
17938 old X509V3 handling code.
17942 * New Configure option "rsaref".
17946 * Don't auto-generate pem.h.
17950 * Introduce type-safe ASN.1 SETs.
17954 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17956 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17958 * Introduce type-safe STACKs. This will almost certainly break lots of code
17959 that links with OpenSSL (well at least cause lots of warnings), but fear
17960 not: the conversion is trivial, and it eliminates loads of evil casts. A
17961 few STACKed things have been converted already. Feel free to convert more.
17962 In the fullness of time, I'll do away with the STACK type altogether.
17966 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17967 specified in `<certfile>` by updating the entry in the index.txt file.
17968 This way one no longer has to edit the index.txt file manually for
17969 revoking a certificate. The -revoke option does the gory details now.
17971 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17973 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17974 `-text` option at all and this way the `-noout -text` combination was
17975 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17977 *Ralf S. Engelschall*
17979 * Make sure a corresponding plain text error message exists for the
17980 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17981 verify callback function determined that a certificate was revoked.
17983 *Ralf S. Engelschall*
17985 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17986 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17987 all available ciphers including rc5, which was forgotten until now.
17988 In order to let the testing shell script know which algorithms
17989 are available, a new (up to now undocumented) command
17990 `openssl list-cipher-commands` is used.
17994 * Bugfix: s_client occasionally would sleep in select() when
17995 it should have checked SSL_pending() first.
17999 * New functions DSA_do_sign and DSA_do_verify to provide access to
18000 the raw DSA values prior to ASN.1 encoding.
18004 * Tweaks to Configure
18006 *Niels Poppe <niels@netbox.org>*
18008 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18013 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18017 * New config option to avoid instructions that are illegal on the 80386.
18018 The default code is faster, but requires at least a 486.
18022 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18023 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18024 same as SSL2_VERSION anyway.
18028 * New "-showcerts" option for s_client.
18032 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18033 application. Various cleanups and fixes.
18037 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18038 modify error routines to work internally. Add error codes and PBE init
18039 to library startup routines.
18043 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18044 packing functions to asn1 and evp. Changed function names and error
18045 codes along the way.
18049 * PKCS12 integration: and so it begins... First of several patches to
18050 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18051 objects to objects.h
18055 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18056 and display support for Thawte strong extranet extension.
18060 * Add LinuxPPC support.
18062 *Jeff Dubrule <igor@pobox.org>*
18064 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18065 bn_div_words in alpha.s.
18067 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18069 * Make sure the RSA OAEP test is skipped under -DRSAref because
18070 OAEP isn't supported when OpenSSL is built with RSAref.
18072 *Ulf Moeller <ulf@fitug.de>*
18074 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18075 so they no longer are missing under -DNOPROTO.
18077 *Soren S. Jorvang <soren@t.dk>*
18079 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18081 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18082 doesn't work when the session is reused. Coming soon!
18086 * Fix a security hole, that allows sessions to be reused in the wrong
18087 context thus bypassing client cert protection! All software that uses
18088 client certs and session caches in multiple contexts NEEDS PATCHING to
18089 allow session reuse! A fuller solution is in the works.
18091 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18093 * Some more source tree cleanups (removed obsolete files
18094 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18095 permission on "config" script to be executable) and a fix for the INSTALL
18098 *Ulf Moeller <ulf@fitug.de>*
18100 * Remove some legacy and erroneous uses of malloc, free instead of
18103 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18105 * Make rsa_oaep_test return non-zero on error.
18107 *Ulf Moeller <ulf@fitug.de>*
18109 * Add support for native Solaris shared libraries. Configure
18110 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18111 if someone would make that last step automatic.
18113 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18115 * ctx_size was not built with the right compiler during "make links". Fixed.
18119 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18120 except NULL ciphers". This means the default cipher list will no longer
18121 enable NULL ciphers. They need to be specifically enabled e.g. with
18122 the string "DEFAULT:eNULL".
18126 * Fix to RSA private encryption routines: if p < q then it would
18127 occasionally produce an invalid result. This will only happen with
18128 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18132 * Be less restrictive and allow also `perl util/perlpath.pl
18133 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18134 because this way one can also use an interpreter named `perl5` (which is
18135 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18136 installed as `perl`).
18138 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18140 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18142 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18144 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18145 advapi32.lib to Win32 build and change the pem test comparison
18146 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18147 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18148 and crypto/des/ede_cbcm_enc.c.
18152 * DES quad checksum was broken on big-endian architectures. Fixed.
