4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
25 listed here are only a brief description.
26 The migration guide contains more detailed information related to new features,
27 breaking changes, and mappings for the large list of deprecated functions.
29 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
31 ### Changes between 3.0.1 and 3.0.2 [xx XXX xxxx]
33 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
38 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
40 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
41 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
42 verify a certificate supplied by a server. That function may return a
43 negative return value to indicate an internal error (for example out of
44 memory). Such a negative return value is mishandled by OpenSSL and will cause
45 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
46 success and a subsequent call to SSL_get_error() to return the value
47 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
48 returned by OpenSSL if the application has previously called
49 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
50 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
51 totally unexpected and applications may not behave correctly as a result. The
52 exact behaviour will depend on the application but it could result in
53 crashes, infinite loops or other similar incorrect responses.
55 This issue is made more serious in combination with a separate bug in OpenSSL
56 3.0 that will cause X509_verify_cert() to indicate an internal error when
57 processing a certificate chain. This will occur where a certificate does not
58 include the Subject Alternative Name extension but where a Certificate
59 Authority has enforced name constraints. This issue can occur even with valid
65 * Corrected a few file name and file reference bugs in the build,
66 installation and setup scripts, which lead to installation verification
67 failures. Slightly enhanced the installation verification script.
71 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
76 * Fixed PVK encoder to properly query for the passphrase.
80 * Multiple fixes in the OSSL_HTTP API functions.
84 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
85 OSSL_PARAM_INTEGER data type and return error on negative numbers
86 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
87 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
91 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
95 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
99 * Multiple threading fixes.
103 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
107 * Allow fetching an operation from the provider that owns an unexportable key
108 as a fallback if that is still allowed by the property query.
112 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
114 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
119 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
120 S390X capability vector to zero. This simplifies testing of different code
121 paths on S390X architecture.
125 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
126 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
127 SP 800-38D". The communication will fail at this point.
131 * The EC_GROUP_clear_free() function is deprecated as there is nothing
132 confidential in EC_GROUP data.
136 * The byte order mark (BOM) character is ignored if encountered at the
137 beginning of a PEM-formatted file.
141 * Added CMS support for the Russian GOST algorithms.
145 * Due to move of the implementation of cryptographic operations
146 to the providers, validation of various operation parameters can
147 be postponed until the actual operation is executed where previously
148 it happened immediately when an operation parameter was set.
150 For example when setting an unsupported curve with
151 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
152 fail but later keygen operations with the EVP_PKEY_CTX will fail.
154 *OpenSSL team members and many third party contributors*
156 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
157 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
158 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
159 instead to retrieve these algorithms from a provider.
163 * On build targets where the multilib postfix is set in the build
164 configuration the libdir directory was changing based on whether
165 the lib directory with the multilib postfix exists on the system
166 or not. This unpredictable behavior was removed and eventual
167 multilib postfix is now always added to the default libdir. Use
168 `--libdir=lib` to override the libdir if adding the postfix is
173 * The triple DES key wrap functionality now conforms to RFC 3217 but is
174 no longer interoperable with OpenSSL 1.1.1.
178 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
179 function codes, this function can only cause problems for calling
184 * Add a configurable flag to output date formats as ISO 8601. Does not
185 change the default date format.
189 * Version of MSVC earlier than 1300 could get link warnings, which could
190 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
191 Support for this flag has been removed.
195 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
196 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
197 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
198 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
199 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
203 * The signatures of the functions to get and set options on SSL and
204 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
205 Some source code changes may be required.
209 * The public definitions of conf_method_st and conf_st have been
210 deprecated. They will be made opaque in a future release.
212 *Rich Salz and Tomáš Mráz*
214 * Client-initiated renegotiation is disabled by default. To allow it, use
215 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
216 flag, or the "ClientRenegotiation" config parameter as appropriate.
220 * Add "abspath" and "includedir" pragma's to config files, to prevent,
221 or modify relative pathname inclusion.
225 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
226 validated. Please consult the README-FIPS and
227 README-PROVIDERS files, as well as the migration guide.
229 *OpenSSL team members and many third party contributors*
231 * For the key types DH and DHX the allowed settable parameters are now different.
235 * The openssl commands that read keys, certificates, and CRLs now
236 automatically detect the PEM or DER format of the input files.
238 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
240 * Added enhanced PKCS#12 APIs which accept a library context.
244 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
248 * Added support for Kernel TLS (KTLS).
250 *Boris Pismenny, John Baldwin and Andrew Gallatin*
252 * Support for RFC 5746 secure renegotiation is now required by default for
253 SSL or TLS connections to succeed.
257 * The signature of the `copy` functional parameter of the
258 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
259 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
260 the signature of the `pub_decode` functional parameter of the
261 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
262 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
266 * The error return values from some control calls (ctrl) have changed.
270 * A public key check is now performed during EVP_PKEY_derive_set_peer().
274 * Many functions in the EVP_ namespace that are getters of values from
275 implementations or contexts were renamed to include get or get0 in their
276 names. Old names are provided as macro aliases for compatibility and
281 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
282 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
283 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
288 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
291 * The output from the command line applications may have minor
296 * The output from numerous "printing" may have minor changes.
300 * Windows thread synchronization uses read/write primitives (SRWLock) when
301 supported by the OS, otherwise CriticalSection continues to be used.
305 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
306 work on read only BIO source/sinks that do not support these functions.
307 This allows piping or redirection of a file BIO using stdin to be buffered
308 into memory. This is used internally in OSSL_DECODER_from_bio().
312 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
313 this function would return one of the values OSSL_STORE_INFO_NAME,
314 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
315 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
316 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
317 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
318 using this function should be amended to handle the changed return value.
322 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
323 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
324 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
325 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
326 contains more than one certificate identifier: This means that all
327 certificates referenced there MUST be part of the validation chain.
331 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
332 RC5, DESX and DES have been moved to the legacy provider.
336 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
337 RIPEMD-160 have been moved to the legacy provider.
341 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
346 * The deprecated functions EVP_PKEY_get0_RSA(),
347 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
348 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
349 well as the similarly named "get1" functions behave differently in
354 * A number of functions handling low-level keys or engines were deprecated
355 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
356 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
357 EVP_PKEY_get0_siphash().
361 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
362 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
363 will need to load the legacy crypto provider. This includes these PBE
364 algorithms which use this KDF:
365 - NID_pbeWithMD2AndDES_CBC
366 - NID_pbeWithMD5AndDES_CBC
367 - NID_pbeWithSHA1AndRC2_CBC
368 - NID_pbeWithMD2AndRC2_CBC
369 - NID_pbeWithMD5AndRC2_CBC
370 - NID_pbeWithSHA1AndDES_CBC
374 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
375 BIO_debug_callback() functions.
379 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
380 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
384 * The RAND_METHOD APIs have been deprecated.
388 * The SRP APIs have been deprecated.
392 * Add a compile time option to prevent the caching of provider fetched
393 algorithms. This is enabled by including the no-cached-fetch option
394 at configuration time.
398 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
399 count of PKCS12_DEFAULT_ITER.
401 *Tomáš Mráz and Sahana Prasad*
403 * The openssl speed command does not use low-level API calls anymore.
407 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
410 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
412 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
416 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
417 providers may supply their own group implementations (using either the "key
418 exchange" or the "key encapsulation" methods) which will automatically be
419 detected and used by libssl.
421 *Matt Caswell, Nicola Tuveri*
423 * The undocumented function X509_certificate_type() has been deprecated;
427 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
431 * Removed RSA padding mode for SSLv23 (which was only used for
432 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
433 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
438 * Deprecated the obsolete X9.31 RSA key generation related functions.
440 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
441 is not allowed to return a value > 1, this is no more taken as failure.
443 *Viktor Dukhovni and David von Oheimb*
445 * Deprecated the obsolete X9.31 RSA key generation related functions
446 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
447 BN_X931_generate_prime_ex().
451 * The default key generation method for the regular 2-prime RSA keys was
452 changed to the FIPS 186-4 B.3.6 method.
456 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
460 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
464 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
465 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
467 *Rich Salz, Richard Levitte, and David von Oheimb*
469 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
473 * Deprecated `OCSP_parse_url()`.
477 * Validation of SM2 keys has been separated from the validation of regular EC
482 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
483 switches: a validation failure triggers an early exit, returning a failure
484 exit status to the parent process.
488 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
489 to ignore unknown ciphers.
493 * The `-cipher-commands` and `-digest-commands` options
494 of the command line utility `list` have been deprecated.
495 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
499 * Added convenience functions for generating asymmetric key pairs:
500 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
501 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
505 * All of the low level EC_KEY functions have been deprecated.
507 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
509 * Deprecated all the libcrypto and libssl error string loading
514 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
515 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
520 * The `-crypt` option to the `passwd` command line tool has been removed.
524 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
529 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
533 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
534 EVP_PKEY_get1_tls_encodedpoint().
538 * The security callback, which can be customised by application code, supports
539 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
540 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
544 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
545 interface. Their functionality remains unchanged.
549 * Added new option for 'openssl list', '-providers', which will display the
550 list of loaded providers, their names, version and status. It optionally
551 displays their gettable parameters.
555 * Removed EVP_PKEY_set_alias_type().
559 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
560 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
564 * Changed all "STACK" functions to be macros instead of inline functions. Macro
565 parameters are still checked for type safety at compile time via helper
570 * Remove the RAND_DRBG API
572 *Paul Dale and Matthias St. Pierre*
574 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
575 as well as actual hostnames.
579 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
580 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
581 conversely, silently ignore DTLS protocol version bounds when configuring
582 TLS-based contexts. The commands can be repeated to set bounds of both
583 types. The same applies with the corresponding "min_protocol" and
584 "max_protocol" command-line switches, in case some application uses both TLS
587 SSL_CTX instances that are created for a fixed protocol version (e.g.
588 `TLSv1_server_method()`) also silently ignore version bounds. Previously
589 attempts to apply bounds to these protocol versions would result in an
590 error. Now only the "version-flexible" SSL_CTX instances are subject to
591 limits in configuration files in command-line options.
595 * Deprecated the `ENGINE` API. Engines should be replaced with providers
600 * Reworked the recorded ERR codes to make better space for system errors.
601 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
602 given code is a system error (true) or an OpenSSL error (false).
606 * Reworked the test perl framework to better allow parallel testing.
608 *Nicola Tuveri and David von Oheimb*
610 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
611 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
615 * 'Configure' has been changed to figure out the configuration target if
616 none is given on the command line. Consequently, the 'config' script is
617 now only a mere wrapper. All documentation is changed to only mention
620 *Rich Salz and Richard Levitte*
622 * Added a library context `OSSL_LIB_CTX` that applications as well as
623 other libraries can use to form a separate context within which
624 libcrypto operations are performed.
628 * Added various `_ex` functions to the OpenSSL API that support using
629 a non-default `OSSL_LIB_CTX`.
633 * Handshake now fails if Extended Master Secret extension is dropped
638 * Dropped interactive mode from the `openssl` program.
642 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
644 *David von Oheimb and Shane Lontis*
646 * Deprecated `EC_METHOD_get_field_type()`.
650 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
651 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
652 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
656 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
660 * Add CAdES-BES signature verification support, mostly derived
661 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
663 *Filipe Raimundo da Silva*
665 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
669 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
670 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
674 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
678 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
679 EC_KEY_precompute_mult().
683 * Deprecated EC_POINTs_mul().
687 * Removed FIPS_mode() and FIPS_mode_set().
691 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
695 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
696 EC_POINT_get_Jprojective_coordinates_GFp().
700 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
701 arrays to be more easily constructed via a series of utility functions.
702 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
703 the various push functions and finally convert to a passable OSSL_PARAM
704 array using OSSL_PARAM_BLD_to_param().
708 * The security strength of SHA1 and MD5 based signatures in TLS has been
713 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
714 contain a provider side internal key.
718 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
722 * Project text documents not yet having a proper file name extension
723 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
724 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
728 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
729 have been converted to Markdown with the goal to produce documents
730 which not only look pretty when viewed online in the browser, but
731 remain well readable inside a plain text editor.
733 To achieve this goal, a 'minimalistic' Markdown style has been applied
734 which avoids formatting elements that interfere too much with the
735 reading flow in the text file. For example, it
737 * avoids [ATX headings][] and uses [setext headings][] instead
738 (which works for `<h1>` and `<h2>` headings only).
739 * avoids [inline links][] and uses [reference links][] instead.
740 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
742 [ATX headings]: https://github.github.com/gfm/#atx-headings
743 [setext headings]: https://github.github.com/gfm/#setext-headings
744 [inline links]: https://github.github.com/gfm/#inline-link
745 [reference links]: https://github.github.com/gfm/#reference-link
746 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
747 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
749 *Matthias St. Pierre*
751 * The test suite is changed to preserve results of each test recipe.
752 A new directory test-runs/ with subdirectories named like the
753 test recipes are created in the build tree for this purpose.
757 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
758 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
759 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
761 *David von Oheimb, Martin Peylo*
763 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
764 It supports arbitrary request and response content types, GET redirection,
765 TLS, connections via HTTP(S) proxies, connections and exchange via
766 user-defined BIOs (allowing implicit connections), persistent connections,
767 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
768 The legacy OCSP-focused (and only partly documented) API
769 is retained for backward compatibility, while most of it is deprecated.
773 * Added `util/check-format.pl`, a tool for checking adherence to the
774 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
775 The checks performed are incomplete and yield some false positives.
776 Still the tool should be useful for detecting most typical glitches.
780 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
781 If domain name resolution yields multiple IP addresses all of them are tried
782 after `connect()` failures.
786 * All of the low level RSA functions have been deprecated.
790 * X509 certificates signed using SHA1 are no longer allowed at security
795 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
796 modified to use PKEY APIs. These commands are now in maintenance mode
797 and no new features will be added to them.
801 * The command line utility rsautl has been deprecated.
805 * The command line utilities genrsa and rsa have been modified to use PKEY
806 APIs. They now write PKCS#8 keys by default. These commands are now in
807 maintenance mode and no new features will be added to them.
811 * All of the low level DH functions have been deprecated.
813 *Paul Dale and Matt Caswell*
815 * All of the low level DSA functions have been deprecated.
819 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
820 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
824 * Deprecated low level ECDH and ECDSA functions.
828 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
832 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
833 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
834 a new formulation to include all the things it can be used for,
835 as well as words of caution.
839 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
843 * All of the low level HMAC functions have been deprecated.
845 *Paul Dale and David von Oheimb*
847 * Over two thousand fixes were made to the documentation, including:
848 - Common options (such as -rand/-writerand, TLS version control, etc)
849 were refactored and point to newly-enhanced descriptions in openssl.pod.
850 - Added style conformance for all options (with help from Richard Levitte),
851 documented all reported missing options, added a CI build to check
852 that all options are documented and that no unimplemented options
854 - Documented some internals, such as all use of environment variables.
855 - Addressed all internal broken L<> references.
859 * All of the low level CMAC functions have been deprecated.
863 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
864 functions have been deprecated.
866 *Paul Dale and David von Oheimb*
868 * Corrected the documentation of the return values from the `EVP_DigestSign*`
869 set of functions. The documentation mentioned negative values for some
870 errors, but this was never the case, so the mention of negative values
873 Code that followed the documentation and thereby check with something
874 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
878 * All of the low level cipher functions have been deprecated.
880 *Matt Caswell and Paul Dale*
882 * Removed include/openssl/opensslconf.h.in and replaced it with
883 include/openssl/configuration.h.in, which differs in not including
884 <openssl/macros.h>. A short header include/openssl/opensslconf.h
885 was added to include both.
887 This allows internal hacks where one might need to modify the set
888 of configured macros, for example this if deprecated symbols are
889 still supposed to be available internally:
891 #include <openssl/configuration.h>
893 #undef OPENSSL_NO_DEPRECATED
894 #define OPENSSL_SUPPRESS_DEPRECATED
896 #include <openssl/macros.h>
898 This should not be used by applications that use the exported
899 symbols, as that will lead to linking errors.
903 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
904 used in exponentiation with 512-bit moduli. No EC algorithms are
905 affected. Analysis suggests that attacks against 2-prime RSA1024,
906 3-prime RSA1536, and DSA1024 as a result of this defect would be very
907 difficult to perform and are not believed likely. Attacks against DH512
908 are considered just feasible. However, for an attack the target would
909 have to re-use the DH512 private key, which is not recommended anyway.
910 Also applications directly using the low-level API BN_mod_exp may be
911 affected if they use BN_FLG_CONSTTIME.
916 * Most memory-debug features have been deprecated, and the functionality
917 replaced with no-ops.
921 * Added documentation for the STACK API.
925 * Introduced a new method type and API, OSSL_ENCODER, to represent
926 generic encoders. These do the same sort of job that PEM writers
927 and d2i functions do, but with support for methods supplied by
928 providers, and the possibility for providers to support other
933 * Introduced a new method type and API, OSSL_DECODER, to represent
934 generic decoders. These do the same sort of job that PEM readers
935 and i2d functions do, but with support for methods supplied by
936 providers, and the possibility for providers to support other
941 * Added a .pragma directive to the syntax of configuration files, to
942 allow varying behavior in a supported and predictable manner.
943 Currently added pragma:
947 This allows dollar signs to be a keyword character unless it's
948 followed by a opening brace or parenthesis. This is useful for
949 platforms where dollar signs are commonly used in names, such as
950 volume names and system directory names on VMS.
954 * Added functionality to create an EVP_PKEY from user data.
958 * Change the interpretation of the '--api' configuration option to
959 mean that this is a desired API compatibility level with no
960 further meaning. The previous interpretation, that this would
961 also mean to remove all deprecated symbols up to and including
962 the given version, no requires that 'no-deprecated' is also used
963 in the configuration.
965 When building applications, the desired API compatibility level
966 can be set with the OPENSSL_API_COMPAT macro like before. For
967 API compatibility version below 3.0, the old style numerical
968 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
969 For version 3.0 and on, the value is expected to be the decimal
970 value calculated from the major and minor version like this:
972 MAJOR * 10000 + MINOR * 100
976 -DOPENSSL_API_COMPAT=30000 For 3.0
977 -DOPENSSL_API_COMPAT=30200 For 3.2
979 To hide declarations that are deprecated up to and including the
980 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
981 given when building the application as well.
985 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
986 access to certificate and CRL stores via URIs and OSSL_STORE
989 This adds the following functions:
991 - X509_LOOKUP_store()
992 - X509_STORE_load_file()
993 - X509_STORE_load_path()
994 - X509_STORE_load_store()
995 - SSL_add_store_cert_subjects_to_stack()
996 - SSL_CTX_set_default_verify_store()
997 - SSL_CTX_load_verify_file()
998 - SSL_CTX_load_verify_dir()
999 - SSL_CTX_load_verify_store()
1003 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1004 The presence of this system service is determined at run-time.
1008 * Added functionality to create an EVP_PKEY context based on data
1009 for methods from providers. This takes an algorithm name and a
1010 property query string and simply stores them, with the intent
1011 that any operation that uses this context will use those strings
1012 to fetch the needed methods implicitly, thereby making the port
1013 of application written for pre-3.0 OpenSSL easier.
1017 * The undocumented function NCONF_WIN32() has been deprecated; for
1018 conversion details see the HISTORY section of doc/man5/config.pod
1022 * Introduced the new functions EVP_DigestSignInit_ex() and
1023 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1024 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1025 pages for further details.
1029 * Over two thousand fixes were made to the documentation, including:
1030 adding missing command flags, better style conformance, documentation
1033 *Rich Salz, Richard Levitte*
1035 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1036 X25519, X448, Ed25519 and Ed448.
1040 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1045 * Deprecated the public definition of `ERR_STATE` as well as the function
1046 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1051 * Added ERR functionality to give callers access to the stored function
1052 names that have replaced the older function code based functions.
1054 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1055 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1056 ERR_peek_error_all() and ERR_peek_last_error_all().
1058 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1059 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1060 ERR_func_error_string().
1064 * Extended testing to be verbose for failing tests only. The make variables
1065 VERBOSE_FAILURE or VF can be used to enable this:
1067 $ make VF=1 test # Unix
1068 $ mms /macro=(VF=1) test ! OpenVMS
1069 $ nmake VF=1 test # Windows
1073 * Added the `-copy_extensions` option to the `x509` command for use with
1074 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1075 all extensions in the request are copied to the certificate or vice versa.
1077 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1079 * Added the `-copy_extensions` option to the `req` command for use with
1080 `-x509`. When given with the `copy` or `copyall` argument,
1081 all extensions in the certification request are copied to the certificate.
1085 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1086 they generate are by default RFC 5280 compliant in the following sense:
1087 There is a subjectKeyIdentifier extension with a hash value of the public key
1088 and for not self-signed certs there is an authorityKeyIdentifier extension
1089 with a keyIdentifier field or issuer information identifying the signing key.
1090 This is done unless some configuration overrides the new default behavior,
1091 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1095 * Added several checks to `X509_verify_cert()` according to requirements in
1096 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1097 (which may be done by using the CLI option `-x509_strict`):
1098 * The basicConstraints of CA certificates must be marked critical.
1099 * CA certificates must explicitly include the keyUsage extension.
1100 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1101 * The issuer name of any certificate must not be empty.
1102 * The subject name of CA certs, certs with keyUsage crlSign,
1103 and certs without subjectAlternativeName must not be empty.
1104 * If a subjectAlternativeName extension is given it must not be empty.
1105 * The signatureAlgorithm field and the cert signature must be consistent.
1106 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1107 must not be marked critical.
1108 * The authorityKeyIdentifier must be given for X.509v3 certs
1109 unless they are self-signed.
1110 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1114 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1115 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1119 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1120 used even when parsing explicit parameters, when loading a encoded key
1121 or calling `EC_GROUP_new_from_ecpkparameters()`/
1122 `EC_GROUP_new_from_ecparameters()`.
1123 This prevents bypass of security hardening and performance gains,
1124 especially for curves with specialized EC_METHODs.
1125 By default, if a key encoded with explicit parameters is loaded and later
1126 encoded, the output is still encoded with explicit parameters, even if
1127 internally a "named" EC_GROUP is used for computation.
1131 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1132 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1133 NULL. After this change, only the cofactor parameter can be NULL. It also
1134 does some minimal sanity checks on the passed order.
1139 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1140 An attack is simple, if the first CMS_recipientInfo is valid but the
1141 second CMS_recipientInfo is chosen ciphertext. If the second
1142 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1143 encryption key will be replaced by garbage, and the message cannot be
1144 decoded, but if the RSA decryption fails, the correct encryption key is
1145 used and the recipient will not notice the attack.
1146 As a work around for this potential attack the length of the decrypted
1147 key must be equal to the cipher default key length, in case the
1148 certifiate is not given and all recipientInfo are tried out.
1149 The old behaviour can be re-enabled in the CMS code by setting the
1150 CMS_DEBUG_DECRYPT flag.
1154 * Early start up entropy quality from the DEVRANDOM seed source has been
1155 improved for older Linux systems. The RAND subsystem will wait for
1156 /dev/random to be producing output before seeding from /dev/urandom.
1157 The seeded state is stored for future library initialisations using
1158 a system global shared memory segment. The shared memory identifier
1159 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1160 the desired value. The default identifier is 114.
1164 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1165 when primes for RSA keys are computed.
1166 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1167 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1168 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1169 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1170 This avoids possible fingerprinting of newly generated RSA modules.
1174 * Correct the extended master secret constant on EBCDIC systems. Without this
1175 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1176 negotiate EMS will fail. Unfortunately this also means that TLS connections
1177 between EBCDIC systems with this fix, and EBCDIC systems without this
1178 fix will fail if they negotiate EMS.
1182 * Changed the library initialisation so that the config file is now loaded
1183 by default. This was already the case for libssl. It now occurs for both
1184 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1185 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1189 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1190 where the former acts as a replacement for `ERR_put_error()`, and the
1191 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1192 `ERR_raise_data()` adds more flexibility by taking a format string and
1193 an arbitrary number of arguments following it, to be processed with
1198 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1199 to check if a named provider is loaded and available. When called, it
1200 will also activate all fallback providers if such are still present.
1204 * Enforce a minimum DH modulus size of 512 bits.
1208 * Changed DH parameters to generate the order q subgroup instead of 2q.
1209 Previously generated DH parameters are still accepted by DH_check
1210 but DH_generate_key works around that by clearing bit 0 of the
1211 private key for those. This avoids leaking bit 0 of the private key.
1215 * Significantly reduce secure memory usage by the randomness pools.
1219 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1224 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1225 algorithms. An implementation of a key exchange algorithm can be obtained
1226 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1227 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1228 the older EVP_PKEY_derive_init() function. See the man pages for the new
1229 functions for further details.
1233 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1237 * Removed the function names from error messages and deprecated the
1242 * Removed NextStep support and the macro OPENSSL_UNISTD
1246 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1247 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1248 Also removed "export var as function" capability; we do not export
1249 variables, only functions.
1253 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1254 an error and 1 indicating success. In previous versions of OpenSSL this
1255 was a void type. If a key was set longer than the maximum possible this
1260 * Support SM2 signing and verification schemes with X509 certificate.
1264 * Use SHA256 as the default digest for TS query in the `ts` app.
1268 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1272 * Default cipher lists/suites are now available via a function, the
1273 #defines are deprecated.
1277 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1278 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1279 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1283 * Join the directories crypto/x509 and crypto/x509v3
1287 * Added command 'openssl kdf' that uses the EVP_KDF API.
1291 * Added command 'openssl mac' that uses the EVP_MAC API.
1295 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1296 as default directories. Also added the command 'openssl info'
1297 for scripting purposes.
1301 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1306 * Add prediction resistance to the DRBG reseeding process.
1310 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1311 mandated by IEEE Std 1619-2018.
1315 * Added newline escaping functionality to a filename when using openssl dgst.
1316 This output format is to replicate the output format found in the `*sum`
1317 checksum programs. This aims to preserve backward compatibility.
1319 *Matt Eaton, Richard Levitte, and Paul Dale*
1321 * Removed the heartbeat message in DTLS feature, as it has very
1322 little usage and doesn't seem to fulfill a valuable purpose.
1323 The configuration option is now deprecated.
1327 * Changed the output of 'openssl {digestname} < file' to display the
1328 digest name in its output.
1332 * Added a new generic trace API which provides support for enabling
1333 instrumentation through trace output.
1335 *Richard Levitte & Matthias St. Pierre*
1337 * Added build tests for C++. These are generated files that only do one
1338 thing, to include one public OpenSSL head file each. This tests that
1339 the public header files can be usefully included in a C++ application.
1341 This test isn't enabled by default. It can be enabled with the option
1342 'enable-buildtest-c++'.
1346 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1350 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1354 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1358 * Added KMAC to EVP_MAC.
1362 * Added property based algorithm implementation selection framework to
1367 * Added SCA hardening for modular field inversion in EC_GROUP through
1368 a new dedicated field_inv() pointer in EC_METHOD.
1369 This also addresses a leakage affecting conversions from projective
1370 to affine coordinates.
1372 *Billy Bob Brumley, Nicola Tuveri*
1374 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1375 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1376 those algorithms that were already supported through the EVP_PKEY API
1377 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1378 and scrypt are now wrappers that call EVP_KDF.
1382 * Build devcrypto engine as a dynamic engine.
1384 *Eneas U de Queiroz*
1386 * Add keyed BLAKE2 to EVP_MAC.
1390 * Fix a bug in the computation of the endpoint-pair shared secret used
1391 by DTLS over SCTP. This breaks interoperability with older versions
1392 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1393 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1394 interoperability with such broken implementations. However, enabling
1395 this switch breaks interoperability with correct implementations.
1397 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1398 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1402 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1406 * Changed the license to the Apache License v2.0.
1410 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1412 - Major releases (indicated by incrementing the MAJOR release number)
1413 may introduce incompatible API/ABI changes.
1414 - Minor releases (indicated by incrementing the MINOR release number)
1415 may introduce new features but retain API/ABI compatibility.
1416 - Patch releases (indicated by incrementing the PATCH number)
1417 are intended for bug fixes and other improvements of existing
1418 features only (like improving performance or adding documentation)
1419 and retain API/ABI compatibility.
1423 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1427 * Remove the 'dist' target and add a tarball building script. The
1428 'dist' target has fallen out of use, and it shouldn't be
1429 necessary to configure just to create a source distribution.
1433 * Recreate the OS390-Unix config target. It no longer relies on a
1434 special script like it did for OpenSSL pre-1.1.0.
1438 * Instead of having the source directories listed in Configure, add
1439 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1444 * Add GMAC to EVP_MAC.
1448 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1452 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1453 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1454 to facilitate the continued use of MACs through raw private keys in
1455 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1459 * Deprecate ECDH_KDF_X9_62().
1463 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1464 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1465 are retained for backwards compatibility.
1469 * AES-XTS mode now enforces that its two keys are different to mitigate
1470 the attacked described in "Efficient Instantiations of Tweakable
1471 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1472 Details of this attack can be obtained from:
1473 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1477 * Rename the object files, i.e. give them other names than in previous
1478 versions. Their names now include the name of the final product, as
1479 well as its type mnemonic (bin, lib, shlib).
1483 * Added new option for 'openssl list', '-objects', which will display the
1484 list of built in objects, i.e. OIDs with names.
1488 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1489 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1494 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1495 improves application performance by removing data copies and providing
1496 applications with zero-copy system calls such as sendfile and splice.
1500 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1504 * `PKCS12_parse` now maintains the order of the parsed certificates
1505 when outputting them via `*ca` (rather than reversing it).
1509 * Deprecated pthread fork support methods.
1513 * Added support for FFDHE key exchange in TLS 1.3.
1517 * Added a new concept for OpenSSL plugability: providers. This
1518 functionality is designed to replace the ENGINE API and ENGINE
1519 implementations, and to be much more dynamic, allowing provider
1520 authors to introduce new algorithms among other things, as long as
1521 there's an API that supports the algorithm type.
1523 With this concept comes a new core API for interaction between
1524 libcrypto and provider implementations. Public libcrypto functions
1525 that want to use providers do so through this core API.
1527 The main documentation for this core API is found in
1528 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1529 refer to other manuals describing the API specific for supported
1530 algorithm types (also called operations).
1537 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
1539 * Avoid loading of a dynamic engine twice.
1543 * Prioritise DANE TLSA issuer certs over peer certs
1547 * Fixed random API for MacOS prior to 10.12
1549 These MacOS versions don't support the CommonCrypto APIs
1553 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1555 * Fixed an SM2 Decryption Buffer Overflow.
1557 In order to decrypt SM2 encrypted data an application is expected to
1558 call the API function EVP_PKEY_decrypt(). Typically an application will
1559 call this function twice. The first time, on entry, the "out" parameter
1560 can be NULL and, on exit, the "outlen" parameter is populated with the
1561 buffer size required to hold the decrypted plaintext. The application
1562 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1563 again, but this time passing a non-NULL value for the "out" parameter.
1565 A bug in the implementation of the SM2 decryption code means that the
1566 calculation of the buffer size required to hold the plaintext returned
1567 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1568 size required by the second call. This can lead to a buffer overflow
1569 when EVP_PKEY_decrypt() is called by the application a second time with
1570 a buffer that is too small.
1572 A malicious attacker who is able present SM2 content for decryption to
1573 an application could cause attacker chosen data to overflow the buffer
1574 by up to a maximum of 62 bytes altering the contents of other data held
1575 after the buffer, possibly changing application behaviour or causing
1576 the application to crash. The location of the buffer is application
1577 dependent but is typically heap allocated.
1582 * Fixed various read buffer overruns processing ASN.1 strings
1584 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1585 structure which contains a buffer holding the string data and a field
1586 holding the buffer length. This contrasts with normal C strings which
1587 are repesented as a buffer for the string data which is terminated
1588 with a NUL (0) byte.
1590 Although not a strict requirement, ASN.1 strings that are parsed using
1591 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1592 well as any string whose value has been set with the ASN1_STRING_set()
1593 function will additionally NUL terminate the byte array in the
1594 ASN1_STRING structure.
1596 However, it is possible for applications to directly construct valid
1597 ASN1_STRING structures which do not NUL terminate the byte array by
1598 directly setting the "data" and "length" fields in the ASN1_STRING
1599 array. This can also happen by using the ASN1_STRING_set0() function.
1601 Numerous OpenSSL functions that print ASN.1 data have been found to
1602 assume that the ASN1_STRING byte array will be NUL terminated, even
1603 though this is not guaranteed for strings that have been directly
1604 constructed. Where an application requests an ASN.1 structure to be
1605 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1606 been directly constructed by the application without NUL terminating
1607 the "data" field, then a read buffer overrun can occur.
1609 The same thing can also occur during name constraints processing
1610 of certificates (for example if a certificate has been directly
1611 constructed by the application instead of loading it via the OpenSSL
1612 parsing functions, and the certificate contains non NUL terminated
1613 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1614 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1616 If a malicious actor can cause an application to directly construct an
1617 ASN1_STRING and then process it through one of the affected OpenSSL
1618 functions then this issue could be hit. This might result in a crash
1619 (causing a Denial of Service attack). It could also result in the
1620 disclosure of private memory contents (such as private keys, or
1621 sensitive plaintext).
1626 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1628 * Fixed a problem with verifying a certificate chain when using the
1629 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1630 the certificates present in a certificate chain. It is not set by default.
1632 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1633 the chain that have explicitly encoded elliptic curve parameters was added
1634 as an additional strict check.
1636 An error in the implementation of this check meant that the result of a
1637 previous check to confirm that certificates in the chain are valid CA
1638 certificates was overwritten. This effectively bypasses the check
1639 that non-CA certificates must not be able to issue other certificates.
1641 If a "purpose" has been configured then there is a subsequent opportunity
1642 for checks that the certificate is a valid CA. All of the named "purpose"
1643 values implemented in libcrypto perform this check. Therefore, where
1644 a purpose is set the certificate chain will still be rejected even when the
1645 strict flag has been used. A purpose is set by default in libssl client and
1646 server certificate verification routines, but it can be overridden or
1647 removed by an application.
1649 In order to be affected, an application must explicitly set the
1650 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1651 for the certificate verification or, in the case of TLS client or server
1652 applications, override the default purpose.
1657 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1658 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1659 renegotiation ClientHello omits the signature_algorithms extension (where it
1660 was present in the initial ClientHello), but includes a
1661 signature_algorithms_cert extension then a NULL pointer dereference will
1662 result, leading to a crash and a denial of service attack.
1664 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1665 (which is the default configuration). OpenSSL TLS clients are not impacted by
1669 *Peter Kästle and Samuel Sapalski*
1671 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1673 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1674 create a unique hash value based on the issuer and serial number data
1675 contained within an X509 certificate. However it was failing to correctly
1676 handle any errors that may occur while parsing the issuer field (which might
1677 occur if the issuer field is maliciously constructed). This may subsequently
1678 result in a NULL pointer deref and a crash leading to a potential denial of
1684 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1685 padding mode to correctly check for rollback attacks. This is considered a
1686 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1691 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1692 functions. Previously they could overflow the output length argument in some
1693 cases where the input length is close to the maximum permissable length for
1694 an integer on the platform. In such cases the return value from the function
1695 call would be 1 (indicating success), but the output length value would be
1696 negative. This could cause applications to behave incorrectly or crash.
1701 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1702 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1703 could be exploited in a side channel attack to recover the password. Since
1704 the attack is local host only this is outside of the current OpenSSL
1705 threat model and therefore no CVE is assigned.
1707 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1712 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1714 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1715 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1716 If an attacker can control both items being compared then this could lead
1717 to a possible denial of service attack. OpenSSL itself uses the
1718 GENERAL_NAME_cmp function for two purposes:
1719 1) Comparing CRL distribution point names between an available CRL and a
1720 CRL distribution point embedded in an X509 certificate
1721 2) When verifying that a timestamp response token signer matches the
1722 timestamp authority name (exposed via the API functions
1723 TS_RESP_verify_response and TS_RESP_verify_token)
1728 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1730 * Certificates with explicit curve parameters are now disallowed in
1731 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1735 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1736 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1737 conversely, silently ignore DTLS protocol version bounds when configuring
1738 TLS-based contexts. The commands can be repeated to set bounds of both
1739 types. The same applies with the corresponding "min_protocol" and
1740 "max_protocol" command-line switches, in case some application uses both TLS
1743 SSL_CTX instances that are created for a fixed protocol version (e.g.
1744 TLSv1_server_method()) also silently ignore version bounds. Previously
1745 attempts to apply bounds to these protocol versions would result in an
1746 error. Now only the "version-flexible" SSL_CTX instances are subject to
1747 limits in configuration files in command-line options.
1751 * Handshake now fails if Extended Master Secret extension is dropped
1756 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1758 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1760 * Fixed segmentation fault in SSL_check_chain()
1761 Server or client applications that call the SSL_check_chain() function
1762 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1763 dereference as a result of incorrect handling of the
1764 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1765 or unrecognised signature algorithm is received from the peer. This could
1766 be exploited by a malicious peer in a Denial of Service attack.
1771 * Added AES consttime code for no-asm configurations
1772 an optional constant time support for AES was added
1773 when building openssl for no-asm.
1774 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1775 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1776 At this time this feature is by default disabled.
1777 It will be enabled by default in 3.0.
1781 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1783 * Revert the change of EOF detection while reading in libssl to avoid
1784 regressions in applications depending on the current way of reporting
1785 the EOF. As the existing method is not fully accurate the change to
1786 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1787 branch and will be present in the 3.0 release.
1791 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1792 when primes for RSA keys are computed.
1793 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1794 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1795 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1796 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1797 This avoids possible fingerprinting of newly generated RSA modules.
1801 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1803 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1804 while reading in libssl then we would report an error back to the
1805 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1806 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1807 therefore give a hint as to what went wrong.
1811 * Check that ed25519 and ed448 are allowed by the security level. Previously
1812 signature algorithms not using an MD were not being checked that they were
1813 allowed by the security level.
1817 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1818 was not quite right. The behaviour was not consistent between resumption
1819 and normal handshakes, and also not quite consistent with historical
1820 behaviour. The behaviour in various scenarios has been clarified and
1821 it has been updated to make it match historical behaviour as closely as
1826 * *[VMS only]* The header files that the VMS compilers include automatically,
1827 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1828 that the C++ compiler doesn't understand. This is a shortcoming in the
1829 compiler, but can be worked around with `__cplusplus` guards.
1831 C++ applications that use OpenSSL libraries must be compiled using the
1832 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1833 functions. Otherwise, only functions with symbols of less than 31
1834 characters can be used, as the linker will not be able to successfully
1835 resolve symbols with longer names.
1839 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1840 The presence of this system service is determined at run-time.
1844 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1849 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1851 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1852 number generator (RNG). This was intended to include protection in the
1853 event of a fork() system call in order to ensure that the parent and child
1854 processes did not share the same RNG state. However this protection was not
1855 being used in the default case.
1857 A partial mitigation for this issue is that the output from a high
1858 precision timer is mixed into the RNG state so the likelihood of a parent
1859 and child process sharing state is significantly reduced.
1861 If an application already calls OPENSSL_init_crypto() explicitly using
1862 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1865 *Matthias St. Pierre*
1867 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1868 used even when parsing explicit parameters, when loading a encoded key
1869 or calling `EC_GROUP_new_from_ecpkparameters()`/
1870 `EC_GROUP_new_from_ecparameters()`.
1871 This prevents bypass of security hardening and performance gains,
1872 especially for curves with specialized EC_METHODs.
1873 By default, if a key encoded with explicit parameters is loaded and later
1874 encoded, the output is still encoded with explicit parameters, even if
1875 internally a "named" EC_GROUP is used for computation.
1879 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1880 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1881 NULL. After this change, only the cofactor parameter can be NULL. It also
1882 does some minimal sanity checks on the passed order.
1887 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1888 An attack is simple, if the first CMS_recipientInfo is valid but the
1889 second CMS_recipientInfo is chosen ciphertext. If the second
1890 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1891 encryption key will be replaced by garbage, and the message cannot be
1892 decoded, but if the RSA decryption fails, the correct encryption key is
1893 used and the recipient will not notice the attack.
1894 As a work around for this potential attack the length of the decrypted
1895 key must be equal to the cipher default key length, in case the
1896 certifiate is not given and all recipientInfo are tried out.
1897 The old behaviour can be re-enabled in the CMS code by setting the
1898 CMS_DEBUG_DECRYPT flag.
1903 * Early start up entropy quality from the DEVRANDOM seed source has been
1904 improved for older Linux systems. The RAND subsystem will wait for
1905 /dev/random to be producing output before seeding from /dev/urandom.
1906 The seeded state is stored for future library initialisations using
1907 a system global shared memory segment. The shared memory identifier
1908 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1909 the desired value. The default identifier is 114.
1913 * Correct the extended master secret constant on EBCDIC systems. Without this
1914 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1915 negotiate EMS will fail. Unfortunately this also means that TLS connections
1916 between EBCDIC systems with this fix, and EBCDIC systems without this
1917 fix will fail if they negotiate EMS.
1921 * Use Windows installation paths in the mingw builds
1923 Mingw isn't a POSIX environment per se, which means that Windows
1924 paths should be used for installation.
1929 * Changed DH_check to accept parameters with order q and 2q subgroups.
1930 With order 2q subgroups the bit 0 of the private key is not secret
1931 but DH_generate_key works around that by clearing bit 0 of the
1932 private key for those. This avoids leaking bit 0 of the private key.
1936 * Significantly reduce secure memory usage by the randomness pools.
1940 * Revert the DEVRANDOM_WAIT feature for Linux systems
1942 The DEVRANDOM_WAIT feature added a select() call to wait for the
1943 /dev/random device to become readable before reading from the
1944 /dev/urandom device.
1946 It turned out that this change had negative side effects on
1947 performance which were not acceptable. After some discussion it
1948 was decided to revert this feature and leave it up to the OS
1949 resp. the platform maintainer to ensure a proper initialization
1950 during early boot time.
1952 *Matthias St. Pierre*
1954 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1956 * Add build tests for C++. These are generated files that only do one
1957 thing, to include one public OpenSSL head file each. This tests that
1958 the public header files can be usefully included in a C++ application.
1960 This test isn't enabled by default. It can be enabled with the option
1961 'enable-buildtest-c++'.
1965 * Enable SHA3 pre-hashing for ECDSA and DSA.
1969 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1970 This changes the size when using the `genpkey` command when no size is given.
1971 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1972 generation commands to use 2048 bits by default.
1976 * Reorganize the manual pages to consistently have RETURN VALUES,
1977 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1978 util/fix-doc-nits accordingly.
1980 *Paul Yang, Joshua Lock*
1982 * Add the missing accessor EVP_PKEY_get0_engine()
1986 * Have commands like `s_client` and `s_server` output the signature scheme
1987 along with other cipher suite parameters when debugging.
1991 * Make OPENSSL_config() error agnostic again.
1995 * Do the error handling in RSA decryption constant time.
1999 * Prevent over long nonces in ChaCha20-Poly1305.
2001 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2002 for every encryption operation. RFC 7539 specifies that the nonce value
2003 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2004 and front pads the nonce with 0 bytes if it is less than 12
2005 bytes. However it also incorrectly allows a nonce to be set of up to 16
2006 bytes. In this case only the last 12 bytes are significant and any
2007 additional leading bytes are ignored.
2009 It is a requirement of using this cipher that nonce values are
2010 unique. Messages encrypted using a reused nonce value are susceptible to
2011 serious confidentiality and integrity attacks. If an application changes
2012 the default nonce length to be longer than 12 bytes and then makes a
2013 change to the leading bytes of the nonce expecting the new value to be a
2014 new unique nonce then such an application could inadvertently encrypt
2015 messages with a reused nonce.
2017 Additionally the ignored bytes in a long nonce are not covered by the
2018 integrity guarantee of this cipher. Any application that relies on the
2019 integrity of these ignored leading bytes of a long nonce may be further
2020 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2021 is safe because no such use sets such a long nonce value. However user
2022 applications that use this cipher directly and set a non-default nonce
2023 length to be longer than 12 bytes may be vulnerable.
2025 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2031 * Add DEVRANDOM_WAIT feature for Linux systems
2033 On older Linux systems where the getrandom() system call is not available,
2034 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2035 Contrary to getrandom(), the /dev/urandom device will not block during
2036 early boot when the kernel CSPRNG has not been seeded yet.
2038 To mitigate this known weakness, use select() to wait for /dev/random to
2039 become readable before reading from /dev/urandom.
2041 * Ensure that SM2 only uses SM3 as digest algorithm
2045 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2047 * Change the info callback signals for the start and end of a post-handshake
2048 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2049 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2050 confused by this and assume that a TLSv1.2 renegotiation has started. This
2051 can break KeyUpdate handling. Instead we no longer signal the start and end
2052 of a post handshake message exchange (although the messages themselves are
2053 still signalled). This could break some applications that were expecting
2054 the old signals. However without this KeyUpdate is not usable for many
2059 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2061 * Timing vulnerability in DSA signature generation
2063 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2064 timing side channel attack. An attacker could use variations in the signing
2065 algorithm to recover the private key.
2067 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2072 * Timing vulnerability in ECDSA signature generation
2074 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2075 timing side channel attack. An attacker could use variations in the signing
2076 algorithm to recover the private key.
2078 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2083 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2084 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2085 of two gigabytes and the error handling improved.
2087 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2088 categorized as a normal bug, not a security issue, because the DRBG reseeds
2089 automatically and is fully functional even without additional randomness
2090 provided by the application.
2092 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2094 * Add a new ClientHello callback. Provides a callback interface that gives
2095 the application the ability to adjust the nascent SSL object at the
2096 earliest stage of ClientHello processing, immediately after extensions have
2097 been collected but before they have been processed. In particular, this
2098 callback can adjust the supported TLS versions in response to the contents
2103 * Add SM2 base algorithm support.
2107 * s390x assembly pack: add (improved) hardware-support for the following
2108 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2109 aes-cfb/cfb8, aes-ecb.
2113 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2114 parameter is no longer accepted, as it leads to a corrupt table. NULL
2115 pem_str is reserved for alias entries only.
2119 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2120 step for prime curves. The new implementation is based on formulae from
2121 differential addition-and-doubling in homogeneous projective coordinates
2122 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2123 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2124 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2125 to work in projective coordinates.
2127 *Billy Bob Brumley, Nicola Tuveri*
2129 * Change generating and checking of primes so that the error rate of not
2130 being prime depends on the intended use based on the size of the input.
2131 For larger primes this will result in more rounds of Miller-Rabin.
2132 The maximal error rate for primes with more than 1080 bits is lowered
2135 *Kurt Roeckx, Annie Yousar*
2137 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2141 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2142 moving between systems, and to avoid confusion when a Windows build is
2143 done with mingw vs with MSVC. For POSIX installs, there's still a
2144 symlink or copy named 'tsget' to avoid that confusion as well.
2148 * Revert blinding in ECDSA sign and instead make problematic addition
2149 length-invariant. Switch even to fixed-length Montgomery multiplication.
2153 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2154 step for binary curves. The new implementation is based on formulae from
2155 differential addition-and-doubling in mixed Lopez-Dahab projective
2156 coordinates, modified to independently blind the operands.
2158 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2160 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2161 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2162 EC_METHODs to implement their own specialized "ladder step", to take
2163 advantage of more favorable coordinate systems or more efficient
2164 differential addition-and-doubling algorithms.
2166 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2168 * Modified the random device based seed sources to keep the relevant
2169 file descriptors open rather than reopening them on each access.
2170 This allows such sources to operate in a chroot() jail without
2171 the associated device nodes being available. This behaviour can be
2172 controlled using RAND_keep_random_devices_open().
2176 * Numerous side-channel attack mitigations have been applied. This may have
2177 performance impacts for some algorithms for the benefit of improved
2178 security. Specific changes are noted in this change log by their respective
2183 * AIX shared library support overhaul. Switch to AIX "natural" way of
2184 handling shared libraries, which means collecting shared objects of
2185 different versions and bitnesses in one common archive. This allows to
2186 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2187 doesn't affect the way 3rd party applications are linked, only how
2188 multi-version installation is managed.
2192 * Make ec_group_do_inverse_ord() more robust and available to other
2193 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2194 mitigations are applied to the fallback BN_mod_inverse().
2195 When using this function rather than BN_mod_inverse() directly, new
2196 EC cryptosystem implementations are then safer-by-default.
2200 * Add coordinate blinding for EC_POINT and implement projective
2201 coordinate blinding for generic prime curves as a countermeasure to
2202 chosen point SCA attacks.
2204 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2206 * Add blinding to ECDSA and DSA signatures to protect against side channel
2207 attacks discovered by Keegan Ryan (NCC Group).
2211 * Enforce checking in the `pkeyutl` command to ensure that the input
2212 length does not exceed the maximum supported digest length when performing
2213 a sign, verify or verifyrecover operation.
2217 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2218 I/O in combination with something like select() or poll() will hang. This
2219 can be turned off again using SSL_CTX_clear_mode().
2220 Many applications do not properly handle non-application data records, and
2221 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2222 around the problems in those applications, but can also break some.
2223 It's recommended to read the manpages about SSL_read(), SSL_write(),
2224 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2225 SSL_CTX_set_read_ahead() again.
2229 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2230 now allow empty (zero character) pass phrases.
2234 * Apply blinding to binary field modular inversion and remove patent
2235 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2239 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2240 binary and prime elliptic curves.
2244 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2245 constant time fixed point multiplication.
2249 * Revise elliptic curve scalar multiplication with timing attack
2250 defenses: ec_wNAF_mul redirects to a constant time implementation
2251 when computing fixed point and variable point multiplication (which
2252 in OpenSSL are mostly used with secret scalars in keygen, sign,
2253 ECDH derive operations).
2254 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2257 * Updated CONTRIBUTING
2261 * Updated DRBG / RAND to request nonce and additional low entropy
2262 randomness from the system.
2264 *Matthias St. Pierre*
2266 * Updated 'openssl rehash' to use OpenSSL consistent default.
2270 * Moved the load of the ssl_conf module to libcrypto, which helps
2271 loading engines that libssl uses before libssl is initialised.
2275 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2279 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2281 *Ingo Schwarze, Rich Salz*
2283 * Added output of accepting IP address and port for 'openssl s_server'
2287 * Added a new API for TLSv1.3 ciphersuites:
2288 SSL_CTX_set_ciphersuites()
2289 SSL_set_ciphersuites()
2293 * Memory allocation failures consistently add an error to the error
2298 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2299 in libcrypto when run as setuid/setgid.
2303 * Load any config file by default when libssl is used.
2307 * Added new public header file <openssl/rand_drbg.h> and documentation
2308 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2310 *Matthias St. Pierre*
2312 * QNX support removed (cannot find contributors to get their approval
2313 for the license change).
2317 * TLSv1.3 replay protection for early data has been implemented. See the
2318 SSL_read_early_data() man page for further details.
2322 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2323 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2324 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2325 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2326 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2327 configuration has been separated out. See the ciphers man page or the
2328 SSL_CTX_set_ciphersuites() man page for more information.
2332 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2333 in responder mode now supports the new "-multi" option, which
2334 spawns the specified number of child processes to handle OCSP
2335 requests. The "-timeout" option now also limits the OCSP
2336 responder's patience to wait to receive the full client request
2337 on a newly accepted connection. Child processes are respawned
2338 as needed, and the CA index file is automatically reloaded
2339 when changed. This makes it possible to run the "ocsp" responder
2340 as a long-running service, making the OpenSSL CA somewhat more
2341 feature-complete. In this mode, most diagnostic messages logged
2342 after entering the event loop are logged via syslog(3) rather than
2347 * Added support for X448 and Ed448. Heavily based on original work by
2352 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2353 objects loaded. This adds the functions OSSL_STORE_expect() and
2354 OSSL_STORE_find() as well as needed tools to construct searches and
2355 get the search data out of them.
2359 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2360 version of OpenSSL should review their configuration settings to ensure
2361 that they are still appropriate for TLSv1.3. For further information see:
2362 <https://wiki.openssl.org/index.php/TLS1.3>
2366 * Grand redesign of the OpenSSL random generator
2368 The default RAND method now utilizes an AES-CTR DRBG according to
2369 NIST standard SP 800-90Ar1. The new random generator is essentially
2370 a port of the default random generator from the OpenSSL FIPS 2.0
2371 object module. It is a hybrid deterministic random bit generator
2372 using an AES-CTR bit stream and which seeds and reseeds itself
2373 automatically using trusted system entropy sources.
2375 Some of its new features are:
2376 - Support for multiple DRBG instances with seed chaining.
2377 - The default RAND method makes use of a DRBG.
2378 - There is a public and private DRBG instance.
2379 - The DRBG instances are fork-safe.
2380 - Keep all global DRBG instances on the secure heap if it is enabled.
2381 - The public and private DRBG instance are per thread for lock free
2384 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2386 * Changed Configure so it only says what it does and doesn't dump
2387 so much data. Instead, ./configdata.pm should be used as a script
2388 to display all sorts of configuration data.
2392 * Added processing of "make variables" to Configure.
2396 * Added SHA512/224 and SHA512/256 algorithm support.
2400 * The last traces of Netware support, first removed in 1.1.0, have
2405 * Get rid of Makefile.shared, and in the process, make the processing
2406 of certain files (rc.obj, or the .def/.map/.opt files produced from
2407 the ordinal files) more visible and hopefully easier to trace and
2408 debug (or make silent).
2412 * Make it possible to have environment variable assignments as
2413 arguments to config / Configure.
2417 * Add multi-prime RSA (RFC 8017) support.
2421 * Add SM3 implemented according to GB/T 32905-2016
2422 *Jack Lloyd <jack.lloyd@ribose.com>,*
2423 *Ronald Tse <ronald.tse@ribose.com>,*
2424 *Erick Borsboom <erick.borsboom@ribose.com>*
2426 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2427 as documented in RFC6066.
2428 Based on a patch from Tomasz Moń
2430 *Filipe Raimundo da Silva*
2432 * Add SM4 implemented according to GB/T 32907-2016.
2433 *Jack Lloyd <jack.lloyd@ribose.com>,*
2434 *Ronald Tse <ronald.tse@ribose.com>,*
2435 *Erick Borsboom <erick.borsboom@ribose.com>*
2437 * Reimplement -newreq-nodes and ERR_error_string_n; the
2438 original author does not agree with the license change.
2442 * Add ARIA AEAD TLS support.
2446 * Some macro definitions to support VS6 have been removed. Visual
2447 Studio 6 has not worked since 1.1.0
2451 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2452 without clearing the errors.
2456 * Add "atfork" functions. If building on a system that without
2457 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2458 requirements. The RAND facility now uses/requires this.
2466 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2467 not possible to disable entirely. However, it's still possible to
2468 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2471 To disable, configure with 'no-ui-console'. 'no-ui' is still
2472 possible to use as an alias. Check at compile time with the
2473 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2474 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2478 * Add a STORE module, which implements a uniform and URI based reader of
2479 stores that can contain keys, certificates, CRLs and numerous other
2480 objects. The main API is loosely based on a few stdio functions,
2481 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2482 OSSL_STORE_error and OSSL_STORE_close.
2483 The implementation uses backends called "loaders" to implement arbitrary
2484 URI schemes. There is one built in "loader" for the 'file' scheme.
2488 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2489 then adjusted to work on FreeBSD 8.4 as well.
2490 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2491 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2495 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2496 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2497 error code calls like this:
2499 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2501 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2502 that can be encoded in C. For the foreseeable future, this will only
2505 *Richard Levitte and Tim Hudson*
2507 * Removed BSD cryptodev engine.
2511 * Add a build target 'build_all_generated', to build all generated files
2512 and only that. This can be used to prepare everything that requires
2513 things like perl for a system that lacks perl and then move everything
2514 to that system and do the rest of the build there.
2518 * In the UI interface, make it possible to duplicate the user data. This
2519 can be used by engines that need to retain the data for a longer time
2520 than just the call where this user data is passed.
2524 * Ignore the '-named_curve auto' value for compatibility of applications
2527 *Tomáš Mráz <tmraz@fedoraproject.org>*
2529 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2530 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2531 alerts across multiple records (some of which could be empty). In practice
2532 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2533 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2534 support this at all. Supporting it adds significant complexity to the
2535 record layer, and its removal is unlikely to cause interoperability
2540 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2541 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2542 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2547 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2548 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2550 *Richard Levitte, Andy Polyakov*
2552 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2557 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2558 platform rather than 'mingw'.