18156 * Comment out two functions in bio.h that aren't implemented. Fix up the
18157 Win32 test batch file so it (might) work again. The Win32 test batch file
18158 is horrible: I feel ill....
18162 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18163 in e_os.h. Audit of header files to check ANSI and non ANSI
18164 sections: 10 functions were absent from non ANSI section and not exported
18165 from Windows DLLs. Fixed up libeay.num for new functions.
18169 * Make `openssl version` output lines consistent.
18171 *Ralf S. Engelschall*
18173 * Fix Win32 symbol export lists for BIO functions: Added
18174 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18175 to ms/libeay{16,32}.def.
18177 *Ralf S. Engelschall*
18179 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18180 fine under Unix and passes some trivial tests I've now added. But the
18181 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18182 added to make sure no one expects that this stuff really works in the
18183 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18184 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18187 *Ralf S. Engelschall*
18189 * Fix the generation of two part addresses in perl.
18191 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18193 * Add config entry for Linux on MIPS.
18195 *John Tobey <jtobey@channel1.com>*
18197 * Make links whenever Configure is run, unless we are on Windoze.
18201 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18202 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18207 * Add a useful kludge to allow package maintainers to specify compiler and
18208 other platforms details on the command line without having to patch the
18209 Configure script every time: One now can use
18210 `perl Configure <id>:<details>`,
18211 i.e. platform ids are allowed to have details appended
18212 to them (separated by colons). This is treated as there would be a static
18213 pre-configured entry in Configure's %table under key `<id>` with value
18214 `<details>` and `perl Configure <id>` is called. So, when you want to
18215 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18216 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18217 now, which overrides the FreeBSD-elf entry on-the-fly.
18219 *Ralf S. Engelschall*
18221 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18225 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18226 on the `perl Configure ...` command line. This way one can compile
18227 OpenSSL libraries with Position Independent Code (PIC) which is needed
18228 for linking it into DSOs.
18230 *Ralf S. Engelschall*
18232 * Remarkably, export ciphers were totally broken and no-one had noticed!
18237 * Cleaned up the LICENSE document: The official contact for any license
18238 questions now is the OpenSSL core team under openssl-core@openssl.org.
18239 And add a paragraph about the dual-license situation to make sure people
18240 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18241 to the OpenSSL toolkit.
18243 *Ralf S. Engelschall*
18245 * General source tree makefile cleanups: Made `making xxx in yyy...`
18246 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18247 Additionally cleaned up the `make links` target: Remove unnecessary
18248 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18249 to speed processing and no longer clutter the display with confusing
18250 stuff. Instead only the actually done links are displayed.
18252 *Ralf S. Engelschall*
18254 * Permit null encryption ciphersuites, used for authentication only. It used
18255 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18256 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18261 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18262 signed attributes when verifying signatures (this would break them),
18263 the detached data encoding was wrong and public keys obtained using
18264 X509_get_pubkey() weren't freed.
18268 * Add text documentation for the BUFFER functions. Also added a work around
18269 to a Win95 console bug. This was triggered by the password read stuff: the
18270 last character typed gets carried over to the next fread(). If you were
18271 generating a new cert request using 'req' for example then the last
18272 character of the passphrase would be CR which would then enter the first
18277 * Added the new 'Includes OpenSSL Cryptography Software' button as
18278 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18279 button and can be used by applications based on OpenSSL to show the
18280 relationship to the OpenSSL project.
18282 *Ralf S. Engelschall*
18284 * Remove confusing variables in function signatures in files
18285 ssl/ssl_lib.c and ssl/ssl.h.
18287 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18289 * Don't install bss_file.c under PREFIX/include/
18291 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18293 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18294 functions that return function pointers and has support for NT specific
18295 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18296 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18297 unsigned to signed types: this was killing the Win32 compile.
18301 * Add new certificate file to stack functions,
18302 SSL_add_dir_cert_subjects_to_stack() and
18303 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18304 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18305 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18306 This means that Apache-SSL and similar packages don't have to mess around
18307 to add as many CAs as they want to the preferred list.
18311 * Experiment with doxygen documentation. Currently only partially applied to
18313 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18314 openssl.doxy as the configuration file.
18318 * Get rid of remaining C++-style comments which strict C compilers hate.
18320 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18322 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18323 compiled in by default: it has problems with large keys.
18327 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18328 DH private keys and/or callback functions which directly correspond to
18329 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18330 is needed for applications which have to configure certificates on a
18331 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18333 For the RSA certificate situation is makes no difference, but
18334 for the DSA certificate situation this fixes the "no shared cipher"
18335 problem where the OpenSSL cipher selection procedure failed because the
18336 temporary keys were not overtaken from the context and the API provided
18337 no way to reconfigure them.
18338 The new functions now let applications reconfigure the stuff and they
18339 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18340 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18341 non-public-API function ssl_cert_instantiate() is used as a helper
18342 function and also to reduce code redundancy inside ssl_rsa.c.
18344 *Ralf S. Engelschall*
18346 * Move s_server -dcert and -dkey options out of the undocumented feature
18347 area because they are useful for the DSA situation and should be
18348 recognized by the users.
18350 *Ralf S. Engelschall*
18352 * Fix the cipher decision scheme for export ciphers: the export bits are
18353 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18354 SSL_EXP_MASK. So, the original variable has to be used instead of the
18355 already masked variable.
18357 *Richard Levitte <levitte@stacken.kth.se>*
18359 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18361 *Richard Levitte <levitte@stacken.kth.se>*
18363 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18364 from `int` to `unsigned int` because it is a length and initialized by
18365 EVP_DigestFinal() which expects an `unsigned int *`.
18367 *Richard Levitte <levitte@stacken.kth.se>*
18369 * Don't hard-code path to Perl interpreter on shebang line of Configure
18370 script. Instead use the usual Shell->Perl transition trick.
18372 *Ralf S. Engelschall*
18374 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18375 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18376 -noout -modulus` as it's already the case for `openssl rsa -noout
18377 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18378 currently the public key is printed (a decision which was already done by
18379 `openssl dsa -modulus` in the past) which serves a similar purpose.
18380 Additionally the NO_RSA no longer completely removes the whole -modulus
18381 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18384 *Ralf S. Engelschall*
18386 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18387 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18389 *Arne Ansper <arne@ats.cyber.ee>*
18391 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18392 to be added. Now both 'req' and 'ca' can use new objects defined in the
18397 * Add cool BIO that does syslog (or event log on NT).
18399 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18401 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18402 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18403 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18404 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18408 * Add preliminary config info for new extension code.
18412 * Make RSA_NO_PADDING really use no padding.
18414 *Ulf Moeller <ulf@fitug.de>*
18416 * Generate errors when private/public key check is done.
18420 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18421 for some CRL extensions and new objects added.
18425 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18426 key usage extension and fuller support for authority key id.
18430 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18431 padding method for RSA, which is recommended for new applications in PKCS
18432 #1 v2.0 (RFC 2437, October 1998).
18433 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18434 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18435 against Bleichbacher's attack on RSA.
18436 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18439 * Updates to the new SSL compression code
18441 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18443 * Fix so that the version number in the master secret, when passed
18444 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18445 (because the server will not accept higher), that the version number
18446 is 0x03,0x01, not 0x03,0x00
18448 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18450 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18451 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18452 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18456 * Support for RAW extensions where an arbitrary extension can be
18457 created by including its DER encoding. See `apps/openssl.cnf` for
18462 * Make sure latest Perl versions don't interpret some generated C array
18463 code as Perl array code in the crypto/err/err_genc.pl script.
18465 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18467 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18468 not many people have the assembler. Various Win32 compilation fixes and
18469 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18470 build instructions.
18474 * Modify configure script 'Configure' to automatically create crypto/date.h
18475 file under Win32 and also build pem.h from pem.org. New script
18476 util/mkfiles.pl to create the MINFO file on environments that can't do a
18477 'make files': perl util/mkfiles.pl >MINFO should work.
18481 * Major rework of DES function declarations, in the pursuit of correctness
18482 and purity. As a result, many evil casts evaporated, and some weirdness,
18483 too. You may find this causes warnings in your code. Zapping your evil
18484 casts will probably fix them. Mostly.
18488 * Fix for a typo in asn1.h. Bug fix to object creation script
18489 obj_dat.pl. It considered a zero in an object definition to mean
18490 "end of object": none of the objects in objects.h have any zeros
18491 so it wasn't spotted.
18493 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18495 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18496 Masking (CBCM). In the absence of test vectors, the best I have been able
18497 to do is check that the decrypt undoes the encrypt, so far. Send me test
18498 vectors if you have them.
18502 * Correct calculation of key length for export ciphers (too much space was
18503 allocated for null ciphers). This has not been tested!