2562 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2563 success if they are asked to add an object which already exists
2564 in the store. This change cascades to other functions which load
2565 certificates and CRLs.
2569 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2570 facilitate stack unwinding even from assembly subroutines.
2574 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2575 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2579 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2580 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2581 which is the minimum version we support.
2585 * Certificate time validation (X509_cmp_time) enforces stricter
2586 compliance with RFC 5280. Fractional seconds and timezone offsets
2587 are no longer allowed.
2591 * Add support for ARIA
2595 * s_client will now send the Server Name Indication (SNI) extension by
2596 default unless the new "-noservername" option is used. The server name is
2597 based on the host provided to the "-connect" option unless overridden by
2598 using "-servername".
2602 * Add support for SipHash
2606 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2607 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2608 prevent issues where no progress is being made and the peer continually
2609 sends unrecognised record types, using up resources processing them.
2613 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2614 using the algorithm defined in
2615 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2619 * Heartbeat support has been removed; the ABI is changed for now.
2621 *Richard Levitte, Rich Salz*
2623 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2627 * The RSA "null" method, which was partially supported to avoid patent
2628 issues, has been replaced to always returns NULL.
2635 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2637 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2638 used even when parsing explicit parameters, when loading a encoded key
2639 or calling `EC_GROUP_new_from_ecpkparameters()`/
2640 `EC_GROUP_new_from_ecparameters()`.
2641 This prevents bypass of security hardening and performance gains,
2642 especially for curves with specialized EC_METHODs.
2643 By default, if a key encoded with explicit parameters is loaded and later
2644 encoded, the output is still encoded with explicit parameters, even if
2645 internally a "named" EC_GROUP is used for computation.
2649 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2650 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2651 NULL. After this change, only the cofactor parameter can be NULL. It also
2652 does some minimal sanity checks on the passed order.
2657 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2658 An attack is simple, if the first CMS_recipientInfo is valid but the
2659 second CMS_recipientInfo is chosen ciphertext. If the second
2660 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2661 encryption key will be replaced by garbage, and the message cannot be
2662 decoded, but if the RSA decryption fails, the correct encryption key is
2663 used and the recipient will not notice the attack.
2664 As a work around for this potential attack the length of the decrypted
2665 key must be equal to the cipher default key length, in case the
2666 certifiate is not given and all recipientInfo are tried out.
2667 The old behaviour can be re-enabled in the CMS code by setting the
2668 CMS_DEBUG_DECRYPT flag.
2673 * Use Windows installation paths in the mingw builds
2675 Mingw isn't a POSIX environment per se, which means that Windows
2676 paths should be used for installation.
2681 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2683 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2684 This changes the size when using the `genpkey` command when no size is given.
2685 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2686 generation commands to use 2048 bits by default.
2690 * Prevent over long nonces in ChaCha20-Poly1305.
2692 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2693 for every encryption operation. RFC 7539 specifies that the nonce value
2694 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2695 and front pads the nonce with 0 bytes if it is less than 12
2696 bytes. However it also incorrectly allows a nonce to be set of up to 16
2697 bytes. In this case only the last 12 bytes are significant and any
2698 additional leading bytes are ignored.
2700 It is a requirement of using this cipher that nonce values are
2701 unique. Messages encrypted using a reused nonce value are susceptible to
2702 serious confidentiality and integrity attacks. If an application changes
2703 the default nonce length to be longer than 12 bytes and then makes a
2704 change to the leading bytes of the nonce expecting the new value to be a
2705 new unique nonce then such an application could inadvertently encrypt
2706 messages with a reused nonce.
2708 Additionally the ignored bytes in a long nonce are not covered by the
2709 integrity guarantee of this cipher. Any application that relies on the
2710 integrity of these ignored leading bytes of a long nonce may be further
2711 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2712 is safe because no such use sets such a long nonce value. However user
2713 applications that use this cipher directly and set a non-default nonce
2714 length to be longer than 12 bytes may be vulnerable.
2716 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2722 * Added SCA hardening for modular field inversion in EC_GROUP through
2723 a new dedicated field_inv() pointer in EC_METHOD.
2724 This also addresses a leakage affecting conversions from projective
2725 to affine coordinates.
2727 *Billy Bob Brumley, Nicola Tuveri*
2729 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2730 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2734 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2738 * Remove the 'dist' target and add a tarball building script. The
2739 'dist' target has fallen out of use, and it shouldn't be
2740 necessary to configure just to create a source distribution.
2744 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2746 * Timing vulnerability in DSA signature generation
2748 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2749 timing side channel attack. An attacker could use variations in the signing
2750 algorithm to recover the private key.
2752 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2757 * Timing vulnerability in ECDSA signature generation
2759 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2760 timing side channel attack. An attacker could use variations in the signing
2761 algorithm to recover the private key.
2763 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2768 * Add coordinate blinding for EC_POINT and implement projective
2769 coordinate blinding for generic prime curves as a countermeasure to
2770 chosen point SCA attacks.
2772 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2774 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2776 * Client DoS due to large DH parameter
2778 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2779 malicious server can send a very large prime value to the client. This will
2780 cause the client to spend an unreasonably long period of time generating a
2781 key for this prime resulting in a hang until the client has finished. This
2782 could be exploited in a Denial Of Service attack.
2784 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2789 * Cache timing vulnerability in RSA Key Generation
2791 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2792 a cache timing side channel attack. An attacker with sufficient access to
2793 mount cache timing attacks during the RSA key generation process could
2794 recover the private key.
2796 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2797 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2802 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2803 parameter is no longer accepted, as it leads to a corrupt table. NULL
2804 pem_str is reserved for alias entries only.
2808 * Revert blinding in ECDSA sign and instead make problematic addition
2809 length-invariant. Switch even to fixed-length Montgomery multiplication.
2813 * Change generating and checking of primes so that the error rate of not
2814 being prime depends on the intended use based on the size of the input.
2815 For larger primes this will result in more rounds of Miller-Rabin.
2816 The maximal error rate for primes with more than 1080 bits is lowered
2819 *Kurt Roeckx, Annie Yousar*
2821 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2825 * Add blinding to ECDSA and DSA signatures to protect against side channel
2826 attacks discovered by Keegan Ryan (NCC Group).
2830 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2831 now allow empty (zero character) pass phrases.
2835 * Certificate time validation (X509_cmp_time) enforces stricter
2836 compliance with RFC 5280. Fractional seconds and timezone offsets
2837 are no longer allowed.
2841 * Fixed a text canonicalisation bug in CMS
2843 Where a CMS detached signature is used with text content the text goes
2844 through a canonicalisation process first prior to signing or verifying a
2845 signature. This process strips trailing space at the end of lines, converts
2846 line terminators to CRLF and removes additional trailing line terminators
2847 at the end of a file. A bug in the canonicalisation process meant that
2848 some characters, such as form-feed, were incorrectly treated as whitespace
2849 and removed. This is contrary to the specification (RFC5485). This fix
2850 could mean that detached text data signed with an earlier version of
2851 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2852 signed with a fixed OpenSSL may fail to verify with an earlier version of
2853 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2854 and use the "-binary" flag (for the "cms" command line application) or set
2855 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2859 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2861 * Constructed ASN.1 types with a recursive definition could exceed the stack
2863 Constructed ASN.1 types with a recursive definition (such as can be found
2864 in PKCS7) could eventually exceed the stack given malicious input with
2865 excessive recursion. This could result in a Denial Of Service attack. There
2866 are no such structures used within SSL/TLS that come from untrusted sources
2867 so this is considered safe.
2869 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2875 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2877 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2878 effectively reduced to only comparing the least significant bit of each
2879 byte. This allows an attacker to forge messages that would be considered as
2880 authenticated in an amount of tries lower than that guaranteed by the
2881 security claims of the scheme. The module can only be compiled by the
2882 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2884 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2890 * Add a build target 'build_all_generated', to build all generated files
2891 and only that. This can be used to prepare everything that requires
2892 things like perl for a system that lacks perl and then move everything
2893 to that system and do the rest of the build there.
2897 * Backport SSL_OP_NO_RENGOTIATION
2899 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2900 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2901 changes this is no longer possible in 1.1.0. Therefore the new
2902 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2903 1.1.0 to provide equivalent functionality.
2905 Note that if an application built against 1.1.0h headers (or above) is run
2906 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2907 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2911 * Removed the OS390-Unix config target. It relied on a script that doesn't
2916 * rsaz_1024_mul_avx2 overflow bug on x86_64
2918 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2919 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2920 Analysis suggests that attacks against RSA and DSA as a result of this
2921 defect would be very difficult to perform and are not believed likely.
2922 Attacks against DH1024 are considered just feasible, because most of the
2923 work necessary to deduce information about a private key may be performed
2924 offline. The amount of resources required for such an attack would be
2925 significant. However, for an attack on TLS to be meaningful, the server
2926 would have to share the DH1024 private key among multiple clients, which is
2927 no longer an option since CVE-2016-0701.
2929 This only affects processors that support the AVX2 but not ADX extensions
2930 like Intel Haswell (4th generation).
2932 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2933 was originally found via the OSS-Fuzz project.
2938 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2940 * bn_sqrx8x_internal carry bug on x86_64
2942 There is a carry propagating bug in the x86_64 Montgomery squaring
2943 procedure. No EC algorithms are affected. Analysis suggests that attacks
2944 against RSA and DSA as a result of this defect would be very difficult to
2945 perform and are not believed likely. Attacks against DH are considered just
2946 feasible (although very difficult) because most of the work necessary to
2947 deduce information about a private key may be performed offline. The amount
2948 of resources required for such an attack would be very significant and
2949 likely only accessible to a limited number of attackers. An attacker would
2950 additionally need online access to an unpatched system using the target
2951 private key in a scenario with persistent DH parameters and a private
2952 key that is shared between multiple clients.
2954 This only affects processors that support the BMI1, BMI2 and ADX extensions
2955 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2957 This issue was reported to OpenSSL by the OSS-Fuzz project.
2962 * Malformed X.509 IPAddressFamily could cause OOB read
2964 If an X.509 certificate has a malformed IPAddressFamily extension,
2965 OpenSSL could do a one-byte buffer overread. The most likely result
2966 would be an erroneous display of the certificate in text format.
2968 This issue was reported to OpenSSL by the OSS-Fuzz project.
2973 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2975 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2976 platform rather than 'mingw'.
2980 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2981 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2982 which is the minimum version we support.
2986 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2988 * Encrypt-Then-Mac renegotiation crash
2990 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2991 negotiated where it was not in the original handshake (or vice-versa) then
2992 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2993 and servers are affected.
2995 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3000 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3002 * Truncated packet could crash via OOB read
3004 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3005 cipher is being used, then a truncated packet can cause that host to
3006 perform an out-of-bounds read, usually resulting in a crash.
3008 This issue was reported to OpenSSL by Robert Święcki of Google.
3013 * Bad (EC)DHE parameters cause a client crash
3015 If a malicious server supplies bad parameters for a DHE or ECDHE key
3016 exchange then this can result in the client attempting to dereference a
3017 NULL pointer leading to a client crash. This could be exploited in a Denial
3020 This issue was reported to OpenSSL by Guido Vranken.
3025 * BN_mod_exp may produce incorrect results on x86_64
3027 There is a carry propagating bug in the x86_64 Montgomery squaring
3028 procedure. No EC algorithms are affected. Analysis suggests that attacks
3029 against RSA and DSA as a result of this defect would be very difficult to
3030 perform and are not believed likely. Attacks against DH are considered just
3031 feasible (although very difficult) because most of the work necessary to
3032 deduce information about a private key may be performed offline. The amount
3033 of resources required for such an attack would be very significant and
3034 likely only accessible to a limited number of attackers. An attacker would
3035 additionally need online access to an unpatched system using the target
3036 private key in a scenario with persistent DH parameters and a private
3037 key that is shared between multiple clients. For example this can occur by
3038 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3039 similar to CVE-2015-3193 but must be treated as a separate problem.
3041 This issue was reported to OpenSSL by the OSS-Fuzz project.
3046 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3048 * ChaCha20/Poly1305 heap-buffer-overflow
3050 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3051 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3052 crash. This issue is not considered to be exploitable beyond a DoS.
3054 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3059 * CMS Null dereference
3061 Applications parsing invalid CMS structures can crash with a NULL pointer
3062 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3063 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3064 structure callback if an attempt is made to free certain invalid encodings.
3065 Only CHOICE structures using a callback which do not handle NULL value are
3068 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3073 * Montgomery multiplication may produce incorrect results
3075 There is a carry propagating bug in the Broadwell-specific Montgomery
3076 multiplication procedure that handles input lengths divisible by, but
3077 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3078 and DH private keys are impossible. This is because the subroutine in
3079 question is not used in operations with the private key itself and an input
3080 of the attacker's direct choice. Otherwise the bug can manifest itself as
3081 transient authentication and key negotiation failures or reproducible
3082 erroneous outcome of public-key operations with specially crafted input.
3083 Among EC algorithms only Brainpool P-512 curves are affected and one
3084 presumably can attack ECDH key negotiation. Impact was not analyzed in
3085 detail, because pre-requisites for attack are considered unlikely. Namely
3086 multiple clients have to choose the curve in question and the server has to
3087 share the private key among them, neither of which is default behaviour.
3088 Even then only clients that chose the curve will be affected.
3090 This issue was publicly reported as transient failures and was not
3091 initially recognized as a security issue. Thanks to Richard Morgan for
3092 providing reproducible case.
3097 * Removed automatic addition of RPATH in shared libraries and executables,
3098 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3102 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3104 * Fix Use After Free for large message sizes
3106 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3107 message larger than approx 16k is received then the underlying buffer to
3108 store the incoming message is reallocated and moved. Unfortunately a
3109 dangling pointer to the old location is left which results in an attempt to
3110 write to the previously freed location. This is likely to result in a
3111 crash, however it could potentially lead to execution of arbitrary code.
3113 This issue only affects OpenSSL 1.1.0a.
3115 This issue was reported to OpenSSL by Robert Święcki.
3120 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3122 * OCSP Status Request extension unbounded memory growth
3124 A malicious client can send an excessively large OCSP Status Request
3125 extension. If that client continually requests renegotiation, sending a
3126 large OCSP Status Request extension each time, then there will be unbounded
3127 memory growth on the server. This will eventually lead to a Denial Of
3128 Service attack through memory exhaustion. Servers with a default
3129 configuration are vulnerable even if they do not support OCSP. Builds using
3130 the "no-ocsp" build time option are not affected.
3132 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3137 * SSL_peek() hang on empty record
3139 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3140 sends an empty record. This could be exploited by a malicious peer in a
3141 Denial Of Service attack.
3143 This issue was reported to OpenSSL by Alex Gaynor.
3148 * Excessive allocation of memory in tls_get_message_header() and
3149 dtls1_preprocess_fragment()
3151 A (D)TLS message includes 3 bytes for its length in the header for the
3152 message. This would allow for messages up to 16Mb in length. Messages of
3153 this length are excessive and OpenSSL includes a check to ensure that a
3154 peer is sending reasonably sized messages in order to avoid too much memory
3155 being consumed to service a connection. A flaw in the logic of version
3156 1.1.0 means that memory for the message is allocated too early, prior to
3157 the excessive message length check. Due to way memory is allocated in
3158 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3159 to service a connection. This could lead to a Denial of Service through
3160 memory exhaustion. However, the excessive message length check still takes
3161 place, and this would cause the connection to immediately fail. Assuming
3162 that the application calls SSL_free() on the failed connection in a timely
3163 manner then the 21Mb of allocated memory will then be immediately freed
3164 again. Therefore the excessive memory allocation will be transitory in
3165 nature. This then means that there is only a security impact if:
3167 1) The application does not call SSL_free() in a timely manner in the event
3168 that the connection fails
3170 2) The application is working in a constrained environment where there is
3171 very little free memory
3173 3) The attacker initiates multiple connection attempts such that there are
3174 multiple connections in a state where memory has been allocated for the
3175 connection; SSL_free() has not yet been called; and there is insufficient
3176 memory to service the multiple requests.
3178 Except in the instance of (1) above any Denial Of Service is likely to be
3179 transitory because as soon as the connection fails the memory is
3180 subsequently freed again in the SSL_free() call. However there is an
3181 increased risk during this period of application crashes due to the lack of
3182 memory - which would then mean a more serious Denial of Service.
3184 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3185 (CVE-2016-6307 and CVE-2016-6308)
3189 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3190 had to be removed. Primary reason is that vendor assembler can't
3191 assemble our modules with -KPIC flag. As result it, assembly
3192 support, was not even available as option. But its lack means
3193 lack of side-channel resistant code, which is incompatible with
3194 security by todays standards. Fortunately gcc is readily available
3195 prepackaged option, which we firmly point at...
3199 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3201 * Windows command-line tool supports UTF-8 opt-in option for arguments
3202 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3203 (to any value) allows Windows user to access PKCS#12 file generated
3204 with Windows CryptoAPI and protected with non-ASCII password, as well
3205 as files generated under UTF-8 locale on Linux also protected with
3210 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3211 have been disabled by default and removed from DEFAULT, just like RC4.
3212 See the RC4 item below to re-enable both.
3216 * The method for finding the storage location for the Windows RAND seed file
3217 has changed. First we check %RANDFILE%. If that is not set then we check
3218 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3219 all else fails we fall back to C:\.
3223 * The EVP_EncryptUpdate() function has had its return type changed from void
3224 to int. A return of 0 indicates and error while a return of 1 indicates
3229 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3230 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3231 off the constant time implementation for RSA, DSA and DH have been made
3232 no-ops and deprecated.
3236 * Windows RAND implementation was simplified to only get entropy by
3237 calling CryptGenRandom(). Various other RAND-related tickets
3240 *Joseph Wylie Yandle, Rich Salz*
3242 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3243 and `OPENSSL_LH_`, respectively. The old names are available
3244 with API compatibility. They new names are now completely documented.
3248 * Unify TYPE_up_ref(obj) methods signature.
3249 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3250 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3251 int (instead of void) like all others TYPE_up_ref() methods.
3252 So now these methods also check the return value of CRYPTO_atomic_add(),
3253 and the validity of object reference counter.
3255 *fdasilvayy@gmail.com*
3257 * With Windows Visual Studio builds, the .pdb files are installed
3258 alongside the installed libraries and executables. For a static
3259 library installation, ossl_static.pdb is the associate compiler
3260 generated .pdb file to be used when linking programs.
3264 * Remove openssl.spec. Packaging files belong with the packagers.
3268 * Automatic Darwin/OSX configuration has had a refresh, it will now
3269 recognise x86_64 architectures automatically. You can still decide
3270 to build for a different bitness with the environment variable
3271 KERNEL_BITS (can be 32 or 64), for example:
3273 KERNEL_BITS=32 ./config
3277 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3278 256 bit AES and HMAC with SHA256.
3282 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3286 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3290 * To enable users to have their own config files and build file templates,
3291 Configure looks in the directory indicated by the environment variable
3292 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3293 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3294 name and is used as is.
3298 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3299 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3300 X509_CERT_FILE_CTX was removed.
3304 * "shared" builds are now the default. To create only static libraries use
3305 the "no-shared" Configure option.
3309 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3310 All of these option have not worked for some while and are fundamental
3315 * Make various cleanup routines no-ops and mark them as deprecated. Most
3316 global cleanup functions are no longer required because they are handled
3317 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3318 Explicitly de-initing can cause problems (e.g. where a library that uses
3319 OpenSSL de-inits, but an application is still using it). The affected
3320 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3321 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3322 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3323 COMP_zlib_cleanup().
3327 * --strict-warnings no longer enables runtime debugging options
3328 such as REF_DEBUG. Instead, debug options are automatically
3329 enabled with '--debug' builds.
3331 *Andy Polyakov, Emilia Käsper*
3333 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3334 have been moved out of the public header files. New functions for managing
3335 these have been added.
3339 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3340 objects have been moved out of the public header files. New
3341 functions for managing these have been added.
3345 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3346 have been moved out of the public header files. New functions for managing
3347 these have been added.
3351 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3352 moved out of the public header files. New functions for managing these
3357 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3361 * Removed the mk1mf build scripts.
3365 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3366 it is always safe to #include a header now.
3370 * Removed the aged BC-32 config and all its supporting scripts
3374 * Removed support for Ultrix, Netware, and OS/2.
3378 * Add support for HKDF.
3380 *Alessandro Ghedini*
3382 * Add support for blake2b and blake2s
3386 * Added support for "pipelining". Ciphers that have the
3387 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3388 encryptions/decryptions simultaneously. There are currently no built-in
3389 ciphers with this property but the expectation is that engines will be able
3390 to offer it to significantly improve throughput. Support has been extended
3391 into libssl so that multiple records for a single connection can be
3392 processed in one go (for >=TLS 1.1).
3396 * Added the AFALG engine. This is an async capable engine which is able to
3397 offload work to the Linux kernel. In this initial version it only supports
3398 AES128-CBC. The kernel must be version 4.1.0 or greater.
3402 * OpenSSL now uses a new threading API. It is no longer necessary to
3403 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3404 are two supported threading models: pthreads and windows threads. It is
3405 also possible to configure OpenSSL at compile time for "no-threads". The
3406 old threading API should no longer be used. The functions have been
3407 replaced with "no-op" compatibility macros.
3409 *Alessandro Ghedini, Matt Caswell*
3411 * Modify behavior of ALPN to invoke callback after SNI/servername
3412 callback, such that updates to the SSL_CTX affect ALPN.
3416 * Add SSL_CIPHER queries for authentication and key-exchange.
3420 * Changes to the DEFAULT cipherlist:
3421 - Prefer (EC)DHE handshakes over plain RSA.
3422 - Prefer AEAD ciphers over legacy ciphers.
3423 - Prefer ECDSA over RSA when both certificates are available.
3424 - Prefer TLSv1.2 ciphers/PRF.
3425 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3430 * Change the ECC default curve list to be this, in order: x25519,
3431 secp256r1, secp521r1, secp384r1.
3435 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3436 disabled by default. They can be re-enabled using the
3437 enable-weak-ssl-ciphers option to Configure.
3441 * If the server has ALPN configured, but supports no protocols that the
3442 client advertises, send a fatal "no_application_protocol" alert.
3443 This behaviour is SHALL in RFC 7301, though it isn't universally
3444 implemented by other servers.
3448 * Add X25519 support.
3449 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3450 for public and private key encoding using the format documented in
3451 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3452 key generation and key derivation.
3454 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3459 * Deprecate SRP_VBASE_get_by_user.
3460 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3461 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3462 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3463 seed, even if the seed is configured.
3465 Users should use SRP_VBASE_get1_by_user instead. Note that in
3466 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3467 also that even though configuring the SRP seed attempts to hide
3468 invalid usernames by continuing the handshake with fake
3469 credentials, this behaviour is not constant time and no strong
3470 guarantees are made that the handshake is indistinguishable from
3471 that of a valid user.
3475 * Configuration change; it's now possible to build dynamic engines
3476 without having to build shared libraries and vice versa. This
3477 only applies to the engines in `engines/`, those in `crypto/engine/`
3478 will always be built into libcrypto (i.e. "static").
3480 Building dynamic engines is enabled by default; to disable, use
3481 the configuration option "disable-dynamic-engine".
3483 The only requirements for building dynamic engines are the
3484 presence of the DSO module and building with position independent
3485 code, so they will also automatically be disabled if configuring
3486 with "disable-dso" or "disable-pic".
3488 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3489 are also taken away from openssl/opensslconf.h, as they are
3494 * Configuration change; if there is a known flag to compile
3495 position independent code, it will always be applied on the
3496 libcrypto and libssl object files, and never on the application
3497 object files. This means other libraries that use routines from
3498 libcrypto / libssl can be made into shared libraries regardless
3499 of how OpenSSL was configured.
3501 If this isn't desirable, the configuration options "disable-pic"
3502 or "no-pic" can be used to disable the use of PIC. This will
3503 also disable building shared libraries and dynamic engines.
3507 * Removed JPAKE code. It was experimental and has no wide use.
3511 * The INSTALL_PREFIX Makefile variable has been renamed to
3512 DESTDIR. That makes for less confusion on what this variable
3513 is for. Also, the configuration option --install_prefix is
3518 * Heartbeat for TLS has been removed and is disabled by default
3519 for DTLS; configure with enable-heartbeats. Code that uses the
3520 old #define's might need to be updated.
3522 *Emilia Käsper, Rich Salz*
3524 * Rename REF_CHECK to REF_DEBUG.
3528 * New "unified" build system
3530 The "unified" build system is aimed to be a common system for all
3531 platforms we support. With it comes new support for VMS.
3533 This system builds supports building in a different directory tree
3534 than the source tree. It produces one Makefile (for unix family
3535 or lookalikes), or one descrip.mms (for VMS).
3537 The source of information to make the Makefile / descrip.mms is
3538 small files called 'build.info', holding the necessary
3539 information for each directory with source to compile, and a
3540 template in Configurations, like unix-Makefile.tmpl or
3543 With this change, the library names were also renamed on Windows
3544 and on VMS. They now have names that are closer to the standard
3545 on Unix, and include the major version number, and in certain
3546 cases, the architecture they are built for. See "Notes on shared
3547 libraries" in INSTALL.
3549 We rely heavily on the perl module Text::Template.
3553 * Added support for auto-initialisation and de-initialisation of the library.
3554 OpenSSL no longer requires explicit init or deinit routines to be called,
3555 except in certain circumstances. See the OPENSSL_init_crypto() and
3556 OPENSSL_init_ssl() man pages for further information.
3560 * The arguments to the DTLSv1_listen function have changed. Specifically the
3561 "peer" argument is now expected to be a BIO_ADDR object.
3563 * Rewrite of BIO networking library. The BIO library lacked consistent
3564 support of IPv6, and adding it required some more extensive
3565 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3566 which hold all types of addresses and chains of address information.
3567 It also introduces a new API, with functions like BIO_socket,
3568 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3569 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3570 have been adapted accordingly.
3574 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3579 * CRIME protection: disable compression by default, even if OpenSSL is
3580 compiled with zlib enabled. Applications can still enable compression
3581 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3582 using the SSL_CONF library to configure compression.
3586 * The signature of the session callback configured with
3587 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3588 was explicitly marked as `const unsigned char*` instead of
3593 * Always DPURIFY. Remove the use of uninitialized memory in the
3594 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3598 * Removed many obsolete configuration items, including
3599 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3600 MD2_CHAR, MD2_INT, MD2_LONG
3602 IDEA_SHORT, IDEA_LONG
3603 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3605 *Rich Salz, with advice from Andy Polyakov*
3607 * Many BN internals have been moved to an internal header file.
3609 *Rich Salz with help from Andy Polyakov*
3611 * Configuration and writing out the results from it has changed.
3612 Files such as Makefile include/openssl/opensslconf.h and are now
3613 produced through general templates, such as Makefile.in and
3614 crypto/opensslconf.h.in and some help from the perl module
3617 Also, the center of configuration information is no longer
3618 Makefile. Instead, Configure produces a perl module in
3619 configdata.pm which holds most of the config data (in the hash
3620 table %config), the target data that comes from the target
3621 configuration in one of the `Configurations/*.conf` files (in
3626 * To clarify their intended purposes, the Configure options
3627 --prefix and --openssldir change their semantics, and become more
3628 straightforward and less interdependent.
3630 --prefix shall be used exclusively to give the location INSTALLTOP
3631 where programs, scripts, libraries, include files and manuals are
3632 going to be installed. The default is now /usr/local.
3634 --openssldir shall be used exclusively to give the default
3635 location OPENSSLDIR where certificates, private keys, CRLs are
3636 managed. This is also where the default openssl.cnf gets
3638 If the directory given with this option is a relative path, the
3639 values of both the --prefix value and the --openssldir value will
3640 be combined to become OPENSSLDIR.
3641 The default for --openssldir is INSTALLTOP/ssl.
3643 Anyone who uses --openssldir to specify where OpenSSL is to be
3644 installed MUST change to use --prefix instead.
3648 * The GOST engine was out of date and therefore it has been removed. An up
3649 to date GOST engine is now being maintained in an external repository.
3650 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3651 support for GOST ciphersuites (these are only activated if a GOST engine
3656 * EGD is no longer supported by default; use enable-egd when
3659 *Ben Kaduk and Rich Salz*
3661 * The distribution now has Makefile.in files, which are used to
3662 create Makefile's when Configure is run. *Configure must be run
3663 before trying to build now.*
3667 * The return value for SSL_CIPHER_description() for error conditions
3672 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3674 Obtaining and performing DNSSEC validation of TLSA records is
3675 the application's responsibility. The application provides
3676 the TLSA records of its choice to OpenSSL, and these are then
3677 used to authenticate the peer.
3679 The TLSA records need not even come from DNS. They can, for
3680 example, be used to implement local end-entity certificate or
3681 trust-anchor "pinning", where the "pin" data takes the form
3682 of TLSA records, which can augment or replace verification
3683 based on the usual WebPKI public certification authorities.
3687 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3688 continues to support deprecated interfaces in default builds.
3689 However, applications are strongly advised to compile their
3690 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3691 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3692 or the 1.1.0 releases.
3694 In environments in which all applications have been ported to
3695 not use any deprecated interfaces OpenSSL's Configure script
3696 should be used with the --api=1.1.0 option to entirely remove
3697 support for the deprecated features from the library and
3698 unconditionally disable them in the installed headers.
3699 Essentially the same effect can be achieved with the "no-deprecated"
3700 argument to Configure, except that this will always restrict
3701 the build to just the latest API, rather than a fixed API
3704 As applications are ported to future revisions of the API,
3705 they should update their compile-time OPENSSL_API_COMPAT define
3706 accordingly, but in most cases should be able to continue to
3707 compile with later releases.
3709 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3710 0x10000000L and 0x00908000L, respectively. However those
3711 versions did not support the OPENSSL_API_COMPAT feature, and
3712 so applications are not typically tested for explicit support
3713 of just the undeprecated features of either release.
3717 * Add support for setting the minimum and maximum supported protocol.
3718 It can bet set via the SSL_set_min_proto_version() and
3719 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3720 MaxProtocol. It's recommended to use the new APIs to disable
3721 protocols instead of disabling individual protocols using
3722 SSL_set_options() or SSL_CONF's Protocol. This change also
3723 removes support for disabling TLS 1.2 in the OpenSSL TLS
3724 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3728 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3732 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3733 and integrates ECDSA and ECDH functionality into EC. Implementations can
3734 now redirect key generation and no longer need to convert to or from
3737 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3738 include the ec.h header file instead.
3742 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3743 ciphers who are no longer supported and drops support the ephemeral RSA key
3744 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3748 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3749 opaque. For HMAC_CTX, the following constructors and destructors
3752 HMAC_CTX *HMAC_CTX_new(void);
3753 void HMAC_CTX_free(HMAC_CTX *ctx);
3755 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3756 destroy such methods has been added. See EVP_MD_meth_new(3) and
3757 EVP_CIPHER_meth_new(3) for documentation.
3760 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3761 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3762 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3763 an already created structure.
3764 2) For consistency with the majority of our object creators and
3765 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3766 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3767 for deprecated builds.
3771 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3772 cryptographic operations to be performed asynchronously as long as an
3773 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3774 further details. Libssl has also had this capability integrated with the
3775 introduction of the new mode SSL_MODE_ASYNC and associated error
3776 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3777 pages. This work was developed in partnership with Intel Corp.
3781 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3782 always enabled now. If you want to disable the support you should
3783 exclude it using the list of supported ciphers. This also means that the
3784 "-no_ecdhe" option has been removed from s_server.
3788 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3789 SSL_{CTX_}set1_curves() which can set a list.
3793 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3794 curve you want to support using SSL_{CTX_}set1_curves().
3798 * State machine rewrite. The state machine code has been significantly
3799 refactored in order to remove much duplication of code and solve issues
3800 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3801 further details). This change does have some associated API changes.
3802 Notably the SSL_state() function has been removed and replaced by
3803 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3804 SSL_set_state() has been removed altogether. The previous handshake states
3805 defined in ssl.h and ssl3.h have also been removed.
3809 * All instances of the string "ssleay" in the public API were replaced
3810 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3811 Some error codes related to internal RSA_eay API's were renamed.
3815 * The demo files in crypto/threads were moved to demo/threads.
3819 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3822 *Matt Caswell, Rich Salz*
3824 * New ASN.1 embed macro.
3826 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3827 structure is not allocated: it is part of the parent. That is instead of
3835 This reduces memory fragmentation and make it impossible to accidentally
3836 set a mandatory field to NULL.
3838 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3839 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3840 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3845 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3849 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3850 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3851 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3852 DES and RC4 ciphersuites.
3856 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3857 This changes the decoding behaviour for some invalid messages,
3858 though the change is mostly in the more lenient direction, and
3859 legacy behaviour is preserved as much as possible.
3863 * Fix no-stdio build.
3864 *David Woodhouse <David.Woodhouse@intel.com> and also*
3865 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3867 * New testing framework
3868 The testing framework has been largely rewritten and is now using
3869 perl and the perl modules Test::Harness and an extended variant of
3870 Test::More called OpenSSL::Test to do its work. All test scripts in
3871 test/ have been rewritten into test recipes, and all direct calls to
3872 executables in test/Makefile have become individual recipes using the
3873 simplified testing OpenSSL::Test::Simple.
3875 For documentation on our testing modules, do:
3877 perldoc test/testlib/OpenSSL/Test/Simple.pm
3878 perldoc test/testlib/OpenSSL/Test.pm
3882 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3883 are used; the latter aborts on memory leaks (usually checked on exit).
3884 Some undocumented "set malloc, etc., hooks" functions were removed
3885 and others were changed. All are now documented.
3889 * In DSA_generate_parameters_ex, if the provided seed is too short,
3892 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3894 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3895 from RFC4279, RFC4785, RFC5487, RFC5489.
3897 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3898 original RSA_PSK patch.
3902 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3903 era flag was never set throughout the codebase (only read). Also removed
3904 SSL3_FLAGS_POP_BUFFER which was only used if
3905 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3909 * Changed the default name options in the "ca", "crl", "req" and "x509"
3910 to be "oneline" instead of "compat".
3914 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3915 not aware of clients that still exhibit this bug, and the workaround
3916 hasn't been working properly for a while.
3920 * The return type of BIO_number_read() and BIO_number_written() as well as
3921 the corresponding num_read and num_write members in the BIO structure has
3922 changed from unsigned long to uint64_t. On platforms where an unsigned
3923 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3928 * Given the pervasive nature of TLS extensions it is inadvisable to run
3929 OpenSSL without support for them. It also means that maintaining
3930 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3931 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3935 * Removed support for the two export grade static DH ciphersuites
3936 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3937 were newly added (along with a number of other static DH ciphersuites) to
3938 1.0.2. However the two export ones have *never* worked since they were
3939 introduced. It seems strange in any case to be adding new export
3940 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3944 * Version negotiation has been rewritten. In particular SSLv23_method(),
3945 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3946 and turned into macros which simply call the new preferred function names
3947 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3948 should use the new names instead. Also as part of this change the ssl23.h
3949 header file has been removed.
3953 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3954 code and the associated standard is no longer considered fit-for-purpose.
3958 * RT2547 was closed. When generating a private key, try to make the
3959 output file readable only by the owner. This behavior change might
3960 be noticeable when interacting with other software.
3962 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3967 * Added HTTP GET support to the ocsp command.
3971 * Changed default digest for the dgst and enc commands from MD5 to
3976 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3980 * Added support for TLS extended master secret from
3981 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3982 initial patch which was a great help during development.
3986 * All libssl internal structures have been removed from the public header
3987 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3988 now redundant). Users should not attempt to access internal structures
3989 directly. Instead they should use the provided API functions.
3993 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3994 Access to deprecated functions can be re-enabled by running config with
3995 "enable-deprecated". In addition applications wishing to use deprecated
3996 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3997 will, by default, disable some transitive includes that previously existed
3998 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4002 * Added support for OCB mode. OpenSSL has been granted a patent license
4003 compatible with the OpenSSL license for use of OCB. Details are available
4004 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4005 for OCB can be removed by calling config with no-ocb.
4009 * SSLv2 support has been removed. It still supports receiving a SSLv2
4010 compatible client hello.
4014 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4015 done while fixing the error code for the key-too-small case.
4017 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4019 * CA.sh has been removed; use CA.pl instead.
4023 * Removed old DES API.
4027 * Remove various unsupported platforms:
4033 Sinix/ReliantUNIX RM400
4038 16-bit platforms such as WIN16
4042 * Clean up OPENSSL_NO_xxx #define's
4043 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4044 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4045 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4046 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4047 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4048 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4049 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4050 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4051 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4052 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4056 * Cleaned up dead code
4057 Remove all but one '#ifdef undef' which is to be looked at.
4061 * Clean up calling of xxx_free routines.
4062 Just like free(), fix most of the xxx_free routines to accept
4063 NULL. Remove the non-null checks from callers. Save much code.
4067 * Add secure heap for storage of private keys (when possible).
4068 Add BIO_s_secmem(), CBIGNUM, etc.
4069 Contributed by Akamai Technologies under our Corporate CLA.
4073 * Experimental support for a new, fast, unbiased prime candidate generator,
4074 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4076 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4078 * New output format NSS in the sess_id command line tool. This allows
4079 exporting the session id and the master key in NSS keylog format.
4081 *Martin Kaiser <martin@kaiser.cx>*
4083 * Harmonize version and its documentation. -f flag is used to display
4086 *mancha <mancha1@zoho.com>*
4088 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4089 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4091 *mancha <mancha1@zoho.com>*
4093 * Fix some double frees. These are not thought to be exploitable.
4095 *mancha <mancha1@zoho.com>*
4097 * A missing bounds check in the handling of the TLS heartbeat extension
4098 can be used to reveal up to 64k of memory to a connected client or
4101 Thanks for Neel Mehta of Google Security for discovering this bug and to
4102 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4103 preparing the fix ([CVE-2014-0160])
4105 *Adam Langley, Bodo Moeller*
4107 * Fix for the attack described in the paper "Recovering OpenSSL
4108 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4109 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4110 <http://eprint.iacr.org/2014/140>
4112 Thanks to Yuval Yarom and Naomi Benger for discovering this
4113 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4115 *Yuval Yarom and Naomi Benger*
4117 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4118 this fixes a limitation in previous versions of OpenSSL.
4122 * Experimental encrypt-then-mac support.
4124 Experimental support for encrypt then mac from
4125 draft-gutmann-tls-encrypt-then-mac-02.txt
4127 To enable it set the appropriate extension number (0x42 for the test
4128 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4130 For non-compliant peers (i.e. just about everything) this should have no
4133 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4137 * Add EVP support for key wrapping algorithms, to avoid problems with
4138 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4139 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4140 algorithms and include tests cases.
4144 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4149 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4150 MGF1 digest and OAEP label.
4154 * Make openssl verify return errors.
4156 *Chris Palmer <palmer@google.com> and Ben Laurie*
4158 * New function ASN1_TIME_diff to calculate the difference between two
4159 ASN1_TIME structures or one structure and the current time.
4163 * Update fips_test_suite to support multiple command line options. New
4164 test to induce all self test errors in sequence and check expected
4169 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4170 sign or verify all in one operation.
4174 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4175 test programs and fips_test_suite. Includes functionality to parse
4176 the minimal script output of fipsalgest.pl directly.
4180 * Add authorisation parameter to FIPS_module_mode_set().
4184 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4188 * Use separate DRBG fields for internal and external flags. New function
4189 FIPS_drbg_health_check() to perform on demand health checking. Add
4190 generation tests to fips_test_suite with reduced health check interval to
4191 demonstrate periodic health checking. Add "nodh" option to
4192 fips_test_suite to skip very slow DH test.
4196 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4201 * More extensive health check for DRBG checking many more failure modes.
4202 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4203 combination: call this in fips_test_suite.
4207 * Add support for canonical generation of DSA parameter 'g'. See
4210 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4211 POST to handle HMAC cases.
4215 * Add functions FIPS_module_version() and FIPS_module_version_text()
4216 to return numerical and string versions of the FIPS module number.
4220 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4221 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4222 outside the validated module in the FIPS capable OpenSSL.
4226 * Minor change to DRBG entropy callback semantics. In some cases
4227 there is no multiple of the block length between min_len and
4228 max_len. Allow the callback to return more than max_len bytes
4229 of entropy but discard any extra: it is the callback's responsibility
4230 to ensure that the extra data discarded does not impact the
4231 requested amount of entropy.
4235 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4236 information in FIPS186-3, SP800-57 and SP800-131A.
4240 * CCM support via EVP. Interface is very similar to GCM case except we
4241 must supply all data in one chunk (i.e. no update, final) and the
4242 message length must be supplied if AAD is used. Add algorithm test
4247 * Initial version of POST overhaul. Add POST callback to allow the status
4248 of POST to be monitored and/or failures induced. Modify fips_test_suite
4249 to use callback. Always run all selftests even if one fails.
4253 * XTS support including algorithm test driver in the fips_gcmtest program.
4254 Note: this does increase the maximum key length from 32 to 64 bytes but
4255 there should be no binary compatibility issues as existing applications
4256 will never use XTS mode.
4260 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4261 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4262 performs algorithm blocking for unapproved PRNG types. Also do not
4263 set PRNG type in FIPS_mode_set(): leave this to the application.
4264 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4265 the standard OpenSSL PRNG: set additional data to a date time vector.
4269 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4270 This shouldn't present any incompatibility problems because applications
4271 shouldn't be using these directly and any that are will need to rethink
4272 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4276 * Extensive self tests and health checking required by SP800-90 DRBG.
4277 Remove strength parameter from FIPS_drbg_instantiate and always
4278 instantiate at maximum supported strength.
4282 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4286 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4290 * New function DH_compute_key_padded() to compute a DH key and pad with
4291 leading zeroes if needed: this complies with SP800-56A et al.
4295 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4296 anything, incomplete, subject to change and largely untested at present.
4300 * Modify fipscanisteronly build option to only build the necessary object
4301 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4305 * Add experimental option FIPSSYMS to give all symbols in
4306 fipscanister.o and FIPS or fips prefix. This will avoid
4307 conflicts with future versions of OpenSSL. Add perl script
4308 util/fipsas.pl to preprocess assembly language source files
4309 and rename any affected symbols.
4313 * Add selftest checks and algorithm block of non-fips algorithms in
4314 FIPS mode. Remove DES2 from selftests.
4318 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4319 return internal method without any ENGINE dependencies. Add new
4320 tiny fips sign and verify functions.
4324 * New build option no-ec2m to disable characteristic 2 code.
4328 * New build option "fipscanisteronly". This only builds fipscanister.o
4329 and (currently) associated fips utilities. Uses the file Makefile.fips
4330 instead of Makefile.org as the prototype.
4334 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4335 Update fips_gcmtest to use IV generator.
4339 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4340 setting output buffer to NULL. The `*Final` function must be
4341 called although it will not retrieve any additional data. The tag
4342 can be set or retrieved with a ctrl. The IV length is by default 12
4343 bytes (96 bits) but can be set to an alternative value. If the IV
4344 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4349 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4350 underlying do_cipher function handles all cipher semantics itself
4351 including padding and finalisation. This is useful if (for example)
4352 an ENGINE cipher handles block padding itself. The behaviour of
4353 do_cipher is subtly changed if this flag is set: the return value
4354 is the number of characters written to the output buffer (zero is
4355 no longer an error code) or a negative error code. Also if the
4356 input buffer is NULL and length 0 finalisation should be performed.
4360 * If a candidate issuer certificate is already part of the constructed
4361 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4365 * Improve forward-security support: add functions
4367 void SSL_CTX_set_not_resumable_session_callback(
4368 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4369 void SSL_set_not_resumable_session_callback(
4370 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4372 for use by SSL/TLS servers; the callback function will be called whenever a
4373 new session is created, and gets to decide whether the session may be
4374 cached to make it resumable (return 0) or not (return 1). (As by the
4375 SSL/TLS protocol specifications, the session_id sent by the server will be
4376 empty to indicate that the session is not resumable; also, the server will
4377 not generate RFC 4507 (RFC 5077) session tickets.)
4379 A simple reasonable callback implementation is to return is_forward_secure.
4380 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4381 by the SSL/TLS server library, indicating whether it can provide forward
4384 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4386 * New -verify_name option in command line utilities to set verification
4391 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4392 Add CMAC pkey methods.
4396 * Experimental renegotiation in s_server -www mode. If the client
4397 browses /reneg connection is renegotiated. If /renegcert it is
4398 renegotiated requesting a certificate.
4402 * Add an "external" session cache for debugging purposes to s_server. This
4403 should help trace issues which normally are only apparent in deployed
4404 multi-process servers.
4408 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4409 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4410 BIO_set_cipher() and some obscure PEM functions were changed so they
4411 can now return an error. The RAND changes required a change to the
4412 RAND_METHOD structure.
4416 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4417 a gcc attribute to warn if the result of a function is ignored. This
4418 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4419 whose return value is often ignored.
4423 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4424 These allow SCTs (signed certificate timestamps) to be requested and
4425 validated when establishing a connection.
4427 *Rob Percival <robpercival@google.com>*
4432 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4434 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4435 used even when parsing explicit parameters, when loading a encoded key
4436 or calling `EC_GROUP_new_from_ecpkparameters()`/
4437 `EC_GROUP_new_from_ecparameters()`.
4438 This prevents bypass of security hardening and performance gains,
4439 especially for curves with specialized EC_METHODs.
4440 By default, if a key encoded with explicit parameters is loaded and later
4441 encoded, the output is still encoded with explicit parameters, even if
4442 internally a "named" EC_GROUP is used for computation.
4446 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4447 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4448 NULL. After this change, only the cofactor parameter can be NULL. It also
4449 does some minimal sanity checks on the passed order.
4454 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4455 An attack is simple, if the first CMS_recipientInfo is valid but the
4456 second CMS_recipientInfo is chosen ciphertext. If the second
4457 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4458 encryption key will be replaced by garbage, and the message cannot be
4459 decoded, but if the RSA decryption fails, the correct encryption key is
4460 used and the recipient will not notice the attack.
4461 As a work around for this potential attack the length of the decrypted
4462 key must be equal to the cipher default key length, in case the
4463 certifiate is not given and all recipientInfo are tried out.
4464 The old behaviour can be re-enabled in the CMS code by setting the
4465 CMS_DEBUG_DECRYPT flag.
4470 * Document issue with installation paths in diverse Windows builds
4472 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4473 binaries and run-time config file.
4478 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4480 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4481 This changes the size when using the `genpkey` command when no size is given.
4482 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4483 generation commands to use 2048 bits by default.
4487 * Add FIPS support for Android Arm 64-bit
4489 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4490 Module in Version 2.0.10. For some reason, the corresponding target
4491 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4492 built with FIPS support on Android Arm 64-bit. This omission has been
4495 *Matthias St. Pierre*
4497 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4499 * 0-byte record padding oracle
4501 If an application encounters a fatal protocol error and then calls
4502 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4503 then OpenSSL can respond differently to the calling application if a 0 byte
4504 record is received with invalid padding compared to if a 0 byte record is
4505 received with an invalid MAC. If the application then behaves differently
4506 based on that in a way that is detectable to the remote peer, then this
4507 amounts to a padding oracle that could be used to decrypt data.
4509 In order for this to be exploitable "non-stitched" ciphersuites must be in
4510 use. Stitched ciphersuites are optimised implementations of certain
4511 commonly used ciphersuites. Also the application must call SSL_shutdown()
4512 twice even if a protocol error has occurred (applications should not do
4513 this but some do anyway).
4515 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4516 Aviram, with additional investigation by Steven Collison and Andrew
4517 Hourselt. It was reported to OpenSSL on 10th December 2018.
4522 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4526 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4528 * Microarchitecture timing vulnerability in ECC scalar multiplication
4530 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4531 shown to be vulnerable to a microarchitecture timing side channel attack.
4532 An attacker with sufficient access to mount local timing attacks during
4533 ECDSA signature generation could recover the private key.
4535 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4536 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4542 * Timing vulnerability in DSA signature generation
4544 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4545 timing side channel attack. An attacker could use variations in the signing
4546 algorithm to recover the private key.
4548 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4553 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4554 Module, accidentally introduced while backporting security fixes from the
4555 development branch and hindering the use of ECC in FIPS mode.
4559 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4561 * Client DoS due to large DH parameter
4563 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4564 malicious server can send a very large prime value to the client. This will
4565 cause the client to spend an unreasonably long period of time generating a
4566 key for this prime resulting in a hang until the client has finished. This
4567 could be exploited in a Denial Of Service attack.
4569 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4574 * Cache timing vulnerability in RSA Key Generation
4576 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4577 a cache timing side channel attack. An attacker with sufficient access to
4578 mount cache timing attacks during the RSA key generation process could
4579 recover the private key.
4581 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4582 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4587 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4588 parameter is no longer accepted, as it leads to a corrupt table. NULL
4589 pem_str is reserved for alias entries only.
4593 * Revert blinding in ECDSA sign and instead make problematic addition
4594 length-invariant. Switch even to fixed-length Montgomery multiplication.
4598 * Change generating and checking of primes so that the error rate of not
4599 being prime depends on the intended use based on the size of the input.
4600 For larger primes this will result in more rounds of Miller-Rabin.
4601 The maximal error rate for primes with more than 1080 bits is lowered
4604 *Kurt Roeckx, Annie Yousar*
4606 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4610 * Add blinding to ECDSA and DSA signatures to protect against side channel
4611 attacks discovered by Keegan Ryan (NCC Group).
4615 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4616 now allow empty (zero character) pass phrases.
4620 * Certificate time validation (X509_cmp_time) enforces stricter
4621 compliance with RFC 5280. Fractional seconds and timezone offsets
4622 are no longer allowed.
4626 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4628 * Constructed ASN.1 types with a recursive definition could exceed the stack
4630 Constructed ASN.1 types with a recursive definition (such as can be found
4631 in PKCS7) could eventually exceed the stack given malicious input with
4632 excessive recursion. This could result in a Denial Of Service attack. There
4633 are no such structures used within SSL/TLS that come from untrusted sources
4634 so this is considered safe.
4636 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4642 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4644 * Read/write after SSL object in error state
4646 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4647 mechanism. The intent was that if a fatal error occurred during a handshake
4648 then OpenSSL would move into the error state and would immediately fail if
4649 you attempted to continue the handshake. This works as designed for the
4650 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4651 SSL_connect()), however due to a bug it does not work correctly if
4652 SSL_read() or SSL_write() is called directly. In that scenario, if the
4653 handshake fails then a fatal error will be returned in the initial function
4654 call. If SSL_read()/SSL_write() is subsequently called by the application
4655 for the same SSL object then it will succeed and the data is passed without
4656 being decrypted/encrypted directly from the SSL/TLS record layer.
4658 In order to exploit this issue an application bug would have to be present
4659 that resulted in a call to SSL_read()/SSL_write() being issued after having
4660 already received a fatal error.
4662 This issue was reported to OpenSSL by David Benjamin (Google).
4667 * rsaz_1024_mul_avx2 overflow bug on x86_64
4669 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4670 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4671 Analysis suggests that attacks against RSA and DSA as a result of this
4672 defect would be very difficult to perform and are not believed likely.
4673 Attacks against DH1024 are considered just feasible, because most of the
4674 work necessary to deduce information about a private key may be performed
4675 offline. The amount of resources required for such an attack would be
4676 significant. However, for an attack on TLS to be meaningful, the server
4677 would have to share the DH1024 private key among multiple clients, which is
4678 no longer an option since CVE-2016-0701.
4680 This only affects processors that support the AVX2 but not ADX extensions
4681 like Intel Haswell (4th generation).
4683 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4684 was originally found via the OSS-Fuzz project.
4689 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4691 * bn_sqrx8x_internal carry bug on x86_64
4693 There is a carry propagating bug in the x86_64 Montgomery squaring
4694 procedure. No EC algorithms are affected. Analysis suggests that attacks
4695 against RSA and DSA as a result of this defect would be very difficult to
4696 perform and are not believed likely. Attacks against DH are considered just
4697 feasible (although very difficult) because most of the work necessary to
4698 deduce information about a private key may be performed offline. The amount
4699 of resources required for such an attack would be very significant and
4700 likely only accessible to a limited number of attackers. An attacker would
4701 additionally need online access to an unpatched system using the target
4702 private key in a scenario with persistent DH parameters and a private
4703 key that is shared between multiple clients.
4705 This only affects processors that support the BMI1, BMI2 and ADX extensions
4706 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4708 This issue was reported to OpenSSL by the OSS-Fuzz project.
4713 * Malformed X.509 IPAddressFamily could cause OOB read
4715 If an X.509 certificate has a malformed IPAddressFamily extension,
4716 OpenSSL could do a one-byte buffer overread. The most likely result
4717 would be an erroneous display of the certificate in text format.
4719 This issue was reported to OpenSSL by the OSS-Fuzz project.
4723 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4725 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4726 platform rather than 'mingw'.
4730 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4732 * Truncated packet could crash via OOB read
4734 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4735 cipher is being used, then a truncated packet can cause that host to
4736 perform an out-of-bounds read, usually resulting in a crash.
4738 This issue was reported to OpenSSL by Robert Święcki of Google.
4743 * BN_mod_exp may produce incorrect results on x86_64
4745 There is a carry propagating bug in the x86_64 Montgomery squaring
4746 procedure. No EC algorithms are affected. Analysis suggests that attacks
4747 against RSA and DSA as a result of this defect would be very difficult to
4748 perform and are not believed likely. Attacks against DH are considered just
4749 feasible (although very difficult) because most of the work necessary to
4750 deduce information about a private key may be performed offline. The amount
4751 of resources required for such an attack would be very significant and
4752 likely only accessible to a limited number of attackers. An attacker would
4753 additionally need online access to an unpatched system using the target
4754 private key in a scenario with persistent DH parameters and a private
4755 key that is shared between multiple clients. For example this can occur by
4756 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4757 similar to CVE-2015-3193 but must be treated as a separate problem.
4759 This issue was reported to OpenSSL by the OSS-Fuzz project.
4764 * Montgomery multiplication may produce incorrect results
4766 There is a carry propagating bug in the Broadwell-specific Montgomery
4767 multiplication procedure that handles input lengths divisible by, but
4768 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4769 and DH private keys are impossible. This is because the subroutine in
4770 question is not used in operations with the private key itself and an input
4771 of the attacker's direct choice. Otherwise the bug can manifest itself as
4772 transient authentication and key negotiation failures or reproducible
4773 erroneous outcome of public-key operations with specially crafted input.
4774 Among EC algorithms only Brainpool P-512 curves are affected and one
4775 presumably can attack ECDH key negotiation. Impact was not analyzed in
4776 detail, because pre-requisites for attack are considered unlikely. Namely
4777 multiple clients have to choose the curve in question and the server has to
4778 share the private key among them, neither of which is default behaviour.
4779 Even then only clients that chose the curve will be affected.
4781 This issue was publicly reported as transient failures and was not
4782 initially recognized as a security issue. Thanks to Richard Morgan for
4783 providing reproducible case.
4788 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4789 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4790 prevent issues where no progress is being made and the peer continually
4791 sends unrecognised record types, using up resources processing them.
4795 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4797 * Missing CRL sanity check
4799 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4800 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4801 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4803 This issue only affects the OpenSSL 1.0.2i
4808 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4810 * OCSP Status Request extension unbounded memory growth
4812 A malicious client can send an excessively large OCSP Status Request
4813 extension. If that client continually requests renegotiation, sending a
4814 large OCSP Status Request extension each time, then there will be unbounded
4815 memory growth on the server. This will eventually lead to a Denial Of
4816 Service attack through memory exhaustion. Servers with a default
4817 configuration are vulnerable even if they do not support OCSP. Builds using
4818 the "no-ocsp" build time option are not affected.
4820 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4825 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4828 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4834 * OOB write in MDC2_Update()
4836 An overflow can occur in MDC2_Update() either if called directly or
4837 through the EVP_DigestUpdate() function using MDC2. If an attacker
4838 is able to supply very large amounts of input data after a previous
4839 call to EVP_EncryptUpdate() with a partial block then a length check
4840 can overflow resulting in a heap corruption.
4842 The amount of data needed is comparable to SIZE_MAX which is impractical
4845 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4850 * Malformed SHA512 ticket DoS
4852 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4853 DoS attack where a malformed ticket will result in an OOB read which will
4856 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4857 a custom server callback and ticket lookup mechanism.
4859 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4864 * OOB write in BN_bn2dec()
4866 The function BN_bn2dec() does not check the return value of BN_div_word().
4867 This can cause an OOB write if an application uses this function with an
4868 overly large BIGNUM. This could be a problem if an overly large certificate
4869 or CRL is printed out from an untrusted source. TLS is not affected because
4870 record limits will reject an oversized certificate before it is parsed.
4872 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4877 * OOB read in TS_OBJ_print_bio()
4879 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4880 the total length the OID text representation would use and not the amount
4881 of data written. This will result in OOB reads when large OIDs are
4884 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4889 * Pointer arithmetic undefined behaviour
4891 Avoid some undefined pointer arithmetic
4893 A common idiom in the codebase is to check limits in the following manner:
4896 Where "p" points to some malloc'd data of SIZE bytes and
4899 "len" here could be from some externally supplied data (e.g. from a TLS
4902 The rules of C pointer arithmetic are such that "p + len" is only well
4903 defined where len <= SIZE. Therefore the above idiom is actually
4904 undefined behaviour.
4906 For example this could cause problems if some malloc implementation
4907 provides an address for "p" such that "p + len" actually overflows for
4908 values of len that are too big and therefore p + len < limit.
4910 This issue was reported to OpenSSL by Guido Vranken
4915 * Constant time flag not preserved in DSA signing
4917 Operations in the DSA signing algorithm should run in constant time in
4918 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4919 implementation means that a non-constant time codepath is followed for
4920 certain operations. This has been demonstrated through a cache-timing
4921 attack to be sufficient for an attacker to recover the private DSA key.
4923 This issue was reported by César Pereida (Aalto University), Billy Brumley
4924 (Tampere University of Technology), and Yuval Yarom (The University of
4925 Adelaide and NICTA).
4930 * DTLS buffered message DoS
4932 In a DTLS connection where handshake messages are delivered out-of-order
4933 those messages that OpenSSL is not yet ready to process will be buffered
4934 for later use. Under certain circumstances, a flaw in the logic means that
4935 those messages do not get removed from the buffer even though the handshake
4936 has been completed. An attacker could force up to approx. 15 messages to
4937 remain in the buffer when they are no longer required. These messages will
4938 be cleared when the DTLS connection is closed. The default maximum size for
4939 a message is 100k. Therefore the attacker could force an additional 1500k
4940 to be consumed per connection. By opening many simulataneous connections an
4941 attacker could cause a DoS attack through memory exhaustion.
4943 This issue was reported to OpenSSL by Quan Luo.
4948 * DTLS replay protection DoS
4950 A flaw in the DTLS replay attack protection mechanism means that records
4951 that arrive for future epochs update the replay protection "window" before
4952 the MAC for the record has been validated. This could be exploited by an
4953 attacker by sending a record for the next epoch (which does not have to
4954 decrypt or have a valid MAC), with a very large sequence number. This means
4955 that all subsequent legitimate packets are dropped causing a denial of
4956 service for a specific DTLS connection.
4958 This issue was reported to OpenSSL by the OCAP audit team.
4963 * Certificate message OOB reads
4965 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4966 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4967 theoretical DoS risk but this has not been observed in practice on common
4970 The messages affected are client certificate, client certificate request
4971 and server certificate. As a result the attack can only be performed
4972 against a client or a server which enables client authentication.
4974 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4979 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4981 * Prevent padding oracle in AES-NI CBC MAC check
4983 A MITM attacker can use a padding oracle attack to decrypt traffic
4984 when the connection uses an AES CBC cipher and the server support
4987 This issue was introduced as part of the fix for Lucky 13 padding
4988 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4989 constant time by making sure that always the same bytes are read and
4990 compared against either the MAC or padding bytes. But it no longer
4991 checked that there was enough data to have both the MAC and padding
4994 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4998 * Fix EVP_EncodeUpdate overflow
5000 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5001 Base64 encoding of binary data. If an attacker is able to supply very large
5002 amounts of input data then a length check can overflow resulting in a heap
5005 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5006 the `PEM_write_bio*` family of functions. These are mainly used within the
5007 OpenSSL command line applications, so any application which processes data
5008 from an untrusted source and outputs it as a PEM file should be considered
5009 vulnerable to this issue. User applications that call these APIs directly
5010 with large amounts of untrusted data may also be vulnerable.
5012 This issue was reported by Guido Vranken.
5017 * Fix EVP_EncryptUpdate overflow
5019 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5020 is able to supply very large amounts of input data after a previous call to
5021 EVP_EncryptUpdate() with a partial block then a length check can overflow
5022 resulting in a heap corruption. Following an analysis of all OpenSSL
5023 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5024 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5025 the first called function after an EVP_EncryptInit(), and therefore that
5026 specific call must be safe. The second form is where the length passed to
5027 EVP_EncryptUpdate() can be seen from the code to be some small value and
5028 therefore there is no possibility of an overflow. Since all instances are
5029 one of these two forms, it is believed that there can be no overflows in
5030 internal code due to this problem. It should be noted that
5031 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5032 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5033 of these calls have also been analysed too and it is believed there are no
5034 instances in internal usage where an overflow could occur.
5036 This issue was reported by Guido Vranken.
5041 * Prevent ASN.1 BIO excessive memory allocation
5043 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5044 a short invalid encoding can cause allocation of large amounts of memory
5045 potentially consuming excessive resources or exhausting memory.
5047 Any application parsing untrusted data through d2i BIO functions is
5048 affected. The memory based functions such as d2i_X509() are *not* affected.
5049 Since the memory based functions are used by the TLS library, TLS
5050 applications are not affected.
5052 This issue was reported by Brian Carpenter.
5059 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5060 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5061 in arbitrary stack data being returned in the buffer.
5063 This issue was reported by Guido Vranken.
5068 * Modify behavior of ALPN to invoke callback after SNI/servername
5069 callback, such that updates to the SSL_CTX affect ALPN.
5073 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5078 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5079 methods are enabled and ssl2 is disabled the methods return NULL.
5083 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5085 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5086 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5087 provide any "EXPORT" or "LOW" strength ciphers.
5091 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5092 is by default disabled at build-time. Builds that are not configured with
5093 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5094 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5095 will need to explicitly call either of:
5097 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5099 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5101 as appropriate. Even if either of those is used, or the application
5102 explicitly uses the version-specific SSLv2_method() or its client and
5103 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5104 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5105 ciphers, and SSLv2 56-bit DES are no longer available.
5110 * Fix a double-free in DSA code
5112 A double free bug was discovered when OpenSSL parses malformed DSA private
5113 keys and could lead to a DoS attack or memory corruption for applications
5114 that receive DSA private keys from untrusted sources. This scenario is
5117 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5123 * Disable SRP fake user seed to address a server memory leak.
5125 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5127 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5128 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5129 was changed to ignore the "fake user" SRP seed, even if the seed
5132 Users should use SRP_VBASE_get1_by_user instead. Note that in
5133 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5134 also that even though configuring the SRP seed attempts to hide
5135 invalid usernames by continuing the handshake with fake
5136 credentials, this behaviour is not constant time and no strong
5137 guarantees are made that the handshake is indistinguishable from
5138 that of a valid user.
5143 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5145 In the BN_hex2bn function the number of hex digits is calculated using an
5146 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5147 large values of `i` this can result in `bn_expand` not allocating any
5148 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5149 field as NULL leading to a subsequent NULL ptr deref. For very large values
5150 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5151 In this case memory is allocated to the internal BIGNUM data field, but it
5152 is insufficiently sized leading to heap corruption. A similar issue exists
5153 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5154 is ever called by user applications with very large untrusted hex/dec data.
5155 This is anticipated to be a rare occurrence.
5157 All OpenSSL internal usage of these functions use data that is not expected
5158 to be untrusted, e.g. config file data or application command line
5159 arguments. If user developed applications generate config file data based
5160 on untrusted data then it is possible that this could also lead to security
5161 consequences. This is also anticipated to be rare.
5163 This issue was reported to OpenSSL by Guido Vranken.
5168 * Fix memory issues in `BIO_*printf` functions
5170 The internal `fmtstr` function used in processing a "%s" format string in
5171 the `BIO_*printf` functions could overflow while calculating the length of a
5172 string and cause an OOB read when printing very long strings.
5174 Additionally the internal `doapr_outch` function can attempt to write to an
5175 OOB memory location (at an offset from the NULL pointer) in the event of a
5176 memory allocation failure. In 1.0.2 and below this could be caused where
5177 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5178 could be in processing a very long "%s" format string. Memory leaks can
5181 The first issue may mask the second issue dependent on compiler behaviour.
5182 These problems could enable attacks where large amounts of untrusted data
5183 is passed to the `BIO_*printf` functions. If applications use these functions
5184 in this way then they could be vulnerable. OpenSSL itself uses these
5185 functions when printing out human-readable dumps of ASN.1 data. Therefore
5186 applications that print this data could be vulnerable if the data is from
5187 untrusted sources. OpenSSL command line applications could also be
5188 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5189 as command line arguments.
5191 Libssl is not considered directly vulnerable. Additionally certificates etc
5192 received via remote connections via libssl are also unlikely to be able to
5193 trigger these issues because of message size limits enforced within libssl.
5195 This issue was reported to OpenSSL Guido Vranken.
5200 * Side channel attack on modular exponentiation
5202 A side-channel attack was found which makes use of cache-bank conflicts on
5203 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5204 of RSA keys. The ability to exploit this issue is limited as it relies on
5205 an attacker who has control of code in a thread running on the same
5206 hyper-threaded core as the victim thread which is performing decryptions.
5208 This issue was reported to OpenSSL by Yuval Yarom, The University of
5209 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5210 Nadia Heninger, University of Pennsylvania with more information at
5211 <http://cachebleed.info>.
5216 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5217 if no keysize is specified with default_bits. This fixes an
5218 omission in an earlier change that changed all RSA/DSA key generation
5219 commands to use 2048 bits by default.
5223 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5225 * DH small subgroups
5227 Historically OpenSSL only ever generated DH parameters based on "safe"
5228 primes. More recently (in version 1.0.2) support was provided for
5229 generating X9.42 style parameter files such as those required for RFC 5114
5230 support. The primes used in such files may not be "safe". Where an
5231 application is using DH configured with parameters based on primes that are
5232 not "safe" then an attacker could use this fact to find a peer's private
5233 DH exponent. This attack requires that the attacker complete multiple
5234 handshakes in which the peer uses the same private DH exponent. For example
5235 this could be used to discover a TLS server's private DH exponent if it's
5236 reusing the private DH exponent or it's using a static DH ciphersuite.
5238 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5239 TLS. It is not on by default. If the option is not set then the server
5240 reuses the same private DH exponent for the life of the server process and
5241 would be vulnerable to this attack. It is believed that many popular
5242 applications do set this option and would therefore not be at risk.
5244 The fix for this issue adds an additional check where a "q" parameter is
5245 available (as is the case in X9.42 based parameters). This detects the
5246 only known attack, and is the only possible defense for static DH
5247 ciphersuites. This could have some performance impact.
5249 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5250 default and cannot be disabled. This could have some performance impact.
5252 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5257 * SSLv2 doesn't block disabled ciphers
5259 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5260 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5261 been disabled, provided that the SSLv2 protocol was not also disabled via
5264 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5265 and Sebastian Schinzel.
5270 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5272 * BN_mod_exp may produce incorrect results on x86_64
5274 There is a carry propagating bug in the x86_64 Montgomery squaring
5275 procedure. No EC algorithms are affected. Analysis suggests that attacks
5276 against RSA and DSA as a result of this defect would be very difficult to
5277 perform and are not believed likely. Attacks against DH are considered just
5278 feasible (although very difficult) because most of the work necessary to
5279 deduce information about a private key may be performed offline. The amount
5280 of resources required for such an attack would be very significant and
5281 likely only accessible to a limited number of attackers. An attacker would
5282 additionally need online access to an unpatched system using the target
5283 private key in a scenario with persistent DH parameters and a private
5284 key that is shared between multiple clients. For example this can occur by
5285 default in OpenSSL DHE based SSL/TLS ciphersuites.
5287 This issue was reported to OpenSSL by Hanno Böck.
5292 * Certificate verify crash with missing PSS parameter
5294 The signature verification routines will crash with a NULL pointer
5295 dereference if presented with an ASN.1 signature using the RSA PSS
5296 algorithm and absent mask generation function parameter. Since these
5297 routines are used to verify certificate signature algorithms this can be
5298 used to crash any certificate verification operation and exploited in a
5299 DoS attack. Any application which performs certificate verification is
5300 vulnerable including OpenSSL clients and servers which enable client
5303 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5308 * X509_ATTRIBUTE memory leak
5310 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5311 memory. This structure is used by the PKCS#7 and CMS routines so any
5312 application which reads PKCS#7 or CMS data from untrusted sources is
5313 affected. SSL/TLS is not affected.
5315 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5321 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5322 This changes the decoding behaviour for some invalid messages,
5323 though the change is mostly in the more lenient direction, and
5324 legacy behaviour is preserved as much as possible.
5328 * In DSA_generate_parameters_ex, if the provided seed is too short,
5331 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5333 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5335 * Alternate chains certificate forgery
5337 During certificate verification, OpenSSL will attempt to find an
5338 alternative certificate chain if the first attempt to build such a chain
5339 fails. An error in the implementation of this logic can mean that an
5340 attacker could cause certain checks on untrusted certificates to be
5341 bypassed, such as the CA flag, enabling them to use a valid leaf
5342 certificate to act as a CA and "issue" an invalid certificate.
5344 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5349 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5351 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5352 incompatibility in the handling of HMAC. The previous ABI has now been
5357 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5359 * Malformed ECParameters causes infinite loop
5361 When processing an ECParameters structure OpenSSL enters an infinite loop
5362 if the curve specified is over a specially malformed binary polynomial
5365 This can be used to perform denial of service against any
5366 system which processes public keys, certificate requests or
5367 certificates. This includes TLS clients and TLS servers with
5368 client authentication enabled.
5370 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5375 * Exploitable out-of-bounds read in X509_cmp_time
5377 X509_cmp_time does not properly check the length of the ASN1_TIME
5378 string and can read a few bytes out of bounds. In addition,
5379 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5382 An attacker can use this to craft malformed certificates and CRLs of
5383 various sizes and potentially cause a segmentation fault, resulting in
5384 a DoS on applications that verify certificates or CRLs. TLS clients
5385 that verify CRLs are affected. TLS clients and servers with client
5386 authentication enabled may be affected if they use custom verification
5389 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5390 independently by Hanno Böck.
5395 * PKCS7 crash with missing EnvelopedContent
5397 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5398 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5399 with missing content and trigger a NULL pointer dereference on parsing.
5401 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5402 structures from untrusted sources are affected. OpenSSL clients and
5403 servers are not affected.
5405 This issue was reported to OpenSSL by Michal Zalewski (Google).
5410 * CMS verify infinite loop with unknown hash function
5412 When verifying a signedData message the CMS code can enter an infinite loop
5413 if presented with an unknown hash function OID. This can be used to perform
5414 denial of service against any system which verifies signedData messages using
5416 This issue was reported to OpenSSL by Johannes Bauer.
5421 * Race condition handling NewSessionTicket
5423 If a NewSessionTicket is received by a multi-threaded client when attempting to
5424 reuse a previous ticket then a race condition can occur potentially leading to
5425 a double free of the ticket data.
5430 * Only support 256-bit or stronger elliptic curves with the
5431 'ecdh_auto' setting (server) or by default (client). Of supported
5432 curves, prefer P-256 (both).
5436 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5438 * ClientHello sigalgs DoS fix
5440 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5441 invalid signature algorithms extension a NULL pointer dereference will
5442 occur. This can be exploited in a DoS attack against the server.
5444 This issue was was reported to OpenSSL by David Ramos of Stanford
5448 *Stephen Henson and Matt Caswell*
5450 * Multiblock corrupted pointer fix
5452 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5453 feature only applies on 64 bit x86 architecture platforms that support AES
5454 NI instructions. A defect in the implementation of "multiblock" can cause
5455 OpenSSL's internal write buffer to become incorrectly set to NULL when
5456 using non-blocking IO. Typically, when the user application is using a
5457 socket BIO for writing, this will only result in a failed connection.
5458 However if some other BIO is used then it is likely that a segmentation
5459 fault will be triggered, thus enabling a potential DoS attack.
5461 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5466 * Segmentation fault in DTLSv1_listen fix
5468 The DTLSv1_listen function is intended to be stateless and processes the
5469 initial ClientHello from many peers. It is common for user code to loop
5470 over the call to DTLSv1_listen until a valid ClientHello is received with
5471 an associated cookie. A defect in the implementation of DTLSv1_listen means
5472 that state is preserved in the SSL object from one invocation to the next
5473 that can lead to a segmentation fault. Errors processing the initial
5474 ClientHello can trigger this scenario. An example of such an error could be
5475 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5478 This issue was reported to OpenSSL by Per Allansson.
5483 * Segmentation fault in ASN1_TYPE_cmp fix
5485 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5486 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5487 certificate signature algorithm consistency this can be used to crash any
5488 certificate verification operation and exploited in a DoS attack. Any
5489 application which performs certificate verification is vulnerable including
5490 OpenSSL clients and servers which enable client authentication.
5495 * Segmentation fault for invalid PSS parameters fix
5497 The signature verification routines will crash with a NULL pointer
5498 dereference if presented with an ASN.1 signature using the RSA PSS
5499 algorithm and invalid parameters. Since these routines are used to verify
5500 certificate signature algorithms this can be used to crash any
5501 certificate verification operation and exploited in a DoS attack. Any
5502 application which performs certificate verification is vulnerable including
5503 OpenSSL clients and servers which enable client authentication.
5505 This issue was was reported to OpenSSL by Brian Carpenter.
5510 * ASN.1 structure reuse memory corruption fix
5512 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5513 memory corruption via an invalid write. Such reuse is and has been
5514 strongly discouraged and is believed to be rare.
5516 Applications that parse structures containing CHOICE or ANY DEFINED BY
5517 components may be affected. Certificate parsing (d2i_X509 and related
5518 functions) are however not affected. OpenSSL clients and servers are
5524 * PKCS7 NULL pointer dereferences fix
5526 The PKCS#7 parsing code does not handle missing outer ContentInfo
5527 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5528 missing content and trigger a NULL pointer dereference on parsing.
5530 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5531 otherwise parse PKCS#7 structures from untrusted sources are
5532 affected. OpenSSL clients and servers are not affected.
5534 This issue was reported to OpenSSL by Michal Zalewski (Google).
5539 * DoS via reachable assert in SSLv2 servers fix
5541 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5542 servers that both support SSLv2 and enable export cipher suites by sending
5543 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5545 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5546 (OpenSSL development team).
5551 * Empty CKE with client auth and DHE fix
5553 If client auth is used then a server can seg fault in the event of a DHE
5554 ciphersuite being selected and a zero length ClientKeyExchange message
5555 being sent by the client. This could be exploited in a DoS attack.
5560 * Handshake with unseeded PRNG fix
5562 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5563 with an unseeded PRNG. The conditions are:
5564 - The client is on a platform where the PRNG has not been seeded
5565 automatically, and the user has not seeded manually
5566 - A protocol specific client method version has been used (i.e. not
5567 SSL_client_methodv23)
5568 - A ciphersuite is used that does not require additional random data from
5569 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5571 If the handshake succeeds then the client random that has been used will
5572 have been generated from a PRNG with insufficient entropy and therefore the
5573 output may be predictable.
5575 For example using the following command with an unseeded openssl will
5576 succeed on an unpatched platform:
5578 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5583 * Use After Free following d2i_ECPrivatekey error fix
5585 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5586 could cause a use after free condition. This, in turn, could cause a double
5587 free in several private key parsing functions (such as d2i_PrivateKey
5588 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5589 for applications that receive EC private keys from untrusted
5590 sources. This scenario is considered rare.
5592 This issue was discovered by the BoringSSL project and fixed in their
5598 * X509_to_X509_REQ NULL pointer deref fix
5600 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5601 the certificate key is invalid. This function is rarely used in practice.
5603 This issue was discovered by Brian Carpenter.
5608 * Removed the export ciphers from the DEFAULT ciphers
5612 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5614 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5615 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5616 So far those who have to target multiple platforms would compromise
5617 and argue that binary targeting say ARMv5 would still execute on
5618 ARMv8. "Universal" build resolves this compromise by providing
5619 near-optimal performance even on newer platforms.
5623 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5624 (other platforms pending).
5626 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5628 * Add support for the SignedCertificateTimestampList certificate and
5629 OCSP response extensions from RFC6962.
5633 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5634 for corner cases. (Certain input points at infinity could lead to
5635 bogus results, with non-infinity inputs mapped to infinity too.)
5639 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5640 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5641 common cases are optimized and there still is room for further
5642 improvements. Vector Permutation AES for Altivec is also added.
5646 * Add support for little-endian ppc64 Linux target.
5648 *Marcelo Cerri (IBM)*
5650 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5651 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5652 are optimized and there still is room for further improvements.
5653 Both 32- and 64-bit modes are supported.
5655 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5657 * Improved ARMv7 NEON support.
5661 * Support for SPARC Architecture 2011 crypto extensions, first
5662 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5663 SHA256/512, MD5, GHASH and modular exponentiation.
5665 *Andy Polyakov, David Miller*
5667 * Accelerated modular exponentiation for Intel processors, a.k.a.
5670 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5672 * Support for new and upcoming Intel processors, including AVX2,
5673 BMI and SHA ISA extensions. This includes additional "stitched"
5674 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5677 This work was sponsored by Intel Corp.
5681 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5682 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5683 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5687 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5688 this fixes a limitation in previous versions of OpenSSL.
5692 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5693 MGF1 digest and OAEP label.
5697 * Add EVP support for key wrapping algorithms, to avoid problems with
5698 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5699 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5700 algorithms and include tests cases.
5704 * Add functions to allocate and set the fields of an ECDSA_METHOD
5707 *Douglas E. Engert, Steve Henson*
5709 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5710 difference in days and seconds between two tm or ASN1_TIME structures.
5714 * Add -rev test option to s_server to just reverse order of characters
5715 received by client and send back to server. Also prints an abbreviated
5716 summary of the connection parameters.
5720 * New option -brief for s_client and s_server to print out a brief summary
5721 of connection parameters.
5725 * Add callbacks for arbitrary TLS extensions.
5727 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5729 * New option -crl_download in several openssl utilities to download CRLs
5730 from CRLDP extension in certificates.
5734 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5738 * New function X509_CRL_diff to generate a delta CRL from the difference
5739 of two full CRLs. Add support to "crl" utility.
5743 * New functions to set lookup_crls function and to retrieve
5744 X509_STORE from X509_STORE_CTX.
5748 * Print out deprecated issuer and subject unique ID fields in
5753 * Extend OCSP I/O functions so they can be used for simple general purpose
5754 HTTP as well as OCSP. New wrapper function which can be used to download
5755 CRLs using the OCSP API.
5759 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5763 * `SSL_CONF*` functions. These provide a common framework for application
5764 configuration using configuration files or command lines.
5768 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5769 message callback and prints the results. Needs compile time option
5770 "enable-ssl-trace". New options to s_client and s_server to enable
5775 * New ctrl and macro to retrieve supported points extensions.
5776 Print out extension in s_server and s_client.
5780 * New functions to retrieve certificate signature and signature
5785 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5790 * New Suite B modes for TLS code. These use and enforce the requirements
5791 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5792 only use Suite B curves. The Suite B modes can be set by using the
5793 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5797 * New chain verification flags for Suite B levels of security. Check
5798 algorithms are acceptable when flags are set in X509_verify_cert.
5802 * Make tls1_check_chain return a set of flags indicating checks passed
5803 by a certificate chain. Add additional tests to handle client
5804 certificates: checks for matching certificate type and issuer name
5809 * If an attempt is made to use a signature algorithm not in the peer
5810 preference list abort the handshake. If client has no suitable
5811 signature algorithms in response to a certificate request do not
5812 use the certificate.
5816 * If server EC tmp key is not in client preference list abort handshake.
5820 * Add support for certificate stores in CERT structure. This makes it
5821 possible to have different stores per SSL structure or one store in
5822 the parent SSL_CTX. Include distinct stores for certificate chain
5823 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5824 to build and store a certificate chain in CERT structure: returning
5825 an error if the chain cannot be built: this will allow applications
5826 to test if a chain is correctly configured.
5828 Note: if the CERT based stores are not set then the parent SSL_CTX
5829 store is used to retain compatibility with existing behaviour.
5833 * New function ssl_set_client_disabled to set a ciphersuite disabled
5834 mask based on the current session, check mask when sending client
5835 hello and checking the requested ciphersuite.
5839 * New ctrls to retrieve and set certificate types in a certificate
5840 request message. Print out received values in s_client. If certificate
5841 types is not set with custom values set sensible values based on
5842 supported signature algorithms.
5846 * Support for distinct client and server supported signature algorithms.
5850 * Add certificate callback. If set this is called whenever a certificate
5851 is required by client or server. An application can decide which
5852 certificate chain to present based on arbitrary criteria: for example
5853 supported signature algorithms. Add very simple example to s_server.
5854 This fixes many of the problems and restrictions of the existing client
5855 certificate callback: for example you can now clear an existing
5856 certificate and specify the whole chain.
5860 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5861 the certificate can be used for (if anything). Set valid_flags field
5862 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5863 to have similar checks in it.
5865 Add new "cert_flags" field to CERT structure and include a "strict mode".
5866 This enforces some TLS certificate requirements (such as only permitting
5867 certificate signature algorithms contained in the supported algorithms
5868 extension) which some implementations ignore: this option should be used
5869 with caution as it could cause interoperability issues.
5873 * Update and tidy signature algorithm extension processing. Work out
5874 shared signature algorithms based on preferences and peer algorithms
5875 and print them out in s_client and s_server. Abort handshake if no
5876 shared signature algorithms.
5880 * Add new functions to allow customised supported signature algorithms
5881 for SSL and SSL_CTX structures. Add options to s_client and s_server
5886 * New function SSL_certs_clear() to delete all references to certificates
5887 from an SSL structure. Before this once a certificate had been added
5888 it couldn't be removed.
5892 * Integrate hostname, email address and IP address checking with certificate
5893 verification. New verify options supporting checking in openssl utility.
5897 * Fixes and wildcard matching support to hostname and email checking
5898 functions. Add manual page.
5900 *Florian Weimer (Red Hat Product Security Team)*
5902 * New functions to check a hostname email or IP address against a
5903 certificate. Add options x509 utility to print results of checks against
5908 * Fix OCSP checking.
5910 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5912 * Initial experimental support for explicitly trusted non-root CAs.
5913 OpenSSL still tries to build a complete chain to a root but if an
5914 intermediate CA has a trust setting included that is used. The first
5915 setting is used: whether to trust (e.g., -addtrust option to the x509
5920 * Add -trusted_first option which attempts to find certificates in the
5921 trusted store even if an untrusted chain is also supplied.
5925 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5926 platform support for Linux and Android.
5930 * Support for linux-x32, ILP32 environment in x86_64 framework.
5934 * Experimental multi-implementation support for FIPS capable OpenSSL.
5935 When in FIPS mode the approved implementations are used as normal,
5936 when not in FIPS mode the internal unapproved versions are used instead.
5937 This means that the FIPS capable OpenSSL isn't forced to use the
5938 (often lower performance) FIPS implementations outside FIPS mode.
5942 * Transparently support X9.42 DH parameters when calling
5943 PEM_read_bio_DHparameters. This means existing applications can handle
5944 the new parameter format automatically.
5948 * Initial experimental support for X9.42 DH parameter format: mainly
5949 to support use of 'q' parameter for RFC5114 parameters.
5953 * Add DH parameters from RFC5114 including test data to dhtest.
5957 * Support for automatic EC temporary key parameter selection. If enabled
5958 the most preferred EC parameters are automatically used instead of
5959 hardcoded fixed parameters. Now a server just has to call:
5960 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5961 support ECDH and use the most appropriate parameters.
5965 * Enhance and tidy EC curve and point format TLS extension code. Use
5966 static structures instead of allocation if default values are used.
5967 New ctrls to set curves we wish to support and to retrieve shared curves.
5968 Print out shared curves in s_server. New options to s_server and s_client
5969 to set list of supported curves.
5973 * New ctrls to retrieve supported signature algorithms and
5974 supported curve values as an array of NIDs. Extend openssl utility
5975 to print out received values.
5979 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5980 between NIDs and the more common NIST names such as "P-256". Enhance
5981 ecparam utility and ECC method to recognise the NIST names for curves.
5985 * Enhance SSL/TLS certificate chain handling to support different
5986 chains for each certificate instead of one chain in the parent SSL_CTX.
5990 * Support for fixed DH ciphersuite client authentication: where both
5991 server and client use DH certificates with common parameters.
5995 * Support for fixed DH ciphersuites: those requiring DH server
6000 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6002 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6003 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6004 X509_CINF_get_signature were reverted post internal team review.
6009 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6011 * OCSP Status Request extension unbounded memory growth
6013 A malicious client can send an excessively large OCSP Status Request
6014 extension. If that client continually requests renegotiation, sending a
6015 large OCSP Status Request extension each time, then there will be unbounded
6016 memory growth on the server. This will eventually lead to a Denial Of
6017 Service attack through memory exhaustion. Servers with a default
6018 configuration are vulnerable even if they do not support OCSP. Builds using
6019 the "no-ocsp" build time option are not affected.
6021 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6026 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6029 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6035 * OOB write in MDC2_Update()
6037 An overflow can occur in MDC2_Update() either if called directly or
6038 through the EVP_DigestUpdate() function using MDC2. If an attacker
6039 is able to supply very large amounts of input data after a previous
6040 call to EVP_EncryptUpdate() with a partial block then a length check
6041 can overflow resulting in a heap corruption.
6043 The amount of data needed is comparable to SIZE_MAX which is impractical
6046 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6051 * Malformed SHA512 ticket DoS
6053 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6054 DoS attack where a malformed ticket will result in an OOB read which will
6057 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6058 a custom server callback and ticket lookup mechanism.
6060 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6065 * OOB write in BN_bn2dec()
6067 The function BN_bn2dec() does not check the return value of BN_div_word().
6068 This can cause an OOB write if an application uses this function with an
6069 overly large BIGNUM. This could be a problem if an overly large certificate
6070 or CRL is printed out from an untrusted source. TLS is not affected because
6071 record limits will reject an oversized certificate before it is parsed.
6073 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6078 * OOB read in TS_OBJ_print_bio()
6080 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6081 the total length the OID text representation would use and not the amount
6082 of data written. This will result in OOB reads when large OIDs are
6085 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6090 * Pointer arithmetic undefined behaviour
6092 Avoid some undefined pointer arithmetic
6094 A common idiom in the codebase is to check limits in the following manner:
6097 Where "p" points to some malloc'd data of SIZE bytes and
6100 "len" here could be from some externally supplied data (e.g. from a TLS
6103 The rules of C pointer arithmetic are such that "p + len" is only well
6104 defined where len <= SIZE. Therefore the above idiom is actually
6105 undefined behaviour.
6107 For example this could cause problems if some malloc implementation
6108 provides an address for "p" such that "p + len" actually overflows for
6109 values of len that are too big and therefore p + len < limit.
6111 This issue was reported to OpenSSL by Guido Vranken
6116 * Constant time flag not preserved in DSA signing
6118 Operations in the DSA signing algorithm should run in constant time in
6119 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6120 implementation means that a non-constant time codepath is followed for
6121 certain operations. This has been demonstrated through a cache-timing
6122 attack to be sufficient for an attacker to recover the private DSA key.
6124 This issue was reported by César Pereida (Aalto University), Billy Brumley
6125 (Tampere University of Technology), and Yuval Yarom (The University of
6126 Adelaide and NICTA).
6131 * DTLS buffered message DoS
6133 In a DTLS connection where handshake messages are delivered out-of-order
6134 those messages that OpenSSL is not yet ready to process will be buffered
6135 for later use. Under certain circumstances, a flaw in the logic means that
6136 those messages do not get removed from the buffer even though the handshake
6137 has been completed. An attacker could force up to approx. 15 messages to
6138 remain in the buffer when they are no longer required. These messages will
6139 be cleared when the DTLS connection is closed. The default maximum size for
6140 a message is 100k. Therefore the attacker could force an additional 1500k
6141 to be consumed per connection. By opening many simulataneous connections an
6142 attacker could cause a DoS attack through memory exhaustion.
6144 This issue was reported to OpenSSL by Quan Luo.
6149 * DTLS replay protection DoS
6151 A flaw in the DTLS replay attack protection mechanism means that records
6152 that arrive for future epochs update the replay protection "window" before
6153 the MAC for the record has been validated. This could be exploited by an
6154 attacker by sending a record for the next epoch (which does not have to
6155 decrypt or have a valid MAC), with a very large sequence number. This means
6156 that all subsequent legitimate packets are dropped causing a denial of
6157 service for a specific DTLS connection.
6159 This issue was reported to OpenSSL by the OCAP audit team.
6164 * Certificate message OOB reads
6166 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6167 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6168 theoretical DoS risk but this has not been observed in practice on common
6171 The messages affected are client certificate, client certificate request
6172 and server certificate. As a result the attack can only be performed
6173 against a client or a server which enables client authentication.
6175 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6180 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6182 * Prevent padding oracle in AES-NI CBC MAC check
6184 A MITM attacker can use a padding oracle attack to decrypt traffic
6185 when the connection uses an AES CBC cipher and the server support
6188 This issue was introduced as part of the fix for Lucky 13 padding
6189 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6190 constant time by making sure that always the same bytes are read and
6191 compared against either the MAC or padding bytes. But it no longer
6192 checked that there was enough data to have both the MAC and padding
6195 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6200 * Fix EVP_EncodeUpdate overflow
6202 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6203 Base64 encoding of binary data. If an attacker is able to supply very large
6204 amounts of input data then a length check can overflow resulting in a heap
6207 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6208 the `PEM_write_bio*` family of functions. These are mainly used within the
6209 OpenSSL command line applications, so any application which processes data
6210 from an untrusted source and outputs it as a PEM file should be considered
6211 vulnerable to this issue. User applications that call these APIs directly
6212 with large amounts of untrusted data may also be vulnerable.
6214 This issue was reported by Guido Vranken.
6219 * Fix EVP_EncryptUpdate overflow
6221 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6222 is able to supply very large amounts of input data after a previous call to
6223 EVP_EncryptUpdate() with a partial block then a length check can overflow
6224 resulting in a heap corruption. Following an analysis of all OpenSSL
6225 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6226 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6227 the first called function after an EVP_EncryptInit(), and therefore that
6228 specific call must be safe. The second form is where the length passed to
6229 EVP_EncryptUpdate() can be seen from the code to be some small value and
6230 therefore there is no possibility of an overflow. Since all instances are
6231 one of these two forms, it is believed that there can be no overflows in
6232 internal code due to this problem. It should be noted that
6233 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6234 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6235 of these calls have also been analysed too and it is believed there are no
6236 instances in internal usage where an overflow could occur.
6238 This issue was reported by Guido Vranken.
6243 * Prevent ASN.1 BIO excessive memory allocation
6245 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6246 a short invalid encoding can casuse allocation of large amounts of memory
6247 potentially consuming excessive resources or exhausting memory.
6249 Any application parsing untrusted data through d2i BIO functions is
6250 affected. The memory based functions such as d2i_X509() are *not* affected.
6251 Since the memory based functions are used by the TLS library, TLS
6252 applications are not affected.
6254 This issue was reported by Brian Carpenter.
6261 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6262 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6263 in arbitrary stack data being returned in the buffer.
6265 This issue was reported by Guido Vranken.
6270 * Modify behavior of ALPN to invoke callback after SNI/servername
6271 callback, such that updates to the SSL_CTX affect ALPN.
6275 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6280 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6281 methods are enabled and ssl2 is disabled the methods return NULL.
6285 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6287 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6288 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6289 provide any "EXPORT" or "LOW" strength ciphers.
6293 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6294 is by default disabled at build-time. Builds that are not configured with
6295 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6296 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6297 will need to explicitly call either of:
6299 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6301 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6303 as appropriate. Even if either of those is used, or the application
6304 explicitly uses the version-specific SSLv2_method() or its client and
6305 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6306 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6307 ciphers, and SSLv2 56-bit DES are no longer available.
6312 * Fix a double-free in DSA code
6314 A double free bug was discovered when OpenSSL parses malformed DSA private
6315 keys and could lead to a DoS attack or memory corruption for applications
6316 that receive DSA private keys from untrusted sources. This scenario is
6319 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6325 * Disable SRP fake user seed to address a server memory leak.
6327 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6329 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6330 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6331 was changed to ignore the "fake user" SRP seed, even if the seed
6334 Users should use SRP_VBASE_get1_by_user instead. Note that in
6335 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6336 also that even though configuring the SRP seed attempts to hide
6337 invalid usernames by continuing the handshake with fake
6338 credentials, this behaviour is not constant time and no strong
6339 guarantees are made that the handshake is indistinguishable from
6340 that of a valid user.
6345 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6347 In the BN_hex2bn function the number of hex digits is calculated using an
6348 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6349 large values of `i` this can result in `bn_expand` not allocating any
6350 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6351 field as NULL leading to a subsequent NULL ptr deref. For very large values
6352 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6353 In this case memory is allocated to the internal BIGNUM data field, but it
6354 is insufficiently sized leading to heap corruption. A similar issue exists
6355 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6356 is ever called by user applications with very large untrusted hex/dec data.
6357 This is anticipated to be a rare occurrence.
6359 All OpenSSL internal usage of these functions use data that is not expected
6360 to be untrusted, e.g. config file data or application command line
6361 arguments. If user developed applications generate config file data based
6362 on untrusted data then it is possible that this could also lead to security
6363 consequences. This is also anticipated to be rare.
6365 This issue was reported to OpenSSL by Guido Vranken.
6370 * Fix memory issues in `BIO_*printf` functions
6372 The internal `fmtstr` function used in processing a "%s" format string in
6373 the `BIO_*printf` functions could overflow while calculating the length of a
6374 string and cause an OOB read when printing very long strings.
6376 Additionally the internal `doapr_outch` function can attempt to write to an
6377 OOB memory location (at an offset from the NULL pointer) in the event of a
6378 memory allocation failure. In 1.0.2 and below this could be caused where
6379 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6380 could be in processing a very long "%s" format string. Memory leaks can
6383 The first issue may mask the second issue dependent on compiler behaviour.
6384 These problems could enable attacks where large amounts of untrusted data
6385 is passed to the `BIO_*printf` functions. If applications use these functions
6386 in this way then they could be vulnerable. OpenSSL itself uses these
6387 functions when printing out human-readable dumps of ASN.1 data. Therefore
6388 applications that print this data could be vulnerable if the data is from
6389 untrusted sources. OpenSSL command line applications could also be
6390 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6391 as command line arguments.
6393 Libssl is not considered directly vulnerable. Additionally certificates etc
6394 received via remote connections via libssl are also unlikely to be able to
6395 trigger these issues because of message size limits enforced within libssl.
6397 This issue was reported to OpenSSL Guido Vranken.
6402 * Side channel attack on modular exponentiation
6404 A side-channel attack was found which makes use of cache-bank conflicts on
6405 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6406 of RSA keys. The ability to exploit this issue is limited as it relies on
6407 an attacker who has control of code in a thread running on the same
6408 hyper-threaded core as the victim thread which is performing decryptions.
6410 This issue was reported to OpenSSL by Yuval Yarom, The University of
6411 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6412 Nadia Heninger, University of Pennsylvania with more information at
6413 <http://cachebleed.info>.
6418 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6419 if no keysize is specified with default_bits. This fixes an
6420 omission in an earlier change that changed all RSA/DSA key generation
6421 commands to use 2048 bits by default.
6425 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6427 * Protection for DH small subgroup attacks
6429 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6430 switched on by default and cannot be disabled. This could have some
6435 * SSLv2 doesn't block disabled ciphers
6437 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6438 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6439 been disabled, provided that the SSLv2 protocol was not also disabled via
6442 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6443 and Sebastian Schinzel.
6448 * Reject DH handshakes with parameters shorter than 1024 bits.
6452 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6454 * Certificate verify crash with missing PSS parameter
6456 The signature verification routines will crash with a NULL pointer
6457 dereference if presented with an ASN.1 signature using the RSA PSS
6458 algorithm and absent mask generation function parameter. Since these
6459 routines are used to verify certificate signature algorithms this can be
6460 used to crash any certificate verification operation and exploited in a
6461 DoS attack. Any application which performs certificate verification is
6462 vulnerable including OpenSSL clients and servers which enable client
6465 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6470 * X509_ATTRIBUTE memory leak
6472 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6473 memory. This structure is used by the PKCS#7 and CMS routines so any
6474 application which reads PKCS#7 or CMS data from untrusted sources is
6475 affected. SSL/TLS is not affected.
6477 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6483 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6484 This changes the decoding behaviour for some invalid messages,
6485 though the change is mostly in the more lenient direction, and
6486 legacy behaviour is preserved as much as possible.
6490 * In DSA_generate_parameters_ex, if the provided seed is too short,
6491 use a random seed, as already documented.
6493 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6495 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6497 * Alternate chains certificate forgery
6499 During certificate verfification, OpenSSL will attempt to find an
6500 alternative certificate chain if the first attempt to build such a chain
6501 fails. An error in the implementation of this logic can mean that an
6502 attacker could cause certain checks on untrusted certificates to be
6503 bypassed, such as the CA flag, enabling them to use a valid leaf
6504 certificate to act as a CA and "issue" an invalid certificate.
6506 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6512 * Race condition handling PSK identify hint
6514 If PSK identity hints are received by a multi-threaded client then
6515 the values are wrongly updated in the parent SSL_CTX structure. This can
6516 result in a race condition potentially leading to a double free of the
6522 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6524 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6525 incompatibility in the handling of HMAC. The previous ABI has now been
6528 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6530 * Malformed ECParameters causes infinite loop
6532 When processing an ECParameters structure OpenSSL enters an infinite loop
6533 if the curve specified is over a specially malformed binary polynomial
6536 This can be used to perform denial of service against any
6537 system which processes public keys, certificate requests or
6538 certificates. This includes TLS clients and TLS servers with
6539 client authentication enabled.
6541 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6546 * Exploitable out-of-bounds read in X509_cmp_time
6548 X509_cmp_time does not properly check the length of the ASN1_TIME
6549 string and can read a few bytes out of bounds. In addition,
6550 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6553 An attacker can use this to craft malformed certificates and CRLs of
6554 various sizes and potentially cause a segmentation fault, resulting in
6555 a DoS on applications that verify certificates or CRLs. TLS clients
6556 that verify CRLs are affected. TLS clients and servers with client
6557 authentication enabled may be affected if they use custom verification
6560 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6561 independently by Hanno Böck.
6566 * PKCS7 crash with missing EnvelopedContent
6568 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6569 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6570 with missing content and trigger a NULL pointer dereference on parsing.
6572 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6573 structures from untrusted sources are affected. OpenSSL clients and
6574 servers are not affected.
6576 This issue was reported to OpenSSL by Michal Zalewski (Google).
6581 * CMS verify infinite loop with unknown hash function
6583 When verifying a signedData message the CMS code can enter an infinite loop
6584 if presented with an unknown hash function OID. This can be used to perform
6585 denial of service against any system which verifies signedData messages using
6587 This issue was reported to OpenSSL by Johannes Bauer.
6592 * Race condition handling NewSessionTicket
6594 If a NewSessionTicket is received by a multi-threaded client when attempting to
6595 reuse a previous ticket then a race condition can occur potentially leading to
6596 a double free of the ticket data.
6601 * Reject DH handshakes with parameters shorter than 768 bits.
6603 *Kurt Roeckx and Emilia Kasper*
6605 * dhparam: generate 2048-bit parameters by default.
6607 *Kurt Roeckx and Emilia Kasper*
6609 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6611 * Segmentation fault in ASN1_TYPE_cmp fix
6613 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6614 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6615 certificate signature algorithm consistency this can be used to crash any
6616 certificate verification operation and exploited in a DoS attack. Any
6617 application which performs certificate verification is vulnerable including
6618 OpenSSL clients and servers which enable client authentication.
6623 * ASN.1 structure reuse memory corruption fix
6625 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6626 memory corruption via an invalid write. Such reuse is and has been
6627 strongly discouraged and is believed to be rare.
6629 Applications that parse structures containing CHOICE or ANY DEFINED BY
6630 components may be affected. Certificate parsing (d2i_X509 and related
6631 functions) are however not affected. OpenSSL clients and servers are
6637 * PKCS7 NULL pointer dereferences fix
6639 The PKCS#7 parsing code does not handle missing outer ContentInfo
6640 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6641 missing content and trigger a NULL pointer dereference on parsing.
6643 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6644 otherwise parse PKCS#7 structures from untrusted sources are
6645 affected. OpenSSL clients and servers are not affected.
6647 This issue was reported to OpenSSL by Michal Zalewski (Google).
6652 * DoS via reachable assert in SSLv2 servers fix
6654 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6655 servers that both support SSLv2 and enable export cipher suites by sending
6656 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6658 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6659 (OpenSSL development team).
6664 * Use After Free following d2i_ECPrivatekey error fix
6666 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6667 could cause a use after free condition. This, in turn, could cause a double
6668 free in several private key parsing functions (such as d2i_PrivateKey
6669 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6670 for applications that receive EC private keys from untrusted
6671 sources. This scenario is considered rare.
6673 This issue was discovered by the BoringSSL project and fixed in their
6679 * X509_to_X509_REQ NULL pointer deref fix
6681 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6682 the certificate key is invalid. This function is rarely used in practice.
6684 This issue was discovered by Brian Carpenter.
6689 * Removed the export ciphers from the DEFAULT ciphers
6693 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6695 * Build fixes for the Windows and OpenVMS platforms
6697 *Matt Caswell and Richard Levitte*
6699 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6701 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6702 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6703 dereference. This could lead to a Denial Of Service attack. Thanks to
6704 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6709 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6710 dtls1_buffer_record function under certain conditions. In particular this
6711 could occur if an attacker sent repeated DTLS records with the same
6712 sequence number but for the next epoch. The memory leak could be exploited
6713 by an attacker in a Denial of Service attack through memory exhaustion.
6714 Thanks to Chris Mueller for reporting this issue.
6719 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6720 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6721 method would be set to NULL which could later result in a NULL pointer
6722 dereference. Thanks to Frank Schmirler for reporting this issue.
6727 * Abort handshake if server key exchange message is omitted for ephemeral
6730 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6731 reporting this issue.
6736 * Remove non-export ephemeral RSA code on client and server. This code
6737 violated the TLS standard by allowing the use of temporary RSA keys in
6738 non-export ciphersuites and could be used by a server to effectively
6739 downgrade the RSA key length used to a value smaller than the server
6740 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6741 INRIA or reporting this issue.
6746 * Fixed issue where DH client certificates are accepted without verification.
6747 An OpenSSL server will accept a DH certificate for client authentication
6748 without the certificate verify message. This effectively allows a client to
6749 authenticate without the use of a private key. This only affects servers
6750 which trust a client certificate authority which issues certificates
6751 containing DH keys: these are extremely rare and hardly ever encountered.
6752 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6758 * Ensure that the session ID context of an SSL is updated when its
6759 SSL_CTX is updated via SSL_set_SSL_CTX.
6761 The session ID context is typically set from the parent SSL_CTX,
6762 and can vary with the CTX.
6766 * Fix various certificate fingerprint issues.
6768 By using non-DER or invalid encodings outside the signed portion of a
6769 certificate the fingerprint can be changed without breaking the signature.
6770 Although no details of the signed portion of the certificate can be changed
6771 this can cause problems with some applications: e.g. those using the
6772 certificate fingerprint for blacklists.
6774 1. Reject signatures with non zero unused bits.
6776 If the BIT STRING containing the signature has non zero unused bits reject
6777 the signature. All current signature algorithms require zero unused bits.
6779 2. Check certificate algorithm consistency.
6781 Check the AlgorithmIdentifier inside TBS matches the one in the
6782 certificate signature. NB: this will result in signature failure
6783 errors for some broken certificates.
6785 Thanks to Konrad Kraszewski from Google for reporting this issue.
6787 3. Check DSA/ECDSA signatures use DER.
6789 Re-encode DSA/ECDSA signatures and compare with the original received
6790 signature. Return an error if there is a mismatch.
6792 This will reject various cases including garbage after signature
6793 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6794 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6795 (negative or with leading zeroes).
6797 Further analysis was conducted and fixes were developed by Stephen Henson
6798 of the OpenSSL core team.
6804 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6805 results on some platforms, including x86_64. This bug occurs at random
6806 with a very low probability, and is not known to be exploitable in any
6807 way, though its exact impact is difficult to determine. Thanks to Pieter
6808 Wuille (Blockstream) who reported this issue and also suggested an initial
6809 fix. Further analysis was conducted by the OpenSSL development team and
6810 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6811 the OpenSSL core team.
6816 * Do not resume sessions on the server if the negotiated protocol
6817 version does not match the session's version. Resuming with a different
6818 version, while not strictly forbidden by the RFC, is of questionable
6819 sanity and breaks all known clients.
6821 *David Benjamin, Emilia Käsper*
6823 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6824 early CCS messages during renegotiation. (Note that because
6825 renegotiation is encrypted, this early CCS was not exploitable.)
6829 * Tighten client-side session ticket handling during renegotiation:
6830 ensure that the client only accepts a session ticket if the server sends
6831 the extension anew in the ServerHello. Previously, a TLS client would
6832 reuse the old extension state and thus accept a session ticket if one was
6833 announced in the initial ServerHello.
6835 Similarly, ensure that the client requires a session ticket if one
6836 was advertised in the ServerHello. Previously, a TLS client would
6837 ignore a missing NewSessionTicket message.
6841 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6845 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6846 sends a carefully crafted handshake message, to cause OpenSSL to fail
6847 to free up to 64k of memory causing a memory leak. This could be
6848 exploited in a Denial Of Service attack. This issue affects OpenSSL
6849 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6850 whether SRTP is used or configured. Implementations of OpenSSL that
6851 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6853 The fix was developed by the OpenSSL team.
6858 * Session Ticket Memory Leak.
6860 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6861 integrity of that ticket is first verified. In the event of a session
6862 ticket integrity check failing, OpenSSL will fail to free memory
6863 causing a memory leak. By sending a large number of invalid session
6864 tickets an attacker could exploit this issue in a Denial Of Service
6870 * Build option no-ssl3 is incomplete.
6872 When OpenSSL is configured with "no-ssl3" as a build option, servers
6873 could accept and complete a SSL 3.0 handshake, and clients could be
6874 configured to send them.
6877 *Akamai and the OpenSSL team*
6879 * Add support for TLS_FALLBACK_SCSV.
6880 Client applications doing fallback retries should call
6881 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6884 *Adam Langley, Bodo Moeller*
6886 * Add additional DigestInfo checks.
6888 Re-encode DigestInto in DER and check against the original when
6889 verifying RSA signature: this will reject any improperly encoded
6890 DigestInfo structures.
6892 Note: this is a precautionary measure and no attacks are currently known.
6896 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6898 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6899 SRP code can be overrun an internal buffer. Add sanity check that
6900 g, A, B < N to SRP code.
6902 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6903 Group for discovering this issue.
6908 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6909 TLS 1.0 instead of higher protocol versions when the ClientHello message
6910 is badly fragmented. This allows a man-in-the-middle attacker to force a
6911 downgrade to TLS 1.0 even if both the server and the client support a
6912 higher protocol version, by modifying the client's TLS records.
6914 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6915 researching this issue.
6920 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6921 to a denial of service attack. A malicious server can crash the client
6922 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6923 ciphersuite and sending carefully crafted handshake messages.
6925 Thanks to Felix Gröbert (Google) for discovering and researching this
6931 * By sending carefully crafted DTLS packets an attacker could cause openssl
6932 to leak memory. This can be exploited through a Denial of Service attack.
6933 Thanks to Adam Langley for discovering and researching this issue.
6938 * An attacker can force openssl to consume large amounts of memory whilst
6939 processing DTLS handshake messages. This can be exploited through a
6940 Denial of Service attack.
6941 Thanks to Adam Langley for discovering and researching this issue.
6946 * An attacker can force an error condition which causes openssl to crash
6947 whilst processing DTLS packets due to memory being freed twice. This
6948 can be exploited through a Denial of Service attack.
6949 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6955 * If a multithreaded client connects to a malicious server using a resumed
6956 session and the server sends an ec point format extension it could write
6957 up to 255 bytes to freed memory.
6959 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6965 * A malicious server can crash an OpenSSL client with a null pointer
6966 dereference (read) by specifying an SRP ciphersuite even though it was not
6967 properly negotiated with the client. This can be exploited through a
6968 Denial of Service attack.
6970 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6971 discovering and researching this issue.
6976 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6977 X509_name_oneline, X509_name_print_ex et al. to leak some information
6978 from the stack. Applications may be affected if they echo pretty printing
6979 output to the attacker.
6981 Thanks to Ivan Fratric (Google) for discovering this issue.
6984 *Emilia Käsper, and Steve Henson*
6986 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6987 for corner cases. (Certain input points at infinity could lead to
6988 bogus results, with non-infinity inputs mapped to infinity too.)
6992 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6994 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6995 handshake can force the use of weak keying material in OpenSSL
6996 SSL/TLS clients and servers.
6998 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6999 researching this issue. ([CVE-2014-0224])
7001 *KIKUCHI Masashi, Steve Henson*
7003 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7004 OpenSSL DTLS client the code can be made to recurse eventually crashing
7007 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7010 *Imre Rad, Steve Henson*
7012 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7013 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7014 client or server. This is potentially exploitable to run arbitrary
7015 code on a vulnerable client or server.
7017 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7019 *Jüri Aedla, Steve Henson*
7021 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7022 are subject to a denial of service attack.
7024 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7025 this issue. ([CVE-2014-3470])
7027 *Felix Gröbert, Ivan Fratric, Steve Henson*
7029 * Harmonize version and its documentation. -f flag is used to display
7032 *mancha <mancha1@zoho.com>*
7034 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7035 in i2d_ECPrivateKey.
7037 *mancha <mancha1@zoho.com>*
7039 * Fix some double frees. These are not thought to be exploitable.
7041 *mancha <mancha1@zoho.com>*
7043 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7045 * A missing bounds check in the handling of the TLS heartbeat extension
7046 can be used to reveal up to 64k of memory to a connected client or
7049 Thanks for Neel Mehta of Google Security for discovering this bug and to
7050 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7051 preparing the fix ([CVE-2014-0160])
7053 *Adam Langley, Bodo Moeller*
7055 * Fix for the attack described in the paper "Recovering OpenSSL
7056 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7057 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7058 <http://eprint.iacr.org/2014/140>
7060 Thanks to Yuval Yarom and Naomi Benger for discovering this
7061 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7063 *Yuval Yarom and Naomi Benger*
7065 * TLS pad extension: draft-agl-tls-padding-03
7067 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7068 TLS client Hello record length value would otherwise be > 255 and
7069 less that 512 pad with a dummy extension containing zeroes so it
7070 is at least 512 bytes long.
7072 *Adam Langley, Steve Henson*
7074 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7076 * Fix for TLS record tampering bug. A carefully crafted invalid
7077 handshake could crash OpenSSL with a NULL pointer exception.
7078 Thanks to Anton Johansson for reporting this issues.
7081 * Keep original DTLS digest and encryption contexts in retransmission
7082 structures so we can use the previous session parameters if they need
7083 to be resent. ([CVE-2013-6450])
7087 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7088 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7089 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7090 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7091 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7092 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7094 *Rob Stradling, Adam Langley*
7096 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7098 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7099 supporting platforms or when small records were transferred.