18507 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18508 message is now correct (it understands "crypto" and "ssl" on its
18509 command line). There is also now an "update" option. This will update
18510 the util/ssleay.num and util/libeay.num files with any new functions.
18512 perl util/mkdef.pl crypto ssl update
18513 it will update them.
18517 * Overhauled the Perl interface:
18518 - ported BN stuff to OpenSSL's different BN library
18519 - made the perl/ source tree CVS-aware
18520 - renamed the package from SSLeay to OpenSSL (the files still contain
18521 their history because I've copied them in the repository)
18522 - removed obsolete files (the test scripts will be replaced
18523 by better Test::Harness variants in the future)
18525 *Ralf S. Engelschall*
18527 * First cut for a very conservative source tree cleanup:
18528 1. merge various obsolete readme texts into doc/ssleay.txt
18529 where we collect the old documents and readme texts.
18530 2. remove the first part of files where I'm already sure that we no
18531 longer need them because of three reasons: either they are just temporary
18532 files which were left by Eric or they are preserved original files where
18533 I've verified that the diff is also available in the CVS via "cvs diff
18534 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18535 the crypto/md/ stuff).
18537 *Ralf S. Engelschall*
18539 * More extension code. Incomplete support for subject and issuer alt
18540 name, issuer and authority key id. Change the i2v function parameters
18541 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18542 what that's for :-) Fix to ASN1 macro which messed up
18543 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18547 * Preliminary support for ENUMERATED type. This is largely copied from the
18552 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18554 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18556 * Make sure `make rehash` target really finds the `openssl` program.
18558 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18560 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18561 like to hear about it if this slows down other processors.
18565 * Add CygWin32 platform information to Configure script.
18567 *Alan Batie <batie@aahz.jf.intel.com>*
18569 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18571 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18573 * New program nseq to manipulate netscape certificate sequences
18577 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18582 * Fixes to BN code. Previously the default was to define BN_RECURSION
18583 but the BN code had some problems that would cause failures when
18584 doing certificate verification and some other functions.
18586 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18588 * Add ASN1 and PEM code to support netscape certificate sequences.
18592 * Add ASN1 and PEM code to support netscape certificate sequences.
18596 * Add several PKIX and private extended key usage OIDs.
18600 * Modify the 'ca' program to handle the new extension code. Modify
18601 openssl.cnf for new extension format, add comments.
18605 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18606 and add a sample to openssl.cnf so req -x509 now adds appropriate
18611 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18612 error code, add initial support to X509_print() and x509 application.
18616 * Takes a deep breath and start adding X509 V3 extension support code. Add
18617 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18618 stuff is currently isolated and isn't even compiled yet.
18622 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18623 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18624 Removed the versions check from X509 routines when loading extensions:
18625 this allows certain broken certificates that don't set the version
18626 properly to be processed.
18630 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18631 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18632 can still be regenerated with "make depend".
18636 * Spelling mistake in C version of CAST-128.
18638 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18640 * Changes to the error generation code. The perl script err-code.pl
18641 now reads in the old error codes and retains the old numbers, only
18642 adding new ones if necessary. It also only changes the .err files if new
18643 codes are added. The makefiles have been modified to only insert errors
18644 when needed (to avoid needlessly modifying header files). This is done
18645 by only inserting errors if the .err file is newer than the auto generated
18646 C file. To rebuild all the error codes from scratch (the old behaviour)
18647 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18648 or delete all the .err files.
18652 * CAST-128 was incorrectly implemented for short keys. The C version has
18653 been fixed, but is untested. The assembler versions are also fixed, but
18654 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18655 to regenerate it if needed.
18656 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18657 Hagino <itojun@kame.net>*
18659 * File was opened incorrectly in randfile.c.
18661 *Ulf Möller <ulf@fitug.de>*
18663 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18664 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18665 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18666 al: it's just almost always a UTCTime. Note this patch adds new error
18667 codes so do a "make errors" if there are problems.
18671 * Correct Linux 1 recognition in config.
18673 *Ulf Möller <ulf@fitug.de>*
18675 * Remove pointless MD5 hash when using DSA keys in ca.
18677 *Anonymous <nobody@replay.com>*
18679 * Generate an error if given an empty string as a cert directory. Also
18680 generate an error if handed NULL (previously returned 0 to indicate an
18681 error, but didn't set one).
18683 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18685 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18689 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18690 parameters. This was causing a warning which killed off the Win32 compile.
18694 * Remove C++ style comments from crypto/bn/bn_local.h.