7101 *Andy Polyakov, Steve Henson*
7103 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7105 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7107 This addresses the flaw in CBC record processing discovered by
7108 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7109 at: <http://www.isg.rhul.ac.uk/tls/>
7111 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7112 Security Group at Royal Holloway, University of London
7113 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7114 Emilia Käsper for the initial patch.
7117 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7119 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7120 ciphersuites which can be exploited in a denial of service attack.
7121 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7122 and detecting this bug and to Wolfgang Ettlinger
7123 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7128 * Return an error when checking OCSP signatures when key is NULL.
7129 This fixes a DoS attack. ([CVE-2013-0166])
7133 * Make openssl verify return errors.
7135 *Chris Palmer <palmer@google.com> and Ben Laurie*
7137 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7138 the right response is stapled. Also change SSL_get_certificate()
7139 so it returns the certificate actually sent.
7140 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7142 *Rob Stradling <rob.stradling@comodo.com>*
7144 * Fix possible deadlock when decoding public keys.
7148 * Don't use TLS 1.0 record version number in initial client hello
7153 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7155 * Sanity check record length before skipping explicit IV in TLS
7156 1.2, 1.1 and DTLS to fix DoS attack.
7158 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7159 fuzzing as a service testing platform.
7164 * Initialise tkeylen properly when encrypting CMS messages.
7165 Thanks to Solar Designer of Openwall for reporting this issue.
7169 * In FIPS mode don't try to use composite ciphers as they are not
7174 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7176 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7177 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7178 mean any application compiled against OpenSSL 1.0.0 headers setting
7179 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7180 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7181 0x10000000L Any application which was previously compiled against
7182 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7183 will need to be recompiled as a result. Letting be results in
7184 inability to disable specifically TLS 1.1 and in client context,
7185 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7189 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7190 disable just protocol X, but all protocols above X *if* there are
7191 protocols *below* X still enabled. In more practical terms it means
7192 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7193 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7194 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7199 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7201 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7202 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7203 in CRYPTO_realloc_clean.
7205 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7206 issue and to Adam Langley <agl@chromium.org> for fixing it.
7209 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7211 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7215 * Workarounds for some broken servers that "hang" if a client hello
7216 record length exceeds 255 bytes.
7218 1. Do not use record version number > TLS 1.0 in initial client
7219 hello: some (but not all) hanging servers will now work.
7220 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7221 the number of ciphers sent in the client hello. This should be
7222 set to an even number, such as 50, for example by passing:
7223 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7224 Most broken servers should now work.
7225 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7226 TLS 1.2 client support entirely.
7230 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7234 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7236 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7237 STRING form instead of a DigestInfo.
7241 * The format used for MDC2 RSA signatures is inconsistent between EVP
7242 and the RSA_sign/RSA_verify functions. This was made more apparent when
7243 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7244 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7245 the correct format in RSA_verify so both forms transparently work.
7249 * Some servers which support TLS 1.0 can choke if we initially indicate
7250 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7251 encrypted premaster secret. As a workaround use the maximum permitted
7252 client version in client hello, this should keep such servers happy
7253 and still work with previous versions of OpenSSL.
7257 * Add support for TLS/DTLS heartbeats.
7259 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7261 * Add support for SCTP.
7263 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7265 * Improved PRNG seeding for VOS.
7267 *Paul Green <Paul.Green@stratus.com>*
7269 * Extensive assembler packs updates, most notably:
7271 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7272 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7273 - x86_64: bit-sliced AES implementation;
7274 - ARM: NEON support, contemporary platforms optimizations;
7275 - s390x: z196 support;
7276 - `*`: GHASH and GF(2^m) multiplication implementations;
7280 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7281 (removal of unnecessary code)
7283 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7285 * Add TLS key material exporter from RFC 5705.
7289 * Add DTLS-SRTP negotiation from RFC 5764.
7293 * Add Next Protocol Negotiation,
7294 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7295 disabled with a no-npn flag to config or Configure. Code donated
7298 *Adam Langley <agl@google.com> and Ben Laurie*
7300 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7301 NIST-P256, NIST-P521, with constant-time single point multiplication on
7302 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7303 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7304 Code made available under Apache License version 2.0.
7306 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7307 line to include this in your build of OpenSSL, and run "make depend" (or
7308 "make update"). This enables the following EC_METHODs:
7310 EC_GFp_nistp224_method()
7311 EC_GFp_nistp256_method()
7312 EC_GFp_nistp521_method()
7314 EC_GROUP_new_by_curve_name() will automatically use these (while
7315 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7318 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7320 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7321 all platforms. Move ssize_t definition from e_os.h to the public
7322 header file e_os2.h as it now appears in public header file cms.h
7326 * New -sigopt option to the ca, req and x509 utilities. Additional
7327 signature parameters can be passed using this option and in
7332 * Add RSA PSS signing function. This will generate and set the
7333 appropriate AlgorithmIdentifiers for PSS based on those in the
7334 corresponding EVP_MD_CTX structure. No application support yet.
7338 * Support for companion algorithm specific ASN1 signing routines.
7339 New function ASN1_item_sign_ctx() signs a pre-initialised
7340 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7341 the appropriate parameters.
7345 * Add new algorithm specific ASN1 verification initialisation function
7346 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7347 handling will be the same no matter what EVP_PKEY_METHOD is used.
7348 Add a PSS handler to support verification of PSS signatures: checked
7349 against a number of sample certificates.
7353 * Add signature printing for PSS. Add PSS OIDs.
7355 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7357 * Add algorithm specific signature printing. An individual ASN1 method
7358 can now print out signatures instead of the standard hex dump.
7360 More complex signatures (e.g. PSS) can print out more meaningful
7361 information. Include DSA version that prints out the signature
7366 * Password based recipient info support for CMS library: implementing
7371 * Split password based encryption into PBES2 and PBKDF2 functions. This
7372 neatly separates the code into cipher and PBE sections and is required
7373 for some algorithms that split PBES2 into separate pieces (such as
7374 password based CMS).
7378 * Session-handling fixes:
7379 - Fix handling of connections that are resuming with a session ID,
7380 but also support Session Tickets.
7381 - Fix a bug that suppressed issuing of a new ticket if the client
7382 presented a ticket with an expired session.
7383 - Try to set the ticket lifetime hint to something reasonable.
7384 - Make tickets shorter by excluding irrelevant information.
7385 - On the client side, don't ignore renewed tickets.
7387 *Adam Langley, Bodo Moeller (Google)*
7389 * Fix PSK session representation.
7393 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7395 This work was sponsored by Intel.
7399 * Add GCM support to TLS library. Some custom code is needed to split
7400 the IV between the fixed (from PRF) and explicit (from TLS record)
7401 portions. This adds all GCM ciphersuites supported by RFC5288 and
7402 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7403 add a special AESGCM string for GCM only.
7407 * Expand range of ctrls for AES GCM. Permit setting invocation
7408 field on decrypt and retrieval of invocation field only on encrypt.
7412 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7413 As required by RFC5289 these ciphersuites cannot be used if for
7414 versions of TLS earlier than 1.2.
7418 * For FIPS capable OpenSSL interpret a NULL default public key method
7419 as unset and return the appropriate default but do *not* set the default.
7420 This means we can return the appropriate method in applications that
7421 switch between FIPS and non-FIPS modes.
7425 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7426 ENGINE is used then we cannot handle that in the FIPS module so we
7427 keep original code iff non-FIPS operations are allowed.
7431 * Add -attime option to openssl utilities.
7433 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7435 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7439 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7440 FIPS EC methods unconditionally for now.
7444 * New build option no-ec2m to disable characteristic 2 code.
7448 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7449 all cases can be covered as some introduce binary incompatibilities.
7453 * Redirect RSA operations to FIPS module including keygen,
7454 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7458 * Add similar low-level API blocking to ciphers.
7462 * low-level digest APIs are not approved in FIPS mode: any attempt
7463 to use these will cause a fatal error. Applications that *really* want
7464 to use them can use the `private_*` version instead.
7468 * Redirect cipher operations to FIPS module for FIPS builds.
7472 * Redirect digest operations to FIPS module for FIPS builds.
7476 * Update build system to add "fips" flag which will link in fipscanister.o
7477 for static and shared library builds embedding a signature if needed.
7481 * Output TLS supported curves in preference order instead of numerical
7482 order. This is currently hardcoded for the highest order curves first.
7483 This should be configurable so applications can judge speed vs strength.
7487 * Add TLS v1.2 server support for client authentication.
7491 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7496 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7497 FIPS modules versions.
7501 * Add TLS v1.2 client side support for client authentication. Keep cache
7502 of handshake records longer as we don't know the hash algorithm to use
7503 until after the certificate request message is received.
7507 * Initial TLS v1.2 client support. Add a default signature algorithms
7508 extension including all the algorithms we support. Parse new signature
7509 format in client key exchange. Relax some ECC signing restrictions for
7510 TLS v1.2 as indicated in RFC5246.
7514 * Add server support for TLS v1.2 signature algorithms extension. Switch
7515 to new signature format when needed using client digest preference.
7516 All server ciphersuites should now work correctly in TLS v1.2. No client
7517 support yet and no support for client certificates.
7521 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7522 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7523 ciphersuites. At present only RSA key exchange ciphersuites work with
7524 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7525 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7526 and version checking.
7530 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7531 with this defined it will not be affected by any changes to ssl internal
7532 structures. Add several utility functions to allow openssl application
7533 to work with OPENSSL_NO_SSL_INTERN defined.
7537 * A long standing patch to add support for SRP from EdelWeb (Peter
7538 Sylvester and Christophe Renou) was integrated.
7539 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7540 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7543 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7547 * Permit abbreviated handshakes when renegotiating using the function
7548 SSL_renegotiate_abbreviated().
7550 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7552 * Add call to ENGINE_register_all_complete() to
7553 ENGINE_load_builtin_engines(), so some implementations get used
7554 automatically instead of needing explicit application support.
7558 * Add support for TLS key exporter as described in RFC5705.
7560 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7562 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7563 a few changes are required:
7565 Add SSL_OP_NO_TLSv1_1 flag.
7566 Add TLSv1_1 methods.
7567 Update version checking logic to handle version 1.1.
7568 Add explicit IV handling (ported from DTLS code).
7569 Add command line options to s_client/s_server.
7576 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7578 * X509_ATTRIBUTE memory leak
7580 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7581 memory. This structure is used by the PKCS#7 and CMS routines so any
7582 application which reads PKCS#7 or CMS data from untrusted sources is
7583 affected. SSL/TLS is not affected.
7585 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7591 * Race condition handling PSK identify hint
7593 If PSK identity hints are received by a multi-threaded client then
7594 the values are wrongly updated in the parent SSL_CTX structure. This can
7595 result in a race condition potentially leading to a double free of the
7601 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7603 * Malformed ECParameters causes infinite loop
7605 When processing an ECParameters structure OpenSSL enters an infinite loop
7606 if the curve specified is over a specially malformed binary polynomial
7609 This can be used to perform denial of service against any
7610 system which processes public keys, certificate requests or
7611 certificates. This includes TLS clients and TLS servers with
7612 client authentication enabled.
7614 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7619 * Exploitable out-of-bounds read in X509_cmp_time
7621 X509_cmp_time does not properly check the length of the ASN1_TIME
7622 string and can read a few bytes out of bounds. In addition,
7623 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7626 An attacker can use this to craft malformed certificates and CRLs of
7627 various sizes and potentially cause a segmentation fault, resulting in
7628 a DoS on applications that verify certificates or CRLs. TLS clients
7629 that verify CRLs are affected. TLS clients and servers with client
7630 authentication enabled may be affected if they use custom verification
7633 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7634 independently by Hanno Böck.
7639 * PKCS7 crash with missing EnvelopedContent
7641 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7642 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7643 with missing content and trigger a NULL pointer dereference on parsing.
7645 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7646 structures from untrusted sources are affected. OpenSSL clients and
7647 servers are not affected.
7649 This issue was reported to OpenSSL by Michal Zalewski (Google).
7654 * CMS verify infinite loop with unknown hash function
7656 When verifying a signedData message the CMS code can enter an infinite loop
7657 if presented with an unknown hash function OID. This can be used to perform
7658 denial of service against any system which verifies signedData messages using
7660 This issue was reported to OpenSSL by Johannes Bauer.
7665 * Race condition handling NewSessionTicket
7667 If a NewSessionTicket is received by a multi-threaded client when attempting to
7668 reuse a previous ticket then a race condition can occur potentially leading to
7669 a double free of the ticket data.
7674 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7676 * Segmentation fault in ASN1_TYPE_cmp fix
7678 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7679 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7680 certificate signature algorithm consistency this can be used to crash any
7681 certificate verification operation and exploited in a DoS attack. Any
7682 application which performs certificate verification is vulnerable including
7683 OpenSSL clients and servers which enable client authentication.
7688 * ASN.1 structure reuse memory corruption fix
7690 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7691 memory corruption via an invalid write. Such reuse is and has been
7692 strongly discouraged and is believed to be rare.
7694 Applications that parse structures containing CHOICE or ANY DEFINED BY
7695 components may be affected. Certificate parsing (d2i_X509 and related
7696 functions) are however not affected. OpenSSL clients and servers are
7702 * PKCS7 NULL pointer dereferences fix
7704 The PKCS#7 parsing code does not handle missing outer ContentInfo
7705 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7706 missing content and trigger a NULL pointer dereference on parsing.
7708 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7709 otherwise parse PKCS#7 structures from untrusted sources are
7710 affected. OpenSSL clients and servers are not affected.
7712 This issue was reported to OpenSSL by Michal Zalewski (Google).
7717 * DoS via reachable assert in SSLv2 servers fix
7719 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7720 servers that both support SSLv2 and enable export cipher suites by sending
7721 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7723 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7724 (OpenSSL development team).
7729 * Use After Free following d2i_ECPrivatekey error fix
7731 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7732 could cause a use after free condition. This, in turn, could cause a double
7733 free in several private key parsing functions (such as d2i_PrivateKey
7734 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7735 for applications that receive EC private keys from untrusted
7736 sources. This scenario is considered rare.
7738 This issue was discovered by the BoringSSL project and fixed in their
7744 * X509_to_X509_REQ NULL pointer deref fix
7746 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7747 the certificate key is invalid. This function is rarely used in practice.
7749 This issue was discovered by Brian Carpenter.
7754 * Removed the export ciphers from the DEFAULT ciphers
7758 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7760 * Build fixes for the Windows and OpenVMS platforms
7762 *Matt Caswell and Richard Levitte*
7764 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7766 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7767 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7768 dereference. This could lead to a Denial Of Service attack. Thanks to
7769 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7774 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7775 dtls1_buffer_record function under certain conditions. In particular this
7776 could occur if an attacker sent repeated DTLS records with the same
7777 sequence number but for the next epoch. The memory leak could be exploited
7778 by an attacker in a Denial of Service attack through memory exhaustion.
7779 Thanks to Chris Mueller for reporting this issue.
7784 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7785 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7786 method would be set to NULL which could later result in a NULL pointer
7787 dereference. Thanks to Frank Schmirler for reporting this issue.
7792 * Abort handshake if server key exchange message is omitted for ephemeral
7795 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7796 reporting this issue.
7801 * Remove non-export ephemeral RSA code on client and server. This code
7802 violated the TLS standard by allowing the use of temporary RSA keys in
7803 non-export ciphersuites and could be used by a server to effectively
7804 downgrade the RSA key length used to a value smaller than the server
7805 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7806 INRIA or reporting this issue.
7811 * Fixed issue where DH client certificates are accepted without verification.
7812 An OpenSSL server will accept a DH certificate for client authentication
7813 without the certificate verify message. This effectively allows a client to
7814 authenticate without the use of a private key. This only affects servers
7815 which trust a client certificate authority which issues certificates
7816 containing DH keys: these are extremely rare and hardly ever encountered.
7817 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7823 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7824 results on some platforms, including x86_64. This bug occurs at random
7825 with a very low probability, and is not known to be exploitable in any
7826 way, though its exact impact is difficult to determine. Thanks to Pieter
7827 Wuille (Blockstream) who reported this issue and also suggested an initial
7828 fix. Further analysis was conducted by the OpenSSL development team and
7829 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7830 the OpenSSL core team.
7835 * Fix various certificate fingerprint issues.
7837 By using non-DER or invalid encodings outside the signed portion of a
7838 certificate the fingerprint can be changed without breaking the signature.
7839 Although no details of the signed portion of the certificate can be changed
7840 this can cause problems with some applications: e.g. those using the
7841 certificate fingerprint for blacklists.
7843 1. Reject signatures with non zero unused bits.
7845 If the BIT STRING containing the signature has non zero unused bits reject
7846 the signature. All current signature algorithms require zero unused bits.
7848 2. Check certificate algorithm consistency.
7850 Check the AlgorithmIdentifier inside TBS matches the one in the
7851 certificate signature. NB: this will result in signature failure
7852 errors for some broken certificates.
7854 Thanks to Konrad Kraszewski from Google for reporting this issue.
7856 3. Check DSA/ECDSA signatures use DER.
7858 Reencode DSA/ECDSA signatures and compare with the original received
7859 signature. Return an error if there is a mismatch.
7861 This will reject various cases including garbage after signature
7862 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7863 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7864 (negative or with leading zeroes).
7866 Further analysis was conducted and fixes were developed by Stephen Henson
7867 of the OpenSSL core team.
7873 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7875 * Session Ticket Memory Leak.
7877 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7878 integrity of that ticket is first verified. In the event of a session
7879 ticket integrity check failing, OpenSSL will fail to free memory
7880 causing a memory leak. By sending a large number of invalid session
7881 tickets an attacker could exploit this issue in a Denial Of Service
7887 * Build option no-ssl3 is incomplete.
7889 When OpenSSL is configured with "no-ssl3" as a build option, servers
7890 could accept and complete a SSL 3.0 handshake, and clients could be
7891 configured to send them.
7894 *Akamai and the OpenSSL team*
7896 * Add support for TLS_FALLBACK_SCSV.
7897 Client applications doing fallback retries should call
7898 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7901 *Adam Langley, Bodo Moeller*
7903 * Add additional DigestInfo checks.
7905 Reencode DigestInto in DER and check against the original when
7906 verifying RSA signature: this will reject any improperly encoded
7907 DigestInfo structures.
7909 Note: this is a precautionary measure and no attacks are currently known.
7913 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7915 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7916 to a denial of service attack. A malicious server can crash the client
7917 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7918 ciphersuite and sending carefully crafted handshake messages.
7920 Thanks to Felix Gröbert (Google) for discovering and researching this
7926 * By sending carefully crafted DTLS packets an attacker could cause openssl
7927 to leak memory. This can be exploited through a Denial of Service attack.
7928 Thanks to Adam Langley for discovering and researching this issue.
7933 * An attacker can force openssl to consume large amounts of memory whilst
7934 processing DTLS handshake messages. This can be exploited through a
7935 Denial of Service attack.
7936 Thanks to Adam Langley for discovering and researching this issue.
7941 * An attacker can force an error condition which causes openssl to crash
7942 whilst processing DTLS packets due to memory being freed twice. This
7943 can be exploited through a Denial of Service attack.
7944 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7950 * If a multithreaded client connects to a malicious server using a resumed
7951 session and the server sends an ec point format extension it could write
7952 up to 255 bytes to freed memory.
7954 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7960 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7961 X509_name_oneline, X509_name_print_ex et al. to leak some information
7962 from the stack. Applications may be affected if they echo pretty printing
7963 output to the attacker.
7965 Thanks to Ivan Fratric (Google) for discovering this issue.
7968 *Emilia Käsper, and Steve Henson*
7970 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7971 for corner cases. (Certain input points at infinity could lead to
7972 bogus results, with non-infinity inputs mapped to infinity too.)
7976 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7978 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7979 handshake can force the use of weak keying material in OpenSSL
7980 SSL/TLS clients and servers.
7982 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7983 researching this issue. ([CVE-2014-0224])
7985 *KIKUCHI Masashi, Steve Henson*
7987 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7988 OpenSSL DTLS client the code can be made to recurse eventually crashing
7991 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7994 *Imre Rad, Steve Henson*
7996 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7997 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7998 client or server. This is potentially exploitable to run arbitrary
7999 code on a vulnerable client or server.
8001 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8003 *Jüri Aedla, Steve Henson*
8005 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8006 are subject to a denial of service attack.
8008 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8009 this issue. ([CVE-2014-3470])
8011 *Felix Gröbert, Ivan Fratric, Steve Henson*
8013 * Harmonize version and its documentation. -f flag is used to display
8016 *mancha <mancha1@zoho.com>*
8018 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8019 in i2d_ECPrivateKey.
8021 *mancha <mancha1@zoho.com>*
8023 * Fix some double frees. These are not thought to be exploitable.
8025 *mancha <mancha1@zoho.com>*
8027 * Fix for the attack described in the paper "Recovering OpenSSL
8028 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8029 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8030 <http://eprint.iacr.org/2014/140>
8032 Thanks to Yuval Yarom and Naomi Benger for discovering this
8033 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8035 *Yuval Yarom and Naomi Benger*
8037 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8039 * Keep original DTLS digest and encryption contexts in retransmission
8040 structures so we can use the previous session parameters if they need
8041 to be resent. ([CVE-2013-6450])
8045 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8046 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8047 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8048 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8049 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8050 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8052 *Rob Stradling, Adam Langley*
8054 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8056 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8058 This addresses the flaw in CBC record processing discovered by
8059 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8060 at: <http://www.isg.rhul.ac.uk/tls/>
8062 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8063 Security Group at Royal Holloway, University of London
8064 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8065 Emilia Käsper for the initial patch.
8068 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8070 * Return an error when checking OCSP signatures when key is NULL.
8071 This fixes a DoS attack. ([CVE-2013-0166])
8075 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8076 the right response is stapled. Also change SSL_get_certificate()
8077 so it returns the certificate actually sent.
8078 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8079 (This is a backport)
8081 *Rob Stradling <rob.stradling@comodo.com>*
8083 * Fix possible deadlock when decoding public keys.
8087 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8089 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8092 * Sanity check record length before skipping explicit IV in DTLS
8095 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8096 fuzzing as a service testing platform.
8101 * Initialise tkeylen properly when encrypting CMS messages.
8102 Thanks to Solar Designer of Openwall for reporting this issue.
8106 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8108 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8109 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8110 in CRYPTO_realloc_clean.
8112 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8113 issue and to Adam Langley <agl@chromium.org> for fixing it.
8116 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8118 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8120 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8121 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8122 content decryption and always return the same error. Note: this attack
8123 needs on average 2^20 messages so it only affects automated senders. The
8124 old behaviour can be re-enabled in the CMS code by setting the
8125 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8126 an MMA defence is not necessary.
8127 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8128 this issue. ([CVE-2012-0884])
8132 * Fix CVE-2011-4619: make sure we really are receiving a
8133 client hello before rejecting multiple SGC restarts. Thanks to
8134 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8138 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8140 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8141 Thanks to Antonio Martin, Enterprise Secure Access Research and
8142 Development, Cisco Systems, Inc. for discovering this bug and
8143 preparing a fix. ([CVE-2012-0050])
8147 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8149 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8150 of the Vaudenay padding oracle attack on CBC mode encryption
8151 which enables an efficient plaintext recovery attack against
8152 the OpenSSL implementation of DTLS. Their attack exploits timing
8153 differences arising during decryption processing. A research
8154 paper describing this attack can be found at:
8155 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8156 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8157 Security Group at Royal Holloway, University of London
8158 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8159 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8160 for preparing the fix. ([CVE-2011-4108])
8162 *Robin Seggelmann, Michael Tuexen*
8164 * Clear bytes used for block padding of SSL 3.0 records.
8167 *Adam Langley (Google)*
8169 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8170 Kadianakis <desnacked@gmail.com> for discovering this issue and
8171 Adam Langley for preparing the fix. ([CVE-2011-4619])
8173 *Adam Langley (Google)*
8175 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8177 *Andrey Kulikov <amdeich@gmail.com>*
8179 * Prevent malformed RFC3779 data triggering an assertion failure.
8180 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8181 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8183 *Rob Austein <sra@hactrn.net>*
8185 * Improved PRNG seeding for VOS.
8187 *Paul Green <Paul.Green@stratus.com>*
8189 * Fix ssl_ciph.c set-up race.
8191 *Adam Langley (Google)*
8193 * Fix spurious failures in ecdsatest.c.
8195 *Emilia Käsper (Google)*
8197 * Fix the BIO_f_buffer() implementation (which was mixing different
8198 interpretations of the `..._len` fields).
8200 *Adam Langley (Google)*
8202 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8203 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8204 threads won't reuse the same blinding coefficients.
8206 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8207 lock to call BN_BLINDING_invert_ex, and avoids one use of
8208 BN_BLINDING_update for each BN_BLINDING structure (previously,
8209 the last update always remained unused).
8211 *Emilia Käsper (Google)*
8213 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8215 *Bob Buckholz (Google)*
8217 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8219 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8220 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8222 *Kaspar Brand <ossl@velox.ch>*
8224 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8225 for multi-threaded use of ECDH. ([CVE-2011-3210])
8227 *Adam Langley (Google)*
8229 * Fix x509_name_ex_d2i memory leak on bad inputs.
8233 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8234 signature public key algorithm by using OID xref utilities instead.
8235 Before this you could only use some ECC ciphersuites with SHA1 only.
8239 * Add protection against ECDSA timing attacks as mentioned in the paper
8240 by Billy Bob Brumley and Nicola Tuveri, see:
8241 <http://eprint.iacr.org/2011/232.pdf>
8243 *Billy Bob Brumley and Nicola Tuveri*
8245 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8247 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8249 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8251 * Fix bug in string printing code: if *any* escaping is enabled we must
8252 escape the escape character (backslash) or the resulting string is
8257 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8259 * Disable code workaround for ancient and obsolete Netscape browsers
8260 and servers: an attacker can use it in a ciphersuite downgrade attack.
8261 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8265 * Fixed J-PAKE implementation error, originally discovered by
8266 Sebastien Martini, further info and confirmation from Stefan
8267 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8271 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8273 * Fix extension code to avoid race conditions which can result in a buffer
8274 overrun vulnerability: resumed sessions must not be modified as they can
8275 be shared by multiple threads. CVE-2010-3864
8279 * Fix WIN32 build system to correctly link an ENGINE directory into
8284 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8286 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8289 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8291 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8293 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8294 context. The operation can be customised via the ctrl mechanism in
8295 case ENGINEs want to include additional functionality.
8299 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8303 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8304 output hashes compatible with older versions of OpenSSL.
8306 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8308 * Fix compression algorithm handling: if resuming a session use the
8309 compression algorithm of the resumed session instead of determining
8310 it from client hello again. Don't allow server to change algorithm.
8314 * Add load_crls() function to commands tidying load_certs() too. Add option
8315 to verify utility to allow additional CRLs to be included.
8319 * Update OCSP request code to permit adding custom headers to the request:
8320 some responders need this.
8324 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8327 *Julia Lawall <julia@diku.dk>*
8329 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8330 needlessly dereferenced structures, used obsolete functions and
8331 didn't handle all updated verify codes correctly.
8335 * Disable MD2 in the default configuration.
8339 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8340 indicate the initial BIO being pushed or popped. This makes it possible
8341 to determine whether the BIO is the one explicitly called or as a result
8342 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8343 it handles reference counts correctly and doesn't zero out the I/O bio
8344 when it is not being explicitly popped. WARNING: applications which
8345 included workarounds for the old buggy behaviour will need to be modified
8346 or they could free up already freed BIOs.
8350 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8351 renaming to all platforms (within the 0.9.8 branch, this was
8352 done conditionally on Netware platforms to avoid a name clash).
8354 *Guenter <lists@gknw.net>*
8356 * Add ECDHE and PSK support to DTLS.
8358 *Michael Tuexen <tuexen@fh-muenster.de>*
8360 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8365 * Add "missing" function EVP_MD_flags() (without this the only way to
8366 retrieve a digest flags is by accessing the structure directly. Update
8367 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8368 or cipher is registered as in the "from" argument. Print out all
8369 registered digests in the dgst usage message instead of manually
8370 attempting to work them out.
8374 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8375 this allows the use of compression and extensions. Change default cipher
8376 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8377 by default unless an application cipher string requests it.
8381 * Alter match criteria in PKCS12_parse(). It used to try to use local
8382 key ids to find matching certificates and keys but some PKCS#12 files
8383 don't follow the (somewhat unwritten) rules and this strategy fails.
8384 Now just gather all certificates together and the first private key
8385 then look for the first certificate that matches the key.
8389 * Support use of registered digest and cipher names for dgst and cipher
8390 commands instead of having to add each one as a special case. So now
8397 openssl dgst -sha256 foo
8399 and this works for ENGINE based algorithms too.
8403 * Update Gost ENGINE to support parameter files.
8405 *Victor B. Wagner <vitus@cryptocom.ru>*
8407 * Support GeneralizedTime in ca utility.
8409 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8411 * Enhance the hash format used for certificate directory links. The new
8412 form uses the canonical encoding (meaning equivalent names will work
8413 even if they aren't identical) and uses SHA1 instead of MD5. This form
8414 is incompatible with the older format and as a result c_rehash should
8415 be used to rebuild symbolic links.
8419 * Make PKCS#8 the default write format for private keys, replacing the
8420 traditional format. This form is standardised, more secure and doesn't
8421 include an implicit MD5 dependency.
8425 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8426 committed to OpenSSL should pass this lot as a minimum.
8430 * Add session ticket override functionality for use by EAP-FAST.
8432 *Jouni Malinen <j@w1.fi>*
8434 * Modify HMAC functions to return a value. Since these can be implemented
8435 in an ENGINE errors can occur.
8439 * Type-checked OBJ_bsearch_ex.
8443 * Type-checked OBJ_bsearch. Also some constification necessitated
8444 by type-checking. Still to come: TXT_DB, bsearch(?),
8445 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8450 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8451 seconds to a tm structure directly, instead of going through OS
8452 specific date routines. This avoids any issues with OS routines such
8453 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8454 and X509_time_adj_ex() to cover the extended range. The existing
8455 X509_time_adj() is still usable and will no longer have any date issues.
8459 * Delta CRL support. New use deltas option which will attempt to locate
8460 and search any appropriate delta CRLs available.
8462 This work was sponsored by Google.
8466 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8467 code and add additional score elements. Validate alternate CRL paths
8468 as part of the CRL checking and indicate a new error "CRL path validation
8469 error" in this case. Applications wanting additional details can use
8470 the verify callback and check the new "parent" field. If this is not
8471 NULL CRL path validation is taking place. Existing applications won't
8472 see this because it requires extended CRL support which is off by
8475 This work was sponsored by Google.
8479 * Support for freshest CRL extension.
8481 This work was sponsored by Google.
8485 * Initial indirect CRL support. Currently only supported in the CRLs
8486 passed directly and not via lookup. Process certificate issuer
8487 CRL entry extension and lookup CRL entries by bother issuer name
8488 and serial number. Check and process CRL issuer entry in IDP extension.
8490 This work was sponsored by Google.
8494 * Add support for distinct certificate and CRL paths. The CRL issuer
8495 certificate is validated separately in this case. Only enabled if
8496 an extended CRL support flag is set: this flag will enable additional
8497 CRL functionality in future.
8499 This work was sponsored by Google.
8503 * Add support for policy mappings extension.
8505 This work was sponsored by Google.
8509 * Fixes to pathlength constraint, self issued certificate handling,
8510 policy processing to align with RFC3280 and PKITS tests.
8512 This work was sponsored by Google.
8516 * Support for name constraints certificate extension. DN, email, DNS
8517 and URI types are currently supported.
8519 This work was sponsored by Google.
8523 * To cater for systems that provide a pointer-based thread ID rather
8524 than numeric, deprecate the current numeric thread ID mechanism and
8525 replace it with a structure and associated callback type. This
8526 mechanism allows a numeric "hash" to be extracted from a thread ID in
8527 either case, and on platforms where pointers are larger than 'long',
8528 mixing is done to help ensure the numeric 'hash' is usable even if it
8529 can't be guaranteed unique. The default mechanism is to use "&errno"
8530 as a pointer-based thread ID to distinguish between threads.
8532 Applications that want to provide their own thread IDs should now use
8533 CRYPTO_THREADID_set_callback() to register a callback that will call
8534 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8536 Note that ERR_remove_state() is now deprecated, because it is tied
8537 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8538 to free the current thread's error state should be replaced by
8539 ERR_remove_thread_state(NULL).
8541 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8542 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8543 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8544 application was previously providing a numeric thread callback that
8545 was inappropriate for distinguishing threads, then uniqueness might
8546 have been obtained with &errno that happened immediately in the
8547 intermediate development versions of OpenSSL; this is no longer the
8548 case, the numeric thread callback will now override the automatic use
8551 *Geoff Thorpe, with help from Bodo Moeller*
8553 * Initial support for different CRL issuing certificates. This covers a
8554 simple case where the self issued certificates in the chain exist and
8555 the real CRL issuer is higher in the existing chain.
8557 This work was sponsored by Google.
8561 * Removed effectively defunct crypto/store from the build.
8565 * Revamp of STACK to provide stronger type-checking. Still to come:
8566 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8567 ASN1_STRING, CONF_VALUE.
8571 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8572 RAM on SSL connections. This option can save about 34k per idle SSL.
8576 * Revamp of LHASH to provide stronger type-checking. Still to come:
8577 STACK, TXT_DB, bsearch, qsort.
8581 * Initial support for Cryptographic Message Syntax (aka CMS) based
8582 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8583 support for data, signedData, compressedData, digestedData and
8584 encryptedData, envelopedData types included. Scripts to check against
8585 RFC4134 examples draft and interop and consistency checks of many
8586 content types and variants.
8590 * Add options to enc utility to support use of zlib compression BIO.
8594 * Extend mk1mf to support importing of options and assembly language
8595 files from Configure script, currently only included in VC-WIN32.
8596 The assembly language rules can now optionally generate the source
8597 files from the associated perl scripts.
8601 * Implement remaining functionality needed to support GOST ciphersuites.
8602 Interop testing has been performed using CryptoPro implementations.
8604 *Victor B. Wagner <vitus@cryptocom.ru>*
8606 * s390x assembler pack.
8610 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8615 * Implement Opaque PRF Input TLS extension as specified in
8616 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8617 official specification yet and no extension type assignment by
8618 IANA exists, this extension (for now) will have to be explicitly
8619 enabled when building OpenSSL by providing the extension number
8620 to use. For example, specify an option
8622 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8624 to the "config" or "Configure" script to enable the extension,
8625 assuming extension number 0x9527 (which is a completely arbitrary
8626 and unofficial assignment based on the MD5 hash of the Internet
8627 Draft). Note that by doing so, you potentially lose
8628 interoperability with other TLS implementations since these might
8629 be using the same extension number for other purposes.
8631 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8632 opaque PRF input value to use in the handshake. This will create
8633 an internal copy of the length-'len' string at 'src', and will
8634 return non-zero for success.
8636 To get more control and flexibility, provide a callback function
8639 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8640 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8644 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8647 Callback function 'cb' will be called in handshakes, and is
8648 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8649 Argument 'arg' is for application purposes (the value as given to
8650 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8651 be provided to the callback function). The callback function
8652 has to return non-zero to report success: usually 1 to use opaque
8653 PRF input just if possible, or 2 to enforce use of the opaque PRF
8654 input. In the latter case, the library will abort the handshake
8655 if opaque PRF input is not successfully negotiated.
8657 Arguments 'peerinput' and 'len' given to the callback function
8658 will always be NULL and 0 in the case of a client. A server will
8659 see the client's opaque PRF input through these variables if
8660 available (NULL and 0 otherwise). Note that if the server
8661 provides an opaque PRF input, the length must be the same as the
8662 length of the client's opaque PRF input.
8664 Note that the callback function will only be called when creating
8665 a new session (session resumption can resume whatever was
8666 previously negotiated), and will not be called in SSL 2.0
8667 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8668 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8669 for applications that need to enforce opaque PRF input.
8673 * Update ssl code to support digests other than SHA1+MD5 for handshake
8676 *Victor B. Wagner <vitus@cryptocom.ru>*
8678 * Add RFC4507 support to OpenSSL. This includes the corrections in
8679 RFC4507bis. The encrypted ticket format is an encrypted encoded
8680 SSL_SESSION structure, that way new session features are automatically
8683 If a client application caches session in an SSL_SESSION structure
8684 support is transparent because tickets are now stored in the encoded
8687 The SSL_CTX structure automatically generates keys for ticket
8688 protection in servers so again support should be possible
8689 with no application modification.
8691 If a client or server wishes to disable RFC4507 support then the option
8692 SSL_OP_NO_TICKET can be set.
8694 Add a TLS extension debugging callback to allow the contents of any client
8695 or server extensions to be examined.
8697 This work was sponsored by Google.
8701 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8702 OpenSSL should now compile cleanly on gcc 4.2
8704 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8706 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8707 support including streaming MAC support: this is required for GOST
8708 ciphersuite support.
8710 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8712 * Add option -stream to use PKCS#7 streaming in smime utility. New
8713 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8714 to output in BER and PEM format.
8718 * Experimental support for use of HMAC via EVP_PKEY interface. This
8719 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8720 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8721 ENGINE support for HMAC keys which are unextractable. New -mac and
8722 -macopt options to dgst utility.
8726 * New option -sigopt to dgst utility. Update dgst to use
8727 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8728 alternative signing parameters such as X9.31 or PSS in the dgst
8733 * Change ssl_cipher_apply_rule(), the internal function that does
8734 the work each time a ciphersuite string requests enabling
8735 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8736 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8737 the order of disabled ciphersuites such that those ciphersuites
8738 that most recently went from enabled to disabled not only stay
8739 in order with respect to each other, but also have higher priority
8740 than other disabled ciphersuites the next time ciphersuites are
8743 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8744 the same ciphersuites as with "HIGH" alone, but in a specific
8745 order where the PSK ciphersuites come first (since they are the
8746 most recently disabled ciphersuites when "HIGH" is parsed).
8748 Also, change ssl_create_cipher_list() (using this new
8749 functionality) such that between otherwise identical
8750 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8755 * Change ssl_create_cipher_list() so that it automatically
8756 arranges the ciphersuites in reasonable order before starting
8757 to process the rule string. Thus, the definition for "DEFAULT"
8758 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8759 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8760 This makes it much easier to arrive at a reasonable default order
8761 in applications for which anonymous ciphers are OK (meaning
8762 that you can't actually use DEFAULT).
8764 *Bodo Moeller; suggested by Victor Duchovni*
8766 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8767 processing) into multiple integers instead of setting
8768 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8769 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8770 (These masks as well as the individual bit definitions are hidden
8771 away into the non-exported interface ssl/ssl_locl.h, so this
8772 change to the definition of the SSL_CIPHER structure shouldn't
8773 affect applications.) This give us more bits for each of these
8774 categories, so there is no longer a need to coagulate AES128 and
8775 AES256 into a single algorithm bit, and to coagulate Camellia128
8776 and Camellia256 into a single algorithm bit, which has led to all
8779 Thus, among other things, the kludge introduced in 0.9.7m and
8780 0.9.8e for masking out AES256 independently of AES128 or masking
8781 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8783 With the change, we also introduce new ciphersuite aliases that
8784 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8789 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8790 Use the leftmost N bytes of the signature input if the input is
8791 larger than the prime q (with N being the size in bytes of q).
8795 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8796 it yet and it is largely untested.
8800 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8804 * Initial incomplete changes to avoid need for function casts in OpenSSL
8805 some compilers (gcc 4.2 and later) reject their use. Safestack is
8806 reimplemented. Update ASN1 to avoid use of legacy functions.
8810 * Win32/64 targets are linked with Winsock2.
8814 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8815 to external functions. This can be used to increase CRL handling
8816 efficiency especially when CRLs are very large by (for example) storing
8817 the CRL revoked certificates in a database.
8821 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8822 new CRLs added to a directory can be used. New command line option
8823 -verify_return_error to s_client and s_server. This causes real errors
8824 to be returned by the verify callback instead of carrying on no matter
8825 what. This reflects the way a "real world" verify callback would behave.
8829 * GOST engine, supporting several GOST algorithms and public key formats.
8830 Kindly donated by Cryptocom.
8834 * Partial support for Issuing Distribution Point CRL extension. CRLs
8835 partitioned by DP are handled but no indirect CRL or reason partitioning
8836 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8837 selected via a scoring technique which handles IDP and AKID in CRLs.
8841 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8842 will ultimately be used for all verify operations: this will remove the
8843 X509_STORE dependency on certificate verification and allow alternative
8844 lookup methods. X509_STORE based implementations of these two callbacks.
8848 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8849 Modify get_crl() to find a valid (unexpired) CRL if possible.
8853 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8854 this would be called X509_CRL_cmp() but that name is already used by
8855 a function that just compares CRL issuer names. Cache several CRL
8856 extensions in X509_CRL structure and cache CRLDP in X509.
8860 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8861 this maps equivalent X509_NAME structures into a consistent structure.
8862 Name comparison can then be performed rapidly using memcmp().
8866 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8871 * Allow digests to supply their own micalg string for S/MIME type using
8872 the ctrl EVP_MD_CTRL_MICALG.
8876 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8877 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8878 ctrl. It can then customise the structure before and/or after signing
8883 * New function OBJ_add_sigid() to allow application defined signature OIDs
8884 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8885 to free up any added signature OIDs.
8889 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8890 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8891 digest and cipher tables. New options added to openssl utility:
8892 list-message-digest-algorithms and list-cipher-algorithms.
8896 * Change the array representation of binary polynomials: the list
8897 of degrees of non-zero coefficients is now terminated with -1.
8898 Previously it was terminated with 0, which was also part of the
8899 value; thus, the array representation was not applicable to
8900 polynomials where t^0 has coefficient zero. This change makes
8901 the array representation useful in a more general context.
8905 * Various modifications and fixes to SSL/TLS cipher string
8906 handling. For ECC, the code now distinguishes between fixed ECDH
8907 with RSA certificates on the one hand and with ECDSA certificates
8908 on the other hand, since these are separate ciphersuites. The
8909 unused code for Fortezza ciphersuites has been removed.
8911 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8912 (not "ECDHE"). For consistency with the code for DH
8913 certificates, use of ECDH certificates is now considered ECDH
8914 authentication, not RSA or ECDSA authentication (the latter is
8915 merely the CA's signing algorithm and not actively used in the
8918 The temporary ciphersuite alias "ECCdraft" is no longer
8919 available, and ECC ciphersuites are no longer excluded from "ALL"
8920 and "DEFAULT". The following aliases now exist for RFC 4492
8921 ciphersuites, most of these by analogy with the DH case:
8923 kECDHr - ECDH cert, signed with RSA
8924 kECDHe - ECDH cert, signed with ECDSA
8925 kECDH - ECDH cert (signed with either RSA or ECDSA)
8926 kEECDH - ephemeral ECDH
8927 ECDH - ECDH cert or ephemeral ECDH
8933 AECDH - anonymous ECDH
8934 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8938 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8939 Use correct micalg parameters depending on digest(s) in signed message.
8943 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8944 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8948 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8949 an engine to register a method. Add ENGINE lookups for methods and
8950 functional reference processing.
8954 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8955 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8960 * New -resign option to smime utility. This adds one or more signers
8961 to an existing PKCS#7 signedData structure. Also -md option to use an
8962 alternative message digest algorithm for signing.
8966 * Tidy up PKCS#7 routines and add new functions to make it easier to
8967 create PKCS7 structures containing multiple signers. Update smime
8968 application to support multiple signers.
8972 * New -macalg option to pkcs12 utility to allow setting of an alternative
8977 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8978 Reorganize PBE internals to lookup from a static table using NIDs,
8979 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8980 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8981 PRF which will be automatically used with PBES2.
8985 * Replace the algorithm specific calls to generate keys in "req" with the
8990 * Update PKCS#7 enveloped data routines to use new API. This is now
8991 supported by any public key method supporting the encrypt operation. A
8992 ctrl is added to allow the public key algorithm to examine or modify
8993 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8998 * Add a ctrl to asn1 method to allow a public key algorithm to express
8999 a default digest type to use. In most cases this will be SHA1 but some
9000 algorithms (such as GOST) need to specify an alternative digest. The
9001 return value indicates how strong the preference is 1 means optional and
9002 2 is mandatory (that is it is the only supported type). Modify
9003 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9004 use the default md. Update openssl utilities to use the default digest
9005 type for signing if it is not explicitly indicated.
9009 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9010 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9011 signing method from the key type. This effectively removes the link
9012 between digests and public key types.
9016 * Add an OID cross reference table and utility functions. Its purpose is to
9017 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9018 rsaEncryption. This will allow some of the algorithm specific hackery
9019 needed to use the correct OID to be removed.
9023 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9024 structures for PKCS7_sign(). They are now set up by the relevant public
9029 * Add provisional EC pkey method with support for ECDSA and ECDH.
9033 * Add support for key derivation (agreement) in the API, DH method and
9038 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9039 public and private key formats. As a side effect these add additional
9040 command line functionality not previously available: DSA signatures can be
9041 generated and verified using pkeyutl and DH key support and generation in
9048 *Oliver Tappe <zooey@hirschkaefer.de>*
9050 * New make target "install_html_docs" installs HTML renditions of the
9053 *Oliver Tappe <zooey@hirschkaefer.de>*
9055 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9056 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9057 support key and parameter generation and add initial key generation
9058 functionality for RSA.
9062 * Add functions for main EVP_PKEY_method operations. The undocumented
9063 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9064 `EVP_PKEY_{encrypt,decrypt}_old`.
9068 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9069 key API, doesn't do much yet.
9073 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9074 public key algorithms. New option to openssl utility:
9075 "list-public-key-algorithms" to print out info.
9079 * Implement the Supported Elliptic Curves Extension for
9080 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9084 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9085 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9089 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9090 utilities such as rsa, dsa, dsaparam etc except they process any key
9095 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9096 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9097 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9102 * Initial support for pluggable public key ASN1.
9103 De-spaghettify the public key ASN1 handling. Move public and private
9104 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9105 algorithm specific handling to a single module within the relevant
9106 algorithm directory. Add functions to allow (near) opaque processing
9107 of public and private key structures.
9111 * Implement the Supported Point Formats Extension for
9112 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9116 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9117 for the psk identity [hint] and the psk callback functions to the
9118 SSL_SESSION, SSL and SSL_CTX structure.
9121 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9125 SSL_CTX_use_psk_identity_hint
9126 SSL_get_psk_identity_hint
9127 SSL_get_psk_identity
9128 SSL_use_psk_identity_hint
9130 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9132 * Add RFC 3161 compliant time stamp request creation, response generation
9133 and response verification functionality.
9135 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9137 * Add initial support for TLS extensions, specifically for the server_name
9138 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9139 have new members for a host name. The SSL data structure has an
9140 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9141 stored in that context to allow for session resumption, even after the
9142 SSL has been switched to a new SSL_CTX in reaction to a client's
9143 server_name extension.
9145 New functions (subject to change):
9147 SSL_get_servername()
9148 SSL_get_servername_type()
9151 New CTRL codes and macros (subject to change):
9153 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9154 - SSL_CTX_set_tlsext_servername_callback()
9155 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9156 - SSL_CTX_set_tlsext_servername_arg()
9157 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9159 openssl s_client has a new '-servername ...' option.
9161 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9162 '-key2 ...', '-servername_fatal' (subject to change). This allows
9163 testing the HostName extension for a specific single host name ('-cert'
9164 and '-key' remain fallbacks for handshakes without HostName
9165 negotiation). If the unrecognized_name alert has to be sent, this by
9166 default is a warning; it becomes fatal with the '-servername_fatal'
9169 *Peter Sylvester, Remy Allais, Christophe Renou*
9171 * Whirlpool hash implementation is added.
9175 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9176 bn(64,32). Because of instruction set limitations it doesn't have
9177 any negative impact on performance. This was done mostly in order
9178 to make it possible to share assembler modules, such as bn_mul_mont
9179 implementations, between 32- and 64-bit builds without hassle.
9183 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9184 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9189 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9190 dedicated Montgomery multiplication procedure, is introduced.
9191 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9192 "64-bit" performance on certain 32-bit targets.
9196 * New option SSL_OP_NO_COMP to disable use of compression selectively
9197 in SSL structures. New SSL ctrl to set maximum send fragment size.
9198 Save memory by setting the I/O buffer sizes dynamically instead of
9199 using the maximum available value.
9203 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9204 in addition to the text details.
9208 * Very, very preliminary EXPERIMENTAL support for printing of general
9209 ASN1 structures. This currently produces rather ugly output and doesn't
9210 handle several customised structures at all.
9214 * Integrated support for PVK file format and some related formats such
9215 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9216 these in the 'rsa' and 'dsa' utilities.
9220 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9224 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9225 place for the (very old) "NETSCAPE" format certificates which are now
9226 handled using new ASN1 code equivalents.
9230 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9231 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9232 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9236 * Modify CRL distribution points extension code to print out previously
9237 unsupported fields. Enhance extension setting code to allow setting of
9242 * Add print and set support for Issuing Distribution Point CRL extension.
9246 * Change 'Configure' script to enable Camellia by default.
9253 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9255 * When rejecting SSL/TLS records due to an incorrect version number, never
9256 update s->server with a new major version number. As of
9257 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9258 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9259 the previous behavior could result in a read attempt at NULL when
9260 receiving specific incorrect SSL/TLS records once record payload
9261 protection is active. ([CVE-2010-0740])
9263 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9265 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9266 could be crashed if the relevant tables were not present (e.g. chrooted).
9268 *Tomas Hoger <thoger@redhat.com>*
9270 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9272 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9274 *Martin Olsson, Neel Mehta*
9276 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9277 accommodate for stack sorting, always a write lock!).
9281 * On some versions of WIN32 Heap32Next is very slow. This can cause
9282 excessive delays in the RAND_poll(): over a minute. As a workaround
9283 include a time check in the inner Heap32Next loop too.
9287 * The code that handled flushing of data in SSL/TLS originally used the
9288 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9289 the problem outlined in PR#1949. The fix suggested there however can
9290 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9291 of Apache). So instead simplify the code to flush unconditionally.
9292 This should be fine since flushing with no data to flush is a no op.
9296 * Handle TLS versions 2.0 and later properly and correctly use the
9297 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9298 off ancient servers have a habit of sticking around for a while...
9302 * Modify compression code so it frees up structures without using the
9303 ex_data callbacks. This works around a problem where some applications
9304 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9305 restarting) then use compression (e.g. SSL with compression) later.
9306 This results in significant per-connection memory leaks and
9307 has caused some security issues including CVE-2008-1678 and
9312 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9313 change when encrypting or decrypting.
9317 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9318 connect and renegotiate with servers which do not support RI.
9319 Until RI is more widely deployed this option is enabled by default.
9323 * Add "missing" ssl ctrls to clear options and mode.
9327 * If client attempts to renegotiate and doesn't support RI respond with
9328 a no_renegotiation alert as required by RFC5746. Some renegotiating
9329 TLS clients will continue a connection gracefully when they receive
9330 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9331 waiting for a server hello which it will never receive. Now we treat a
9332 received no_renegotiation alert as a fatal error. This is because
9333 applications requesting a renegotiation might well expect it to succeed
9334 and would have no code in place to handle the server denying it so the
9335 only safe thing to do is to terminate the connection.
9339 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9340 peer supports secure renegotiation and 0 otherwise. Print out peer
9341 renegotiation support in s_client/s_server.
9345 * Replace the highly broken and deprecated SPKAC certification method with
9346 the updated NID creation version. This should correctly handle UTF8.
9350 * Implement RFC5746. Re-enable renegotiation but require the extension
9351 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9352 turns out to be a bad idea. It has been replaced by
9353 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9354 SSL_CTX_set_options(). This is really not recommended unless you
9355 know what you are doing.
9357 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9359 * Fixes to stateless session resumption handling. Use initial_ctx when
9360 issuing and attempting to decrypt tickets in case it has changed during
9361 servername handling. Use a non-zero length session ID when attempting
9362 stateless session resumption: this makes it possible to determine if
9363 a resumption has occurred immediately after receiving server hello
9364 (several places in OpenSSL subtly assume this) instead of later in
9369 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9370 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9371 fixes for a few places where the return code is not checked
9374 *Julia Lawall <julia@diku.dk>*
9376 * Add --strict-warnings option to Configure script to include devteam
9377 warnings in other configurations.
9381 * Add support for --libdir option and LIBDIR variable in makefiles. This
9382 makes it possible to install openssl libraries in locations which
9383 have names other than "lib", for example "/usr/lib64" which some
9386 *Steve Henson, based on patch from Jeremy Utley*
9388 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9389 X690 8.9.12 and can produce some misleading textual output of OIDs.
9391 *Steve Henson, reported by Dan Kaminsky*
9393 * Delete MD2 from algorithm tables. This follows the recommendation in
9394 several standards that it is not used in new applications due to
9395 several cryptographic weaknesses. For binary compatibility reasons
9396 the MD2 API is still compiled in by default.
9400 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9405 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9406 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9409 *Guenter <lists@gknw.net>*
9411 * Fix the server certificate chain building code to use X509_verify_cert(),
9412 it used to have an ad-hoc builder which was unable to cope with anything
9413 other than a simple chain.
9415 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9417 * Don't check self signed certificate signatures in X509_verify_cert()
9418 by default (a flag can override this): it just wastes time without
9419 adding any security. As a useful side effect self signed root CAs
9420 with non-FIPS digests are now usable in FIPS mode.
9424 * In dtls1_process_out_of_seq_message() the check if the current message
9425 is already buffered was missing. For every new message was memory
9426 allocated, allowing an attacker to perform an denial of service attack
9427 with sending out of seq handshake messages until there is no memory
9428 left. Additionally every future message was buffered, even if the
9429 sequence number made no sense and would be part of another handshake.
9430 So only messages with sequence numbers less than 10 in advance will be
9431 buffered. ([CVE-2009-1378])
9433 *Robin Seggelmann, discovered by Daniel Mentz*
9435 * Records are buffered if they arrive with a future epoch to be
9436 processed after finishing the corresponding handshake. There is
9437 currently no limitation to this buffer allowing an attacker to perform
9438 a DOS attack with sending records with future epochs until there is no
9439 memory left. This patch adds the pqueue_size() function to determine
9440 the size of a buffer and limits the record buffer to 100 entries.
9443 *Robin Seggelmann, discovered by Daniel Mentz*
9445 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9446 parent structure is freed. ([CVE-2009-1379])
9450 * Handle non-blocking I/O properly in SSL_shutdown() call.
9452 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9454 * Add `2.5.4.*` OIDs
9456 *Ilya O. <vrghost@gmail.com>*
9458 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9460 * Disable renegotiation completely - this fixes a severe security
9461 problem ([CVE-2009-3555]) at the cost of breaking all
9462 renegotiation. Renegotiation can be re-enabled by setting
9463 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9464 run-time. This is really not recommended unless you know what
9469 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9471 * Don't set val to NULL when freeing up structures, it is freed up by
9472 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9473 zeroing past the valid field. ([CVE-2009-0789])
9475 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9477 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9478 checked correctly. This would allow some invalid signed attributes to
9479 appear to verify correctly. ([CVE-2009-0591])
9481 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9483 * Reject UniversalString and BMPString types with invalid lengths. This
9484 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9485 a legal length. ([CVE-2009-0590])
9489 * Set S/MIME signing as the default purpose rather than setting it
9490 unconditionally. This allows applications to override it at the store
9495 * Permit restricted recursion of ASN1 strings. This is needed in practice
9496 to handle some structures.
9500 * Improve efficiency of mem_gets: don't search whole buffer each time
9503 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9505 * New -hex option for openssl rand.
9509 * Print out UTF8String and NumericString when parsing ASN1.
9513 * Support NumericString type for name components.
9517 * Allow CC in the environment to override the automatically chosen
9518 compiler. Note that nothing is done to ensure flags work with the
9523 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9525 * Properly check EVP_VerifyFinal() and similar return values
9528 *Ben Laurie, Bodo Moeller, Google Security Team*
9530 * Enable TLS extensions by default.
9534 * Allow the CHIL engine to be loaded, whether the application is
9535 multithreaded or not. (This does not release the developer from the
9536 obligation to set up the dynamic locking callbacks.)
9538 *Sander Temme <sander@temme.net>*
9540 * Use correct exit code if there is an error in dgst command.
9542 *Steve Henson; problem pointed out by Roland Dirlewanger*
9544 * Tweak Configure so that you need to say "experimental-jpake" to enable
9545 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9549 * Add experimental JPAKE support, including demo authentication in
9550 s_client and s_server.
9554 * Set the comparison function in v3_addr_canonize().
9556 *Rob Austein <sra@hactrn.net>*
9558 * Add support for XMPP STARTTLS in s_client.
9560 *Philip Paeps <philip@freebsd.org>*
9562 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9563 to ensure that even with this option, only ciphersuites in the
9564 server's preference list will be accepted. (Note that the option
9565 applies only when resuming a session, so the earlier behavior was
9566 just about the algorithm choice for symmetric cryptography.)
9570 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9572 * Fix NULL pointer dereference if a DTLS server received
9573 ChangeCipherSpec as first record ([CVE-2009-1386]).
9577 * Fix a state transition in s3_srvr.c and d1_srvr.c
9578 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9582 * The fix in 0.9.8c that supposedly got rid of unsafe
9583 double-checked locking was incomplete for RSA blinding,
9584 addressing just one layer of what turns out to have been
9585 doubly unsafe triple-checked locking.
9587 So now fix this for real by retiring the MONT_HELPER macro
9588 in crypto/rsa/rsa_eay.c.
9590 *Bodo Moeller; problem pointed out by Marius Schilder*
9592 * Various precautionary measures:
9594 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9596 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9597 (NB: This would require knowledge of the secret session ticket key
9598 to exploit, in which case you'd be SOL either way.)
9600 - Change bn_nist.c so that it will properly handle input BIGNUMs
9601 outside the expected range.
9603 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9606 *Neel Mehta, Bodo Moeller*
9608 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9609 the load fails. Useful for distros.
9611 *Ben Laurie and the FreeBSD team*
9613 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9617 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9621 * Expand ENGINE to support engine supplied SSL client certificate functions.
9623 This work was sponsored by Logica.
9627 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9628 keystores. Support for SSL/TLS client authentication too.
9629 Not compiled unless enable-capieng specified to Configure.
9631 This work was sponsored by Logica.
9635 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9636 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9637 attribute creation routines such as certificate requests and PKCS#12
9642 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9644 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9645 handshake which could lead to a client crash as found using the
9646 Codenomicon TLS test suite ([CVE-2008-1672])
9648 *Steve Henson, Mark Cox*
9650 * Fix double free in TLS server name extensions which could lead to
9651 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9655 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9657 Clear the error queue to ensure that error entries left from
9658 older function calls do not interfere with the correct operation.
9660 *Lutz Jaenicke, Erik de Castro Lopo*
9662 * Remove root CA certificates of commercial CAs:
9664 The OpenSSL project does not recommend any specific CA and does not
9665 have any policy with respect to including or excluding any CA.
9666 Therefore it does not make any sense to ship an arbitrary selection
9667 of root CA certificates with the OpenSSL software.
9671 * RSA OAEP patches to fix two separate invalid memory reads.
9672 The first one involves inputs when 'lzero' is greater than
9673 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9674 before the beginning of from). The second one involves inputs where
9675 the 'db' section contains nothing but zeroes (there is a one-byte
9676 invalid read after the end of 'db').
9678 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9680 * Partial backport from 0.9.9-dev:
9682 Introduce bn_mul_mont (dedicated Montgomery multiplication
9683 procedure) as a candidate for BIGNUM assembler implementation.
9684 While 0.9.9-dev uses assembler for various architectures, only
9685 x86_64 is available by default here in the 0.9.8 branch, and
9686 32-bit x86 is available through a compile-time setting.
9688 To try the 32-bit x86 assembler implementation, use Configure
9689 option "enable-montasm" (which exists only for this backport).
9691 As "enable-montasm" for 32-bit x86 disclaims code stability
9692 anyway, in this constellation we activate additional code
9693 backported from 0.9.9-dev for further performance improvements,
9694 namely BN_from_montgomery_word. (To enable this otherwise,
9695 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9697 *Andy Polyakov (backport partially by Bodo Moeller)*
9699 * Add TLS session ticket callback. This allows an application to set
9700 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9701 values. This is useful for key rollover for example where several key
9702 sets may exist with different names.
9706 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9707 This was broken until now in 0.9.8 releases, such that the only way
9708 a registered ENGINE could be used (assuming it initialises
9709 successfully on the host) was to explicitly set it as the default
9710 for the relevant algorithms. This is in contradiction with 0.9.7
9711 behaviour and the documentation. With this fix, when an ENGINE is
9712 registered into a given algorithm's table of implementations, the
9713 'uptodate' flag is reset so that auto-discovery will be used next
9714 time a new context for that algorithm attempts to select an
9717 *Ian Lister (tweaked by Geoff Thorpe)*
9719 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9720 implementation in the following ways:
9722 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9725 Lack of BER streaming support means one pass streaming processing is
9726 only supported if data is detached: setting the streaming flag is
9727 ignored for embedded content.
9729 CMS support is disabled by default and must be explicitly enabled
9730 with the enable-cms configuration option.
9734 * Update the GMP engine glue to do direct copies between BIGNUM and
9735 mpz_t when openssl and GMP use the same limb size. Otherwise the
9736 existing "conversion via a text string export" trick is still used.
9738 *Paul Sheer <paulsheer@gmail.com>*
9740 * Zlib compression BIO. This is a filter BIO which compressed and
9741 uncompresses any data passed through it.
9745 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9746 RFC3394 compatible AES key wrapping.
9750 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9751 sets string data without copying. X509_ALGOR_set0() and
9752 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9753 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9754 from an X509_ATTRIBUTE structure optionally checking it occurs only
9755 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9760 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9761 to get the expected BN_FLG_CONSTTIME behavior.
9763 *Bodo Moeller (Google)*
9767 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9768 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9769 - added some more tests to do_tests.pl
9770 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9771 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9772 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9773 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9774 - various changes to netware.pl to enable gcc-cross builds on Win32
9776 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9777 - various changes to fix missing prototype warnings
9778 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9779 - added AES, WHIRLPOOL and CPUID assembler code to build files
9780 - added missing AES assembler make rules to mk1mf.pl
9781 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9783 *Guenter Knauf <eflash@gmx.net>*
9785 * Implement certificate status request TLS extension defined in RFC3546.
9786 A client can set the appropriate parameters and receive the encoded
9787 OCSP response via a callback. A server can query the supplied parameters
9788 and set the encoded OCSP response in the callback. Add simplified examples
9789 to s_client and s_server.
9793 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9796 + Binary incompatibility of ssl_ctx_st structure
9797 + DTLS interoperation with non-compliant servers
9798 + Don't call get_session_cb() without proposed session
9799 + Fix ia64 assembler code
9801 *Andy Polyakov, Steve Henson*
9803 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9805 * DTLS Handshake overhaul. There were longstanding issues with
9806 OpenSSL DTLS implementation, which were making it impossible for
9807 RFC 4347 compliant client to communicate with OpenSSL server.
9808 Unfortunately just fixing these incompatibilities would "cut off"
9809 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9810 server keeps tolerating non RFC compliant syntax. The opposite is
9811 not true, 0.9.8f client can not communicate with earlier server.
9812 This update even addresses CVE-2007-4995.
9816 * Changes to avoid need for function casts in OpenSSL: some compilers
9817 (gcc 4.2 and later) reject their use.
9818 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9821 * Add RFC4507 support to OpenSSL. This includes the corrections in
9822 RFC4507bis. The encrypted ticket format is an encrypted encoded
9823 SSL_SESSION structure, that way new session features are automatically
9826 If a client application caches session in an SSL_SESSION structure
9827 support is transparent because tickets are now stored in the encoded
9830 The SSL_CTX structure automatically generates keys for ticket
9831 protection in servers so again support should be possible
9832 with no application modification.
9834 If a client or server wishes to disable RFC4507 support then the option
9835 SSL_OP_NO_TICKET can be set.
9837 Add a TLS extension debugging callback to allow the contents of any client
9838 or server extensions to be examined.
9840 This work was sponsored by Google.
9844 * Add initial support for TLS extensions, specifically for the server_name
9845 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9846 have new members for a host name. The SSL data structure has an
9847 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9848 stored in that context to allow for session resumption, even after the
9849 SSL has been switched to a new SSL_CTX in reaction to a client's
9850 server_name extension.
9852 New functions (subject to change):
9854 SSL_get_servername()
9855 SSL_get_servername_type()
9858 New CTRL codes and macros (subject to change):
9860 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9861 - SSL_CTX_set_tlsext_servername_callback()
9862 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9863 - SSL_CTX_set_tlsext_servername_arg()
9864 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9866 openssl s_client has a new '-servername ...' option.
9868 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9869 '-key2 ...', '-servername_fatal' (subject to change). This allows
9870 testing the HostName extension for a specific single host name ('-cert'
9871 and '-key' remain fallbacks for handshakes without HostName
9872 negotiation). If the unrecognized_name alert has to be sent, this by
9873 default is a warning; it becomes fatal with the '-servername_fatal'
9876 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9878 * Add AES and SSE2 assembly language support to VC++ build.
9882 * Mitigate attack on final subtraction in Montgomery reduction.
9886 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9887 (which previously caused an internal error).
9891 * Squeeze another 10% out of IGE mode when in != out.
9895 * AES IGE mode speedup.
9897 *Dean Gaudet (Google)*
9899 * Add the Korean symmetric 128-bit cipher SEED (see
9900 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9901 add SEED ciphersuites from RFC 4162:
9903 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9904 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9905 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9906 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9908 To minimize changes between patchlevels in the OpenSSL 0.9.8
9909 series, SEED remains excluded from compilation unless OpenSSL
9910 is configured with 'enable-seed'.
9912 *KISA, Bodo Moeller*
9914 * Mitigate branch prediction attacks, which can be practical if a
9915 single processor is shared, allowing a spy process to extract
9916 information. For detailed background information, see
9917 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9918 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9919 and Necessary Software Countermeasures"). The core of the change
9920 are new versions BN_div_no_branch() and
9921 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9922 respectively, which are slower, but avoid the security-relevant
9923 conditional branches. These are automatically called by BN_div()
9924 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9925 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9926 remove a conditional branch.
9928 BN_FLG_CONSTTIME is the new name for the previous
9929 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9930 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9931 in the exponent causes BN_mod_exp_mont() to use the alternative
9932 implementation in BN_mod_exp_mont_consttime().) The old name
9933 remains as a deprecated alias.
9935 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9936 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9937 constant-time implementations for more than just exponentiation.
9938 Here too the old name is kept as a deprecated alias.
9940 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9941 the BN_BLINDING structure gets an independent copy of the
9942 modulus. This means that the previous `BIGNUM *m` argument to
9943 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9944 essentially becomes `const BIGNUM *m`, although we can't actually
9945 change this in the header file before 0.9.9. It allows
9946 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9947 enable BN_FLG_CONSTTIME.
9949 *Matthew D Wood (Intel Corp)*
9951 * In the SSL/TLS server implementation, be strict about session ID
9952 context matching (which matters if an application uses a single
9953 external cache for different purposes). Previously,
9954 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9955 set. This did ensure strict client verification, but meant that,
9956 with applications using a single external cache for quite
9957 different requirements, clients could circumvent ciphersuite
9958 restrictions for a given session ID context by starting a session
9959 in a different context.
9963 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9964 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9965 authentication-only ciphersuites.
9969 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9970 not complete and could lead to a possible single byte overflow
9971 ([CVE-2007-5135]) [Ben Laurie]
9973 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9975 * Since AES128 and AES256 (and similarly Camellia128 and
9976 Camellia256) share a single mask bit in the logic of
9977 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9978 kludge to work properly if AES128 is available and AES256 isn't
9979 (or if Camellia128 is available and Camellia256 isn't).
9983 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9984 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9985 When a point or a seed is encoded in a BIT STRING, we need to
9986 prevent the removal of trailing zero bits to get the proper DER
9987 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9988 of a NamedBitList, for which trailing 0 bits need to be removed.)
9992 * Have SSL/TLS server implementation tolerate "mismatched" record
9993 protocol version while receiving ClientHello even if the
9994 ClientHello is fragmented. (The server can't insist on the
9995 particular protocol version it has chosen before the ServerHello
9996 message has informed the client about his choice.)
10000 * Add RFC 3779 support.
10002 *Rob Austein for ARIN, Ben Laurie*
10004 * Load error codes if they are not already present instead of using a
10005 static variable. This allows them to be cleanly unloaded and reloaded.
10006 Improve header file function name parsing.
10010 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10011 or CAPABILITY handshake as required by RFCs.
10013 *Goetz Babin-Ebell*
10015 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10017 * Introduce limits to prevent malicious keys being able to
10018 cause a denial of service. ([CVE-2006-2940])
10020 *Steve Henson, Bodo Moeller*
10022 * Fix ASN.1 parsing of certain invalid structures that can result
10023 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10025 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10026 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10028 * Fix SSL client code which could crash if connecting to a
10029 malicious SSLv2 server. ([CVE-2006-4343])
10031 *Tavis Ormandy and Will Drewry, Google Security Team*
10033 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10034 match only those. Before that, "AES256-SHA" would be interpreted
10035 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10036 the same strength classification in 0.9.7h) as we currently only
10037 have a single AES bit in the ciphersuite description bitmap.
10038 That change, however, also applied to ciphersuite strings such as
10039 "RC4-MD5" that intentionally matched multiple ciphersuites --
10040 namely, SSL 2.0 ciphersuites in addition to the more common ones
10041 from SSL 3.0/TLS 1.0.
10043 So we change the selection algorithm again: Naming an explicit
10044 ciphersuite selects this one ciphersuite, and any other similar
10045 ciphersuite (same bitmap) from *other* protocol versions.
10046 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10047 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10049 Since SSL 2.0 does not have any ciphersuites for which the
10050 128/256 bit distinction would be relevant, this works for now.
10051 The proper fix will be to use different bits for AES128 and
10052 AES256, which would have avoided the problems from the beginning;
10053 however, bits are scarce, so we can only do this in a new release
10054 (not just a patchlevel) when we can change the SSL_CIPHER
10055 definition to split the single 'unsigned long mask' bitmap into
10056 multiple values to extend the available space.
10060 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10062 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10063 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10065 * Add AES IGE and biIGE modes.
10069 * Change the Unix randomness entropy gathering to use poll() when
10070 possible instead of select(), since the latter has some
10071 undesirable limitations.
10073 *Darryl Miles via Richard Levitte and Bodo Moeller*
10075 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10076 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10077 cannot be implicitly activated as part of, e.g., the "AES" alias.
10078 However, please upgrade to OpenSSL 0.9.9[-dev] for
10079 non-experimental use of the ECC ciphersuites to get TLS extension
10080 support, which is required for curve and point format negotiation
10081 to avoid potential handshake problems.
10085 * Disable rogue ciphersuites:
10087 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10088 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10089 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10091 The latter two were purportedly from
10092 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10095 Also deactivate the remaining ciphersuites from
10096 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10097 unofficial, and the ID has long expired.
10101 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10102 dual-core machines) and other potential thread-safety issues.
10106 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10107 versions), which is now available for royalty-free use
10108 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10109 Also, add Camellia TLS ciphersuites from RFC 4132.
10111 To minimize changes between patchlevels in the OpenSSL 0.9.8
10112 series, Camellia remains excluded from compilation unless OpenSSL
10113 is configured with 'enable-camellia'.
10117 * Disable the padding bug check when compression is in use. The padding
10118 bug check assumes the first packet is of even length, this is not
10119 necessarily true if compression is enabled and can result in false
10120 positives causing handshake failure. The actual bug test is ancient
10121 code so it is hoped that implementations will either have fixed it by
10122 now or any which still have the bug do not support compression.
10126 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10128 * When applying a cipher rule check to see if string match is an explicit
10129 cipher suite and only match that one cipher suite if it is.
10133 * Link in manifests for VC++ if needed.
10135 *Austin Ziegler <halostatue@gmail.com>*
10137 * Update support for ECC-based TLS ciphersuites according to
10138 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10139 TLS extensions, which are supported starting with the 0.9.9
10140 branch, not in the OpenSSL 0.9.8 branch).
10144 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10145 opaque EVP_CIPHER_CTX handling.
10149 * Fixes and enhancements to zlib compression code. We now only use
10150 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10151 to conform with the standards mentioned here:
10152 <http://www.zlib.net/DLL_FAQ.txt>
10153 Static zlib linking now works on Windows and the new --with-zlib-include
10154 --with-zlib-lib options to Configure can be used to supply the location
10155 of the headers and library. Gracefully handle case where zlib library
10160 * Several fixes and enhancements to the OID generation code. The old code
10161 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10162 handle numbers larger than ULONG_MAX, truncated printing and had a
10163 non standard OBJ_obj2txt() behaviour.
10167 * Add support for building of engines under engine/ as shared libraries
10168 under VC++ build system.
10172 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10173 Hopefully, we will not see any false combination of paths any more.
10177 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10179 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10180 (part of SSL_OP_ALL). This option used to disable the
10181 countermeasure against man-in-the-middle protocol-version
10182 rollback in the SSL 2.0 server implementation, which is a bad
10183 idea. ([CVE-2005-2969])
10185 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10186 for Information Security, National Institute of Advanced Industrial
10187 Science and Technology [AIST], Japan)*
10189 * Add two function to clear and return the verify parameter flags.
10193 * Keep cipherlists sorted in the source instead of sorting them at
10194 runtime, thus removing the need for a lock.
10198 * Avoid some small subgroup attacks in Diffie-Hellman.
10200 *Nick Mathewson and Ben Laurie*
10202 * Add functions for well-known primes.
10206 * Extended Windows CE support.
10208 *Satoshi Nakamura and Andy Polyakov*
10210 * Initialize SSL_METHOD structures at compile time instead of during
10211 runtime, thus removing the need for a lock.
10215 * Make PKCS7_decrypt() work even if no certificate is supplied by
10216 attempting to decrypt each encrypted key in turn. Add support to
10221 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10223 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10226 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10230 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10231 key into the same file any more.
10235 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10239 * Add -utf8 command line and config file option to 'ca'.
10241 *Stefan <stf@udoma.org*
10243 * Removed the macro des_crypt(), as it seems to conflict with some
10244 libraries. Use DES_crypt().
10248 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10249 involves renaming the source and generated shared-libs for
10250 both. The engines will accept the corrected or legacy ids
10251 ('ncipher' and '4758_cca' respectively) when binding. NB,
10252 this only applies when building 'shared'.
10254 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10256 * Add attribute functions to EVP_PKEY structure. Modify
10257 PKCS12_create() to recognize a CSP name attribute and
10258 use it. Make -CSP option work again in pkcs12 utility.
10262 * Add new functionality to the bn blinding code:
10263 - automatic re-creation of the BN_BLINDING parameters after
10264 a fixed number of uses (currently 32)
10265 - add new function for parameter creation
10266 - introduce flags to control the update behaviour of the
10267 BN_BLINDING parameters
10268 - hide BN_BLINDING structure
10269 Add a second BN_BLINDING slot to the RSA structure to improve
10270 performance when a single RSA object is shared among several
10275 * Add support for DTLS.
10277 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10279 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10280 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10284 * Remove buggy and incomplete DH cert support from
10285 ssl/ssl_rsa.c and ssl/s3_both.c
10289 * Use SHA-1 instead of MD5 as the default digest algorithm for
10290 the `apps/openssl` commands.
10294 * Compile clean with "-Wall -Wmissing-prototypes
10295 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10296 DEBUG_SAFESTACK must also be set.
10300 * Change ./Configure so that certain algorithms can be disabled by default.
10301 The new counterpiece to "no-xxx" is "enable-xxx".
10303 The patented RC5 and MDC2 algorithms will now be disabled unless
10304 "enable-rc5" and "enable-mdc2", respectively, are specified.
10306 (IDEA remains enabled despite being patented. This is because IDEA
10307 is frequently required for interoperability, and there is no license
10308 fee for non-commercial use. As before, "no-idea" can be used to
10309 avoid this algorithm.)
10313 * Add processing of proxy certificates (see RFC 3820). This work was
10314 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10315 EGEE (Enabling Grids for E-science in Europe).
10319 * RC4 performance overhaul on modern architectures/implementations, such
10320 as Intel P4, IA-64 and AMD64.
10324 * New utility extract-section.pl. This can be used specify an alternative
10325 section number in a pod file instead of having to treat each file as
10326 a separate case in Makefile. This can be done by adding two lines to the
10329 =for comment openssl_section:XXX
10331 The blank line is mandatory.
10335 * New arguments -certform, -keyform and -pass for s_client and s_server
10336 to allow alternative format key and certificate files and passphrase
10341 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10342 update associated structures and add various utility functions.
10344 Add new policy related verify parameters, include policy checking in
10345 standard verify code. Enhance 'smime' application with extra parameters
10346 to support policy checking and print out.
10350 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10351 Nehemiah processors. These extensions support AES encryption in hardware
10352 as well as RNG (though RNG support is currently disabled).
10354 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10356 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10360 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10362 *Andy Polyakov and a number of other people*
10364 * Improved PowerPC platform support. Most notably BIGNUM assembler
10365 implementation contributed by IBM.
10367 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10369 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10370 exponent rather than 'unsigned long'. There is a corresponding change to
10371 the new 'rsa_keygen' element of the RSA_METHOD structure.
10373 *Jelte Jansen, Geoff Thorpe*
10375 * Functionality for creating the initial serial number file is now
10376 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10378 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10379 number file to 1, which is bound to cause problems. To avoid
10380 the problems while respecting compatibility between different 0.9.7
10381 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10382 CA.pl for serial number initialization. With the new release 0.9.8,
10383 we can fix the problem directly in the 'ca' utility.)
10387 * Reduced header interdependencies by declaring more opaque objects in
10388 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10389 give fewer recursive includes, which could break lazy source code - so
10390 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10391 developers should define this symbol when building and using openssl to
10392 ensure they track the recommended behaviour, interfaces, [etc], but
10393 backwards-compatible behaviour prevails when this isn't defined.
10397 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10401 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10402 This will generate a random key of the appropriate length based on the
10403 cipher context. The EVP_CIPHER can provide its own random key generation
10404 routine to support keys of a specific form. This is used in the des and
10405 3des routines to generate a key of the correct parity. Update S/MIME
10406 code to use new functions and hence generate correct parity DES keys.
10407 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10408 valid (weak or incorrect parity).
10412 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10413 as looking them up. This is useful when the verified structure may contain
10414 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10415 present unless the new PKCS7_NO_CRL flag is asserted.
10419 * Extend ASN1 oid configuration module. It now additionally accepts the
10422 shortName = some long name, 1.2.3.4
10426 * Reimplemented the BN_CTX implementation. There is now no more static
10427 limitation on the number of variables it can handle nor the depth of the
10428 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10429 information can now expand as required, and rather than having a single
10430 static array of bignums, BN_CTX now uses a linked-list of such arrays
10431 allowing it to expand on demand whilst maintaining the usefulness of
10432 BN_CTX's "bundling".
10436 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10437 to allow all RSA operations to function using a single BN_CTX.
10441 * Preliminary support for certificate policy evaluation and checking. This
10442 is initially intended to pass the tests outlined in "Conformance Testing
10443 of Relying Party Client Certificate Path Processing Logic" v1.07.
10447 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10448 remained unused and not that useful. A variety of other little bignum
10449 tweaks and fixes have also been made continuing on from the audit (see
10454 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10455 associated ASN1, EVP and SSL functions and old ASN1 macros.
10459 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10460 and this should never fail. So the return value from the use of
10461 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10462 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10466 * BN_CTX_get() should return zero-valued bignums, providing the same
10467 initialised value as BN_new().
10469 *Geoff Thorpe, suggested by Ulf Möller*
10471 * Support for inhibitAnyPolicy certificate extension.
10475 * An audit of the BIGNUM code is underway, for which debugging code is
10476 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10477 is considered valid when processing BIGNUMs, and causes execution to
10478 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10479 further steps are taken to deliberately pollute unused data in BIGNUM
10480 structures to try and expose faulty code further on. For now, openssl will
10481 (in its default mode of operation) continue to tolerate the inconsistent
10482 forms that it has tolerated in the past, but authors and packagers should
10483 consider trying openssl and their own applications when compiled with
10484 these debugging symbols defined. It will help highlight potential bugs in
10485 their own code, and will improve the test coverage for OpenSSL itself. At
10486 some point, these tighter rules will become openssl's default to improve
10487 maintainability, though the assert()s and other overheads will remain only
10488 in debugging configurations. See bn.h for more details.
10490 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10492 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10493 that can only be obtained through BN_CTX_new() (which implicitly
10494 initialises it). The presence of this function only made it possible
10495 to overwrite an existing structure (and cause memory leaks).
10499 * Because of the callback-based approach for implementing LHASH as a
10500 template type, lh_insert() adds opaque objects to hash-tables and
10501 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10502 to clean up those corresponding objects before destroying the hash table
10503 (and losing the object pointers). So some over-zealous constifications in
10504 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10505 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10506 prototyped to have "const" restrictions on the object pointers they are
10507 given (and so aren't required to cast them away any more).
10511 * The tmdiff.h API was so ugly and minimal that our own timing utility
10512 (speed) prefers to use its own implementation. The two implementations
10513 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10514 its object type properly exposed (MS_TM) instead of casting to/from
10515 `char *`. This may still change yet if someone realises MS_TM and
10517 aren't necessarily the greatest nomenclatures - but this is what was used
10518 internally to the implementation so I've used that for now.
10522 * Ensure that deprecated functions do not get compiled when
10523 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10524 the self-tests were still using deprecated key-generation functions so
10525 these have been updated also.
10529 * Reorganise PKCS#7 code to separate the digest location functionality
10530 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10531 New function PKCS7_set_digest() to set the digest type for PKCS#7
10532 digestedData type. Add additional code to correctly generate the
10533 digestedData type and add support for this type in PKCS7 initialization
10538 * New function PKCS7_set0_type_other() this initializes a PKCS7
10539 structure of type "other".
10543 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10544 sure the loop does correctly stop and breaking ("division by zero")
10545 modulus operations are not performed. The (pre-generated) prime
10546 table crypto/bn/bn_prime.h was already correct, but it could not be
10547 re-generated on some platforms because of the "division by zero"
10548 situation in the script.
10550 *Ralf S. Engelschall*
10552 * Update support for ECC-based TLS ciphersuites according to
10553 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10554 SHA-1 now is only used for "small" curves (where the
10555 representation of a field element takes up to 24 bytes); for
10556 larger curves, the field element resulting from ECDH is directly
10557 used as premaster secret.
10559 *Douglas Stebila (Sun Microsystems Laboratories)*
10561 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10562 curve secp160r1 to the tests.
10564 *Douglas Stebila (Sun Microsystems Laboratories)*
10566 * Add the possibility to load symbols globally with DSO.
10568 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10570 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10571 control of the error stack.
10575 * Add support for STORE in ENGINE.
10579 * Add the STORE type. The intention is to provide a common interface
10580 to certificate and key stores, be they simple file-based stores, or
10581 HSM-type store, or LDAP stores, or...
10582 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10586 * Add a generic structure called OPENSSL_ITEM. This can be used to
10587 pass a list of arguments to any function as well as provide a way
10588 for a function to pass data back to the caller.
10592 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10593 works like BUF_strdup() but can be used to duplicate a portion of
10594 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10599 * Add the function sk_find_ex() which works like sk_find(), but will
10600 return an index to an element even if an exact match couldn't be
10601 found. The index is guaranteed to point at the element where the
10602 searched-for key would be inserted to preserve sorting order.
10606 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10607 takes an extra flags argument for optional functionality. Currently,
10608 the following flags are defined:
10610 OBJ_BSEARCH_VALUE_ON_NOMATCH
10611 This one gets OBJ_bsearch_ex() to return a pointer to the first
10612 element where the comparing function returns a negative or zero
10615 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10616 This one gets OBJ_bsearch_ex() to return a pointer to the first
10617 element where the comparing function returns zero. This is useful
10618 if there are more than one element where the comparing function
10623 * Make it possible to create self-signed certificates with 'openssl ca'
10624 in such a way that the self-signed certificate becomes part of the
10625 CA database and uses the same mechanisms for serial number generation
10626 as all other certificate signing. The new flag '-selfsign' enables
10627 this functionality. Adapt CA.sh and CA.pl.in.
10631 * Add functionality to check the public key of a certificate request
10632 against a given private. This is useful to check that a certificate
10633 request can be signed by that key (self-signing).
10637 * Make it possible to have multiple active certificates with the same
10638 subject in the CA index file. This is done only if the keyword
10639 'unique_subject' is set to 'no' in the main CA section (default
10640 if 'CA_default') of the configuration file. The value is saved
10641 with the database itself in a separate index attribute file,
10642 named like the index file with '.attr' appended to the name.
10646 * Generate multi-valued AVAs using '+' notation in config files for
10651 * Support for nameConstraints certificate extension.
10655 * Support for policyConstraints certificate extension.
10659 * Support for policyMappings certificate extension.
10663 * Make sure the default DSA_METHOD implementation only uses its
10664 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10665 and change its own handlers to be NULL so as to remove unnecessary
10666 indirection. This lets alternative implementations fallback to the
10667 default implementation more easily.
10671 * Support for directoryName in GeneralName related extensions
10676 * Make it possible to link applications using Makefile.shared.
10677 Make that possible even when linking against static libraries!
10681 * Support for single pass processing for S/MIME signing. This now
10682 means that S/MIME signing can be done from a pipe, in addition
10683 cleartext signing (multipart/signed type) is effectively streaming
10684 and the signed data does not need to be all held in memory.
10686 This is done with a new flag PKCS7_STREAM. When this flag is set
10687 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10688 is done after the data is output (and digests calculated) in
10689 SMIME_write_PKCS7().
10693 * Add full support for -rpath/-R, both in shared libraries and
10694 applications, at least on the platforms where it's known how
10699 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10700 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10701 will now compute a table of multiples of the generator that
10702 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10703 faster (notably in the case of a single point multiplication,
10704 scalar * generator).
10706 *Nils Larsch, Bodo Moeller*
10708 * IPv6 support for certificate extensions. The various extensions
10709 which use the IP:a.b.c.d can now take IPv6 addresses using the
10710 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10715 * Added an ENGINE that implements RSA by performing private key
10716 exponentiations with the GMP library. The conversions to and from
10717 GMP's mpz_t format aren't optimised nor are any montgomery forms
10718 cached, and on x86 it appears OpenSSL's own performance has caught up.
10719 However there are likely to be other architectures where GMP could
10720 provide a boost. This ENGINE is not built in by default, but it can be
10721 specified at Configure time and should be accompanied by the necessary
10722 linker additions, eg;
10723 ./config -DOPENSSL_USE_GMP -lgmp
10727 * "openssl engine" will not display ENGINE/DSO load failure errors when
10728 testing availability of engines with "-t" - the old behaviour is
10729 produced by increasing the feature's verbosity with "-tt".
10733 * ECDSA routines: under certain error conditions uninitialized BN objects
10734 could be freed. Solution: make sure initialization is performed early
10735 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10740 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10741 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10742 software implementations. For DSA and DH, parameter generation can
10743 also be overridden by providing the appropriate method callbacks.
10747 * Change the "progress" mechanism used in key-generation and
10748 primality testing to functions that take a new BN_GENCB pointer in
10749 place of callback/argument pairs. The new API functions have `_ex`
10750 postfixes and the older functions are reimplemented as wrappers for
10751 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10752 declarations of the old functions to help (graceful) attempts to
10753 migrate to the new functions. Also, the new key-generation API
10754 functions operate on a caller-supplied key-structure and return
10755 success/failure rather than returning a key or NULL - this is to
10756 help make "keygen" another member function of RSA_METHOD etc.
10758 Example for using the new callback interface:
10760 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10761 void *my_arg = ...;
10764 BN_GENCB_set(&my_cb, my_callback, my_arg);
10766 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10767 /* For the meaning of a, b in calls to my_callback(), see the
10768 * documentation of the function that calls the callback.
10769 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10770 * my_callback should return 1 if it wants BN_is_prime_ex()
10771 * to continue, or 0 to stop.
10776 * Change the ZLIB compression method to be stateful, and make it
10777 available to TLS with the number defined in
10778 draft-ietf-tls-compression-04.txt.
10782 * Add the ASN.1 structures and functions for CertificatePair, which
10783 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10785 CertificatePair ::= SEQUENCE {
10786 forward [0] Certificate OPTIONAL,
10787 reverse [1] Certificate OPTIONAL,
10788 -- at least one of the pair shall be present -- }
10790 Also implement the PEM functions to read and write certificate
10791 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10793 This needed to be defined, mostly for the sake of the LDAP
10794 attribute crossCertificatePair, but may prove useful elsewhere as
10799 * Make it possible to inhibit symlinking of shared libraries in
10800 Makefile.shared, for Cygwin's sake.
10804 * Extend the BIGNUM API by creating a function
10805 void BN_set_negative(BIGNUM *a, int neg);
10806 and a macro that behave like
10807 int BN_is_negative(const BIGNUM *a);
10809 to avoid the need to access 'a->neg' directly in applications.
10813 * Implement fast modular reduction for pseudo-Mersenne primes
10814 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10815 EC_GROUP_new_curve_GFp() will now automatically use this
10818 *Nils Larsch <nla@trustcenter.de>*
10820 * Add new lock type (CRYPTO_LOCK_BN).
10824 * Change the ENGINE framework to automatically load engines
10825 dynamically from specific directories unless they could be
10826 found to already be built in or loaded. Move all the
10827 current engines except for the cryptodev one to a new
10828 directory engines/.
10829 The engines in engines/ are built as shared libraries if
10830 the "shared" options was given to ./Configure or ./config.
10831 Otherwise, they are inserted in libcrypto.a.
10832 /usr/local/ssl/engines is the default directory for dynamic
10833 engines, but that can be overridden at configure time through
10834 the usual use of --prefix and/or --openssldir, and at run
10835 time with the environment variable OPENSSL_ENGINES.
10837 *Geoff Thorpe and Richard Levitte*
10839 * Add Makefile.shared, a helper makefile to build shared
10840 libraries. Adapt Makefile.org.
10844 * Add version info to Win32 DLLs.
10846 *Peter 'Luna' Runestig" <peter@runestig.com>*
10848 * Add new 'medium level' PKCS#12 API. Certificates and keys
10849 can be added using this API to created arbitrary PKCS#12
10850 files while avoiding the low-level API.
10852 New options to PKCS12_create(), key or cert can be NULL and
10853 will then be omitted from the output file. The encryption
10854 algorithm NIDs can be set to -1 for no encryption, the mac
10855 iteration count can be set to 0 to omit the mac.
10857 Enhance pkcs12 utility by making the -nokeys and -nocerts
10858 options work when creating a PKCS#12 file. New option -nomac
10859 to omit the mac, NONE can be set for an encryption algorithm.
10860 New code is modified to use the enhanced PKCS12_create()
10861 instead of the low-level API.
10865 * Extend ASN1 encoder to support indefinite length constructed
10866 encoding. This can output sequences tags and octet strings in
10867 this form. Modify pk7_asn1.c to support indefinite length
10868 encoding. This is experimental and needs additional code to
10869 be useful, such as an ASN1 bio and some enhanced streaming
10872 Extend template encode functionality so that tagging is passed
10873 down to the template encoder.
10877 * Let 'openssl req' fail if an argument to '-newkey' is not
10878 recognized instead of using RSA as a default.
10882 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10883 As these are not official, they are not included in "ALL";
10884 the "ECCdraft" ciphersuite group alias can be used to select them.
10886 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10888 * Add ECDH engine support.
10890 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10892 * Add ECDH in new directory crypto/ecdh/.
10894 *Douglas Stebila (Sun Microsystems Laboratories)*
10896 * Let BN_rand_range() abort with an error after 100 iterations
10897 without success (which indicates a broken PRNG).
10901 * Change BN_mod_sqrt() so that it verifies that the input value
10902 is really the square of the return value. (Previously,
10903 BN_mod_sqrt would show GIGO behaviour.)
10907 * Add named elliptic curves over binary fields from X9.62, SECG,
10908 and WAP/WTLS; add OIDs that were still missing.
10910 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10912 * Extend the EC library for elliptic curves over binary fields
10913 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10916 EC_GF2m_simple_method
10920 EC_GROUP_new_curve_GF2m
10921 EC_GROUP_set_curve_GF2m
10922 EC_GROUP_get_curve_GF2m
10923 EC_POINT_set_affine_coordinates_GF2m
10924 EC_POINT_get_affine_coordinates_GF2m
10925 EC_POINT_set_compressed_coordinates_GF2m
10927 Point compression for binary fields is disabled by default for
10928 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10931 As binary polynomials are represented as BIGNUMs, various members
10932 of the EC_GROUP and EC_POINT data structures can be shared
10933 between the implementations for prime fields and binary fields;
10934 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10935 are essentially identical to their `..._GFp` counterparts.
10936 (For simplicity, the `..._GFp` prefix has been dropped from
10937 various internal method names.)
10939 An internal 'field_div' method (similar to 'field_mul' and
10940 'field_sqr') has been added; this is used only for binary fields.
10942 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10944 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10945 through methods ('mul', 'precompute_mult').
10947 The generic implementations (now internally called 'ec_wNAF_mul'
10948 and 'ec_wNAF_precomputed_mult') remain the default if these
10949 methods are undefined.
10951 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10953 * New function EC_GROUP_get_degree, which is defined through
10954 EC_METHOD. For curves over prime fields, this returns the bit
10955 length of the modulus.
10957 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10959 * New functions EC_GROUP_dup, EC_POINT_dup.
10960 (These simply call ..._new and ..._copy).
10962 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10964 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10965 Polynomials are represented as BIGNUMs (where the sign bit is not
10966 used) in the following functions [macros]:
10969 BN_GF2m_sub [= BN_GF2m_add]
10970 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10971 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10972 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10974 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10975 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10976 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10977 BN_GF2m_cmp [= BN_ucmp]
10979 (Note that only the 'mod' functions are actually for fields GF(2^m).
10980 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10982 For some functions, an the irreducible polynomial defining a
10983 field can be given as an 'unsigned int[]' with strictly
10984 decreasing elements giving the indices of those bits that are set;
10985 i.e., p[] represents the polynomial
10986 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10988 p[0] > p[1] > ... > p[k] = 0.
10989 This applies to the following functions:
10992 BN_GF2m_mod_mul_arr
10993 BN_GF2m_mod_sqr_arr
10994 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10995 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10996 BN_GF2m_mod_exp_arr
10997 BN_GF2m_mod_sqrt_arr
10998 BN_GF2m_mod_solve_quad_arr
11002 Conversion can be performed by the following functions:
11007 bntest.c has additional tests for binary polynomial arithmetic.
11009 Two implementations for BN_GF2m_mod_div() are available.
11010 The default algorithm simply uses BN_GF2m_mod_inv() and
11011 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11012 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11013 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11015 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11017 * Add new error code 'ERR_R_DISABLED' that can be used when some
11018 functionality is disabled at compile-time.
11020 *Douglas Stebila <douglas.stebila@sun.com>*
11022 * Change default behaviour of 'openssl asn1parse' so that more
11023 information is visible when viewing, e.g., a certificate:
11025 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11026 mode the content of non-printable OCTET STRINGs is output in a
11027 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11028 avoid the appearance of a printable string.
11030 *Nils Larsch <nla@trustcenter.de>*
11032 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11034 EC_GROUP_set_asn1_flag()
11035 EC_GROUP_get_asn1_flag()
11036 EC_GROUP_set_point_conversion_form()
11037 EC_GROUP_get_point_conversion_form()
11038 These control ASN1 encoding details:
11039 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11040 has been set to OPENSSL_EC_NAMED_CURVE.
11041 - Points are encoded in uncompressed form by default; options for
11042 asn1_for are as for point2oct, namely
11043 POINT_CONVERSION_COMPRESSED
11044 POINT_CONVERSION_UNCOMPRESSED
11045 POINT_CONVERSION_HYBRID
11047 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11049 EC_GROUP_set_seed()
11050 EC_GROUP_get0_seed()
11051 EC_GROUP_get_seed_len()
11052 This is used only for ASN1 purposes (so far).
11054 *Nils Larsch <nla@trustcenter.de>*
11056 * Add 'field_type' member to EC_METHOD, which holds the NID
11057 of the appropriate field type OID. The new function
11058 EC_METHOD_get_field_type() returns this value.
11060 *Nils Larsch <nla@trustcenter.de>*
11063 EC_POINT_point2bn()
11064 EC_POINT_bn2point()
11065 EC_POINT_point2hex()
11066 EC_POINT_hex2point()
11067 providing useful interfaces to EC_POINT_point2oct() and
11068 EC_POINT_oct2point().
11070 *Nils Larsch <nla@trustcenter.de>*
11072 * Change internals of the EC library so that the functions
11073 EC_GROUP_set_generator()
11074 EC_GROUP_get_generator()
11075 EC_GROUP_get_order()
11076 EC_GROUP_get_cofactor()
11077 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11078 to methods, which would lead to unnecessary code duplication when
11079 adding different types of curves.
11081 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11083 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11084 arithmetic, and such that modified wNAFs are generated
11085 (which avoid length expansion in many cases).
11089 * Add a function EC_GROUP_check_discriminant() (defined via
11090 EC_METHOD) that verifies that the curve discriminant is non-zero.
11092 Add a function EC_GROUP_check() that makes some sanity tests
11093 on a EC_GROUP, its generator and order. This includes
11094 EC_GROUP_check_discriminant().
11096 *Nils Larsch <nla@trustcenter.de>*
11098 * Add ECDSA in new directory crypto/ecdsa/.
11100 Add applications 'openssl ecparam' and 'openssl ecdsa'
11101 (these are based on 'openssl dsaparam' and 'openssl dsa').
11103 ECDSA support is also included in various other files across the
11104 library. Most notably,
11105 - 'openssl req' now has a '-newkey ecdsa:file' option;
11106 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11107 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11108 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11109 them suitable for ECDSA where domain parameters must be
11110 extracted before the specific public key;
11111 - ECDSA engine support has been added.
11113 *Nils Larsch <nla@trustcenter.de>*
11115 * Include some named elliptic curves, and add OIDs from X9.62,
11116 SECG, and WAP/WTLS. Each curve can be obtained from the new
11118 EC_GROUP_new_by_curve_name(),
11119 and the list of available named curves can be obtained with
11120 EC_get_builtin_curves().
11121 Also add a 'curve_name' member to EC_GROUP objects, which can be
11123 EC_GROUP_set_curve_name()
11124 EC_GROUP_get_curve_name()
11126 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11128 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11129 was actually never needed) and in BN_mul(). The removal in BN_mul()
11130 required a small change in bn_mul_part_recursive() and the addition
11131 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11132 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11133 bn_sub_words() and bn_add_words() except they take arrays with
11138 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11140 * Cleanse PEM buffers before freeing them since they may contain
11143 *Benjamin Bennett <ben@psc.edu>*
11145 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11146 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11147 authentication-only ciphersuites.
11151 * Since AES128 and AES256 share a single mask bit in the logic of
11152 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11153 kludge to work properly if AES128 is available and AES256 isn't.
11157 * Expand security boundary to match 1.1.1 module.
11161 * Remove redundant features: hash file source, editing of test vectors
11162 modify fipsld to use external fips_premain.c signature.
11166 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11167 run algorithm test programs.
11171 * Make algorithm test programs more tolerant of whitespace.
11175 * Have SSL/TLS server implementation tolerate "mismatched" record
11176 protocol version while receiving ClientHello even if the
11177 ClientHello is fragmented. (The server can't insist on the
11178 particular protocol version it has chosen before the ServerHello
11179 message has informed the client about his choice.)
11183 * Load error codes if they are not already present instead of using a
11184 static variable. This allows them to be cleanly unloaded and reloaded.
11188 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11190 * Introduce limits to prevent malicious keys being able to
11191 cause a denial of service. ([CVE-2006-2940])
11193 *Steve Henson, Bodo Moeller*
11195 * Fix ASN.1 parsing of certain invalid structures that can result
11196 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11198 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11199 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11201 * Fix SSL client code which could crash if connecting to a
11202 malicious SSLv2 server. ([CVE-2006-4343])
11204 *Tavis Ormandy and Will Drewry, Google Security Team*
11206 * Change ciphersuite string processing so that an explicit
11207 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11208 will no longer include "AES128-SHA"), and any other similar
11209 ciphersuite (same bitmap) from *other* protocol versions (so that
11210 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11211 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11212 changes from 0.9.8b and 0.9.8d.
11216 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11218 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11219 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11221 * Change the Unix randomness entropy gathering to use poll() when
11222 possible instead of select(), since the latter has some
11223 undesirable limitations.
11225 *Darryl Miles via Richard Levitte and Bodo Moeller*
11227 * Disable rogue ciphersuites:
11229 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11230 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11231 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11233 The latter two were purportedly from
11234 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11237 Also deactivate the remaining ciphersuites from
11238 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11239 unofficial, and the ID has long expired.
11243 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11244 dual-core machines) and other potential thread-safety issues.
11248 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11250 * Adapt fipsld and the build system to link against the validated FIPS
11251 module in FIPS mode.
11255 * Fixes for VC++ 2005 build under Windows.
11259 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11260 from a Windows bash shell such as MSYS. It is autodetected from the
11261 "config" script when run from a VC++ environment. Modify standard VC++
11262 build to use fipscanister.o from the GNU make build.
11266 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11268 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11269 The value now differs depending on if you build for FIPS or not.
11270 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11271 safely run with a non-FIPSed libcrypto, as it may crash because of
11272 the difference induced by this change.
11276 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11278 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11279 (part of SSL_OP_ALL). This option used to disable the
11280 countermeasure against man-in-the-middle protocol-version
11281 rollback in the SSL 2.0 server implementation, which is a bad
11282 idea. ([CVE-2005-2969])
11284 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11285 for Information Security, National Institute of Advanced Industrial
11286 Science and Technology [AIST, Japan)]*
11288 * Minimal support for X9.31 signatures and PSS padding modes. This is
11289 mainly for FIPS compliance and not fully integrated at this stage.
11293 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11294 the exponentiation using a fixed-length exponent. (Otherwise,
11295 the information leaked through timing could expose the secret key
11296 after many signatures; cf. Bleichenbacher's attack on DSA with
11301 * Make a new fixed-window mod_exp implementation the default for
11302 RSA, DSA, and DH private-key operations so that the sequence of
11303 squares and multiplies and the memory access pattern are
11304 independent of the particular secret key. This will mitigate
11305 cache-timing and potential related attacks.
11307 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11308 and this is automatically used by BN_mod_exp_mont() if the new flag
11309 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11310 will use this BN flag for private exponents unless the flag
11311 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11312 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11314 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11316 * Change the client implementation for SSLv23_method() and
11317 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11318 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11319 (Previously, the SSL 2.0 backwards compatible Client Hello
11320 message format would be used even with SSL_OP_NO_SSLv2.)
11324 * Add support for smime-type MIME parameter in S/MIME messages which some
11329 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11330 a threadsafe manner. Modify rsa code to use new function and add calls
11331 to dsa and dh code (which had race conditions before).
11335 * Include the fixed error library code in the C error file definitions
11336 instead of fixing them up at runtime. This keeps the error code
11337 structures constant.
11341 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11343 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11346 * Fixes for newer kerberos headers. NB: the casts are needed because
11347 the 'length' field is signed on one version and unsigned on another
11348 with no (?) obvious way to tell the difference, without these VC++
11349 complains. Also the "definition" of FAR (blank) is no longer included
11350 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11351 some needed definitions.
11355 * Undo Cygwin change.
11359 * Added support for proxy certificates according to RFC 3820.
11360 Because they may be a security thread to unaware applications,
11361 they must be explicitly allowed in run-time. See
11362 docs/HOWTO/proxy_certificates.txt for further information.
11366 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11368 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11369 server and client random values. Previously
11370 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11371 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11373 This change has negligible security impact because:
11375 1. Server and client random values still have 24 bytes of pseudo random
11378 2. Server and client random values are sent in the clear in the initial
11381 3. The master secret is derived using the premaster secret (48 bytes in
11382 size for static RSA ciphersuites) as well as client server and random
11385 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11388 *Stephen Henson, reported by UK NISCC*
11390 * Use Windows randomness collection on Cygwin.
11394 * Fix hang in EGD/PRNGD query when communication socket is closed
11395 prematurely by EGD/PRNGD.
11397 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11399 * Prompt for pass phrases when appropriate for PKCS12 input format.
11403 * Back-port of selected performance improvements from development
11404 branch, as well as improved support for PowerPC platforms.
11408 * Add lots of checks for memory allocation failure, error codes to indicate
11409 failure and freeing up memory if a failure occurs.
11411 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11413 * Add new -passin argument to dgst.
11417 * Perform some character comparisons of different types in X509_NAME_cmp:
11418 this is needed for some certificates that re-encode DNs into UTF8Strings
11419 (in violation of RFC3280) and can't or won't issue name rollover
11424 * Make an explicit check during certificate validation to see that
11425 the CA setting in each certificate on the chain is correct. As a
11426 side effect always do the following basic checks on extensions,
11427 not just when there's an associated purpose to the check:
11429 - if there is an unhandled critical extension (unless the user
11430 has chosen to ignore this fault)
11431 - if the path length has been exceeded (if one is set at all)
11432 - that certain extensions fit the associated purpose (if one has
11437 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11439 * Avoid a race condition when CRLs are checked in a multi threaded
11440 environment. This would happen due to the reordering of the revoked
11441 entries during signature checking and serial number lookup. Now the
11442 encoding is cached and the serial number sort performed under a lock.
11443 Add new STACK function sk_is_sorted().
11447 * Add Delta CRL to the extension code.
11451 * Various fixes to s3_pkt.c so alerts are sent properly.
11453 *David Holmes <d.holmes@f5.com>*
11455 * Reduce the chances of duplicate issuer name and serial numbers (in
11456 violation of RFC3280) using the OpenSSL certificate creation utilities.
11457 This is done by creating a random 64 bit value for the initial serial
11458 number when a serial number file is created or when a self signed
11459 certificate is created using 'openssl req -x509'. The initial serial
11460 number file is created using 'openssl x509 -next_serial' in CA.pl
11461 rather than being initialized to 1.
11465 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11467 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11468 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11470 *Joe Orton, Steve Henson*
11472 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11475 *Joe Orton, Steve Henson*
11477 * Make it possible to have multiple active certificates with the same
11478 subject in the CA index file. This is done only if the keyword
11479 'unique_subject' is set to 'no' in the main CA section (default
11480 if 'CA_default') of the configuration file. The value is saved
11481 with the database itself in a separate index attribute file,
11482 named like the index file with '.attr' appended to the name.
11486 * X509 verify fixes. Disable broken certificate workarounds when
11487 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11488 keyUsage extension present. Don't accept CRLs with unhandled critical
11489 extensions: since verify currently doesn't process CRL extensions this
11490 rejects a CRL with *any* critical extensions. Add new verify error codes
11495 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11496 A clarification of RFC2560 will require the use of OCTET STRINGs and
11497 some implementations cannot handle the current raw format. Since OpenSSL
11498 copies and compares OCSP nonces as opaque blobs without any attempt at
11499 parsing them this should not create any compatibility issues.
11503 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11504 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11505 this HMAC (and other) operations are several times slower than OpenSSL
11510 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11512 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11514 * Use the correct content when signing type "other".
11518 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11520 * Fix various bugs revealed by running the NISCC test suite:
11522 Stop out of bounds reads in the ASN1 code when presented with
11523 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11525 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11527 If verify callback ignores invalid public key errors don't try to check
11528 certificate signature with the NULL public key.
11532 * New -ignore_err option in ocsp application to stop the server
11533 exiting on the first error in a request.
11537 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11538 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11543 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11544 extra data after the compression methods not only for TLS 1.0
11545 but also for SSL 3.0 (as required by the specification).
11547 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11549 * Change X509_certificate_type() to mark the key as exported/exportable
11550 when it's 512 *bits* long, not 512 bytes.
11554 * Change AES_cbc_encrypt() so it outputs exact multiple of
11555 blocks during encryption.
11559 * Various fixes to base64 BIO and non blocking I/O. On write
11560 flushes were not handled properly if the BIO retried. On read
11561 data was not being buffered properly and had various logic bugs.
11562 This also affects blocking I/O when the data being decoded is a
11567 * Various S/MIME bugfixes and compatibility changes:
11568 output correct application/pkcs7 MIME type if
11569 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11570 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11571 of files as .eml work). Correctly handle very long lines in MIME
11576 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11578 * Countermeasure against the Klima-Pokorny-Rosa extension of
11579 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11580 a protocol version number mismatch like a decryption error
11581 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11585 * Turn on RSA blinding by default in the default implementation
11586 to avoid a timing attack. Applications that don't want it can call
11587 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11588 They would be ill-advised to do so in most cases.
11590 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11592 * Change RSA blinding code so that it works when the PRNG is not
11593 seeded (in this case, the secret RSA exponent is abused as
11594 an unpredictable seed -- if it is not unpredictable, there
11595 is no point in blinding anyway). Make RSA blinding thread-safe
11596 by remembering the creator's thread ID in rsa->blinding and
11597 having all other threads use local one-time blinding factors
11598 (this requires more computation than sharing rsa->blinding, but
11599 avoids excessive locking; and if an RSA object is not shared
11600 between threads, blinding will still be very fast).
11604 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11605 ENGINE as defaults for all supported algorithms irrespective of
11606 the 'flags' parameter. 'flags' is now honoured, so applications
11607 should make sure they are passing it correctly.
11611 * Target "mingw" now allows native Windows code to be generated in
11612 the Cygwin environment as well as with the MinGW compiler.
11616 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11618 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11619 via timing by performing a MAC computation even if incorrect
11620 block cipher padding has been found. This is a countermeasure
11621 against active attacks where the attacker has to distinguish
11622 between bad padding and a MAC verification error. ([CVE-2003-0078])
11624 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11625 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11626 Martin Vuagnoux (EPFL, Ilion)*
11628 * Make the no-err option work as intended. The intention with no-err
11629 is not to have the whole error stack handling routines removed from
11630 libcrypto, it's only intended to remove all the function name and
11631 reason texts, thereby removing some of the footprint that may not
11632 be interesting if those errors aren't displayed anyway.
11634 NOTE: it's still possible for any application or module to have its
11635 own set of error texts inserted. The routines are there, just not
11636 used by default when no-err is given.
11640 * Add support for FreeBSD on IA64.
11642 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11644 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11645 Kerberos function mit_des_cbc_cksum(). Before this change,
11646 the value returned by DES_cbc_cksum() was like the one from
11647 mit_des_cbc_cksum(), except the bytes were swapped.
11649 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11651 * Allow an application to disable the automatic SSL chain building.
11652 Before this a rather primitive chain build was always performed in
11653 ssl3_output_cert_chain(): an application had no way to send the
11654 correct chain if the automatic operation produced an incorrect result.
11656 Now the chain builder is disabled if either:
11658 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11660 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11662 The reasoning behind this is that an application would not want the
11663 auto chain building to take place if extra chain certificates are
11664 present and it might also want a means of sending no additional
11665 certificates (for example the chain has two certificates and the
11670 * Add the possibility to build without the ENGINE framework.
11672 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11674 * Under Win32 gmtime() can return NULL: check return value in
11675 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11679 * DSA routines: under certain error conditions uninitialized BN objects
11680 could be freed. Solution: make sure initialization is performed early
11681 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11682 Nils Larsch <nla@trustcenter.de> via PR#459)
11686 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11687 checked on reconnect on the client side, therefore session resumption
11688 could still fail with a "ssl session id is different" error. This
11689 behaviour is masked when SSL_OP_ALL is used due to
11690 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11691 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11692 followup to PR #377.
11696 * IA-32 assembler support enhancements: unified ELF targets, support
11697 for SCO/Caldera platforms, fix for Cygwin shared build.
11701 * Add support for FreeBSD on sparc64. As a consequence, support for
11702 FreeBSD on non-x86 processors is separate from x86 processors on
11703 the config script, much like the NetBSD support.