18696 *Neil Costigan <neil.costigan@celocom.com>*
18698 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18699 based on a text string, looking up short and long names and finally
18700 "dot" format. The "dot" format stuff didn't work. Added new function
18701 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18702 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18703 OID is not part of the table.
18707 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18708 X509_LOOKUP_by_alias().
18712 * Sort openssl functions by name.
18716 * Get the `gendsa` command working and add it to the `list` command. Remove
18717 encryption from sample DSA keys (in case anyone is interested the password
18722 * Make *all* `*_free` functions accept a NULL pointer.
18724 *Frans Heymans <fheymans@isaserver.be>*
18726 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18729 *Anonymous <nobody@replay.com>*
18731 * s_server should send the CAfile as acceptable CAs, not its own cert.
18733 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18735 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18737 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18739 * Temp key "for export" tests were wrong in s3_srvr.c.
18741 *Anonymous <nobody@replay.com>*
18743 * Add prototype for temp key callback functions
18744 SSL_CTX_set_tmp_{rsa,dh}_callback().
18748 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18749 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18753 * X509_name_add_entry() freed the wrong thing after an error.
18755 *Arne Ansper <arne@ats.cyber.ee>*
18757 * rsa_eay.c would attempt to free a NULL context.
18759 *Arne Ansper <arne@ats.cyber.ee>*
18761 * BIO_s_socket() had a broken should_retry() on Windoze.
18763 *Arne Ansper <arne@ats.cyber.ee>*
18765 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18767 *Arne Ansper <arne@ats.cyber.ee>*
18769 * Make sure the already existing X509_STORE->depth variable is initialized
18770 in X509_STORE_new(), but document the fact that this variable is still
18771 unused in the certificate verification process.
18773 *Ralf S. Engelschall*
18775 * Fix the various library and `apps/` files to free up pkeys obtained from
18776 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18780 * Fix reference counting in X509_PUBKEY_get(). This makes
18781 demos/maurice/example2.c work, amongst others, probably.
18783 *Steve Henson and Ben Laurie*
18785 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18786 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18787 are no longer created. This way we have a single and consistent command
18788 line interface `openssl <command>`, similar to `cvs <command>`.
18790 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18792 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18793 BIT STRING wrapper always have zero unused bits.
18797 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18801 * Make the top-level INSTALL documentation easier to understand.
18805 * Makefiles updated to exit if an error occurs in a sub-directory
18806 make (including if user presses ^C) [Paul Sutton]
18808 * Make Montgomery context stuff explicit in RSA data structure.
18812 * Fix build order of pem and err to allow for generated pem.h.
18816 * Fix renumbering bug in X509_NAME_delete_entry().
18820 * Enhanced the err-ins.pl script so it makes the error library number
18821 global and can add a library name. This is needed for external ASN1 and
18822 other error libraries.
18826 * Fixed sk_insert which never worked properly.
18830 * Fix ASN1 macros so they can handle indefinite length constructed
18831 EXPLICIT tags. Some non standard certificates use these: they can now
18836 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18837 into a single doc/ssleay.txt bundle. This way the information is still
18838 preserved but no longer messes up this directory. Now it's new room for
18839 the new set of documentation files.
18841 *Ralf S. Engelschall*
18843 * SETs were incorrectly DER encoded. This was a major pain, because they
18844 shared code with SEQUENCEs, which aren't coded the same. This means that
18845 almost everything to do with SETs or SEQUENCEs has either changed name or
18846 number of arguments.
18848 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18850 * Fix test data to work with the above.
18854 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18855 was already fixed by Eric for 0.9.1 it seems.
18857 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18859 * Autodetect FreeBSD3.
18863 * Fix various bugs in Configure. This affects the following platforms:
18867 unixware-2.0-pentium
18872 * Eliminate generated files from CVS. Reorder tests to regenerate files
18873 before they are needed.
18877 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18881 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18883 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18884 changed SSLeay to OpenSSL in version strings.
18886 *Ralf S. Engelschall*
18888 * Some fixups to the top-level documents.
18892 * Fixed the nasty bug where rsaref.h was not found under compile-time
18893 because the symlink to include/ was missing.
18895 *Ralf S. Engelschall*
18897 * Incorporated the popular no-RSA/DSA-only patches
18898 which allow to compile a RSA-free SSLeay.
18900 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18902 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18903 when "ssleay" is still not found.
18905 *Ralf S. Engelschall*
18907 * Added more platforms to Configure: Cray T3E, HPUX 11,
18909 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18911 * Updated the README file.