11705 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11707 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11709 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11712 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11713 code (06) was taken as the first octet of the session ID and the last
11714 octet was ignored consequently. As a result SSLv2 client side session
11715 caching could not have worked due to the session ID mismatch between
11717 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11722 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11723 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11728 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11729 seems that in spite of existing for more than a year, many application
11730 author have done nothing to provide the necessary callbacks, which
11731 means that this particular engine will not work properly anywhere.
11732 This is a very unfortunate situation which forces us, in the name
11733 of usability, to give the hw_ncipher.c a static lock, which is part
11735 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11736 appear in 0.9.8 or later. We EXPECT application authors to have
11737 dealt properly with this when 0.9.8 is released (unless we actually
11738 make such changes in the libcrypto locking code that changes will
11739 have to be made anyway).
11743 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11744 octets have been read, EOF or an error occurs. Without this change
11745 some truncated ASN1 structures will not produce an error.
11749 * Disable Heimdal support, since it hasn't been fully implemented.
11750 Still give the possibility to force the use of Heimdal, but with
11751 warnings and a request that patches get sent to openssl-dev.
11755 * Add the VC-CE target, introduce the WINCE sysname, and add
11756 INSTALL.WCE and appropriate conditionals to make it build.
11758 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11760 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11761 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11762 edit numbers of the version.
11764 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11766 * Introduce safe string copy and catenation functions
11767 (BUF_strlcpy() and BUF_strlcat()).
11769 *Ben Laurie (CHATS) and Richard Levitte*
11771 * Avoid using fixed-size buffers for one-line DNs.
11773 *Ben Laurie (CHATS)*
11775 * Add BUF_MEM_grow_clean() to avoid information leakage when
11776 resizing buffers containing secrets, and use where appropriate.
11778 *Ben Laurie (CHATS)*
11780 * Avoid using fixed size buffers for configuration file location.
11782 *Ben Laurie (CHATS)*
11784 * Avoid filename truncation for various CA files.
11786 *Ben Laurie (CHATS)*
11788 * Use sizeof in preference to magic numbers.
11790 *Ben Laurie (CHATS)*
11792 * Avoid filename truncation in cert requests.
11794 *Ben Laurie (CHATS)*
11796 * Add assertions to check for (supposedly impossible) buffer
11799 *Ben Laurie (CHATS)*
11801 * Don't cache truncated DNS entries in the local cache (this could
11802 potentially lead to a spoofing attack).
11804 *Ben Laurie (CHATS)*
11806 * Fix various buffers to be large enough for hex/decimal
11807 representations in a platform independent manner.
11809 *Ben Laurie (CHATS)*
11811 * Add CRYPTO_realloc_clean() to avoid information leakage when
11812 resizing buffers containing secrets, and use where appropriate.
11814 *Ben Laurie (CHATS)*
11816 * Add BIO_indent() to avoid much slightly worrying code to do
11819 *Ben Laurie (CHATS)*
11821 * Convert sprintf()/BIO_puts() to BIO_printf().
11823 *Ben Laurie (CHATS)*
11825 * buffer_gets() could terminate with the buffer only half
11828 *Ben Laurie (CHATS)*
11830 * Add assertions to prevent user-supplied crypto functions from
11831 overflowing internal buffers by having large block sizes, etc.
11833 *Ben Laurie (CHATS)*
11835 * New OPENSSL_assert() macro (similar to assert(), but enabled
11838 *Ben Laurie (CHATS)*
11840 * Eliminate unused copy of key in RC4.
11842 *Ben Laurie (CHATS)*
11844 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11846 *Ben Laurie (CHATS)*
11848 * Fix off-by-one error in EGD path.
11850 *Ben Laurie (CHATS)*
11852 * If RANDFILE path is too long, ignore instead of truncating.
11854 *Ben Laurie (CHATS)*
11856 * Eliminate unused and incorrectly sized X.509 structure
11859 *Ben Laurie (CHATS)*
11861 * Eliminate unused and dangerous function knumber().
11863 *Ben Laurie (CHATS)*
11865 * Eliminate unused and dangerous structure, KSSL_ERR.
11867 *Ben Laurie (CHATS)*
11869 * Protect against overlong session ID context length in an encoded
11870 session object. Since these are local, this does not appear to be
11873 *Ben Laurie (CHATS)*
11875 * Change from security patch (see 0.9.6e below) that did not affect
11876 the 0.9.6 release series:
11878 Remote buffer overflow in SSL3 protocol - an attacker could
11879 supply an oversized master key in Kerberos-enabled versions.
11882 *Ben Laurie (CHATS)*
11884 * Change the SSL kerb5 codes to match RFC 2712.
11888 * Make -nameopt work fully for req and add -reqopt switch.
11890 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11892 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11894 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11896 * Make sure tests can be performed even if the corresponding algorithms
11897 have been removed entirely. This was also the last step to make
11898 OpenSSL compilable with DJGPP under all reasonable conditions.
11900 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11902 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11903 to allow version independent disabling of normally unselected ciphers,
11904 which may be activated as a side-effect of selecting a single cipher.
11906 (E.g., cipher list string "RSA" enables ciphersuites that are left
11907 out of "ALL" because they do not provide symmetric encryption.
11908 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11910 *Lutz Jaenicke, Bodo Moeller*
11912 * Add appropriate support for separate platform-dependent build
11913 directories. The recommended way to make a platform-dependent
11914 build directory is the following (tested on Linux), maybe with
11917 # Place yourself outside of the OpenSSL source tree. In
11918 # this example, the environment variable OPENSSL_SOURCE
11919 # is assumed to contain the absolute OpenSSL source directory.
11920 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11921 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11922 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11923 mkdir -p `dirname $F`
11924 ln -s $OPENSSL_SOURCE/$F $F
11927 To be absolutely sure not to disturb the source tree, a "make clean"
11928 is a good thing. If it isn't successful, don't worry about it,
11929 it probably means the source directory is very clean.
11933 * Make sure any ENGINE control commands make local copies of string
11934 pointers passed to them whenever necessary. Otherwise it is possible
11935 the caller may have overwritten (or deallocated) the original string
11936 data when a later ENGINE operation tries to use the stored values.
11938 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11940 * Improve diagnostics in file reading and command-line digests.
11942 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11944 * Add AES modes CFB and OFB to the object database. Correct an
11945 error in AES-CFB decryption.
11949 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11950 allows existing EVP_CIPHER_CTX structures to be reused after
11951 calling `EVP_*Final()`. This behaviour is used by encryption
11952 BIOs and some applications. This has the side effect that
11953 applications must explicitly clean up cipher contexts with
11954 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11958 * Check the values of dna and dnb in bn_mul_recursive before calling
11959 bn_mul_comba (a non zero value means the a or b arrays do not contain
11960 n2 elements) and fallback to bn_mul_normal if either is not zero.
11964 * Fix escaping of non-ASCII characters when using the -subj option
11965 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11969 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11970 form for "surname", serialNumber has no short form.
11971 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11972 therefore remove "mail" short name for "internet 7".
11973 The OID for unique identifiers in X509 certificates is
11974 x500UniqueIdentifier, not uniqueIdentifier.
11975 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11979 * Add an "init" command to the ENGINE config module and auto initialize
11980 ENGINEs. Without any "init" command the ENGINE will be initialized
11981 after all ctrl commands have been executed on it. If init=1 the
11982 ENGINE is initialized at that point (ctrls before that point are run
11983 on the uninitialized ENGINE and after on the initialized one). If
11984 init=0 then the ENGINE will not be initialized at all.
11988 * Fix the 'app_verify_callback' interface so that the user-defined
11989 argument is actually passed to the callback: In the
11990 SSL_CTX_set_cert_verify_callback() prototype, the callback
11991 declaration has been changed from
11994 int (*cb)(X509_STORE_CTX *,void *);
11995 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11996 i=s->ctx->app_verify_callback(&ctx)
11997 has been changed into
11998 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12000 To update applications using SSL_CTX_set_cert_verify_callback(),
12001 a dummy argument can be added to their callback functions.
12003 *D. K. Smetters <smetters@parc.xerox.com>*
12005 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12007 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12009 * Add and OPENSSL_LOAD_CONF define which will cause
12010 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12011 This allows older applications to transparently support certain
12012 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12013 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12014 load the config file and OPENSSL_add_all_algorithms_conf() which will
12015 always load it have also been added.
12019 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12020 Adjust NIDs and EVP layer.
12022 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12024 * Config modules support in openssl utility.
12026 Most commands now load modules from the config file,
12027 though in a few (such as version) this isn't done
12028 because it couldn't be used for anything.
12030 In the case of ca and req the config file used is
12031 the same as the utility itself: that is the -config
12032 command line option can be used to specify an
12037 * Move default behaviour from OPENSSL_config(). If appname is NULL
12038 use "openssl_conf" if filename is NULL use default openssl config file.
12042 * Add an argument to OPENSSL_config() to allow the use of an alternative
12043 config section name. Add a new flag to tolerate a missing config file
12044 and move code to CONF_modules_load_file().
12048 * Support for crypto accelerator cards from Accelerated Encryption
12049 Processing, www.aep.ie. (Use engine 'aep')
12050 The support was copied from 0.9.6c [engine] and adapted/corrected
12051 to work with the new engine framework.
12053 *AEP Inc. and Richard Levitte*
12055 * Support for SureWare crypto accelerator cards from Baltimore
12056 Technologies. (Use engine 'sureware')
12057 The support was copied from 0.9.6c [engine] and adapted
12058 to work with the new engine framework.
12062 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12063 make the newer ENGINE framework commands for the CHIL engine work.
12065 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12067 * Make it possible to produce shared libraries on ReliantUNIX.
12069 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12071 * Add the configuration target debug-linux-ppro.
12072 Make 'openssl rsa' use the general key loading routines
12073 implemented in `apps.c`, and make those routines able to
12074 handle the key format FORMAT_NETSCAPE and the variant
12077 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12079 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12081 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12083 * Add -keyform to rsautl, and document -engine.
12085 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12087 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12088 BIO_R_NO_SUCH_FILE error code rather than the generic
12089 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12093 * Add new functions
12094 ERR_peek_last_error
12095 ERR_peek_last_error_line
12096 ERR_peek_last_error_line_data.
12097 These are similar to
12099 ERR_peek_error_line
12100 ERR_peek_error_line_data,
12101 but report on the latest error recorded rather than the first one
12102 still in the error queue.
12104 *Ben Laurie, Bodo Moeller*
12106 * default_algorithms option in ENGINE config module. This allows things
12108 default_algorithms = ALL
12109 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12113 * Preliminary ENGINE config module.
12117 * New experimental application configuration code.
12121 * Change the AES code to follow the same name structure as all other
12122 symmetric ciphers, and behave the same way. Move everything to
12123 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12125 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12127 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12129 *Ben Laurie and Theo de Raadt*
12131 * Add option to output public keys in req command.
12133 *Massimiliano Pala madwolf@openca.org*
12135 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12136 (up to about 10% better than before for P-192 and P-224).
12140 * New functions/macros
12142 SSL_CTX_set_msg_callback(ctx, cb)
12143 SSL_CTX_set_msg_callback_arg(ctx, arg)
12144 SSL_set_msg_callback(ssl, cb)
12145 SSL_set_msg_callback_arg(ssl, arg)
12147 to request calling a callback function
12149 void cb(int write_p, int version, int content_type,
12150 const void *buf, size_t len, SSL *ssl, void *arg)
12152 whenever a protocol message has been completely received
12153 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12154 protocol version according to which the SSL library interprets
12155 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12156 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12157 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12158 specification (change_cipher_spec(20), alert(21), handshake(22)).
12159 'buf' and 'len' point to the actual message, 'ssl' to the
12160 SSL object, and 'arg' is the application-defined value set by
12161 SSL[_CTX]_set_msg_callback_arg().
12163 'openssl s_client' and 'openssl s_server' have new '-msg' options
12164 to enable a callback that displays all protocol messages.
12168 * Change the shared library support so shared libraries are built as
12169 soon as the corresponding static library is finished, and thereby get
12170 openssl and the test programs linked against the shared library.
12171 This still only happens when the keyword "shard" has been given to
12172 the configuration scripts.
12174 NOTE: shared library support is still an experimental thing, and
12175 backward binary compatibility is still not guaranteed.
12177 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12179 * Add support for Subject Information Access extension.
12181 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12183 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12184 additional bytes when new memory had to be allocated, not just
12185 when reusing an existing buffer.
12189 * New command line and configuration option 'utf8' for the req command.
12190 This allows field values to be specified as UTF8 strings.
12194 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12195 runs for the former and machine-readable output for the latter.
12199 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12200 of the e-mail address in the DN (i.e., it will go into a certificate
12201 extension only). The new configuration file option 'email_in_dn = no'
12202 has the same effect.
12204 *Massimiliano Pala madwolf@openca.org*
12206 * Change all functions with names starting with `des_` to be starting
12207 with `DES_` instead. Add wrappers that are compatible with libdes,
12208 but are named `_ossl_old_des_*`. Finally, add macros that map the
12209 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12210 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12211 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12214 Since we provide two compatibility mappings, the user needs to
12215 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12216 compatibility is desired. The default (i.e., when that macro
12217 isn't defined) is OpenSSL 0.9.6c compatibility.
12219 There are also macros that enable and disable the support of old
12220 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12221 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12222 are defined, the default will apply: to support the old des routines.
12224 In either case, one must include openssl/des.h to get the correct
12225 definitions. Do not try to just include openssl/des_old.h, that
12228 NOTE: This is a major break of an old API into a new one. Software
12229 authors are encouraged to switch to the `DES_` style functions. Some
12230 time in the future, des_old.h and the libdes compatibility functions
12231 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12232 default), and then completely removed.
12236 * Test for certificates which contain unsupported critical extensions.
12237 If such a certificate is found during a verify operation it is
12238 rejected by default: this behaviour can be overridden by either
12239 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12240 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12241 X509_supported_extension() has also been added which returns 1 if a
12242 particular extension is supported.
12246 * Modify the behaviour of EVP cipher functions in similar way to digests
12247 to retain compatibility with existing code.
12251 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12252 compatibility with existing code. In particular the 'ctx' parameter does
12253 not have to be to be initialized before the call to EVP_DigestInit() and
12254 it is tidied up after a call to EVP_DigestFinal(). New function
12255 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12256 EVP_MD_CTX_copy() changed to not require the destination to be
12257 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12258 requires the destination to be valid.
12260 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12261 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12265 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12266 so that complete 'Handshake' protocol structures are kept in memory
12267 instead of overwriting 'msg_type' and 'length' with 'body' data.
12271 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12273 *Massimo Santin via Richard Levitte*
12275 * Major restructuring to the underlying ENGINE code. This includes
12276 reduction of linker bloat, separation of pure "ENGINE" manipulation
12277 (initialisation, etc) from functionality dealing with implementations
12278 of specific crypto interfaces. This change also introduces integrated
12279 support for symmetric ciphers and digest implementations - so ENGINEs
12280 can now accelerate these by providing EVP_CIPHER and EVP_MD
12281 implementations of their own. This is detailed in
12282 [crypto/engine/README.md](crypto/engine/README.md)
12283 as it couldn't be adequately described here. However, there are a few
12284 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12285 were changed in the original introduction of ENGINE code have now
12286 reverted back - the hooking from this code to ENGINE is now a good
12287 deal more passive and at run-time, operations deal directly with
12288 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12289 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12290 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12291 they were not being used by the framework as there is no concept of a
12292 BIGNUM_METHOD and they could not be generalised to the new
12293 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12294 ENGINE_cpy() has been removed as it cannot be consistently defined in
12299 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12303 * Change mkdef.pl to sort symbols that get the same entry number,
12304 and make sure the automatically generated functions `ERR_load_*`
12305 become part of libeay.num as well.
12309 * New function SSL_renegotiate_pending(). This returns true once
12310 renegotiation has been requested (either SSL_renegotiate() call
12311 or HelloRequest/ClientHello received from the peer) and becomes
12312 false once a handshake has been completed.
12313 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12314 sends a HelloRequest, but does not ensure that a handshake takes
12315 place. SSL_renegotiate_pending() is useful for checking if the
12316 client has followed the request.)
12320 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12321 By default, clients may request session resumption even during
12322 renegotiation (if session ID contexts permit); with this option,
12323 session resumption is possible only in the first handshake.
12325 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12326 more bits available for options that should not be part of
12327 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12331 * Add some demos for certificate and certificate request creation.
12335 * Make maximum certificate chain size accepted from the peer application
12336 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12337 "Douglas E. Engert" <deengert@anl.gov>.
12341 * Add support for shared libraries for Unixware-7
12342 (Boyd Lynn Gerber <gerberb@zenez.com>).
12346 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12347 be done prior to destruction. Use this to unload error strings from
12348 ENGINEs that load their own error strings. NB: This adds two new API
12349 functions to "get" and "set" this destroy handler in an ENGINE.
12353 * Alter all existing ENGINE implementations (except "openssl" and
12354 "openbsd") to dynamically instantiate their own error strings. This
12355 makes them more flexible to be built both as statically-linked ENGINEs
12356 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12357 Also, add stub code to each that makes building them as self-contained
12358 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12362 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12363 implementations into applications that are completely implemented in
12364 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12365 commands that can be used to configure what shared-library to load and
12366 to control aspects of the way it is handled. Also, made an update to
12367 the [README-Engine.md](README-Engine.md) file
12368 that brings its information up-to-date and
12369 provides some information and instructions on the "dynamic" ENGINE
12370 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12374 * Make it possible to unload ranges of ERR strings with a new
12375 "ERR_unload_strings" function.
12379 * Add a copy() function to EVP_MD.
12383 * Make EVP_MD routines take a context pointer instead of just the
12384 md_data void pointer.
12388 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12389 that the digest can only process a single chunk of data
12390 (typically because it is provided by a piece of
12391 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12392 is only going to provide a single chunk of data, and hence the
12393 framework needn't accumulate the data for oneshot drivers.
12397 * As with "ERR", make it possible to replace the underlying "ex_data"
12398 functions. This change also alters the storage and management of global
12399 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12400 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12401 index counters. The API functions that use this state have been changed
12402 to take a "class_index" rather than pointers to the class's local STACK
12403 and counter, and there is now an API function to dynamically create new
12404 classes. This centralisation allows us to (a) plug a lot of the
12405 thread-safety problems that existed, and (b) makes it possible to clean
12406 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12407 such data would previously have always leaked in application code and
12408 workarounds were in place to make the memory debugging turn a blind eye
12409 to it. Application code that doesn't use this new function will still
12410 leak as before, but their memory debugging output will announce it now
12411 rather than letting it slide.
12413 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12414 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12415 has a return value to indicate success or failure.
12419 * Make it possible to replace the underlying "ERR" functions such that the
12420 global state (2 LHASH tables and 2 locks) is only used by the "default"
12421 implementation. This change also adds two functions to "get" and "set"
12422 the implementation prior to it being automatically set the first time
12423 any other ERR function takes place. Ie. an application can call "get",
12424 pass the return value to a module it has just loaded, and that module
12425 can call its own "set" function using that value. This means the
12426 module's "ERR" operations will use (and modify) the error state in the
12427 application and not in its own statically linked copy of OpenSSL code.
12431 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12432 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12433 the operation, and provides a more encapsulated way for external code
12434 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12435 to use these functions rather than manually incrementing the counts.
12437 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12441 * Add EVP test program.
12445 * Add symmetric cipher support to ENGINE. Expect the API to change!
12449 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12450 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12451 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12452 These allow a CRL to be built without having to access X509_CRL fields
12453 directly. Modify 'ca' application to use new functions.
12457 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12458 bug workarounds. Rollback attack detection is a security feature.
12459 The problem will only arise on OpenSSL servers when TLSv1 is not
12460 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12461 Software authors not wanting to support TLSv1 will have special reasons
12462 for their choice and can explicitly enable this option.
12464 *Bodo Moeller, Lutz Jaenicke*
12466 * Rationalise EVP so it can be extended: don't include a union of
12467 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12468 (similar to those existing for EVP_CIPHER_CTX).
12473 EVP_MD_CTX_init(&md); /* new function call */
12474 EVP_DigestInit(&md, EVP_sha1());
12475 EVP_DigestUpdate(&md, in, len);
12476 EVP_DigestFinal(&md, out, NULL);
12477 EVP_MD_CTX_cleanup(&md); /* new function call */
12481 * Make DES key schedule conform to the usual scheme, as well as
12482 correcting its structure. This means that calls to DES functions
12483 now have to pass a pointer to a des_key_schedule instead of a
12484 plain des_key_schedule (which was actually always a pointer
12487 des_key_schedule ks;
12489 des_set_key_checked(..., &ks);
12490 des_ncbc_encrypt(..., &ks, ...);
12492 (Note that a later change renames 'des_...' into 'DES_...'.)
12496 * Initial reduction of linker bloat: the use of some functions, such as
12497 PEM causes large amounts of unused functions to be linked in due to
12498 poor organisation. For example pem_all.c contains every PEM function
12499 which has a knock on effect of linking in large amounts of (unused)
12500 ASN1 code. Grouping together similar functions and splitting unrelated
12501 functions prevents this.
12505 * Cleanup of EVP macros.
12509 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12510 correct `_ecb suffix`.
12514 * Add initial OCSP responder support to ocsp application. The
12515 revocation information is handled using the text based index
12516 use by the ca application. The responder can either handle
12517 requests generated internally, supplied in files (for example
12518 via a CGI script) or using an internal minimal server.
12522 * Add configuration choices to get zlib compression for TLS.
12526 * Changes to Kerberos SSL for RFC 2712 compliance:
12527 1. Implemented real KerberosWrapper, instead of just using
12528 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12529 2. Implemented optional authenticator field of KerberosWrapper.
12531 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12532 and authenticator structs; see crypto/krb5/.
12534 Generalized Kerberos calls to support multiple Kerberos libraries.
12535 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12536 via Richard Levitte*
12538 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12539 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12540 values for each of the key sizes rather than having just
12541 parameters (and 'speed' generating keys each time).
12545 * Speed up EVP routines.
12548 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12549 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12550 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12551 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12553 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12554 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12555 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12558 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12560 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12564 * Added the OS2-EMX target.
12566 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12568 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12569 New functions to support `NCONF` routines in extension code.
12570 New function `CONF_set_nconf()`
12571 to allow functions which take an `NCONF` to also handle the old `LHASH`
12572 structure: this means that the old `CONF` compatible routines can be
12573 retained (in particular w.rt. extensions) without having to duplicate the
12574 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12578 * Enhance the general user interface with mechanisms for inner control
12579 and with possibilities to have yes/no kind of prompts.
12583 * Change all calls to low-level digest routines in the library and
12584 applications to use EVP. Add missing calls to HMAC_cleanup() and
12585 don't assume HMAC_CTX can be copied using memcpy().
12587 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12589 * Add the possibility to control engines through control names but with
12590 arbitrary arguments instead of just a string.
12591 Change the key loaders to take a UI_METHOD instead of a callback
12592 function pointer. NOTE: this breaks binary compatibility with earlier
12593 versions of OpenSSL [engine].
12594 Adapt the nCipher code for these new conditions and add a card insertion
12599 * Enhance the general user interface with mechanisms to better support
12600 dialog box interfaces, application-defined prompts, the possibility
12601 to use defaults (for example default passwords from somewhere else)
12602 and interrupts/cancellations.
12606 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12607 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12611 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12612 tidy up some unnecessarily weird code in 'sk_new()').
12614 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12616 * Change the key loading routines for ENGINEs to use the same kind
12617 callback (pem_password_cb) as all other routines that need this
12622 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12623 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12624 than this minimum value is recommended.
12628 * New random seeder for OpenVMS, using the system process statistics
12629 that are easily reachable.
12633 * Windows apparently can't transparently handle global
12634 variables defined in DLLs. Initialisations such as:
12636 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12638 won't compile. This is used by the any applications that need to
12639 declare their own ASN1 modules. This was fixed by adding the option
12640 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12641 needed for static libraries under Win32.
12645 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12646 setting of purpose and trust fields. New X509_STORE trust and
12647 purpose functions and tidy up setting in other SSL functions.
12651 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12652 structure. These are inherited by X509_STORE_CTX when it is
12653 initialised. This allows various defaults to be set in the
12654 X509_STORE structure (such as flags for CRL checking and custom
12655 purpose or trust settings) for functions which only use X509_STORE_CTX
12656 internally such as S/MIME.
12658 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12659 trust settings if they are not set in X509_STORE. This allows X509_STORE
12660 purposes and trust (in S/MIME for example) to override any set by default.
12662 Add command line options for CRL checking to smime, s_client and s_server
12667 * Initial CRL based revocation checking. If the CRL checking flag(s)
12668 are set then the CRL is looked up in the X509_STORE structure and
12669 its validity and signature checked, then if the certificate is found
12670 in the CRL the verify fails with a revoked error.
12672 Various new CRL related callbacks added to X509_STORE_CTX structure.
12674 Command line options added to 'verify' application to support this.
12676 This needs some additional work, such as being able to handle multiple
12677 CRLs with different times, extension based lookup (rather than just
12678 by subject name) and ultimately more complete V2 CRL extension
12683 * Add a general user interface API (crypto/ui/). This is designed
12684 to replace things like des_read_password and friends (backward
12685 compatibility functions using this new API are provided).
12686 The purpose is to remove prompting functions from the DES code
12687 section as well as provide for prompting through dialog boxes in
12688 a window system and the like.
12692 * Add "ex_data" support to ENGINE so implementations can add state at a
12693 per-structure level rather than having to store it globally.
12697 * Make it possible for ENGINE structures to be copied when retrieved by
12698 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12699 This causes the "original" ENGINE structure to act like a template,
12700 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12701 operational state can be localised to each ENGINE structure, despite the
12702 fact they all share the same "methods". New ENGINE structures returned in
12703 this case have no functional references and the return value is the single
12704 structural reference. This matches the single structural reference returned
12705 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12710 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12711 needs to match any other type at all we need to manually clear the
12716 * Changes to the "openssl engine" utility to include;
12717 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12718 about an ENGINE's available control commands.
12719 - executing control commands from command line arguments using the
12720 '-pre' and '-post' switches. '-post' is only used if '-t' is
12721 specified and the ENGINE is successfully initialised. The syntax for
12722 the individual commands are colon-separated, for example;
12723 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12727 * New dynamic control command support for ENGINEs. ENGINEs can now
12728 declare their own commands (numbers), names (strings), descriptions,
12729 and input types for run-time discovery by calling applications. A
12730 subset of these commands are implicitly classed as "executable"
12731 depending on their input type, and only these can be invoked through
12732 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12733 can be based on user input, config files, etc). The distinction is
12734 that "executable" commands cannot return anything other than a boolean
12735 result and can only support numeric or string input, whereas some
12736 discoverable commands may only be for direct use through
12737 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12738 pointers, or other custom uses. The "executable" commands are to
12739 support parameterisations of ENGINE behaviour that can be
12740 unambiguously defined by ENGINEs and used consistently across any
12741 OpenSSL-based application. Commands have been added to all the
12742 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12743 control over shared-library paths without source code alterations.
12747 * Changed all ENGINE implementations to dynamically allocate their
12748 ENGINEs rather than declaring them statically. Apart from this being
12749 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12750 this also allows the implementations to compile without using the
12751 internal engine_int.h header.
12755 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12756 'const' value. Any code that should be able to modify a RAND_METHOD
12757 should already have non-const pointers to it (ie. they should only
12758 modify their own ones).
12762 * Made a variety of little tweaks to the ENGINE code.
12763 - "atalla" and "ubsec" string definitions were moved from header files
12764 to C code. "nuron" string definitions were placed in variables
12765 rather than hard-coded - allowing parameterisation of these values
12766 later on via ctrl() commands.
12767 - Removed unused "#if 0"'d code.
12768 - Fixed engine list iteration code so it uses ENGINE_free() to release
12769 structural references.
12770 - Constified the RAND_METHOD element of ENGINE structures.
12771 - Constified various get/set functions as appropriate and added
12772 missing functions (including a catch-all ENGINE_cpy that duplicates
12773 all ENGINE values onto a new ENGINE except reference counts/state).
12774 - Removed NULL parameter checks in get/set functions. Setting a method
12775 or function to NULL is a way of cancelling out a previously set
12776 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12777 and doesn't justify the extra error symbols and code.
12778 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12779 flags from engine_int.h to engine.h.
12780 - Changed prototypes for ENGINE handler functions (init(), finish(),
12781 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12785 * Implement binary inversion algorithm for BN_mod_inverse in addition
12786 to the algorithm using long division. The binary algorithm can be
12787 used only if the modulus is odd. On 32-bit systems, it is faster
12788 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12789 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12790 up to 450 bits. In 64-bit environments, the binary algorithm
12791 appears to be advantageous for much longer moduli; here we use it
12792 for moduli up to 2048 bits.
12796 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12797 could not support the combine flag in choice fields.
12801 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12802 extensions from a certificate request to the certificate.
12806 * Allow multiple 'certopt' and 'nameopt' options to be separated
12807 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12808 file: this allows the display of the certificate about to be
12809 signed to be customised, to allow certain fields to be included
12810 or excluded and extension details. The old system didn't display
12811 multicharacter strings properly, omitted fields not in the policy
12812 and couldn't display additional details such as extensions.
12816 * Function EC_POINTs_mul for multiple scalar multiplication
12817 of an arbitrary number of elliptic curve points
12818 \sum scalars[i]*points[i],
12819 optionally including the generator defined for the EC_GROUP:
12820 scalar*generator + \sum scalars[i]*points[i].
12822 EC_POINT_mul is a simple wrapper function for the typical case
12823 that the point list has just one item (besides the optional
12828 * First EC_METHODs for curves over GF(p):
12830 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12831 operations and provides various method functions that can also
12832 operate with faster implementations of modular arithmetic.
12834 EC_GFp_mont_method() reuses most functions that are part of
12835 EC_GFp_simple_method, but uses Montgomery arithmetic.
12837 *Bodo Moeller; point addition and point doubling
12838 implementation directly derived from source code provided by
12839 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12841 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12842 crypto/ec/ec_lib.c):
12844 Curves are EC_GROUP objects (with an optional group generator)
12845 based on EC_METHODs that are built into the library.
12847 Points are EC_POINT objects based on EC_GROUP objects.
12849 Most of the framework would be able to handle curves over arbitrary
12850 finite fields, but as there are no obvious types for fields other
12851 than GF(p), some functions are limited to that for now.
12855 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12856 that the file contains a complete HTTP response.
12860 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12861 change the def and num file printf format specifier from "%-40sXXX"
12862 to "%-39s XXX". The latter will always guarantee a space after the
12863 field while the former will cause them to run together if the field
12864 is 40 of more characters long.
12868 * Constify the cipher and digest 'method' functions and structures
12869 and modify related functions to take constant EVP_MD and EVP_CIPHER
12874 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12875 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12879 * Modify `EVP_Digest*()` routines so they now return values. Although the
12880 internal software routines can never fail additional hardware versions
12885 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12887 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12888 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12891 ERR_R_NESTED_ASN1_ERROR
12893 ERR_R_MISSING_ASN1_EOS
12894 were 4 .. 9, conflicting with
12895 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12897 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12898 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12900 Add new error code 'ERR_R_INTERNAL_ERROR'.
12904 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12909 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12910 sets the subject name for a new request or supersedes the
12911 subject name in a given request. Formats that can be parsed are
12912 'CN=Some Name, OU=myOU, C=IT'
12914 'CN=Some Name/OU=myOU/C=IT'.
12916 Add options '-batch' and '-verbose' to 'openssl req'.
12918 *Massimiliano Pala <madwolf@hackmasters.net>*
12920 * Introduce the possibility to access global variables through
12921 functions on platform were that's the best way to handle exporting
12922 global variables in shared libraries. To enable this functionality,
12923 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12924 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12925 is normally done by Configure or something similar).
12927 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12928 in the source file (foo.c) like this:
12930 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12931 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12933 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12934 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12936 OPENSSL_DECLARE_GLOBAL(int,foo);
12937 #define foo OPENSSL_GLOBAL_REF(foo)
12938 OPENSSL_DECLARE_GLOBAL(double,bar);
12939 #define bar OPENSSL_GLOBAL_REF(bar)
12941 The #defines are very important, and therefore so is including the
12942 header file everywhere where the defined globals are used.
12944 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12945 of ASN.1 items, but that structure is a bit different.
12947 The largest change is in util/mkdef.pl which has been enhanced with
12948 better and easier to understand logic to choose which symbols should
12949 go into the Windows .def files as well as a number of fixes and code
12950 cleanup (among others, algorithm keywords are now sorted
12951 lexicographically to avoid constant rewrites).
12955 * In BN_div() keep a copy of the sign of 'num' before writing the
12956 result to 'rm' because if rm==num the value will be overwritten
12957 and produce the wrong result if 'num' is negative: this caused
12958 problems with BN_mod() and BN_nnmod().
12962 * Function OCSP_request_verify(). This checks the signature on an
12963 OCSP request and verifies the signer certificate. The signer
12964 certificate is just checked for a generic purpose and OCSP request
12969 * Add OCSP_check_validity() function to check the validity of OCSP
12970 responses. OCSP responses are prepared in real time and may only
12971 be a few seconds old. Simply checking that the current time lies
12972 between thisUpdate and nextUpdate max reject otherwise valid responses
12973 caused by either OCSP responder or client clock inaccuracy. Instead
12974 we allow thisUpdate and nextUpdate to fall within a certain period of
12975 the current time. The age of the response can also optionally be
12976 checked. Two new options -validity_period and -status_age added to
12981 * If signature or public key algorithm is unrecognized print out its
12982 OID rather that just UNKNOWN.
12986 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12987 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12988 ID to be generated from the issuer certificate alone which can then be
12989 passed to OCSP_id_issuer_cmp().
12993 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12994 ASN1 modules to export functions returning ASN1_ITEM pointers
12995 instead of the ASN1_ITEM structures themselves. This adds several
12996 new macros which allow the underlying ASN1 function/structure to
12997 be accessed transparently. As a result code should not use ASN1_ITEM
12998 references directly (such as &X509_it) but instead use the relevant
12999 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13000 use of the new ASN1 code on platforms where exporting structures
13001 is problematical (for example in shared libraries) but exporting
13002 functions returning pointers to structures is not.
13006 * Add support for overriding the generation of SSL/TLS session IDs.
13007 These callbacks can be registered either in an SSL_CTX or per SSL.
13008 The purpose of this is to allow applications to control, if they wish,
13009 the arbitrary values chosen for use as session IDs, particularly as it
13010 can be useful for session caching in multiple-server environments. A
13011 command-line switch for testing this (and any client code that wishes
13012 to use such a feature) has been added to "s_server".
13014 *Geoff Thorpe, Lutz Jaenicke*
13016 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13017 of the form `#if defined(...) || defined(...) || ...` and
13018 `#if !defined(...) && !defined(...) && ...`. This also avoids
13019 the growing number of special cases it was previously handling.
13023 * Make all configuration macros available for application by making
13024 sure they are available in opensslconf.h, by giving them names starting
13025 with `OPENSSL_` to avoid conflicts with other packages and by making
13026 sure e_os2.h will cover all platform-specific cases together with
13028 Additionally, it is now possible to define configuration/platform-
13029 specific names (called "system identities"). In the C code, these
13030 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13031 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13032 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13037 * New option -set_serial to 'req' and 'x509' this allows the serial
13038 number to use to be specified on the command line. Previously self
13039 signed certificates were hard coded with serial number 0 and the
13040 CA options of 'x509' had to use a serial number in a file which was
13045 * New options to 'ca' utility to support V2 CRL entry extensions.
13046 Currently CRL reason, invalidity date and hold instruction are
13047 supported. Add new CRL extensions to V3 code and some new objects.
13051 * New function EVP_CIPHER_CTX_set_padding() this is used to
13052 disable standard block padding (aka PKCS#5 padding) in the EVP
13053 API, which was previously mandatory. This means that the data is
13054 not padded in any way and so the total length much be a multiple
13055 of the block size, otherwise an error occurs.
13059 * Initial (incomplete) OCSP SSL support.
13063 * New function OCSP_parse_url(). This splits up a URL into its host,
13064 port and path components: primarily to parse OCSP URLs. New -url
13065 option to ocsp utility.
13069 * New nonce behavior. The return value of OCSP_check_nonce() now
13070 reflects the various checks performed. Applications can decide
13071 whether to tolerate certain situations such as an absent nonce
13072 in a response when one was present in a request: the ocsp application
13073 just prints out a warning. New function OCSP_add1_basic_nonce()
13074 this is to allow responders to include a nonce in a response even if
13075 the request is nonce-less.
13079 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13080 skipped when using openssl x509 multiple times on a single input file,
13081 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13085 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13086 set string type: to handle setting ASN1_TIME structures. Fix ca
13087 utility to correctly initialize revocation date of CRLs.
13091 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13092 the clients preferred ciphersuites and rather use its own preferences.
13093 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13094 Internet Explorer by ensuring unchanged hash method during stepup.
13095 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13099 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13100 to aes and add a new 'exist' option to print out symbols that don't
13105 * Additional options to ocsp utility to allow flags to be set and
13106 additional certificates supplied.
13110 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13111 OCSP client a number of certificate to only verify the response
13116 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13117 handle the new API. Currently only ECB, CBC modes supported. Add new
13120 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13121 Encryption Standard (AES) Ciphersuites for Transport Layer
13122 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13123 not enabled by default and were not part of the "ALL" ciphersuite
13124 alias because they were not yet official; they could be
13125 explicitly requested by specifying the "AESdraft" ciphersuite
13126 group alias. In the final release of OpenSSL 0.9.7, the group
13127 alias is called "AES" and is part of "ALL".)
13129 *Ben Laurie, Steve Henson, Bodo Moeller*
13131 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13132 request to response.
13136 * Functions for OCSP responders. OCSP_request_onereq_count(),
13137 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13138 extract information from a certificate request. OCSP_response_create()
13139 creates a response and optionally adds a basic response structure.
13140 OCSP_basic_add1_status() adds a complete single response to a basic
13141 response and returns the OCSP_SINGLERESP structure just added (to allow
13142 extensions to be included for example). OCSP_basic_add1_cert() adds a
13143 certificate to a basic response and OCSP_basic_sign() signs a basic
13144 response with various flags. New helper functions ASN1_TIME_check()
13145 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13146 (converts ASN1_TIME to GeneralizedTime).
13150 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13151 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13152 structure from a certificate. X509_pubkey_digest() digests the public_key
13153 contents: this is used in various key identifiers.
13157 * Make sk_sort() tolerate a NULL argument.
13159 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13161 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13162 passed by the function are trusted implicitly. If any of them signed the
13163 response then it is assumed to be valid and is not verified.
13167 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13168 to data. This was previously part of the PKCS7 ASN1 code. This
13169 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13170 *Steve Henson, reported by Kenneth R. Robinette
13171 <support@securenetterm.com>*
13173 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13174 routines: without these tracing memory leaks is very painful.
13175 Fix leaks in PKCS12 and PKCS7 routines.
13179 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13180 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13181 effectively meant GeneralizedTime would never be used. Now it
13182 is initialised to -1 but X509_time_adj() now has to check the value
13183 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13184 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13185 *Steve Henson, reported by Kenneth R. Robinette
13186 <support@securenetterm.com>*
13188 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13189 result in a zero length in the ASN1_INTEGER structure which was
13190 not consistent with the structure when d2i_ASN1_INTEGER() was used
13191 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13192 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13193 where it did not print out a minus for negative ASN1_INTEGER.
13197 * Add summary printout to ocsp utility. The various functions which
13198 convert status values to strings have been renamed to:
13199 OCSP_response_status_str(), OCSP_cert_status_str() and
13200 OCSP_crl_reason_str() and are no longer static. New options
13201 to verify nonce values and to disable verification. OCSP response
13202 printout format cleaned up.
13206 * Add additional OCSP certificate checks. These are those specified
13207 in RFC2560. This consists of two separate checks: the CA of the
13208 certificate being checked must either be the OCSP signer certificate
13209 or the issuer of the OCSP signer certificate. In the latter case the
13210 OCSP signer certificate must contain the OCSP signing extended key
13211 usage. This check is performed by attempting to match the OCSP
13212 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13213 in the OCSP_CERTID structures of the response.
13217 * Initial OCSP certificate verification added to OCSP_basic_verify()
13218 and related routines. This uses the standard OpenSSL certificate
13219 verify routines to perform initial checks (just CA validity) and
13220 to obtain the certificate chain. Then additional checks will be
13221 performed on the chain. Currently the root CA is checked to see
13222 if it is explicitly trusted for OCSP signing. This is used to set
13223 a root CA as a global signing root: that is any certificate that
13224 chains to that CA is an acceptable OCSP signing certificate.
13228 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13229 extensions from a separate configuration file.
13230 As when reading extensions from the main configuration file,
13231 the '-extensions ...' option may be used for specifying the
13234 *Massimiliano Pala <madwolf@comune.modena.it>*
13236 * New OCSP utility. Allows OCSP requests to be generated or
13237 read. The request can be sent to a responder and the output
13238 parsed, outputted or printed in text form. Not complete yet:
13239 still needs to check the OCSP response validity.
13243 * New subcommands for 'openssl ca':
13244 `openssl ca -status <serial>` prints the status of the cert with
13245 the given serial number (according to the index file).
13246 `openssl ca -updatedb` updates the expiry status of certificates
13249 *Massimiliano Pala <madwolf@comune.modena.it>*
13251 * New '-newreq-nodes' command option to CA.pl. This is like
13252 '-newreq', but calls 'openssl req' with the '-nodes' option
13253 so that the resulting key is not encrypted.
13255 *Damien Miller <djm@mindrot.org>*
13257 * New configuration for the GNU Hurd.
13259 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13261 * Initial code to implement OCSP basic response verify. This
13262 is currently incomplete. Currently just finds the signer's
13263 certificate and verifies the signature on the response.
13267 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13268 value of OPENSSLDIR. This is available via the new '-d' option
13269 to 'openssl version', and is also included in 'openssl version -a'.
13273 * Allowing defining memory allocation callbacks that will be given
13274 file name and line number information in additional arguments
13275 (a `const char*` and an int). The basic functionality remains, as
13276 well as the original possibility to just replace malloc(),
13277 realloc() and free() by functions that do not know about these
13278 additional arguments. To register and find out the current
13279 settings for extended allocation functions, the following
13280 functions are provided:
13282 CRYPTO_set_mem_ex_functions
13283 CRYPTO_set_locked_mem_ex_functions
13284 CRYPTO_get_mem_ex_functions
13285 CRYPTO_get_locked_mem_ex_functions
13287 These work the same way as CRYPTO_set_mem_functions and friends.
13288 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13289 extended allocation function is enabled.
13290 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13291 a conventional allocation function is enabled.
13293 *Richard Levitte, Bodo Moeller*
13295 * Finish off removing the remaining LHASH function pointer casts.
13296 There should no longer be any prototype-casting required when using
13297 the LHASH abstraction, and any casts that remain are "bugs". See
13298 the callback types and macros at the head of lhash.h for details
13299 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13303 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13304 If /dev/[u]random devices are not available or do not return enough
13305 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13307 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13308 /etc/entropy will be queried once each in this sequence, querying stops
13309 when enough entropy was collected without querying more sockets.
13313 * Change the Unix RAND_poll() variant to be able to poll several
13314 random devices, as specified by DEVRANDOM, until a sufficient amount
13315 of data has been collected. We spend at most 10 ms on each file
13316 (select timeout) and read in non-blocking mode. DEVRANDOM now
13317 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13318 (previously it was just the string "/dev/urandom"), so on typical
13319 platforms the 10 ms delay will never occur.
13320 Also separate out the Unix variant to its own file, rand_unix.c.
13321 For VMS, there's a currently-empty rand_vms.c.
13325 * Move OCSP client related routines to ocsp_cl.c. These
13326 provide utility functions which an application needing
13327 to issue a request to an OCSP responder and analyse the
13328 response will typically need: as opposed to those which an
13329 OCSP responder itself would need which will be added later.
13331 OCSP_request_sign() signs an OCSP request with an API similar
13332 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13333 response. OCSP_response_get1_basic() extracts basic response
13334 from response. OCSP_resp_find_status(): finds and extracts status
13335 information from an OCSP_CERTID structure (which will be created
13336 when the request structure is built). These are built from lower
13337 level functions which work on OCSP_SINGLERESP structures but
13338 won't normally be used unless the application wishes to examine
13339 extensions in the OCSP response for example.
13341 Replace nonce routines with a pair of functions.
13342 OCSP_request_add1_nonce() adds a nonce value and optionally
13343 generates a random value. OCSP_check_nonce() checks the
13344 validity of the nonce in an OCSP response.
13348 * Change function OCSP_request_add() to OCSP_request_add0_id().
13349 This doesn't copy the supplied OCSP_CERTID and avoids the
13350 need to free up the newly created id. Change return type
13351 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13352 This can then be used to add extensions to the request.
13353 Deleted OCSP_request_new(), since most of its functionality
13354 is now in OCSP_REQUEST_new() (and the case insensitive name
13355 clash) apart from the ability to set the request name which
13356 will be added elsewhere.
13360 * Update OCSP API. Remove obsolete extensions argument from
13361 various functions. Extensions are now handled using the new
13362 OCSP extension code. New simple OCSP HTTP function which
13363 can be used to send requests and parse the response.
13367 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13368 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13369 uses the special reorder version of SET OF to sort the attributes
13370 and reorder them to match the encoded order. This resolves a long
13371 standing problem: a verify on a PKCS7 structure just after signing
13372 it used to fail because the attribute order did not match the
13373 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13374 it uses the received order. This is necessary to tolerate some broken
13375 software that does not order SET OF. This is handled by encoding
13376 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13377 to produce the required SET OF.
13381 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13382 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13383 files to get correct declarations of the ASN.1 item variables.
13387 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13388 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13389 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13390 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13391 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13392 ASN1_ITEM and no wrapper functions.
13396 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13397 replace the old function pointer based I/O routines. Change most of
13398 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13402 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13403 lines, recognize more "algorithms" that can be deselected, and make
13404 it complain about algorithm deselection that isn't recognised.
13408 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13409 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13410 to use new functions. Add NO_ASN1_OLD which can be set to remove
13411 some old style ASN1 functions: this can be used to determine if old
13412 code will still work when these eventually go away.
13416 * New extension functions for OCSP structures, these follow the
13417 same conventions as certificates and CRLs.
13421 * New function X509V3_add1_i2d(). This automatically encodes and
13422 adds an extension. Its behaviour can be customised with various
13423 flags to append, replace or delete. Various wrappers added for
13424 certificates and CRLs.
13428 * Fix to avoid calling the underlying ASN1 print routine when
13429 an extension cannot be parsed. Correct a typo in the
13430 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13434 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13435 entries for variables.
13439 * Add functionality to `apps/openssl.c` for detecting locking
13440 problems: As the program is single-threaded, all we have
13441 to do is register a locking callback using an array for
13442 storing which locks are currently held by the program.
13446 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13447 SSL_get_ex_data_X509_STORE_idx(), which is used in
13448 ssl_verify_cert_chain() and thus can be called at any time
13449 during TLS/SSL handshakes so that thread-safety is essential.
13450 Unfortunately, the ex_data design is not at all suited
13451 for multi-threaded use, so it probably should be abolished.
13455 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13457 *Broadcom, tweaked and integrated by Geoff Thorpe*
13459 * Move common extension printing code to new function
13460 X509V3_print_extensions(). Reorganise OCSP print routines and
13461 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13465 * New function X509_signature_print() to remove duplication in some
13470 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13471 set (this was treated exactly the same as SET OF previously). This
13472 is used to reorder the STACK representing the structure to match the
13473 encoding. This will be used to get round a problem where a PKCS7
13474 structure which was signed could not be verified because the STACK
13475 order did not reflect the encoded order.
13479 * Reimplement the OCSP ASN1 module using the new code.
13483 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13484 for its ASN1 operations. The old style function pointers still exist
13485 for now but they will eventually go away.
13489 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13490 completely replaces the old ASN1 functionality with a table driven
13491 encoder and decoder which interprets an ASN1_ITEM structure describing
13492 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13493 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13494 has also been converted to the new form.
13498 * Change BN_mod_exp_recp so that negative moduli are tolerated
13499 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13500 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13501 for negative moduli.
13505 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13506 of not touching the result's sign bit.
13510 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13515 * Changed the LHASH code to use prototypes for callbacks, and created
13516 macros to declare and implement thin (optionally static) functions
13517 that provide type-safety and avoid function pointer casting for the
13518 type-specific callbacks.
13522 * Added Kerberos Cipher Suites to be used with TLS, as written in
13524 *Veers Staats <staatsvr@asc.hpc.mil>,
13525 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13527 * Reformat the FAQ so the different questions and answers can be divided
13528 in sections depending on the subject.
13532 * Have the zlib compression code load ZLIB.DLL dynamically under
13537 * New function BN_mod_sqrt for computing square roots modulo a prime
13538 (using the probabilistic Tonelli-Shanks algorithm unless
13539 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13540 be handled deterministically).
13542 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13544 * Make BN_mod_inverse faster by explicitly handling small quotients
13545 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13546 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13550 * New function BN_kronecker.
13554 * Fix BN_gcd so that it works on negative inputs; the result is
13555 positive unless both parameters are zero.
13556 Previously something reasonably close to an infinite loop was
13557 possible because numbers could be growing instead of shrinking
13558 in the implementation of Euclid's algorithm.
13562 * Fix BN_is_word() and BN_is_one() macros to take into account the
13563 sign of the number in question.
13565 Fix BN_is_word(a,w) to work correctly for w == 0.
13567 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13568 because its test if the absolute value of 'a' equals 'w'.
13569 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13570 it exists mostly for use in the implementations of BN_is_zero(),
13571 BN_is_one(), and BN_is_word().
13575 * New function BN_swap.
13579 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13580 the exponentiation functions are more likely to produce reasonable
13581 results on negative inputs.
13585 * Change BN_mod_mul so that the result is always non-negative.
13586 Previously, it could be negative if one of the factors was negative;
13587 I don't think anyone really wanted that behaviour.
13591 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13592 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13593 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13594 and add new functions:
13603 BN_mod_lshift1_quick
13605 BN_mod_lshift_quick
13607 These functions always generate non-negative results.
13609 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13610 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13612 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13613 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13614 be reduced modulo `m`.
13616 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13619 The following entry accidentally appeared in the CHANGES file
13620 distributed with OpenSSL 0.9.7. The modifications described in
13621 it do *not* apply to OpenSSL 0.9.7.
13623 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13624 was actually never needed) and in BN_mul(). The removal in BN_mul()
13625 required a small change in bn_mul_part_recursive() and the addition
13626 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13627 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13628 bn_sub_words() and bn_add_words() except they take arrays with
13634 * In 'openssl passwd', verify passwords read from the terminal
13635 unless the '-salt' option is used (which usually means that
13636 verification would just waste user's time since the resulting
13637 hash is going to be compared with some given password hash)
13638 or the new '-noverify' option is used.
13640 This is an incompatible change, but it does not affect
13641 non-interactive use of 'openssl passwd' (passwords on the command
13642 line, '-stdin' option, '-in ...' option) and thus should not
13643 cause any problems.
13647 * Remove all references to RSAref, since there's no more need for it.
13651 * Make DSO load along a path given through an environment variable
13652 (SHLIB_PATH) with shl_load().
13656 * Constify the ENGINE code as a result of BIGNUM constification.
13657 Also constify the RSA code and most things related to it. In a
13658 few places, most notable in the depth of the ASN.1 code, ugly
13659 casts back to non-const were required (to be solved at a later
13664 * Make it so the openssl application has all engines loaded by default.
13668 * Constify the BIGNUM routines a little more.
13672 * Add the following functions:
13674 ENGINE_load_cswift()
13676 ENGINE_load_atalla()
13677 ENGINE_load_nuron()
13678 ENGINE_load_builtin_engines()
13680 That way, an application can itself choose if external engines that
13681 are built-in in OpenSSL shall ever be used or not. The benefit is
13682 that applications won't have to be linked with libdl or other dso
13683 libraries unless it's really needed.
13685 Changed 'openssl engine' to load all engines on demand.
13686 Changed the engine header files to avoid the duplication of some
13687 declarations (they differed!).
13691 * 'openssl engine' can now list capabilities.
13695 * Better error reporting in 'openssl engine'.
13699 * Never call load_dh_param(NULL) in s_server.
13703 * Add engine application. It can currently list engines by name and
13704 identity, and test if they are actually available.
13708 * Improve RPM specification file by forcing symbolic linking and making
13709 sure the installed documentation is also owned by root.root.
13711 *Damien Miller <djm@mindrot.org>*
13713 * Give the OpenSSL applications more possibilities to make use of
13714 keys (public as well as private) handled by engines.
13718 * Add OCSP code that comes from CertCo.
13722 * Add VMS support for the Rijndael code.
13726 * Added untested support for Nuron crypto accelerator.
13730 * Add support for external cryptographic devices. This code was
13731 previously distributed separately as the "engine" branch.
13733 *Geoff Thorpe, Richard Levitte*
13735 * Rework the filename-translation in the DSO code. It is now possible to
13736 have far greater control over how a "name" is turned into a filename
13737 depending on the operating environment and any oddities about the
13738 different shared library filenames on each system.
13742 * Support threads on FreeBSD-elf in Configure.
13746 * Fix for SHA1 assembly problem with MASM: it produces
13747 warnings about corrupt line number information when assembling
13748 with debugging information. This is caused by the overlapping
13751 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13754 NCONF_get_number() has no error checking at all. As a replacement,
13755 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13756 promoted strongly. The old NCONF_get_number is kept around for
13757 binary backward compatibility.
13758 Make it possible for methods to load from something other than a BIO,
13759 by providing a function pointer that is given a name instead of a BIO.
13760 For example, this could be used to load configuration data from an
13765 * Fix for non blocking accept BIOs. Added new I/O special reason
13766 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13767 with non blocking I/O was not possible because no retry code was
13768 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13773 * Added the beginnings of Rijndael support.
13777 * Fix for bug in DirectoryString mask setting. Add support for
13778 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13779 to allow certificate printing to more controllable, additional
13780 'certopt' option to 'x509' to allow new printing options to be
13785 * Clean old EAY MD5 hack from e_os.h.
13789 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13791 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13792 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13794 *Joe Orton, Steve Henson*
13796 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13798 * Fix additional bug revealed by the NISCC test suite:
13800 Stop bug triggering large recursion when presented with
13801 certain ASN.1 tags ([CVE-2003-0851])
13805 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13807 * Fix various bugs revealed by running the NISCC test suite:
13809 Stop out of bounds reads in the ASN1 code when presented with
13810 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13812 If verify callback ignores invalid public key errors don't try to check
13813 certificate signature with the NULL public key.
13817 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13818 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13823 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13824 extra data after the compression methods not only for TLS 1.0
13825 but also for SSL 3.0 (as required by the specification).
13827 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13829 * Change X509_certificate_type() to mark the key as exported/exportable
13830 when it's 512 *bits* long, not 512 bytes.
13834 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13836 * Countermeasure against the Klima-Pokorny-Rosa extension of
13837 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13838 a protocol version number mismatch like a decryption error
13839 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13843 * Turn on RSA blinding by default in the default implementation
13844 to avoid a timing attack. Applications that don't want it can call
13845 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13846 They would be ill-advised to do so in most cases.
13848 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13850 * Change RSA blinding code so that it works when the PRNG is not
13851 seeded (in this case, the secret RSA exponent is abused as
13852 an unpredictable seed -- if it is not unpredictable, there
13853 is no point in blinding anyway). Make RSA blinding thread-safe
13854 by remembering the creator's thread ID in rsa->blinding and
13855 having all other threads use local one-time blinding factors
13856 (this requires more computation than sharing rsa->blinding, but
13857 avoids excessive locking; and if an RSA object is not shared
13858 between threads, blinding will still be very fast).
13862 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13864 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13865 via timing by performing a MAC computation even if incorrect
13866 block cipher padding has been found. This is a countermeasure
13867 against active attacks where the attacker has to distinguish
13868 between bad padding and a MAC verification error. ([CVE-2003-0078])
13870 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13871 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13872 Martin Vuagnoux (EPFL, Ilion)*
13874 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13876 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13877 memory from its contents. This is done with a counter that will
13878 place alternating values in each byte. This can be used to solve
13879 two issues: 1) the removal of calls to memset() by highly optimizing
13880 compilers, and 2) cleansing with other values than 0, since those can
13881 be read through on certain media, for example a swap space on disk.
13885 * Bugfix: client side session caching did not work with external caching,
13886 because the session->cipher setting was not restored when reloading
13887 from the external cache. This problem was masked, when
13888 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13889 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13893 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13894 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13896 *Zeev Lieber <zeev-l@yahoo.com>*
13898 * Undo an undocumented change introduced in 0.9.6e which caused
13899 repeated calls to OpenSSL_add_all_ciphers() and
13900 OpenSSL_add_all_digests() to be ignored, even after calling
13905 * Change the default configuration reader to deal with last line not
13906 being properly terminated.
13910 * Change X509_NAME_cmp() so it applies the special rules on handling
13911 DN values that are of type PrintableString, as well as RDNs of type
13912 emailAddress where the value has the type ia5String.
13914 *stefank@valicert.com via Richard Levitte*
13916 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13917 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13918 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13919 the bitwise-OR of the two for use by the majority of applications
13920 wanting this behaviour, and update the docs. The documented
13921 behaviour and actual behaviour were inconsistent and had been
13922 changing anyway, so this is more a bug-fix than a behavioural
13925 *Geoff Thorpe, diagnosed by Nadav Har'El*
13927 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13928 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13932 * Fix initialization code race conditions in
13933 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13934 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13935 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13936 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13937 ssl2_get_cipher_by_char(),
13938 ssl3_get_cipher_by_char().
13940 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13942 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13943 the cached sessions are flushed, as the remove_cb() might use ex_data
13944 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13945 (see [openssl.org #212]).
13947 *Geoff Thorpe, Lutz Jaenicke*
13949 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13950 length, instead of the encoding length to d2i_ASN1_OBJECT.
13954 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13956 * [In 0.9.6g-engine release:]
13957 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13959 *Lynn Gazis <lgazis@rainbow.com>*
13961 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13963 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13964 and get fix the header length calculation.
13965 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13966 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13968 * Use proper error handling instead of 'assertions' in buffer
13969 overflow checks added in 0.9.6e. This prevents DoS (the
13970 assertions could call abort()).
13972 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13974 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13976 * Add various sanity checks to asn1_get_length() to reject
13977 the ASN1 length bytes if they exceed sizeof(long), will appear
13978 negative or the content length exceeds the length of the
13981 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13983 * Fix cipher selection routines: ciphers without encryption had no flags
13984 for the cipher strength set and where therefore not handled correctly
13985 by the selection routines (PR #130).
13989 * Fix EVP_dsa_sha macro.
13994 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13995 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13996 that was added in OpenSSL 0.9.6d.
13998 As the countermeasure turned out to be incompatible with some
13999 broken SSL implementations, the new option is part of SSL_OP_ALL.
14000 SSL_OP_ALL is usually employed when compatibility with weird SSL
14001 implementations is desired (e.g. '-bugs' option to 's_client' and
14002 's_server'), so the new option is automatically set in many
14007 * Changes in security patch:
14009 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14010 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14011 Air Force Materiel Command, USAF, under agreement number
14014 * Add various sanity checks to asn1_get_length() to reject
14015 the ASN1 length bytes if they exceed sizeof(long), will appear
14016 negative or the content length exceeds the length of the
14017 supplied buffer. ([CVE-2002-0659])
14019 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14021 * Assertions for various potential buffer overflows, not known to
14022 happen in practice.
14024 *Ben Laurie (CHATS)*
14026 * Various temporary buffers to hold ASCII versions of integers were
14027 too small for 64 bit platforms. ([CVE-2002-0655])
14028 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14030 * Remote buffer overflow in SSL3 protocol - an attacker could
14031 supply an oversized session ID to a client. ([CVE-2002-0656])
14033 *Ben Laurie (CHATS)*
14035 * Remote buffer overflow in SSL2 protocol - an attacker could
14036 supply an oversized client master key. ([CVE-2002-0656])
14038 *Ben Laurie (CHATS)*
14040 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14042 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14043 encoded as NULL) with id-dsa-with-sha1.
14045 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14047 * Check various `X509_...()` return values in `apps/req.c`.
14049 *Nils Larsch <nla@trustcenter.de>*
14051 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14052 an end-of-file condition would erroneously be flagged, when the CRLF
14053 was just at the end of a processed block. The bug was discovered when
14054 processing data through a buffering memory BIO handing the data to a
14055 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14056 <ptsekov@syntrex.com> and Nedelcho Stanev.
14060 * Implement a countermeasure against a vulnerability recently found
14061 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14062 before application data chunks to avoid the use of known IVs
14063 with data potentially chosen by the attacker.
14067 * Fix length checks in ssl3_get_client_hello().
14071 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14072 to prevent ssl3_read_internal() from incorrectly assuming that
14073 ssl3_read_bytes() found application data while handshake
14074 processing was enabled when in fact s->s3->in_read_app_data was
14075 merely automatically cleared during the initial handshake.
14077 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14079 * Fix object definitions for Private and Enterprise: they were not
14080 recognized in their shortname (=lowercase) representation. Extend
14081 obj_dat.pl to issue an error when using undefined keywords instead
14082 of silently ignoring the problem (Svenning Sorensen
14083 <sss@sss.dnsalias.net>).
14087 * Fix DH_generate_parameters() so that it works for 'non-standard'
14088 generators, i.e. generators other than 2 and 5. (Previously, the
14089 code did not properly initialise the 'add' and 'rem' values to
14090 BN_generate_prime().)
14092 In the new general case, we do not insist that 'generator' is
14093 actually a primitive root: This requirement is rather pointless;
14094 a generator of the order-q subgroup is just as good, if not
14099 * Map new X509 verification errors to alerts. Discovered and submitted by
14100 Tom Wu <tom@arcot.com>.
14104 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14105 returning non-zero before the data has been completely received
14106 when using non-blocking I/O.
14108 *Bodo Moeller; problem pointed out by John Hughes*
14110 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14112 *Ben Laurie, Lutz Jaenicke*
14114 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14115 Yoram Zahavi <YoramZ@gilian.com>).
14119 * Add information about CygWin 1.3 and on, and preserve proper
14120 configuration for the versions before that.
14122 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14124 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14125 check whether we deal with a copy of a session and do not delete from
14126 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14127 <izhar@checkpoint.com>.
14131 * Do not store session data into the internal session cache, if it
14132 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14133 flag is set). Proposed by Aslam <aslam@funk.com>.
14137 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14142 * [In 0.9.6d-engine release:]
14143 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14145 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14147 * Add the configuration target linux-s390x.
14149 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14151 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14152 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14153 variable as an indication that a ClientHello message has been
14154 received. As the flag value will be lost between multiple
14155 invocations of ssl3_accept when using non-blocking I/O, the
14156 function may not be aware that a handshake has actually taken
14157 place, thus preventing a new session from being added to the
14160 To avoid this problem, we now set s->new_session to 2 instead of
14161 using a local variable.
14163 *Lutz Jaenicke, Bodo Moeller*
14165 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14166 if the SSL_R_LENGTH_MISMATCH error is detected.
14168 *Geoff Thorpe, Bodo Moeller*
14170 * New 'shared_ldflag' column in Configure platform table.
14174 * Fix EVP_CIPHER_mode macro.
14176 *"Dan S. Camper" <dan@bti.net>*
14178 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14179 type, we must throw them away by setting rr->length to 0.
14181 *D P Chang <dpc@qualys.com>*
14183 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14185 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14186 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14187 worked incorrectly for those cases where range = `10..._2` and
14188 `3*range` is two bits longer than range.)
14192 * Only add signing time to PKCS7 structures if it is not already
14197 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14198 OBJ_ld_ce should be OBJ_id_ce.
14199 Also some ip-pda OIDs in crypto/objects/objects.txt were
14200 incorrect (cf. RFC 3039).
14202 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14204 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14205 returns early because it has nothing to do.
14207 *Andy Schneider <andy.schneider@bjss.co.uk>*
14209 * [In 0.9.6c-engine release:]
14210 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14212 *Andy Schneider <andy.schneider@bjss.co.uk>*
14214 * [In 0.9.6c-engine release:]
14215 Add support for Cryptographic Appliance's keyserver technology.
14216 (Use engine 'keyclient')
14218 *Cryptographic Appliances and Geoff Thorpe*
14220 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14221 is called via tools/c89.sh because arguments have to be
14222 rearranged (all '-L' options must appear before the first object
14225 *Richard Shapiro <rshapiro@abinitio.com>*
14227 * [In 0.9.6c-engine release:]
14228 Add support for Broadcom crypto accelerator cards, backported
14231 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14233 * [In 0.9.6c-engine release:]
14234 Add support for SureWare crypto accelerator cards from
14235 Baltimore Technologies. (Use engine 'sureware')
14237 *Baltimore Technologies and Mark Cox*
14239 * [In 0.9.6c-engine release:]
14240 Add support for crypto accelerator cards from Accelerated
14241 Encryption Processing, www.aep.ie. (Use engine 'aep')
14243 *AEP Inc. and Mark Cox*
14245 * Add a configuration entry for gcc on UnixWare.
14247 *Gary Benson <gbenson@redhat.com>*
14249 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14250 messages are stored in a single piece (fixed-length part and
14251 variable-length part combined) and fix various bugs found on the way.
14255 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14256 instead. BIO_gethostbyname() does not know what timeouts are
14257 appropriate, so entries would stay in cache even when they have
14259 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14261 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14262 faced with a pathologically small ClientHello fragment that does
14263 not contain client_version: Instead of aborting with an error,
14264 simply choose the highest available protocol version (i.e.,
14265 TLS 1.0 unless it is disabled). In practice, ClientHello
14266 messages are never sent like this, but this change gives us
14267 strictly correct behaviour at least for TLS.
14271 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14272 never resets s->method to s->ctx->method when called from within
14273 one of the SSL handshake functions.
14275 *Bodo Moeller; problem pointed out by Niko Baric*
14277 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14278 (sent using the client's version number) if client_version is
14279 smaller than the protocol version in use. Also change
14280 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14281 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14282 the client will at least see that alert.
14286 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14291 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14292 client receives HelloRequest while in a handshake.
14294 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14296 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14297 should end in 'break', not 'goto end' which circumvents various
14298 cleanups done in state SSL_ST_OK. But session related stuff
14299 must be disabled for SSL_ST_OK in the case that we just sent a
14302 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14303 before just sending a HelloRequest.
14305 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14307 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14308 reveal whether illegal block cipher padding was found or a MAC
14309 verification error occurred. (Neither SSLerr() codes nor alerts
14310 are directly visible to potential attackers, but the information
14311 may leak via logfiles.)
14313 Similar changes are not required for the SSL 2.0 implementation
14314 because the number of padding bytes is sent in clear for SSL 2.0,
14315 and the extra bytes are just ignored. However ssl/s2_pkt.c
14316 failed to verify that the purported number of padding bytes is in
14321 * Add OpenUNIX-8 support including shared libraries
14322 (Boyd Lynn Gerber <gerberb@zenez.com>).
14326 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14327 'wristwatch attack' using huge encoding parameters (cf.
14328 James H. Manger's CRYPTO 2001 paper). Note that the
14329 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14330 encoding parameters and hence was not vulnerable.
14334 * BN_sqr() bug fix.
14336 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14338 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14339 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14340 followed by modular reduction.
14342 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14344 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14345 equivalent based on BN_pseudo_rand() instead of BN_rand().
14349 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14350 This function was broken, as the check for a new client hello message
14351 to handle SGC did not allow these large messages.
14352 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14356 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14360 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14361 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14365 * Rework the configuration and shared library support for Tru64 Unix.
14366 The configuration part makes use of modern compiler features and
14367 still retains old compiler behavior for those that run older versions
14368 of the OS. The shared library support part includes a variant that
14369 uses the RPATH feature, and is available through the special
14370 configuration target "alpha-cc-rpath", which will never be selected
14373 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14375 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14376 with the same message size as in ssl3_get_certificate_request().
14377 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14378 messages might inadvertently be reject as too long.
14380 *Petr Lampa <lampa@fee.vutbr.cz>*
14382 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14386 * Modified SSL library such that the verify_callback that has been set
14387 specifically for an SSL object with SSL_set_verify() is actually being
14388 used. Before the change, a verify_callback set with this function was
14389 ignored and the verify_callback() set in the SSL_CTX at the time of
14390 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14391 to allow the necessary settings.
14395 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14396 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14397 done automatically (in contradiction to the requirements of the C
14398 standard). This made problems when used from OpenSSH.
14402 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14403 dh->length and always used
14405 BN_rand_range(priv_key, dh->p).
14407 BN_rand_range() is not necessary for Diffie-Hellman, and this
14408 specific range makes Diffie-Hellman unnecessarily inefficient if
14409 dh->length (recommended exponent length) is much smaller than the
14410 length of dh->p. We could use BN_rand_range() if the order of
14411 the subgroup was stored in the DH structure, but we only have
14416 BN_rand(priv_key, l, ...)
14418 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14425 RSA_eay_public_encrypt
14426 RSA_eay_private_decrypt
14427 RSA_eay_private_encrypt (signing)
14428 RSA_eay_public_decrypt (signature verification)
14430 (default implementations for RSA_public_encrypt,
14431 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14432 always reject numbers >= n.
14436 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14437 to synchronize access to 'locking_thread'. This is necessary on
14438 systems where access to 'locking_thread' (an 'unsigned long'
14439 variable) is not atomic.
14443 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14444 *before* setting the 'crypto_lock_rand' flag. The previous code had
14445 a race condition if 0 is a valid thread ID.
14447 *Travis Vitek <vitek@roguewave.com>*
14449 * Add support for shared libraries under Irix.
14451 *Albert Chin-A-Young <china@thewrittenword.com>*
14453 * Add configuration option to build on Linux on both big-endian and
14454 little-endian MIPS.
14456 *Ralf Baechle <ralf@uni-koblenz.de>*
14458 * Add the possibility to create shared libraries on HP-UX.
14462 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14464 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14465 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14466 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14467 PRNG state recovery was possible based on the output of
14468 one PRNG request appropriately sized to gain knowledge on
14469 'md' followed by enough consecutive 1-byte PRNG requests
14470 to traverse all of 'state'.
14472 1. When updating 'md_local' (the current thread's copy of 'md')
14473 during PRNG output generation, hash all of the previous
14474 'md_local' value, not just the half used for PRNG output.
14476 2. Make the number of bytes from 'state' included into the hash
14477 independent from the number of PRNG bytes requested.
14479 The first measure alone would be sufficient to avoid
14480 Markku-Juhani's attack. (Actually it had never occurred
14481 to me that the half of 'md_local' used for chaining was the
14482 half from which PRNG output bytes were taken -- I had always
14483 assumed that the secret half would be used.) The second
14484 measure makes sure that additional data from 'state' is never
14485 mixed into 'md_local' in small portions; this heuristically
14486 further strengthens the PRNG.
14490 * Fix crypto/bn/asm/mips3.s.
14494 * When only the key is given to "enc", the IV is undefined. Print out
14495 an error message in this case.
14499 * Handle special case when X509_NAME is empty in X509 printing routines.
14503 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14504 positive and less than q.
14508 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14509 used: it isn't thread safe and the add_lock_callback should handle
14512 *Paul Rose <Paul.Rose@bridge.com>*
14514 * Verify that incoming data obeys the block size in
14515 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14521 *Ulf Möller, Bodo Möller*
14523 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14524 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14525 when fixing the server behaviour for backwards-compatible 'client
14526 hello' messages. (Note that the attack is impractical against
14527 SSL 3.0 and TLS 1.0 anyway because length and version checking
14528 means that the probability of guessing a valid ciphertext is
14529 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14532 Before 0.9.5, the countermeasure (hide the error by generating a
14533 random 'decryption result') did not work properly because
14534 ERR_clear_error() was missing, meaning that SSL_get_error() would
14535 detect the supposedly ignored error.
14537 Both problems are now fixed.
14541 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14542 (previously it was 1024).
14546 * Fix for compatibility mode trust settings: ignore trust settings
14547 unless some valid trust or reject settings are present.
14551 * Fix for blowfish EVP: its a variable length cipher.
14555 * Fix various bugs related to DSA S/MIME verification. Handle missing
14556 parameters in DSA public key structures and return an error in the
14557 DSA routines if parameters are absent.
14561 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14562 in the current directory if neither $RANDFILE nor $HOME was set.
14563 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14564 caused some confusion to Windows users who haven't defined $HOME.
14565 Thus RAND_file_name() is changed again: e_os.h can define a
14566 DEFAULT_HOME, which will be used if $HOME is not set.
14567 For Windows, we use "C:"; on other platforms, we still require
14568 environment variables.
14570 * Move 'if (!initialized) RAND_poll()' into regions protected by
14571 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14572 having multiple threads call RAND_poll() concurrently.
14576 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14577 combination of a flag and a thread ID variable.
14578 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14579 flag), *other* threads can enter ssleay_add_bytes without obeying
14580 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14581 that they do not hold after the first thread unsets add_do_not_lock).
14585 * Change bctest again: '-x' expressions are not available in all
14586 versions of 'test'.
14590 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14592 * Fix a couple of memory leaks in PKCS7_dataDecode()
14594 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14596 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14597 the default extension for executables, if any. Also, make the perl
14598 scripts that use symlink() to test if it really exists and use "cp"
14599 if it doesn't. All this made OpenSSL compilable and installable in
14604 * Fix for asn1_GetSequence() for indefinite length constructed data.
14605 If SEQUENCE is length is indefinite just set c->slen to the total
14606 amount of data available.
14608 *Steve Henson, reported by shige@FreeBSD.org*
14610 *This change does not apply to 0.9.7.*
14612 * Change bctest to avoid here-documents inside command substitution
14613 (workaround for FreeBSD /bin/sh bug).
14614 For compatibility with Ultrix, avoid shell functions (introduced
14615 in the bctest version that searches along $PATH).
14619 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14620 with des_encrypt() defined on some operating systems, like Solaris
14625 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14626 On the Importance of Eliminating Errors in Cryptographic
14627 Computations, J. Cryptology 14 (2001) 2, 101-119,
14628 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14632 * MIPS assembler BIGNUM division bug fix.
14636 * Disabled incorrect Alpha assembler code.
14640 * Fix PKCS#7 decode routines so they correctly update the length
14641 after reading an EOC for the EXPLICIT tag.
14645 *This change does not apply to 0.9.7.*
14647 * Fix bug in PKCS#12 key generation routines. This was triggered
14648 if a 3DES key was generated with a 0 initial byte. Include
14649 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14650 (but broken) behaviour.
14654 * Enhance bctest to search for a working bc along $PATH and print
14657 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14659 * Fix memory leaks in err.c: free err_data string if necessary;
14660 don't write to the wrong index in ERR_set_error_data.
14664 * Implement ssl23_peek (analogous to ssl23_read), which previously
14669 * Replace rdtsc with `_emit` statements for VC++ version 5.
14671 *Jeremy Cooper <jeremy@baymoo.org>*
14673 * Make it possible to reuse SSLv2 sessions.
14677 * In copy_email() check for >= 0 as a return value for
14678 X509_NAME_get_index_by_NID() since 0 is a valid index.
14680 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14682 * Avoid coredump with unsupported or invalid public keys by checking if
14683 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14684 PKCS7_verify() fails with non detached data.
14688 * Don't use getenv in library functions when run as setuid/setgid.
14689 New function OPENSSL_issetugid().
14693 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14694 due to incorrect handling of multi-threading:
14696 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14698 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14700 3. Count how many times MemCheck_off() has been called so that
14701 nested use can be treated correctly. This also avoids
14702 inband-signalling in the previous code (which relied on the
14703 assumption that thread ID 0 is impossible).
14707 * Add "-rand" option also to s_client and s_server.
14711 * Fix CPU detection on Irix 6.x.
14712 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14713 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14715 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14720 *This change does not apply to 0.9.7.*
14722 * Use the cached encoding of an X509_NAME structure rather than
14723 copying it. This is apparently the reason for the libsafe "errors"
14724 but the code is actually correct.
14728 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14729 Bleichenbacher's DSA attack.
14730 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14731 to be set and top=0 forces the highest bit to be set; top=-1 is new
14732 and leaves the highest bit random.
14734 *Ulf Moeller, Bodo Moeller*
14736 * In the `NCONF_...`-based implementations for `CONF_...` queries
14737 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14738 a temporary CONF structure with the data component set to NULL
14739 (which gives segmentation faults in lh_retrieve).
14740 Instead, use NULL for the CONF pointer in CONF_get_string and
14741 CONF_get_number (which may use environment variables) and directly
14742 return NULL from CONF_get_section.
14746 * Fix potential buffer overrun for EBCDIC.
14750 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14751 keyUsage if basicConstraints absent for a CA.
14755 * Make SMIME_write_PKCS7() write mail header values with a format that
14756 is more generally accepted (no spaces before the semicolon), since
14757 some programs can't parse those values properly otherwise. Also make
14758 sure BIO's that break lines after each write do not create invalid
14763 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14764 macros previously used would not encode an empty SEQUENCE OF
14765 and break the signature.
14769 *This change does not apply to 0.9.7.*
14771 * Zero the premaster secret after deriving the master secret in
14776 * Add some EVP_add_digest_alias registrations (as found in
14777 OpenSSL_add_all_digests()) to SSL_library_init()
14778 aka OpenSSL_add_ssl_algorithms(). This provides improved
14779 compatibility with peers using X.509 certificates
14780 with unconventional AlgorithmIdentifier OIDs.
14784 * Fix for Irix with NO_ASM.
14786 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14788 * ./config script fixes.
14790 *Ulf Moeller, Richard Levitte*
14792 * Fix 'openssl passwd -1'.
14796 * Change PKCS12_key_gen_asc() so it can cope with non null
14797 terminated strings whose length is passed in the passlen
14798 parameter, for example from PEM callbacks. This was done
14799 by adding an extra length parameter to asc2uni().
14801 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14803 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14804 call failed, free the DSA structure.
14808 * Fix to uni2asc() to cope with zero length Unicode strings.
14809 These are present in some PKCS#12 files.
14813 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14814 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14815 when writing a 32767 byte record.
14817 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14819 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14820 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14822 (RSA objects have a reference count access to which is protected
14823 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14824 so they are meant to be shared between threads.)
14825 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14826 "Reddie, Steven" <Steven.Reddie@ca.com>*
14828 * Fix a deadlock in CRYPTO_mem_leaks().
14832 * Use better test patterns in bntest.
14836 * rand_win.c fix for Borland C.
14840 * BN_rshift bugfix for n == 0.
14844 * Add a 'bctest' script that checks for some known 'bc' bugs
14845 so that 'make test' does not abort just because 'bc' is broken.
14849 * Store verify_result within SSL_SESSION also for client side to
14850 avoid potential security hole. (Re-used sessions on the client side
14851 always resulted in verify_result==X509_V_OK, not using the original
14852 result of the server certificate verification.)
14856 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14857 SSL3_RT_APPLICATION_DATA, return 0.
14858 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14863 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14864 releases, have been re-implemented by renaming the previous
14865 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14866 and ssl3_read_internal, respectively, and adding 'peek' parameters
14867 to them. The new ssl[23]_{read,peek} functions are calls to
14868 ssl[23]_read_internal with the 'peek' flag set appropriately.
14869 A 'peek' parameter has also been added to ssl3_read_bytes, which
14870 does the actual work for ssl3_read_internal.
14874 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14875 the method-specific "init()" handler. Also clean up ex_data after
14876 calling the method-specific "finish()" handler. Previously, this was
14877 happening the other way round.
14881 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14882 The previous value, 12, was not always sufficient for BN_mod_exp().
14886 * Make sure that shared libraries get the internal name engine with
14887 the full version number and not just 0. This should mark the
14888 shared libraries as not backward compatible. Of course, this should
14889 be changed again when we can guarantee backward binary compatibility.
14893 * Fix typo in get_cert_by_subject() in by_dir.c
14895 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14897 * Rework the system to generate shared libraries:
14899 - Make note of the expected extension for the shared libraries and
14900 if there is a need for symbolic links from for example libcrypto.so.0
14901 to libcrypto.so.0.9.7. There is extended info in Configure for
14904 - Make as few rebuilds of the shared libraries as possible.
14906 - Still avoid linking the OpenSSL programs with the shared libraries.
14908 - When installing, install the shared libraries separately from the
14913 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14915 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14916 and not in SSL_clear because the latter is also used by the
14917 accept/connect functions; previously, the settings made by
14918 SSL_set_read_ahead would be lost during the handshake.
14920 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14922 * Correct util/mkdef.pl to be selective about disabled algorithms.
14923 Previously, it would create entries for disabled algorithms no
14928 * Added several new manual pages for SSL_* function.
14932 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14934 * In ssl23_get_client_hello, generate an error message when faced
14935 with an initial SSL 3.0/TLS record that is too small to contain the
14936 first two bytes of the ClientHello message, i.e. client_version.
14937 (Note that this is a pathologic case that probably has never happened
14938 in real life.) The previous approach was to use the version number
14939 from the record header as a substitute; but our protocol choice
14940 should not depend on that one because it is not authenticated
14941 by the Finished messages.
14945 * More robust randomness gathering functions for Windows.
14947 *Jeffrey Altman <jaltman@columbia.edu>*
14949 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14950 not set then we don't setup the error code for issuer check errors
14951 to avoid possibly overwriting other errors which the callback does
14952 handle. If an application does set the flag then we assume it knows
14953 what it is doing and can handle the new informational codes
14958 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14959 a general "ANY" type, as such it should be able to decode anything
14960 including tagged types. However it didn't check the class so it would
14961 wrongly interpret tagged types in the same way as their universal
14962 counterpart and unknown types were just rejected. Changed so that the
14963 tagged and unknown types are handled in the same way as a SEQUENCE:
14964 that is the encoding is stored intact. There is also a new type
14965 "V_ASN1_OTHER" which is used when the class is not universal, in this
14966 case we have no idea what the actual type is so we just lump them all
14971 * On VMS, stdout may very well lead to a file that is written to
14972 in a record-oriented fashion. That means that every write() will
14973 write a separate record, which will be read separately by the
14974 programs trying to read from it. This can be very confusing.
14976 The solution is to put a BIO filter in the way that will buffer
14977 text until a linefeed is reached, and then write everything a
14978 line at a time, so every record written will be an actual line,
14979 not chunks of lines and not (usually doesn't happen, but I've
14980 seen it once) several lines in one record. BIO_f_linebuffer() is
14983 Currently, it's a VMS-only method, because that's where it has
14984 been tested well enough.
14988 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14989 it can return incorrect results.
14990 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14991 but it was in 0.9.6-beta[12].)
14995 * Disable the check for content being present when verifying detached
14996 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14997 include zero length content when signing messages.
15001 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15002 BIO_ctrl (for BIO pairs).
15006 * Add DSO method for VMS.
15010 * Bug fix: Montgomery multiplication could produce results with the
15015 * Add RPM specification openssl.spec and modify it to build three
15016 packages. The default package contains applications, application
15017 documentation and run-time libraries. The devel package contains
15018 include files, static libraries and function documentation. The
15019 doc package contains the contents of the doc directory. The original
15020 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15024 * Add a large number of documentation files for many SSL routines.
15026 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15028 * Add a configuration entry for Sony News 4.
15030 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15032 * Don't set the two most significant bits to one when generating a
15033 random number < q in the DSA library.
15037 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15038 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15039 the underlying transport is blocking) if a handshake took place.
15040 (The default behaviour is needed by applications such as s_client
15041 and s_server that use select() to determine when to use SSL_read;
15042 but for applications that know in advance when to expect data, it
15043 just makes things more complicated.)
15047 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15052 * Add a few more EBCDIC conditionals that make `req` and `x509`
15053 work better on such systems.
15055 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15057 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15058 Update PKCS12_parse() so it copies the friendlyName and the
15059 keyid to the certificates aux info.
15063 * Fix bug in PKCS7_verify() which caused an infinite loop
15064 if there was more than one signature.
15066 *Sven Uszpelkat <su@celocom.de>*
15068 * Major change in util/mkdef.pl to include extra information
15069 about each symbol, as well as presenting variables as well
15070 as functions. This change means that there's n more need
15071 to rebuild the .num files when some algorithms are excluded.
15075 * Allow the verify time to be set by an application,
15076 rather than always using the current time.
15080 * Phase 2 verify code reorganisation. The certificate
15081 verify code now looks up an issuer certificate by a
15082 number of criteria: subject name, authority key id
15083 and key usage. It also verifies self signed certificates
15084 by the same criteria. The main comparison function is
15085 X509_check_issued() which performs these checks.
15087 Lot of changes were necessary in order to support this
15088 without completely rewriting the lookup code.
15090 Authority and subject key identifier are now cached.
15092 The LHASH 'certs' is X509_STORE has now been replaced
15093 by a STACK_OF(X509_OBJECT). This is mainly because an
15094 LHASH can't store or retrieve multiple objects with
15095 the same hash value.
15097 As a result various functions (which were all internal
15098 use only) have changed to handle the new X509_STORE
15099 structure. This will break anything that messed round
15100 with X509_STORE internally.
15102 The functions X509_STORE_add_cert() now checks for an
15103 exact match, rather than just subject name.
15105 The X509_STORE API doesn't directly support the retrieval
15106 of multiple certificates matching a given criteria, however
15107 this can be worked round by performing a lookup first
15108 (which will fill the cache with candidate certificates)
15109 and then examining the cache for matches. This is probably
15110 the best we can do without throwing out X509_LOOKUP
15111 entirely (maybe later...).
15113 The X509_VERIFY_CTX structure has been enhanced considerably.
15115 All certificate lookup operations now go via a get_issuer()
15116 callback. Although this currently uses an X509_STORE it
15117 can be replaced by custom lookups. This is a simple way
15118 to bypass the X509_STORE hackery necessary to make this
15119 work and makes it possible to use more efficient techniques
15120 in future. A very simple version which uses a simple
15121 STACK for its trusted certificate store is also provided
15122 using X509_STORE_CTX_trusted_stack().
15124 The verify_cb() and verify() callbacks now have equivalents
15125 in the X509_STORE_CTX structure.
15127 X509_STORE_CTX also has a 'flags' field which can be used
15128 to customise the verify behaviour.
15132 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15133 excludes S/MIME capabilities.
15137 * When a certificate request is read in keep a copy of the
15138 original encoding of the signed data and use it when outputting
15139 again. Signatures then use the original encoding rather than
15140 a decoded, encoded version which may cause problems if the
15141 request is improperly encoded.
15145 * For consistency with other BIO_puts implementations, call
15146 buffer_write(b, ...) directly in buffer_puts instead of calling
15149 In BIO_puts, increment b->num_write as in BIO_write.
15151 *Peter.Sylvester@EdelWeb.fr*
15153 * Fix BN_mul_word for the case where the word is 0. (We have to use
15154 BN_zero, we may not return a BIGNUM with an array consisting of
15155 words set to zero.)
15159 * Avoid calling abort() from within the library when problems are
15160 detected, except if preprocessor symbols have been defined
15161 (such as REF_CHECK, BN_DEBUG etc.).
15165 * New openssl application 'rsautl'. This utility can be
15166 used for low-level RSA operations. DER public key
15167 BIO/fp routines also added.
15171 * New Configure entry and patches for compiling on QNX 4.
15173 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15175 * A demo state-machine implementation was sponsored by
15176 Nuron (<http://www.nuron.com/>) and is now available in
15177 demos/state_machine.
15181 * New options added to the 'dgst' utility for signature
15182 generation and verification.
15186 * Unrecognized PKCS#7 content types are now handled via a
15187 catch all ASN1_TYPE structure. This allows unsupported
15188 types to be stored as a "blob" and an application can
15189 encode and decode it manually.
15193 * Fix various signed/unsigned issues to make a_strex.c
15194 compile under VC++.
15196 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15198 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15199 length if passed a buffer. ASN1_INTEGER_to_BN failed
15200 if passed a NULL BN and its argument was negative.
15202 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15204 * Modification to PKCS#7 encoding routines to output definite
15205 length encoding. Since currently the whole structures are in
15206 memory there's not real point in using indefinite length
15207 constructed encoding. However if OpenSSL is compiled with
15208 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15212 * Added BIO_vprintf() and BIO_vsnprintf().
15216 * Added more prefixes to parse for in the strings written
15217 through a logging bio, to cover all the levels that are available
15218 through syslog. The prefixes are now:
15220 PANIC, EMERG, EMR => LOG_EMERG
15221 ALERT, ALR => LOG_ALERT
15222 CRIT, CRI => LOG_CRIT
15223 ERROR, ERR => LOG_ERR
15224 WARNING, WARN, WAR => LOG_WARNING
15225 NOTICE, NOTE, NOT => LOG_NOTICE
15226 INFO, INF => LOG_INFO
15227 DEBUG, DBG => LOG_DEBUG
15229 and as before, if none of those prefixes are present at the
15230 beginning of the string, LOG_ERR is chosen.
15232 On Win32, the `LOG_*` levels are mapped according to this:
15234 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15235 LOG_WARNING => EVENTLOG_WARNING_TYPE
15236 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15240 * Made it possible to reconfigure with just the configuration
15241 argument "reconf" or "reconfigure". The command line arguments
15242 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15243 and are retrieved from there when reconfiguring.
15249 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15251 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15255 * The obj_dat.pl script was messing up the sorting of object
15256 names. The reason was that it compared the quoted version
15257 of strings as a result "OCSP" > "OCSP Signing" because
15258 " > SPACE. Changed script to store unquoted versions of
15259 names and add quotes on output. It was also omitting some
15260 names from the lookup table if they were given a default
15261 value (that is if SN is missing it is given the same
15262 value as LN and vice versa), these are now added on the
15263 grounds that if an object has a name we should be able to
15264 look it up. Finally added warning output when duplicate
15265 short or long names are found.
15269 * Changes needed for Tandem NSK.
15271 *Scott Uroff <scott@xypro.com>*
15273 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15274 RSA_padding_check_SSLv23(), special padding was never detected
15275 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15276 version rollback attacks was not effective.
15278 In s23_clnt.c, don't use special rollback-attack detection padding
15279 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15280 client; similarly, in s23_srvr.c, don't do the rollback check if
15281 SSL 2.0 is the only protocol enabled in the server.
15285 * Make it possible to get hexdumps of unprintable data with 'openssl
15286 asn1parse'. By implication, the functions ASN1_parse_dump() and
15287 BIO_dump_indent() are added.
15291 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15292 these print out strings and name structures based on various
15293 flags including RFC2253 support and proper handling of
15294 multibyte characters. Added options to the 'x509' utility
15295 to allow the various flags to be set.
15299 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15300 Also change the functions X509_cmp_current_time() and
15301 X509_gmtime_adj() work with an ASN1_TIME structure,
15302 this will enable certificates using GeneralizedTime in validity
15303 dates to be checked.
15307 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15308 negative public key encodings) on by default,
15309 NO_NEG_PUBKEY_BUG can be set to disable it.
15313 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15314 content octets. An i2c_ASN1_OBJECT is unnecessary because
15315 the encoding can be trivially obtained from the structure.
15319 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15320 not read locks (`CRYPTO_r_[un]lock`).
15324 * A first attempt at creating official support for shared
15325 libraries through configuration. I've kept it so the
15326 default is static libraries only, and the OpenSSL programs
15327 are always statically linked for now, but there are
15328 preparations for dynamic linking in place.
15329 This has been tested on Linux and Tru64.
15333 * Randomness polling function for Win9x, as described in:
15334 Peter Gutmann, Software Generation of Practically Strong
15339 * Fix so PRNG is seeded in req if using an already existing
15344 * New options to smime application. -inform and -outform
15345 allow alternative formats for the S/MIME message including
15346 PEM and DER. The -content option allows the content to be
15347 specified separately. This should allow things like Netscape
15348 form signing output easier to verify.
15352 * Fix the ASN1 encoding of tags using the 'long form'.
15356 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15357 STRING types. These convert content octets to and from the
15358 underlying type. The actual tag and length octets are
15359 already assumed to have been read in and checked. These
15360 are needed because all other string types have virtually
15361 identical handling apart from the tag. By having versions
15362 of the ASN1 functions that just operate on content octets
15363 IMPLICIT tagging can be handled properly. It also allows
15364 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15365 and ASN1_INTEGER are identical apart from the tag.
15369 * Change the handling of OID objects as follows:
15371 - New object identifiers are inserted in objects.txt, following
15372 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15373 - objects.pl is used to process obj_mac.num and create a new
15375 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15378 This is currently kind of a hack, and the perl code in objects.pl
15379 isn't very elegant, but it works as I intended. The simplest way
15380 to check that it worked correctly is to look in obj_dat.h and
15381 check the array nid_objs and make sure the objects haven't moved
15382 around (this is important!). Additions are OK, as well as
15383 consistent name changes.
15387 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15391 * Addition of the command line parameter '-rand file' to 'openssl req'.
15392 The given file adds to whatever has already been seeded into the
15393 random pool through the RANDFILE configuration file option or
15394 environment variable, or the default random state file.
15398 * mkstack.pl now sorts each macro group into lexical order.
15399 Previously the output order depended on the order the files
15400 appeared in the directory, resulting in needless rewriting
15405 * Patches to make OpenSSL compile under Win32 again. Mostly
15406 work arounds for the VC++ problem that it treats func() as
15407 func(void). Also stripped out the parts of mkdef.pl that
15408 added extra typesafe functions: these no longer exist.
15412 * Reorganisation of the stack code. The macros are now all
15413 collected in safestack.h . Each macro is defined in terms of
15414 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15415 DEBUG_SAFESTACK is now handled in terms of function casts,
15416 this has the advantage of retaining type safety without the
15417 use of additional functions. If DEBUG_SAFESTACK is not defined
15418 then the non typesafe macros are used instead. Also modified the
15419 mkstack.pl script to handle the new form. Needs testing to see
15420 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15421 the default if no major problems. Similar behaviour for ASN1_SET_OF
15422 and PKCS12_STACK_OF.
15426 * When some versions of IIS use the 'NET' form of private key the
15427 key derivation algorithm is different. Normally MD5(password) is
15428 used as a 128 bit RC4 key. In the modified case
15429 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15430 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15431 as the old Netscape_RSA functions except they have an additional
15432 'sgckey' parameter which uses the modified algorithm. Also added
15433 an -sgckey command line option to the rsa utility. Thanks to
15434 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15435 algorithm to openssl-dev.
15439 * The evp_local.h macros were using 'c.##kname' which resulted in
15440 invalid expansion on some systems (SCO 5.0.5 for example).
15441 Corrected to 'c.kname'.
15443 *Phillip Porch <root@theporch.com>*
15445 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15446 a STACK of email addresses from a certificate or request, these look
15447 in the subject name and the subject alternative name extensions and
15448 omit any duplicate addresses.
15452 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15453 This makes DSA verification about 2 % faster.
15457 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15458 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15459 plus overhead for 1024 bit moduli).
15460 This makes exponentiations about 0.5 % faster for 1024 bit
15461 exponents (as measured by "openssl speed rsa2048").
15465 * Rename memory handling macros to avoid conflicts with other
15467 Malloc => OPENSSL_malloc
15468 Malloc_locked => OPENSSL_malloc_locked
15469 Realloc => OPENSSL_realloc
15470 Free => OPENSSL_free
15474 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15475 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15479 * CygWin32 support.
15481 *John Jarvie <jjarvie@newsguy.com>*
15483 * The type-safe stack code has been rejigged. It is now only compiled
15484 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15485 by default all type-specific stack functions are "#define"d back to
15486 standard stack functions. This results in more streamlined output
15487 but retains the type-safety checking possibilities of the original
15492 * The STACK code has been cleaned up, and certain type declarations
15493 that didn't make a lot of sense have been brought in line. This has
15494 also involved a cleanup of sorts in safestack.h to more correctly
15495 map type-safe stack functions onto their plain stack counterparts.
15496 This work has also resulted in a variety of "const"ifications of
15497 lots of the code, especially `_cmp` operations which should normally
15498 be prototyped with "const" parameters anyway.
15502 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15503 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15504 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15505 where all of 'md' is used each time the PRNG is used, but 'state'
15506 is used only indexed by a cyclic counter. As entropy may not be
15507 well distributed from the beginning, 'md' is important as a
15508 chaining variable. However, the output function chains only half
15509 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15510 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15511 in all of 'state' being rewritten, with the new values depending
15512 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15516 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15517 the handshake is continued after ssl_verify_cert_chain();
15518 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15519 can lead to 'unexplainable' connection aborts later.
15521 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15523 * Major EVP API cipher revision.
15524 Add hooks for extra EVP features. This allows various cipher
15525 parameters to be set in the EVP interface. Support added for variable
15526 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15527 setting of RC2 and RC5 parameters.
15529 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15532 Remove lots of duplicated code from the EVP library. For example *every*
15533 cipher init() function handles the 'iv' in the same way according to the
15534 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15535 for CFB and OFB modes they zero ctx->num.
15537 New functionality allows removal of S/MIME code RC2 hack.
15539 Most of the routines have the same form and so can be declared in terms
15542 By shifting this to the top level EVP_CipherInit() it can be removed from
15543 all individual ciphers. If the cipher wants to handle IVs or keys
15544 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15547 Change lots of functions like EVP_EncryptUpdate() to now return a
15548 value: although software versions of the algorithms cannot fail
15549 any installed hardware versions can.
15553 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15554 this option is set, tolerate broken clients that send the negotiated
15555 protocol version number instead of the requested protocol version
15560 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15561 i.e. non-zero for export ciphersuites, zero otherwise.
15562 Previous versions had this flag inverted, inconsistent with
15563 rsa_tmp_cb (..._TMP_RSA_CB).
15565 *Bodo Moeller; problem reported by Amit Chopra*
15567 * Add missing DSA library text string. Work around for some IIS
15568 key files with invalid SEQUENCE encoding.
15572 * Add a document (doc/standards.txt) that list all kinds of standards
15573 and so on that are implemented in OpenSSL.
15577 * Enhance c_rehash script. Old version would mishandle certificates
15578 with the same subject name hash and wouldn't handle CRLs at all.
15579 Added -fingerprint option to crl utility, to support new c_rehash
15584 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15588 * Fix for SSL server purpose checking. Server checking was
15589 rejecting certificates which had extended key usage present
15590 but no ssl client purpose.
15592 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15594 * Make PKCS#12 code work with no password. The PKCS#12 spec
15595 is a little unclear about how a blank password is handled.
15596 Since the password in encoded as a BMPString with terminating
15597 double NULL a zero length password would end up as just the
15598 double NULL. However no password at all is different and is
15599 handled differently in the PKCS#12 key generation code. NS
15600 treats a blank password as zero length. MSIE treats it as no
15601 password on export: but it will try both on import. We now do
15602 the same: PKCS12_parse() tries zero length and no password if
15603 the password is set to "" or NULL (NULL is now a valid password:
15604 it wasn't before) as does the pkcs12 application.
15608 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15609 perror when PEM_read_bio_X509_REQ fails, the error message must
15610 be obtained from the error queue.
15614 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15615 it in ERR_remove_state if appropriate, and change ERR_get_state
15616 accordingly to avoid race conditions (this is necessary because
15617 thread_hash is no longer constant once set).
15621 * Bugfix for linux-elf makefile.one.
15625 * RSA_get_default_method() will now cause a default
15626 RSA_METHOD to be chosen if one doesn't exist already.
15627 Previously this was only set during a call to RSA_new()
15628 or RSA_new_method(NULL) meaning it was possible for
15629 RSA_get_default_method() to return NULL.
15633 * Added native name translation to the existing DSO code
15634 that will convert (if the flag to do so is set) filenames
15635 that are sufficiently small and have no path information
15636 into a canonical native form. Eg. "blah" converted to
15637 "libblah.so" or "blah.dll" etc.
15641 * New function ERR_error_string_n(e, buf, len) which is like
15642 ERR_error_string(e, buf), but writes at most 'len' bytes
15643 including the 0 terminator. For ERR_error_string_n, 'buf'
15646 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15648 * CONF library reworked to become more general. A new CONF
15649 configuration file reader "class" is implemented as well as a
15650 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15651 old `CONF_*` functions are still there, but are reimplemented to
15652 work in terms of the new functions. Also, a set of functions
15653 to handle the internal storage of the configuration data is
15654 provided to make it easier to write new configuration file
15655 reader "classes" (I can definitely see something reading a
15656 configuration file in XML format, for example), called `_CONF_*`,
15657 or "the configuration storage API"...
15659 The new configuration file reading functions are:
15661 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15662 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15664 NCONF_default, NCONF_WIN32
15666 NCONF_dump_fp, NCONF_dump_bio
15668 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15669 NCONF_new creates a new CONF object. This works in the same way
15670 as other interfaces in OpenSSL, like the BIO interface.
15671 `NCONF_dump_*` dump the internal storage of the configuration file,
15672 which is useful for debugging. All other functions take the same
15673 arguments as the old `CONF_*` functions with the exception of the
15674 first that must be a `CONF *` instead of a `LHASH *`.
15676 To make it easier to use the new classes with the old `CONF_*` functions,
15677 the function CONF_set_default_method is provided.
15681 * Add '-tls1' option to 'openssl ciphers', which was already
15682 mentioned in the documentation but had not been implemented.
15683 (This option is not yet really useful because even the additional
15684 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15688 * Initial DSO code added into libcrypto for letting OpenSSL (and
15689 OpenSSL-based applications) load shared libraries and bind to
15690 them in a portable way.
15692 *Geoff Thorpe, with contributions from Richard Levitte*
15694 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15696 * Make sure _lrotl and _lrotr are only used with MSVC.
15698 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15699 (the default implementation of RAND_status).
15701 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15702 to '-clrext' (= clear extensions), as intended and documented.
15703 *Bodo Moeller; inconsistency pointed out by Michael Attili
15704 <attili@amaxo.com>*
15706 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15707 was larger than the MD block size.
15709 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15711 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15712 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15713 using the passed key: if the passed key was a private key the result
15714 of X509_print(), for example, would be to print out all the private key
15719 * des_quad_cksum() byte order bug fix.
15720 *Ulf Möller, using the problem description in krb4-0.9.7, where
15721 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15723 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15726 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15728 * For easily testing in shell scripts whether some command
15729 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15730 returns with exit code 0 iff no command of the given name is available.
15731 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15732 the output goes to stdout and nothing is printed to stderr.
15733 Additional arguments are always ignored.
15735 Since for each cipher there is a command of the same name,
15736 the 'no-cipher' compilation switches can be tested this way.
15738 ('openssl no-XXX' is not able to detect pseudo-commands such
15739 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15743 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15747 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15748 is set; it will be thrown away anyway because each handshake creates
15750 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15751 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15752 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15753 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15757 * New s_client option -ign_eof: EOF at stdin is ignored, and
15758 'Q' and 'R' lose their special meanings (quit/renegotiate).
15759 This is part of what -quiet does; unlike -quiet, -ign_eof
15760 does not suppress any output.
15764 * Add compatibility options to the purpose and trust code. The
15765 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15766 accepts a certificate or CA, this was the previous behaviour,
15767 with all the associated security issues.
15769 X509_TRUST_COMPAT is the old trust behaviour: only and
15770 automatically trust self signed roots in certificate store. A
15771 new trust setting X509_TRUST_DEFAULT is used to specify that
15772 a purpose has no associated trust setting and it should instead
15773 use the value in the default purpose.
15777 * Fix the PKCS#8 DSA private key code so it decodes keys again
15778 and fix a memory leak.
15782 * In util/mkerr.pl (which implements 'make errors'), preserve
15783 reason strings from the previous version of the .c file, as
15784 the default to have only downcase letters (and digits) in
15785 automatically generated reasons codes is not always appropriate.
15789 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15790 using strerror. Previously, ERR_reason_error_string() returned
15791 library names as reason strings for SYSerr; but SYSerr is a special
15792 case where small numbers are errno values, not library numbers.
15796 * Add '-dsaparam' option to 'openssl dhparam' application. This
15797 converts DSA parameters into DH parameters. (When creating parameters,
15798 DSA_generate_parameters is used.)
15802 * Include 'length' (recommended exponent length) in C code generated
15803 by 'openssl dhparam -C'.
15807 * The second argument to set_label in perlasm was already being used
15808 so couldn't be used as a "file scope" flag. Moved to third argument
15813 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15814 instead of RAND_bytes for encryption IVs and salts.
15818 * Include RAND_status() into RAND_METHOD instead of implementing
15819 it only for md_rand.c Otherwise replacing the PRNG by calling
15820 RAND_set_rand_method would be impossible.
15824 * Don't let DSA_generate_key() enter an infinite loop if the random
15825 number generation fails.
15829 * New 'rand' application for creating pseudo-random output.
15833 * Added configuration support for Linux/IA64
15835 *Rolf Haberrecker <rolf@suse.de>*
15837 * Assembler module support for Mingw32.
15841 * Shared library support for HPUX (in shlib/).
15843 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15845 * Shared library support for Solaris gcc.
15847 *Lutz Behnke <behnke@trustcenter.de>*
15849 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15851 * PKCS7_encrypt() was adding text MIME headers twice because they
15852 were added manually and by SMIME_crlf_copy().
15856 * In bntest.c don't call BN_rand with zero bits argument.
15858 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15860 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15861 case was implemented. This caused BN_div_recp() to fail occasionally.
15865 * Add an optional second argument to the set_label() in the perl
15866 assembly language builder. If this argument exists and is set
15867 to 1 it signals that the assembler should use a symbol whose
15868 scope is the entire file, not just the current function. This
15869 is needed with MASM which uses the format label:: for this scope.
15871 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15873 * Change the ASN1 types so they are typedefs by default. Before
15874 almost all types were #define'd to ASN1_STRING which was causing
15875 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15880 * Change names of new functions to the new get1/get0 naming
15881 convention: After 'get1', the caller owns a reference count
15882 and has to call `..._free`; 'get0' returns a pointer to some
15883 data structure without incrementing reference counters.
15884 (Some of the existing 'get' functions increment a reference
15885 counter, some don't.)
15886 Similarly, 'set1' and 'add1' functions increase reference
15887 counters or duplicate objects.
15891 * Allow for the possibility of temp RSA key generation failure:
15892 the code used to assume it always worked and crashed on failure.
15896 * Fix potential buffer overrun problem in BIO_printf().
15897 *Ulf Möller, using public domain code by Patrick Powell; problem
15898 pointed out by David Sacerdote <das33@cornell.edu>*
15900 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15901 RAND_egd() and RAND_status(). In the command line application,
15902 the EGD socket can be specified like a seed file using RANDFILE
15907 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15908 Some CAs (e.g. Verisign) distribute certificates in this form.
15912 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15913 list to exclude them. This means that no special compilation option
15914 is needed to use anonymous DH: it just needs to be included in the
15919 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15920 EVP_MD_type. The old functionality is available in a new macro called
15921 EVP_MD_md(). Change code that uses it and update docs.
15925 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15926 where the `void *` argument is replaced by a function pointer argument.
15927 Previously `void *` was abused to point to functions, which works on
15928 many platforms, but is not correct. As these functions are usually
15929 called by macros defined in OpenSSL header files, most source code
15930 should work without changes.
15934 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15935 sections with information on -D... compiler switches used for
15936 compiling the library so that applications can see them. To enable
15937 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15938 must be defined. E.g.,
15939 #define OPENSSL_ALGORITHM_DEFINES
15940 #include <openssl/opensslconf.h>
15941 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15943 *Richard Levitte, Ulf and Bodo Möller*
15945 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15950 * Change the 'other' type in certificate aux info to a STACK_OF
15951 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15952 the required ASN1 format: arbitrary types determined by an OID.
15956 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15957 argument to 'req'. This is not because the function is newer or
15958 better than others it just uses the work 'NEW' in the certificate
15959 request header lines. Some software needs this.
15963 * Reorganise password command line arguments: now passwords can be
15964 obtained from various sources. Delete the PEM_cb function and make
15965 it the default behaviour: i.e. if the callback is NULL and the
15966 usrdata argument is not NULL interpret it as a null terminated pass
15967 phrase. If usrdata and the callback are NULL then the pass phrase
15968 is prompted for as usual.
15972 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15973 the support is automatically enabled. The resulting binaries will
15974 autodetect the card and use it if present.
15976 *Ben Laurie and Compaq Inc.*
15978 * Work around for Netscape hang bug. This sends certificate request
15979 and server done in one record. Since this is perfectly legal in the
15980 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15981 the bugs/SSLv3 entry for more info.
15985 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15989 * Add -rand argument to smime and pkcs12 applications and read/write
15994 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15998 * Add command line password options to the remaining applications.
16002 * Bug fix for BN_div_recp() for numerators with an even number of
16007 * More tests in bntest.c, and changed test_bn output.
16011 * ./config recognizes MacOS X now.
16015 * Bug fix for BN_div() when the first words of num and divisor are
16016 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16020 * Add support for various broken PKCS#8 formats, and command line
16021 options to produce them.
16025 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16026 get temporary BIGNUMs from a BN_CTX.
16030 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16035 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16036 include a #define from the old name to the new. The original intent
16037 was that statically linked binaries could for example just call
16038 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16039 link with digests. This never worked because SSLeay_add_all_digests()
16040 and SSLeay_add_all_ciphers() were in the same source file so calling
16041 one would link with the other. They are now in separate source files.
16045 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16049 * Use a less unusual form of the Miller-Rabin primality test (it used
16050 a binary algorithm for exponentiation integrated into the Miller-Rabin
16051 loop, our standard modexp algorithms are faster).
16055 * Support for the EBCDIC character set completed.
16057 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16059 * Source code cleanups: use const where appropriate, eliminate casts,
16060 use `void *` instead of `char *` in lhash.
16064 * Bugfix: ssl3_send_server_key_exchange was not restartable
16065 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16066 this the server could overwrite ephemeral keys that the client
16071 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16072 using 50 iterations of the Rabin-Miller test.
16074 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16075 iterations of the Rabin-Miller test as required by the appendix
16076 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16077 As BN_is_prime_fasttest includes trial division, DSA parameter
16078 generation becomes much faster.
16080 This implies a change for the callback functions in DSA_is_prime
16081 and DSA_generate_parameters: The callback function is called once
16082 for each positive witness in the Rabin-Miller test, not just
16083 occasionally in the inner loop; and the parameters to the
16084 callback function now provide an iteration count for the outer
16085 loop rather than for the current invocation of the inner loop.
16086 DSA_generate_parameters additionally can call the callback
16087 function with an 'iteration count' of -1, meaning that a
16088 candidate has passed the trial division test (when q is generated
16089 from an application-provided seed, trial division is skipped).
16093 * New function BN_is_prime_fasttest that optionally does trial
16094 division before starting the Rabin-Miller test and has
16095 an additional BN_CTX * argument (whereas BN_is_prime always
16096 has to allocate at least one BN_CTX).
16097 'callback(1, -1, cb_arg)' is called when a number has passed the
16098 trial division stage.
16102 * Fix for bug in CRL encoding. The validity dates weren't being handled
16107 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16111 * New function BN_pseudo_rand().
16115 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16116 bignum version of BN_from_montgomery() with the working code from
16117 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16122 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16123 made it impossible to use the same SSL_SESSION data structure in
16124 SSL2 clients in multiple threads.
16128 * The return value of RAND_load_file() no longer counts bytes obtained
16129 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16130 to seed the PRNG (previously an explicit byte count was required).
16132 *Ulf Möller, Bodo Möller*
16134 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16135 used `char *` instead of `void *` and had casts all over the place.
16139 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16143 * Retain source code compatibility for BN_prime_checks macro:
16144 BN_is_prime(..., BN_prime_checks, ...) now uses
16145 BN_prime_checks_for_size to determine the appropriate number of
16146 Rabin-Miller iterations.
16150 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16151 DH_CHECK_P_NOT_SAFE_PRIME.
16152 (Check if this is true? OpenPGP calls them "strong".)