18913 *Ralf S. Engelschall*
18915 * Added various .cvsignore files in the CVS repository subdirs
18916 to make a "cvs update" really silent.
18918 *Ralf S. Engelschall*
18920 * Recompiled the error-definition header files and added
18921 missing symbols to the Win32 linker tables.
18923 *Ralf S. Engelschall*
18925 * Cleaned up the top-level documents;
18926 o new files: CHANGES and LICENSE
18927 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18928 o merged COPYRIGHT into LICENSE
18929 o removed obsolete TODO file
18930 o renamed MICROSOFT to INSTALL.W32
18932 *Ralf S. Engelschall*
18934 * Removed dummy files from the 0.9.1b source tree:
18935 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18936 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18937 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18938 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18939 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18941 *Ralf S. Engelschall*
18943 * Added various platform portability fixes.
18947 * The Genesis of the OpenSSL rpject:
18948 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18949 Young and Tim J. Hudson created while they were working for C2Net until
18952 *The OpenSSL Project*
18954 ### Changes between 0.9.0b and 0.9.1b [not released]
18956 * Updated a few CA certificates under certs/
18960 * Changed some BIGNUM api stuff.
18964 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18965 DGUX x86, Linux Alpha, etc.
18969 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18970 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18975 * Add -strparse option to asn1pars program which parses nested
18978 *Dr Stephen Henson <shenson@bigfoot.com>*
18980 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18984 * DSA fix for "ca" program.
18988 * Added "-genkey" option to "dsaparam" program.
18992 * Added RIPE MD160 (rmd160) message digest.
18996 * Added -a (all) option to "ssleay version" command.
19000 * Added PLATFORM define which is the id given to Configure.
19004 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19008 * Extended the ASN.1 parser routines.
19012 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19016 * Added a BN_CTX to the BN library.
19020 * Fixed the weak key values in DES library
19024 * Changed API in EVP library for cipher aliases.
19028 * Added support for RC2/64bit cipher.
19032 * Converted the lhash library to the crypto/mem.c functions.
19036 * Added more recognized ASN.1 object ids.
19040 * Added more RSA padding checks for SSL/TLS.
19044 * Added BIO proxy/filter functionality.
19048 * Added extra_certs to SSL_CTX which can be used
19049 send extra CA certificates to the client in the CA cert chain sending
19050 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19054 * Now Fortezza is denied in the authentication phase because
19055 this is key exchange mechanism is not supported by SSLeay at all.
19059 * Additional PKCS1 checks.
19063 * Support the string "TLSv1" for all TLS v1 ciphers.
19067 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19068 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19072 * Fixed a few memory leaks.
19076 * Fixed various code and comment typos.
19080 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19081 bytes sent in the client random.
19083 *Edward Bishop <ebishop@spyglass.com>*
19087 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19088 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19089 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19090 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19091 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19092 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19093 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19094 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19095 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19096 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19097 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19098 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19099 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19100 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19101 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19102 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19103 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19104 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19105 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19106 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19107 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19108 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19109 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19110 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19111 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19112 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19113 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19114 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19115 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19116 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19117 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19118 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19119 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19120 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19121 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19122 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19123 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19124 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19125 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19126 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19127 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19128 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19129 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19130 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19131 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19132 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19133 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19134 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19135 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19136 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19137 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19138 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19139 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19140 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19141 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19142 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19143 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19144 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19145 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19146 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19147 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19148 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19149 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19150 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19151 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19152 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19153 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19154 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19155 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19156 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19157 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19158 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19159 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19160 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19161 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19162 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19163 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19164 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19165 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19166 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19167 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19168 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19169 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19170 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19171 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19172 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19173 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19174 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19175 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19176 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19177 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19178 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19179 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19180 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19181 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19182 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19183 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19184 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19185 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19186 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19187 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19188 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19189 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19190 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19191 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19192 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19193 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19194 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19195 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19196 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19197 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19198 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19199 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19200 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19201 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19202 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19203 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19204 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19205 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19206 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19207 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19208 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19209 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19210 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19211 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19212 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19213 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19214 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19215 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19216 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19217 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19218 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19219 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19220 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19221 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19222 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19223 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19224 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19225 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19226 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19227 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19228 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19229 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19230 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19231 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19232 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19233 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19234 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19235 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19236 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19237 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19238 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19239 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19240 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19241 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19242 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19243 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19244 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19245 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19246 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19247 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19248 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655