16156 * Merge the functionality of "dh" and "gendh" programs into a new program
16157 "dhparam". The old programs are retained for now but will handle DH keys
16158 (instead of parameters) in future.
16162 * Make the ciphers, s_server and s_client programs check the return values
16163 when a new cipher list is set.
16167 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16168 ciphers. Before when the 56bit ciphers were enabled the sorting was
16171 The syntax for the cipher sorting has been extended to support sorting by
16172 cipher-strength (using the strength_bits hard coded in the tables).
16173 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16175 Fix a bug in the cipher-command parser: when supplying a cipher command
16176 string with an "undefined" symbol (neither command nor alphanumeric
16177 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16178 an error is flagged.
16180 Due to the strength-sorting extension, the code of the
16181 ssl_create_cipher_list() function was completely rearranged. I hope that
16182 the readability was also increased :-)
16184 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16186 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16187 for the first serial number and places 2 in the serial number file. This
16188 avoids problems when the root CA is created with serial number zero and
16189 the first user certificate has the same issuer name and serial number
16194 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16195 the new code. Add documentation for this stuff.
16199 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16200 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16201 structures and behave in an analogous way to the X509v3 functions:
16202 they shouldn't be called directly but wrapper functions should be used
16205 So we also now have some wrapper functions that call the X509at functions
16206 when passed certificate requests. (TO DO: similar things can be done with
16207 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16208 things. Some of these need some d2i or i2d and print functionality
16209 because they handle more complex structures.)
16213 * Add missing #ifndefs that caused missing symbols when building libssl
16214 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16215 NO_RSA in `ssl/s2*.c`.
16217 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16219 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16220 has a return value which indicates the quality of the random data
16221 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16222 error queue. New function RAND_pseudo_bytes() generates output that is
16223 guaranteed to be unique but not unpredictable. RAND_add is like
16224 RAND_seed, but takes an extra argument for an entropy estimate
16225 (RAND_seed always assumes full entropy).
16229 * Do more iterations of Rabin-Miller probable prime test (specifically,
16230 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16231 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16232 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16233 false-positive rate of at most 2^-80 for random input.
16237 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16241 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16242 in the 0.9.5 release), this returns the chain
16243 from an X509_CTX structure with a dup of the stack and all
16244 the X509 reference counts upped: so the stack will exist
16245 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16248 Also make SSL_SESSION_print() print out the verify return
16253 * Add manpage for the pkcs12 command. Also change the default
16254 behaviour so MAC iteration counts are used unless the new
16255 -nomaciter option is used. This improves file security and
16256 only older versions of MSIE (4.0 for example) need it.
16260 * Honor the no-xxx Configure options when creating .DEF files.
16264 * Add PKCS#10 attributes to field table: challengePassword,
16265 unstructuredName and unstructuredAddress. These are taken from
16266 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16267 international characters are used.
16269 More changes to X509_ATTRIBUTE code: allow the setting of types
16270 based on strings. Remove the 'loc' parameter when adding
16271 attributes because these will be a SET OF encoding which is sorted
16276 * Initial changes to the 'req' utility to allow request generation
16277 automation. This will allow an application to just generate a template
16278 file containing all the field values and have req construct the
16281 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16282 used all over the place including certificate requests and PKCS#7
16283 structures. They are currently handled manually where necessary with
16284 some primitive wrappers for PKCS#7. The new functions behave in a
16285 manner analogous to the X509 extension functions: they allow
16286 attributes to be looked up by NID and added.
16288 Later something similar to the X509V3 code would be desirable to
16289 automatically handle the encoding, decoding and printing of the
16290 more complex types. The string types like challengePassword can
16291 be handled by the string table functions.
16293 Also modified the multi byte string table handling. Now there is
16294 a 'global mask' which masks out certain types. The table itself
16295 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16296 is useful when for example there is only one permissible type
16297 (as in countryName) and using the mask might result in no valid
16302 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16303 SSL_get_peer_finished to allow applications to obtain the latest
16304 Finished messages sent to the peer or expected from the peer,
16305 respectively. (SSL_get_peer_finished is usually the Finished message
16306 actually received from the peer, otherwise the protocol will be aborted.)
16308 As the Finished message are message digests of the complete handshake
16309 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16310 be used for external authentication procedures when the authentication
16311 provided by SSL/TLS is not desired or is not enough.
16315 * Enhanced support for Alpha Linux is added. Now ./config checks if
16316 the host supports BWX extension and if Compaq C is present on the
16317 $PATH. Just exploiting of the BWX extension results in 20-30%
16318 performance kick for some algorithms, e.g. DES and RC4 to mention
16319 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16324 * Add support for MS "fast SGC". This is arguably a violation of the
16325 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16326 weak crypto and after checking the certificate is SGC a second one
16327 with strong crypto. MS SGC stops the first handshake after receiving
16328 the server certificate message and sends a second client hello. Since
16329 a server will typically do all the time consuming operations before
16330 expecting any further messages from the client (server key exchange
16331 is the most expensive) there is little difference between the two.
16333 To get OpenSSL to support MS SGC we have to permit a second client
16334 hello message after we have sent server done. In addition we have to
16335 reset the MAC if we do get this second client hello.
16339 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16340 if a DER encoded private key is RSA or DSA traditional format. Changed
16341 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16342 format DER encoded private key. Newer code should use PKCS#8 format which
16343 has the key type encoded in the ASN1 structure. Added DER private key
16344 support to pkcs8 application.
16348 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16349 ciphersuites has been selected (as required by the SSL 3/TLS 1
16350 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16351 is set, we interpret this as a request to violate the specification
16352 (the worst that can happen is a handshake failure, and 'correct'
16353 behaviour would result in a handshake failure anyway).
16357 * In SSL_CTX_add_session, take into account that there might be multiple
16358 SSL_SESSION structures with the same session ID (e.g. when two threads
16359 concurrently obtain them from an external cache).
16360 The internal cache can handle only one SSL_SESSION with a given ID,
16361 so if there's a conflict, we now throw out the old one to achieve
16366 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16367 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16368 some routines that use cipher OIDs: some ciphers do not have OIDs
16369 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16374 * Simplify the trust setting structure and code. Now we just have
16375 two sequences of OIDs for trusted and rejected settings. These will
16376 typically have values the same as the extended key usage extension
16377 and any application specific purposes.
16379 The trust checking code now has a default behaviour: it will just
16380 check for an object with the same NID as the passed id. Functions can
16381 be provided to override either the default behaviour or the behaviour
16382 for a given id. SSL client, server and email already have functions
16383 in place for compatibility: they check the NID and also return "trusted"
16384 if the certificate is self signed.
16388 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16389 traditional format into an EVP_PKEY structure.
16393 * Add a password callback function PEM_cb() which either prompts for
16394 a password if usr_data is NULL or otherwise assumes it is a null
16395 terminated password. Allow passwords to be passed on command line
16396 environment or config files in a few more utilities.
16400 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16401 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16402 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16403 Update documentation.
16407 * Support for ASN1 "NULL" type. This could be handled before by using
16408 ASN1_TYPE but there wasn't any function that would try to read a NULL
16409 and produce an error if it couldn't. For compatibility we also have
16410 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16411 don't allocate anything because they don't need to.
16415 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16418 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16420 * Rebuild of the memory allocation routines used by OpenSSL code and
16421 possibly others as well. The purpose is to make an interface that
16422 provide hooks so anyone can build a separate set of allocation and
16423 deallocation routines to be used by OpenSSL, for example memory
16424 pool implementations, or something else, which was previously hard
16425 since Malloc(), Realloc() and Free() were defined as macros having
16426 the values malloc, realloc and free, respectively (except for Win32
16427 compilations). The same is provided for memory debugging code.
16428 OpenSSL already comes with functionality to find memory leaks, but
16429 this gives people a chance to debug other memory problems.
16431 With these changes, a new set of functions and macros have appeared:
16433 CRYPTO_set_mem_debug_functions() [F]
16434 CRYPTO_get_mem_debug_functions() [F]
16435 CRYPTO_dbg_set_options() [F]
16436 CRYPTO_dbg_get_options() [F]
16437 CRYPTO_malloc_debug_init() [M]
16439 The memory debug functions are NULL by default, unless the library
16440 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16441 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16442 gives the standard debugging functions that come with OpenSSL) or
16443 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16444 provided by the library user) must be used. When the standard
16445 debugging functions are used, CRYPTO_dbg_set_options can be used to
16446 request additional information:
16447 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16448 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16450 Also, things like CRYPTO_set_mem_functions will always give the
16451 expected result (the new set of functions is used for allocation
16452 and deallocation) at all times, regardless of platform and compiler
16455 To finish it up, some functions that were never use in any other
16456 way than through macros have a new API and new semantic:
16458 CRYPTO_dbg_malloc()
16459 CRYPTO_dbg_realloc()
16462 All macros of value have retained their old syntax.
16464 *Richard Levitte and Bodo Moeller*
16466 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16467 ordering of SMIMECapabilities wasn't in "strength order" and there
16468 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16473 * Some ASN1 types with illegal zero length encoding (INTEGER,
16474 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16476 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16478 * Merge in my S/MIME library for OpenSSL. This provides a simple
16479 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16480 functionality to handle multipart/signed properly) and a utility
16481 called 'smime' to call all this stuff. This is based on code I
16482 originally wrote for Celo who have kindly allowed it to be
16483 included in OpenSSL.
16487 * Add variants des_set_key_checked and des_set_key_unchecked of
16488 des_set_key (aka des_key_sched). Global variable des_check_key
16489 decides which of these is called by des_set_key; this way
16490 des_check_key behaves as it always did, but applications and
16491 the library itself, which was buggy for des_check_key == 1,
16492 have a cleaner way to pick the version they need.
16496 * New function PKCS12_newpass() which changes the password of a
16501 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16502 dynamic mix. In both cases the ids can be used as an index into the
16503 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16504 functions so they accept a list of the field values and the
16505 application doesn't need to directly manipulate the X509_TRUST
16510 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16515 * Modify the way the V3 extension code looks up extensions. This now
16516 works in a similar way to the object code: we have some "standard"
16517 extensions in a static table which is searched with OBJ_bsearch()
16518 and the application can add dynamic ones if needed. The file
16519 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16520 updated whenever a new extension is added to the core code and kept
16521 in ext_nid order. There is a simple program 'tabtest.c' which checks
16522 this. New extensions are not added too often so this file can readily
16523 be maintained manually.
16525 There are two big advantages in doing things this way. The extensions
16526 can be looked up immediately and no longer need to be "added" using
16527 X509V3_add_standard_extensions(): this function now does nothing.
16528 Side note: I get *lots* of email saying the extension code doesn't
16529 work because people forget to call this function.
16530 Also no dynamic allocation is done unless new extensions are added:
16531 so if we don't add custom extensions there is no need to call
16532 X509V3_EXT_cleanup().
16536 * Modify enc utility's salting as follows: make salting the default. Add a
16537 magic header, so unsalted files fail gracefully instead of just decrypting
16538 to garbage. This is because not salting is a big security hole, so people
16539 should be discouraged from doing it.
16543 * Fixes and enhancements to the 'x509' utility. It allowed a message
16544 digest to be passed on the command line but it only used this
16545 parameter when signing a certificate. Modified so all relevant
16546 operations are affected by the digest parameter including the
16547 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16548 DSA key was used because it didn't fix the digest.
16552 * Initial certificate chain verify code. Currently tests the untrusted
16553 certificates for consistency with the verify purpose (which is set
16554 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16556 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16557 this is because it will reject chains with invalid extensions whereas
16558 every previous version of OpenSSL and SSLeay made no checks at all.
16560 Trust code: checks the root CA for the relevant trust settings. Trust
16561 settings have an initial value consistent with the verify purpose: e.g.
16562 if the verify purpose is for SSL client use it expects the CA to be
16563 trusted for SSL client use. However the default value can be changed to
16564 permit custom trust settings: one example of this would be to only trust
16565 certificates from a specific "secure" set of CAs.
16567 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16568 which should be used for version portability: especially since the
16569 verify structure is likely to change more often now.
16571 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16572 to set them. If not set then assume SSL clients will verify SSL servers
16575 Two new options to the verify program: -untrusted allows a set of
16576 untrusted certificates to be passed in and -purpose which sets the
16577 intended purpose of the certificate. If a purpose is set then the
16578 new chain verify code is used to check extension consistency.
16582 * Support for the authority information access extension.
16586 * Modify RSA and DSA PEM read routines to transparently handle
16587 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16588 public keys in a format compatible with certificate
16589 SubjectPublicKeyInfo structures. Unfortunately there were already
16590 functions called *_PublicKey_* which used various odd formats so
16591 these are retained for compatibility: however the DSA variants were
16592 never in a public release so they have been deleted. Changed dsa/rsa
16593 utilities to handle the new format: note no releases ever handled public
16594 keys so we should be OK.
16596 The primary motivation for this change is to avoid the same fiasco
16597 that dogs private keys: there are several incompatible private key
16598 formats some of which are standard and some OpenSSL specific and
16599 require various evil hacks to allow partial transparent handling and
16600 even then it doesn't work with DER formats. Given the option anything
16601 other than PKCS#8 should be dumped: but the other formats have to
16602 stay in the name of compatibility.
16604 With public keys and the benefit of hindsight one standard format
16605 is used which works with EVP_PKEY, RSA or DSA structures: though
16606 it clearly returns an error if you try to read the wrong kind of key.
16608 Added a -pubkey option to the 'x509' utility to output the public key.
16609 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16610 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16611 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16612 that do the same as the `EVP_PKEY_assign_*()` except they up the
16613 reference count of the added key (they don't "swallow" the
16618 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16619 CRLs would fail if the file contained no certificates or no CRLs:
16620 added a new function to read in both types and return the number
16621 read: this means that if none are read it will be an error. The
16622 DER versions of the certificate and CRL reader would always fail
16623 because it isn't possible to mix certificates and CRLs in DER format
16624 without choking one or the other routine. Changed this to just read
16625 a certificate: this is the best we can do. Also modified the code
16626 in `apps/verify.c` to take notice of return codes: it was previously
16627 attempting to read in certificates from NULL pointers and ignoring
16628 any errors: this is one reason why the cert and CRL reader seemed
16629 to work. It doesn't check return codes from the default certificate
16630 routines: these may well fail if the certificates aren't installed.
16634 * Code to support otherName option in GeneralName.
16638 * First update to verify code. Change the verify utility
16639 so it warns if it is passed a self signed certificate:
16640 for consistency with the normal behaviour. X509_verify
16641 has been modified to it will now verify a self signed
16642 certificate if *exactly* the same certificate appears
16643 in the store: it was previously impossible to trust a
16644 single self signed certificate. This means that:
16645 openssl verify ss.pem
16646 now gives a warning about a self signed certificate but
16647 openssl verify -CAfile ss.pem ss.pem
16652 * For servers, store verify_result in SSL_SESSION data structure
16653 (and add it to external session representation).
16654 This is needed when client certificate verifications fails,
16655 but an application-provided verification callback (set by
16656 SSL_CTX_set_cert_verify_callback) allows accepting the session
16657 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16658 but returns 1): When the session is reused, we have to set
16659 ssl->verify_result to the appropriate error code to avoid
16662 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16664 * Fix a bug in the new PKCS#7 code: it didn't consider the
16665 case in PKCS7_dataInit() where the signed PKCS7 structure
16666 didn't contain any existing data because it was being created.
16668 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16670 * Add a salt to the key derivation routines in enc.c. This
16671 forms the first 8 bytes of the encrypted file. Also add a
16672 -S option to allow a salt to be input on the command line.
16676 * New function X509_cmp(). Oddly enough there wasn't a function
16677 to compare two certificates. We do this by working out the SHA1
16678 hash and comparing that. X509_cmp() will be needed by the trust
16683 * SSL_get1_session() is like SSL_get_session(), but increments
16684 the reference count in the SSL_SESSION returned.
16686 *Geoff Thorpe <geoff@eu.c2.net>*
16688 * Fix for 'req': it was adding a null to request attributes.
16689 Also change the X509_LOOKUP and X509_INFO code to handle
16690 certificate auxiliary information.
16694 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16699 * Add the possibility to add extra information to the memory leak
16700 detecting output, to form tracebacks, showing from where each
16701 allocation was originated: CRYPTO_push_info("constant string") adds
16702 the string plus current file name and line number to a per-thread
16703 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16704 is like calling CYRPTO_pop_info() until the stack is empty.
16705 Also updated memory leak detection code to be multi-thread-safe.
16709 * Add options -text and -noout to pkcs7 utility and delete the
16710 encryption options which never did anything. Update docs.
16714 * Add options to some of the utilities to allow the pass phrase
16715 to be included on either the command line (not recommended on
16716 OSes like Unix) or read from the environment. Update the
16717 manpages and fix a few bugs.
16721 * Add a few manpages for some of the openssl commands.
16725 * Fix the -revoke option in ca. It was freeing up memory twice,
16726 leaking and not finding already revoked certificates.
16730 * Extensive changes to support certificate auxiliary information.
16731 This involves the use of X509_CERT_AUX structure and X509_AUX
16732 functions. An X509_AUX function such as PEM_read_X509_AUX()
16733 can still read in a certificate file in the usual way but it
16734 will also read in any additional "auxiliary information". By
16735 doing things this way a fair degree of compatibility can be
16736 retained: existing certificates can have this information added
16737 using the new 'x509' options.
16739 Current auxiliary information includes an "alias" and some trust
16740 settings. The trust settings will ultimately be used in enhanced
16741 certificate chain verification routines: currently a certificate
16742 can only be trusted if it is self signed and then it is trusted
16747 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16748 The problem was that one of the replacement routines had not been working
16749 since SSLeay releases. For now the offending routine has been replaced
16750 with non-optimised assembler. Even so, this now gives around 95%
16751 performance improvement for 1024 bit RSA signs.
16755 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16756 handling. Most clients have the effective key size in bits equal to
16757 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16758 A few however don't do this and instead use the size of the decrypted key
16759 to determine the RC2 key length and the AlgorithmIdentifier to determine
16760 the effective key length. In this case the effective key length can still
16761 be 40 bits but the key length can be 168 bits for example. This is fixed
16762 by manually forcing an RC2 key into the EVP_PKEY structure because the
16763 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16764 the key length and effective key length are equal.
16768 * Add a bunch of functions that should simplify the creation of
16769 X509_NAME structures. Now you should be able to do:
16770 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16771 and have it automatically work out the correct field type and fill in
16772 the structures. The more adventurous can try:
16773 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16774 and it will (hopefully) work out the correct multibyte encoding.
16778 * Change the 'req' utility to use the new field handling and multibyte
16779 copy routines. Before the DN field creation was handled in an ad hoc
16780 way in req, ca, and x509 which was rather broken and didn't support
16781 BMPStrings or UTF8Strings. Since some software doesn't implement
16782 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16783 using the dirstring_type option. See the new comment in the default
16784 openssl.cnf for more info.
16788 * Make crypto/rand/md_rand.c more robust:
16789 - Assure unique random numbers after fork().
16790 - Make sure that concurrent threads access the global counter and
16791 md serializably so that we never lose entropy in them
16792 or use exactly the same state in multiple threads.
16793 Access to the large state is not always serializable because
16794 the additional locking could be a performance killer, and
16795 md should be large enough anyway.
16799 * New file `apps/app_rand.c` with commonly needed functionality
16800 for handling the random seed file.
16802 Use the random seed file in some applications that previously did not:
16804 dsaparam -genkey (which also ignored its '-rand' option),
16807 x509 (when signing).
16808 Except on systems with /dev/urandom, it is crucial to have a random
16809 seed file at least for key creation, DSA signing, and for DH exchanges;
16810 for RSA signatures we could do without one.
16812 gendh and gendsa (unlike genrsa) used to read only the first byte
16813 of each file listed in the '-rand' option. The function as previously
16814 found in genrsa is now in app_rand.c and is used by all programs
16815 that support '-rand'.
16819 * In RAND_write_file, use mode 0600 for creating files;
16820 don't just chmod when it may be too late.
16824 * Report an error from X509_STORE_load_locations
16825 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16829 * New function ASN1_mbstring_copy() this copies a string in either
16830 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16831 into an ASN1_STRING type. A mask of permissible types is passed
16832 and it chooses the "minimal" type to use or an error if not type
16837 * Add function equivalents to the various macros in asn1.h. The old
16838 macros are retained with an `M_` prefix. Code inside the library can
16839 use the `M_` macros. External code (including the openssl utility)
16840 should *NOT* in order to be "shared library friendly".
16844 * Add various functions that can check a certificate's extensions
16845 to see if it usable for various purposes such as SSL client,
16846 server or S/MIME and CAs of these types. This is currently
16847 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16848 verification. Also added a -purpose flag to x509 utility to
16849 print out all the purposes.
16853 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16858 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16859 for, obtain and decode and extension and obtain its critical flag.
16860 This allows all the necessary extension code to be handled in a
16861 single function call.
16865 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16866 platforms. See crypto/rc4/rc4_enc.c for further details.
16870 * New -noout option to asn1parse. This causes no output to be produced
16871 its main use is when combined with -strparse and -out to extract data
16872 from a file (which may not be in ASN.1 format).
16876 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16877 when producing the local key id.
16879 *Richard Levitte <levitte@stacken.kth.se>*
16881 * New option -dhparam in s_server. This allows a DH parameter file to be
16882 stated explicitly. If it is not stated then it tries the first server
16883 certificate file. The previous behaviour hard coded the filename
16888 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16889 a public key to be input or output. For example:
16890 openssl rsa -in key.pem -pubout -out pubkey.pem
16891 Also added necessary DSA public key functions to handle this.
16895 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16896 in the message. This was handled by allowing
16897 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16899 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16901 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16902 to the end of the strings whereas this didn't. This would cause problems
16903 if strings read with d2i_ASN1_bytes() were later modified.
16905 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16907 * Fix for base64 decode bug. When a base64 bio reads only one line of
16908 data and it contains EOF it will end up returning an error. This is
16909 caused by input 46 bytes long. The cause is due to the way base64
16910 BIOs find the start of base64 encoded data. They do this by trying a
16911 trial decode on each line until they find one that works. When they
16912 do a flag is set and it starts again knowing it can pass all the
16913 data directly through the decoder. Unfortunately it doesn't reset
16914 the context it uses. This means that if EOF is reached an attempt
16915 is made to pass two EOFs through the context and this causes the
16916 resulting error. This can also cause other problems as well. As is
16917 usual with these problems it takes *ages* to find and the fix is
16918 trivial: move one line.
16920 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16922 * Ugly workaround to get s_client and s_server working under Windows. The
16923 old code wouldn't work because it needed to select() on sockets and the
16924 tty (for keypresses and to see if data could be written). Win32 only
16925 supports select() on sockets so we select() with a 1s timeout on the
16926 sockets and then see if any characters are waiting to be read, if none
16927 are present then we retry, we also assume we can always write data to
16928 the tty. This isn't nice because the code then blocks until we've
16929 received a complete line of data and it is effectively polling the
16930 keyboard at 1s intervals: however it's quite a bit better than not
16931 working at all :-) A dedicated Windows application might handle this
16932 with an event loop for example.
16936 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16937 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16938 will be called when RSA_sign() and RSA_verify() are used. This is useful
16939 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16940 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16941 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16942 This necessitated the support of an extra signature type NID_md5_sha1
16943 for SSL signatures and modifications to the SSL library to use it instead
16944 of calling RSA_public_decrypt() and RSA_private_encrypt().
16948 * Add new -verify -CAfile and -CApath options to the crl program, these
16949 will lookup a CRL issuers certificate and verify the signature in a
16950 similar way to the verify program. Tidy up the crl program so it
16951 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16952 less strict. It will now permit CRL extensions even if it is not
16953 a V2 CRL: this will allow it to tolerate some broken CRLs.
16957 * Initialize all non-automatic variables each time one of the openssl
16958 sub-programs is started (this is necessary as they may be started
16959 multiple times from the "OpenSSL>" prompt).
16961 *Lennart Bang, Bodo Moeller*
16963 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16964 removing all other RSA functionality (this is what NO_RSA does). This
16965 is so (for example) those in the US can disable those operations covered
16966 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16971 * Non-copying interface to BIO pairs.
16972 (still largely untested)
16976 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16977 ASCII string. This was handled independently in various places before.
16981 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16982 UTF8 strings a character at a time.
16986 * Use client_version from client hello to select the protocol
16987 (s23_srvr.c) and for RSA client key exchange verification
16988 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16992 * Add various utility functions to handle SPKACs, these were previously
16993 handled by poking round in the structure internals. Added new function
16994 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16995 print, verify and generate SPKACs. Based on an original idea from
16996 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17000 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17004 * Allow the config file extension section to be overwritten on the
17005 command line. Based on an original idea from Massimiliano Pala
17006 <madwolf@comune.modena.it>. The new option is called -extensions
17007 and can be applied to ca, req and x509. Also -reqexts to override
17008 the request extensions in req and -crlexts to override the crl extensions
17013 * Add new feature to the SPKAC handling in ca. Now you can include
17014 the same field multiple times by preceding it by "XXXX." for example:
17017 this is the same syntax as used in the req config file.
17021 * Allow certificate extensions to be added to certificate requests. These
17022 are specified in a 'req_extensions' option of the req section of the
17023 config file. They can be printed out with the -text option to req but
17024 are otherwise ignored at present.
17028 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17029 data read consists of only the final block it would not decrypted because
17030 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17031 A misplaced 'break' also meant the decrypted final block might not be
17032 copied until the next read.
17036 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17037 a few extra parameters to the DH structure: these will be useful if
17038 for example we want the value of 'q' or implement X9.42 DH.
17042 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17043 provides hooks that allow the default DSA functions or functions on a
17044 "per key" basis to be replaced. This allows hardware acceleration and
17045 hardware key storage to be handled without major modification to the
17046 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17047 associated functions.
17051 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17052 as "read only": it can't be written to and the buffer it points to will
17053 not be freed. Reading from a read only BIO is much more efficient than
17054 a normal memory BIO. This was added because there are several times when
17055 an area of memory needs to be read from a BIO. The previous method was
17056 to create a memory BIO and write the data to it, this results in two
17057 copies of the data and an O(n^2) reading algorithm. There is a new
17058 function BIO_new_mem_buf() which creates a read only memory BIO from
17059 an area of memory. Also modified the PKCS#7 routines to use read only
17064 * Bugfix: ssl23_get_client_hello did not work properly when called in
17065 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17066 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17067 but a retry condition occurred while trying to read the rest.
17071 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17072 NID_pkcs7_encrypted by default: this was wrong since this should almost
17073 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17074 the encrypted data type: this is a more sensible place to put it and it
17075 allows the PKCS#12 code to be tidied up that duplicated this
17080 * Changed obj_dat.pl script so it takes its input and output files on
17081 the command line. This should avoid shell escape redirection problems
17086 * Initial support for certificate extension requests, these are included
17087 in things like Xenroll certificate requests. Included functions to allow
17088 extensions to be obtained and added.
17092 * -crlf option to s_client and s_server for sending newlines as
17093 CRLF (as required by many protocols).
17097 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17099 * Install libRSAglue.a when OpenSSL is built with RSAref.
17101 *Ralf S. Engelschall*
17103 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17105 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17107 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17112 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17113 DH parameters/keys (q is lost during that conversion, but the resulting
17114 DH parameters contain its length).
17116 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17117 much faster than DH_generate_parameters (which creates parameters
17118 where `p = 2*q + 1`), and also the smaller q makes DH computations
17119 much more efficient (160-bit exponentiation instead of 1024-bit
17120 exponentiation); so this provides a convenient way to support DHE
17121 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17122 utter importance to use
17123 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17125 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17126 when such DH parameters are used, because otherwise small subgroup
17127 attacks may become possible!
17131 * Avoid memory leak in i2d_DHparams.
17135 * Allow the -k option to be used more than once in the enc program:
17136 this allows the same encrypted message to be read by multiple recipients.
17140 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17141 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17142 it will always use the numerical form of the OID, even if it has a short
17147 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17148 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17149 otherwise bn_mod_exp was called. In the case of hardware keys for example
17150 no private key components need be present and it might store extra data
17151 in the RSA structure, which cannot be accessed from bn_mod_exp.
17152 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17153 private key operations.
17157 * Added support for SPARC Linux.
17161 * pem_password_cb function type incompatibly changed from
17162 typedef int pem_password_cb(char *buf, int size, int rwflag);
17164 ....(char *buf, int size, int rwflag, void *userdata);
17165 so that applications can pass data to their callbacks:
17166 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17167 additional void * argument, which is just handed through whenever
17168 the password callback is called.
17170 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17172 New function SSL_CTX_set_default_passwd_cb_userdata.
17174 Compatibility note: As many C implementations push function arguments
17175 onto the stack in reverse order, the new library version is likely to
17176 interoperate with programs that have been compiled with the old
17177 pem_password_cb definition (PEM_whatever takes some data that
17178 happens to be on the stack as its last argument, and the callback
17179 just ignores this garbage); but there is no guarantee whatsoever that
17182 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17183 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17184 problems not only on Windows, but also on some Unix platforms.
17185 To avoid problematic command lines, these definitions are now in an
17186 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17187 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17191 * MIPS III/IV assembler module is reimplemented.
17195 * More DES library cleanups: remove references to srand/rand and
17196 delete an unused file.
17200 * Add support for the free Netwide assembler (NASM) under Win32,
17201 since not many people have MASM (ml) and it can be hard to obtain.
17202 This is currently experimental but it seems to work OK and pass all
17203 the tests. Check out INSTALL.W32 for info.
17207 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17208 without temporary keys kept an extra copy of the server key,
17209 and connections with temporary keys did not free everything in case
17214 * New function RSA_check_key and new openssl rsa option -check
17215 for verifying the consistency of RSA keys.
17217 *Ulf Moeller, Bodo Moeller*
17219 * Various changes to make Win32 compile work:
17220 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17221 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17222 comparison" warnings.
17223 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17227 * Add a debugging option to PKCS#5 v2 key generation function: when
17228 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17229 derived keys are printed to stderr.
17233 * Copy the flags in ASN1_STRING_dup().
17235 *Roman E. Pavlov <pre@mo.msk.ru>*
17237 * The x509 application mishandled signing requests containing DSA
17238 keys when the signing key was also DSA and the parameters didn't match.
17240 It was supposed to omit the parameters when they matched the signing key:
17241 the verifying software was then supposed to automatically use the CA's
17242 parameters if they were absent from the end user certificate.
17244 Omitting parameters is no longer recommended. The test was also
17245 the wrong way round! This was probably due to unusual behaviour in
17246 EVP_cmp_parameters() which returns 1 if the parameters match.
17247 This meant that parameters were omitted when they *didn't* match and
17248 the certificate was useless. Certificates signed with 'ca' didn't have
17251 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17253 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17254 The interface is as follows:
17255 Applications can use
17256 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17257 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17258 "off" is now the default.
17259 The library internally uses
17260 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17261 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17262 to disable memory-checking temporarily.
17264 Some inconsistent states that previously were possible (and were
17265 even the default) are now avoided.
17267 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17268 with each memory chunk allocated; this is occasionally more helpful
17269 than just having a counter.
17271 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17273 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17278 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17279 which largely parallels "options", but is for changing API behaviour,
17280 whereas "options" are about protocol behaviour.
17281 Initial "mode" flags are:
17283 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17284 a single record has been written.
17285 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17286 retries use the same buffer location.
17287 (But all of the contents must be
17292 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17295 * Fix problems with no-hmac etc.
17297 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17299 * New functions RSA_get_default_method(), RSA_set_method() and
17300 RSA_get_method(). These allows replacement of RSA_METHODs without having
17301 to mess around with the internals of an RSA structure.
17305 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17306 Also really enable memory leak checks in openssl.c and in some
17309 *Chad C. Mulligan, Bodo Moeller*
17311 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17312 up the length of negative integers. This has now been simplified to just
17313 store the length when it is first determined and use it later, rather
17314 than trying to keep track of where data is copied and updating it to
17316 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17318 * Add a new function PKCS7_signatureVerify. This allows the verification
17319 of a PKCS#7 signature but with the signing certificate passed to the
17320 function itself. This contrasts with PKCS7_dataVerify which assumes the
17321 certificate is present in the PKCS#7 structure. This isn't always the
17322 case: certificates can be omitted from a PKCS#7 structure and be
17323 distributed by "out of band" means (such as a certificate database).
17327 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17328 function prototypes in pem.h, also change util/mkdef.pl to add the
17329 necessary function names.
17333 * mk1mf.pl (used by Windows builds) did not properly read the
17334 options set by Configure in the top level Makefile, and Configure
17335 was not even able to write more than one option correctly.
17336 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17340 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17341 file to be loaded from a BIO or FILE pointer. The BIO version will
17342 for example allow memory BIOs to contain config info.
17346 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17347 Whoever hopes to achieve shared-library compatibility across versions
17348 must use this, not the compile-time macro.
17349 (Exercise 0.9.4: Which is the minimum library version required by
17351 Note: All this applies only to multi-threaded programs, others don't
17356 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17357 through a BIO pair triggered the default case, i.e.
17358 SSLerr(...,SSL_R_UNKNOWN_STATE).
17362 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17363 can use the SSL library even if none of the specific BIOs is
17368 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17369 for the encoded length.
17371 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17373 * Add initial documentation of the X509V3 functions.
17377 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17378 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17379 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17380 secure PKCS#8 private key format with a high iteration count.
17384 * Fix determination of Perl interpreter: A perl or perl5
17385 *directory* in $PATH was also accepted as the interpreter.
17387 *Ralf S. Engelschall*
17389 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17390 wrong with it but it was very old and did things like calling
17391 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17392 unusual formatting.
17396 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17397 to use the new extension code.
17401 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17402 with macros. This should make it easier to change their form, add extra
17403 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17408 * Add to configuration table a new entry that can specify an alternative
17409 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17410 according to Mark Crispin <MRC@Panda.COM>.
17414 * DES CBC did not update the IV. Weird.
17418 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17419 Changing the behaviour of the former might break existing programs --
17420 where IV updating is needed, des_ncbc_encrypt can be used.
17423 * When bntest is run from "make test" it drives bc to check its
17424 calculations, as well as internally checking them. If an internal check
17425 fails, it needs to cause bc to give a non-zero result or make test carries
17426 on without noticing the failure. Fixed.
17430 * DES library cleanups.
17434 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17435 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17436 ciphers. NOTE: although the key derivation function has been verified
17437 against some published test vectors it has not been extensively tested
17438 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17443 * Instead of "mkdir -p", which is not fully portable, use new
17444 Perl script "util/mkdir-p.pl".
17448 * Rewrite the way password based encryption (PBE) is handled. It used to
17449 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17450 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17451 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17452 the 'parameter' field of the AlgorithmIdentifier is passed to the
17453 underlying key generation function so it must do its own ASN1 parsing.
17454 This has also changed the EVP_PBE_CipherInit() function which now has a
17455 'parameter' argument instead of literal salt and iteration count values
17456 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17460 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17461 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17462 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17463 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17464 value was just used as a "magic string" and not used directly its
17465 value doesn't matter.
17469 * Introduce some semblance of const correctness to BN. Shame C doesn't
17474 * "linux-sparc64" configuration (ultrapenguin).
17476 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17477 "linux-sparc" configuration.
17479 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17481 * config now generates no-xxx options for missing ciphers.
17485 * Support the EBCDIC character set (work in progress).
17486 File ebcdic.c not yet included because it has a different license.
17488 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17490 * Support BS2000/OSD-POSIX.
17492 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17494 * Make callbacks for key generation use `void *` instead of `char *`.
17498 * Make S/MIME samples compile (not yet tested).
17502 * Additional typesafe stacks.
17506 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17510 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17512 * New configuration variant "sco5-gcc".
17514 * Updated some demos.
17516 *Sean O Riordain, Wade Scholine*
17518 * Add missing BIO_free at exit of pkcs12 application.
17522 * Fix memory leak in conf.c.
17526 * Updates for Win32 to assembler version of MD5.
17530 * Set #! path to perl in `apps/der_chop` to where we found it
17531 instead of using a fixed path.
17535 * SHA library changes for irix64-mips4-cc.
17539 * Improvements for VMS support.
17543 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17545 * Bignum library bug fix. IRIX 6 passes "make test" now!
17546 This also avoids the problems with SC4.2 and unpatched SC5.
17548 *Andy Polyakov <appro@fy.chalmers.se>*
17550 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17551 These are required because of the typesafe stack would otherwise break
17552 existing code. If old code used a structure member which used to be STACK
17553 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17554 sk_num or sk_value it would produce an error because the num, data members
17555 are not present in STACK_OF. Now it just produces a warning. sk_set
17556 replaces the old method of assigning a value to sk_value
17557 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17558 that does this will no longer work (and should use sk_set instead) but
17559 this could be regarded as a "questionable" behaviour anyway.
17563 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17564 correctly handle encrypted S/MIME data.
17568 * Change type of various DES function arguments from des_cblock
17569 (which means, in function argument declarations, pointer to char)
17570 to des_cblock * (meaning pointer to array with 8 char elements),
17571 which allows the compiler to do more typechecking; it was like
17572 that back in SSLeay, but with lots of ugly casts.
17574 Introduce new type const_des_cblock.
17578 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17579 problems: find RecipientInfo structure that matches recipient certificate
17580 and initialise the ASN1 structures properly based on passed cipher.
17584 * Belatedly make the BN tests actually check the results.
17588 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17589 to and from BNs: it was completely broken. New compilation option
17590 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17591 key elements as negative integers.
17595 * Reorganize and speed up MD5.
17597 *Andy Polyakov <appro@fy.chalmers.se>*
17601 *Richard Levitte <richard@levitte.org>*
17603 * New option -out to asn1parse to allow the parsed structure to be
17604 output to a file. This is most useful when combined with the -strparse
17605 option to examine the output of things like OCTET STRINGS.
17609 * Make SSL library a little more fool-proof by not requiring any longer
17610 that `SSL_set_{accept,connect}_state` be called before
17611 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17612 in many applications because usually everything *appeared* to work as
17613 intended anyway -- now it really works as intended).
17617 * Move openssl.cnf out of lib/.
17621 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17622 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17623 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17625 *Ralf S. Engelschall*
17627 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17628 handle PKCS#7 enveloped data properly.
17630 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17632 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17633 copying pointers. The cert_st handling is changed by this in
17634 various ways (and thus what used to be known as ctx->default_cert
17635 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17636 any longer when s->cert does not give us what we need).
17637 ssl_cert_instantiate becomes obsolete by this change.
17638 As soon as we've got the new code right (possibly it already is?),
17639 we have solved a couple of bugs of the earlier code where s->cert
17640 was used as if it could not have been shared with other SSL structures.
17642 Note that using the SSL API in certain dirty ways now will result
17643 in different behaviour than observed with earlier library versions:
17644 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17645 does not influence s as it used to.
17647 In order to clean up things more thoroughly, inside SSL_SESSION
17648 we don't use CERT any longer, but a new structure SESS_CERT
17649 that holds per-session data (if available); currently, this is
17650 the peer's certificate chain and, for clients, the server's certificate
17651 and temporary key. CERT holds only those values that can have
17652 meaningful defaults in an SSL_CTX.
17656 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17657 from the internal representation. Various PKCS#7 fixes: remove some
17658 evil casts and set the enc_dig_alg field properly based on the signing
17663 * Allow PKCS#12 password to be set from the command line or the
17664 environment. Let 'ca' get its config file name from the environment
17665 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17670 * Allow certificate policies extension to use an IA5STRING for the
17671 organization field. This is contrary to the PKIX definition but
17672 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17677 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17678 without disallowing inline assembler and the like for non-pedantic builds.
17682 * Support Borland C++ builder.
17684 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17690 * SHA-1 cleanups and performance enhancements.
17692 *Andy Polyakov <appro@fy.chalmers.se>*
17694 * Sparc v8plus assembler for the bignum library.
17696 *Andy Polyakov <appro@fy.chalmers.se>*
17698 * Accept any -xxx and +xxx compiler options in Configure.
17702 * Update HPUX configuration.
17706 * Add missing `sk_<type>_unshift()` function to safestack.h
17708 *Ralf S. Engelschall*
17710 * New function SSL_CTX_use_certificate_chain_file that sets the
17711 "extra_cert"s in addition to the certificate. (This makes sense
17712 only for "PEM" format files, as chains as a whole are not
17717 * Support verify_depth from the SSL API.
17718 x509_vfy.c had what can be considered an off-by-one-error:
17719 Its depth (which was not part of the external interface)
17720 was actually counting the number of certificates in a chain;
17721 now it really counts the depth.
17725 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17726 instead of X509err, which often resulted in confusing error
17727 messages since the error codes are not globally unique
17728 (e.g. an alleged error in ssl3_accept when a certificate
17729 didn't match the private key).
17731 * New function SSL_CTX_set_session_id_context that allows to set a default
17732 value (so that you don't need SSL_set_session_id_context for each
17733 connection using the SSL_CTX).
17737 * OAEP decoding bug fix.
17741 * Support INSTALL_PREFIX for package builders, as proposed by
17746 * New Configure options "threads" and "no-threads". For systems
17747 where the proper compiler options are known (currently Solaris
17748 and Linux), "threads" is the default.
17752 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17756 * Install various scripts to $(OPENSSLDIR)/misc, not to
17757 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17758 such as /usr/local/bin.
17762 * "make linux-shared" to build shared libraries.
17764 *Niels Poppe <niels@netbox.org>*
17766 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17770 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17771 extension adding in x509 utility.
17775 * Remove NOPROTO sections and error code comments.
17779 * Partial rewrite of the DEF file generator to now parse the ANSI
17784 * New Configure options --prefix=DIR and --openssldir=DIR.
17788 * Complete rewrite of the error code script(s). It is all now handled
17789 by one script at the top level which handles error code gathering,
17790 header rewriting and C source file generation. It should be much better
17791 than the old method: it now uses a modified version of Ulf's parser to
17792 read the ANSI prototypes in all header files (thus the old K&R definitions
17793 aren't needed for error creation any more) and do a better job of
17794 translating function codes into names. The old 'ASN1 error code embedded
17795 in a comment' is no longer necessary and it doesn't use .err files which
17796 have now been deleted. Also the error code call doesn't have to appear all
17797 on one line (which resulted in some large lines...).
17801 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17805 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17806 0 (which usually indicates a closed connection), but continue reading.
17810 * Fix some race conditions.
17814 * Add support for CRL distribution points extension. Add Certificate
17815 Policies and CRL distribution points documentation.
17819 * Move the autogenerated header file parts to crypto/opensslconf.h.
17823 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17824 8 of keying material. Merlin has also confirmed interop with this fix
17825 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17827 *Merlin Hughes <merlin@baltimore.ie>*
17829 * Fix lots of warnings.
17831 *Richard Levitte <levitte@stacken.kth.se>*
17833 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17834 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17836 *Richard Levitte <levitte@stacken.kth.se>*
17838 * Fix problems with sizeof(long) == 8.
17840 *Andy Polyakov <appro@fy.chalmers.se>*
17842 * Change functions to ANSI C.
17846 * Fix typos in error codes.
17848 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17850 * Remove defunct assembler files from Configure.
17854 * SPARC v8 assembler BIGNUM implementation.
17856 *Andy Polyakov <appro@fy.chalmers.se>*
17858 * Support for Certificate Policies extension: both print and set.
17859 Various additions to support the r2i method this uses.
17863 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17864 return a const string when you are expecting an allocated buffer.
17868 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17869 types DirectoryString and DisplayText.
17873 * Add code to allow r2i extensions to access the configuration database,
17874 add an LHASH database driver and add several ctx helper functions.
17878 * Fix an evil bug in bn_expand2() which caused various BN functions to
17879 fail when they extended the size of a BIGNUM.
17883 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17884 support typesafe stack.
17888 * Fix typo in SSL_[gs]et_options().
17890 *Nils Frostberg <nils@medcom.se>*
17892 * Delete various functions and files that belonged to the (now obsolete)
17893 old X509V3 handling code.
17897 * New Configure option "rsaref".
17901 * Don't auto-generate pem.h.
17905 * Introduce type-safe ASN.1 SETs.
17909 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17911 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17913 * Introduce type-safe STACKs. This will almost certainly break lots of code
17914 that links with OpenSSL (well at least cause lots of warnings), but fear
17915 not: the conversion is trivial, and it eliminates loads of evil casts. A
17916 few STACKed things have been converted already. Feel free to convert more.
17917 In the fullness of time, I'll do away with the STACK type altogether.
17921 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17922 specified in `<certfile>` by updating the entry in the index.txt file.
17923 This way one no longer has to edit the index.txt file manually for
17924 revoking a certificate. The -revoke option does the gory details now.
17926 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17928 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17929 `-text` option at all and this way the `-noout -text` combination was
17930 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17932 *Ralf S. Engelschall*
17934 * Make sure a corresponding plain text error message exists for the
17935 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17936 verify callback function determined that a certificate was revoked.
17938 *Ralf S. Engelschall*
17940 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17941 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17942 all available ciphers including rc5, which was forgotten until now.
17943 In order to let the testing shell script know which algorithms
17944 are available, a new (up to now undocumented) command
17945 `openssl list-cipher-commands` is used.
17949 * Bugfix: s_client occasionally would sleep in select() when
17950 it should have checked SSL_pending() first.
17954 * New functions DSA_do_sign and DSA_do_verify to provide access to
17955 the raw DSA values prior to ASN.1 encoding.
17959 * Tweaks to Configure
17961 *Niels Poppe <niels@netbox.org>*
17963 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17968 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17972 * New config option to avoid instructions that are illegal on the 80386.
17973 The default code is faster, but requires at least a 486.
17977 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17978 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17979 same as SSL2_VERSION anyway.
17983 * New "-showcerts" option for s_client.
17987 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17988 application. Various cleanups and fixes.
17992 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17993 modify error routines to work internally. Add error codes and PBE init
17994 to library startup routines.
17998 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17999 packing functions to asn1 and evp. Changed function names and error
18000 codes along the way.
18004 * PKCS12 integration: and so it begins... First of several patches to
18005 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18006 objects to objects.h
18010 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18011 and display support for Thawte strong extranet extension.
18015 * Add LinuxPPC support.
18017 *Jeff Dubrule <igor@pobox.org>*
18019 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18020 bn_div_words in alpha.s.
18022 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18024 * Make sure the RSA OAEP test is skipped under -DRSAref because
18025 OAEP isn't supported when OpenSSL is built with RSAref.
18027 *Ulf Moeller <ulf@fitug.de>*
18029 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18030 so they no longer are missing under -DNOPROTO.
18032 *Soren S. Jorvang <soren@t.dk>*
18034 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18036 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18037 doesn't work when the session is reused. Coming soon!
18041 * Fix a security hole, that allows sessions to be reused in the wrong
18042 context thus bypassing client cert protection! All software that uses
18043 client certs and session caches in multiple contexts NEEDS PATCHING to
18044 allow session reuse! A fuller solution is in the works.
18046 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18048 * Some more source tree cleanups (removed obsolete files
18049 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18050 permission on "config" script to be executable) and a fix for the INSTALL
18053 *Ulf Moeller <ulf@fitug.de>*
18055 * Remove some legacy and erroneous uses of malloc, free instead of
18058 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18060 * Make rsa_oaep_test return non-zero on error.
18062 *Ulf Moeller <ulf@fitug.de>*
18064 * Add support for native Solaris shared libraries. Configure
18065 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18066 if someone would make that last step automatic.
18068 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18070 * ctx_size was not built with the right compiler during "make links". Fixed.
18074 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18075 except NULL ciphers". This means the default cipher list will no longer
18076 enable NULL ciphers. They need to be specifically enabled e.g. with
18077 the string "DEFAULT:eNULL".
18081 * Fix to RSA private encryption routines: if p < q then it would
18082 occasionally produce an invalid result. This will only happen with
18083 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18087 * Be less restrictive and allow also `perl util/perlpath.pl
18088 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18089 because this way one can also use an interpreter named `perl5` (which is
18090 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18091 installed as `perl`).
18093 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18095 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18097 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18099 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18100 advapi32.lib to Win32 build and change the pem test comparison
18101 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18102 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18103 and crypto/des/ede_cbcm_enc.c.
18107 * DES quad checksum was broken on big-endian architectures. Fixed.
18111 * Comment out two functions in bio.h that aren't implemented. Fix up the
18112 Win32 test batch file so it (might) work again. The Win32 test batch file
18113 is horrible: I feel ill....
18117 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18118 in e_os.h. Audit of header files to check ANSI and non ANSI
18119 sections: 10 functions were absent from non ANSI section and not exported
18120 from Windows DLLs. Fixed up libeay.num for new functions.
18124 * Make `openssl version` output lines consistent.
18126 *Ralf S. Engelschall*
18128 * Fix Win32 symbol export lists for BIO functions: Added
18129 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18130 to ms/libeay{16,32}.def.
18132 *Ralf S. Engelschall*
18134 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18135 fine under Unix and passes some trivial tests I've now added. But the
18136 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18137 added to make sure no one expects that this stuff really works in the
18138 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18139 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18142 *Ralf S. Engelschall*
18144 * Fix the generation of two part addresses in perl.
18146 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18148 * Add config entry for Linux on MIPS.
18150 *John Tobey <jtobey@channel1.com>*
18152 * Make links whenever Configure is run, unless we are on Windoze.
18156 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18157 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18162 * Add a useful kludge to allow package maintainers to specify compiler and
18163 other platforms details on the command line without having to patch the
18164 Configure script every time: One now can use
18165 `perl Configure <id>:<details>`,
18166 i.e. platform ids are allowed to have details appended
18167 to them (separated by colons). This is treated as there would be a static
18168 pre-configured entry in Configure's %table under key `<id>` with value
18169 `<details>` and `perl Configure <id>` is called. So, when you want to
18170 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18171 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18172 now, which overrides the FreeBSD-elf entry on-the-fly.
18174 *Ralf S. Engelschall*
18176 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18180 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18181 on the `perl Configure ...` command line. This way one can compile
18182 OpenSSL libraries with Position Independent Code (PIC) which is needed
18183 for linking it into DSOs.
18185 *Ralf S. Engelschall*
18187 * Remarkably, export ciphers were totally broken and no-one had noticed!
18192 * Cleaned up the LICENSE document: The official contact for any license
18193 questions now is the OpenSSL core team under openssl-core@openssl.org.
18194 And add a paragraph about the dual-license situation to make sure people
18195 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18196 to the OpenSSL toolkit.
18198 *Ralf S. Engelschall*
18200 * General source tree makefile cleanups: Made `making xxx in yyy...`
18201 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18202 Additionally cleaned up the `make links` target: Remove unnecessary
18203 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18204 to speed processing and no longer clutter the display with confusing
18205 stuff. Instead only the actually done links are displayed.
18207 *Ralf S. Engelschall*
18209 * Permit null encryption ciphersuites, used for authentication only. It used
18210 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18211 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18216 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18217 signed attributes when verifying signatures (this would break them),
18218 the detached data encoding was wrong and public keys obtained using
18219 X509_get_pubkey() weren't freed.
18223 * Add text documentation for the BUFFER functions. Also added a work around
18224 to a Win95 console bug. This was triggered by the password read stuff: the
18225 last character typed gets carried over to the next fread(). If you were
18226 generating a new cert request using 'req' for example then the last
18227 character of the passphrase would be CR which would then enter the first
18232 * Added the new 'Includes OpenSSL Cryptography Software' button as
18233 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18234 button and can be used by applications based on OpenSSL to show the
18235 relationship to the OpenSSL project.
18237 *Ralf S. Engelschall*
18239 * Remove confusing variables in function signatures in files
18240 ssl/ssl_lib.c and ssl/ssl.h.
18242 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18244 * Don't install bss_file.c under PREFIX/include/
18246 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18248 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18249 functions that return function pointers and has support for NT specific
18250 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18251 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18252 unsigned to signed types: this was killing the Win32 compile.
18256 * Add new certificate file to stack functions,
18257 SSL_add_dir_cert_subjects_to_stack() and
18258 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18259 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18260 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18261 This means that Apache-SSL and similar packages don't have to mess around
18262 to add as many CAs as they want to the preferred list.
18266 * Experiment with doxygen documentation. Currently only partially applied to
18268 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18269 openssl.doxy as the configuration file.
18273 * Get rid of remaining C++-style comments which strict C compilers hate.
18275 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18277 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18278 compiled in by default: it has problems with large keys.
18282 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18283 DH private keys and/or callback functions which directly correspond to
18284 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18285 is needed for applications which have to configure certificates on a
18286 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18288 For the RSA certificate situation is makes no difference, but
18289 for the DSA certificate situation this fixes the "no shared cipher"
18290 problem where the OpenSSL cipher selection procedure failed because the
18291 temporary keys were not overtaken from the context and the API provided
18292 no way to reconfigure them.
18293 The new functions now let applications reconfigure the stuff and they
18294 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18295 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18296 non-public-API function ssl_cert_instantiate() is used as a helper
18297 function and also to reduce code redundancy inside ssl_rsa.c.
18299 *Ralf S. Engelschall*
18301 * Move s_server -dcert and -dkey options out of the undocumented feature
18302 area because they are useful for the DSA situation and should be
18303 recognized by the users.
18305 *Ralf S. Engelschall*
18307 * Fix the cipher decision scheme for export ciphers: the export bits are
18308 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18309 SSL_EXP_MASK. So, the original variable has to be used instead of the
18310 already masked variable.
18312 *Richard Levitte <levitte@stacken.kth.se>*
18314 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18316 *Richard Levitte <levitte@stacken.kth.se>*
18318 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18319 from `int` to `unsigned int` because it is a length and initialized by
18320 EVP_DigestFinal() which expects an `unsigned int *`.
18322 *Richard Levitte <levitte@stacken.kth.se>*
18324 * Don't hard-code path to Perl interpreter on shebang line of Configure
18325 script. Instead use the usual Shell->Perl transition trick.
18327 *Ralf S. Engelschall*
18329 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18330 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18331 -noout -modulus` as it's already the case for `openssl rsa -noout
18332 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18333 currently the public key is printed (a decision which was already done by
18334 `openssl dsa -modulus` in the past) which serves a similar purpose.
18335 Additionally the NO_RSA no longer completely removes the whole -modulus
18336 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18339 *Ralf S. Engelschall*
18341 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18342 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18344 *Arne Ansper <arne@ats.cyber.ee>*
18346 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18347 to be added. Now both 'req' and 'ca' can use new objects defined in the
18352 * Add cool BIO that does syslog (or event log on NT).
18354 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18356 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18357 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18358 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18359 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18363 * Add preliminary config info for new extension code.
18367 * Make RSA_NO_PADDING really use no padding.
18369 *Ulf Moeller <ulf@fitug.de>*
18371 * Generate errors when private/public key check is done.
18375 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18376 for some CRL extensions and new objects added.
18380 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18381 key usage extension and fuller support for authority key id.
18385 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18386 padding method for RSA, which is recommended for new applications in PKCS
18387 #1 v2.0 (RFC 2437, October 1998).
18388 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18389 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18390 against Bleichbacher's attack on RSA.
18391 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18394 * Updates to the new SSL compression code
18396 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18398 * Fix so that the version number in the master secret, when passed
18399 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18400 (because the server will not accept higher), that the version number
18401 is 0x03,0x01, not 0x03,0x00
18403 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18405 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18406 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18407 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18411 * Support for RAW extensions where an arbitrary extension can be
18412 created by including its DER encoding. See `apps/openssl.cnf` for
18417 * Make sure latest Perl versions don't interpret some generated C array
18418 code as Perl array code in the crypto/err/err_genc.pl script.
18420 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18422 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18423 not many people have the assembler. Various Win32 compilation fixes and
18424 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18425 build instructions.
18429 * Modify configure script 'Configure' to automatically create crypto/date.h
18430 file under Win32 and also build pem.h from pem.org. New script
18431 util/mkfiles.pl to create the MINFO file on environments that can't do a
18432 'make files': perl util/mkfiles.pl >MINFO should work.
18436 * Major rework of DES function declarations, in the pursuit of correctness
18437 and purity. As a result, many evil casts evaporated, and some weirdness,
18438 too. You may find this causes warnings in your code. Zapping your evil
18439 casts will probably fix them. Mostly.
18443 * Fix for a typo in asn1.h. Bug fix to object creation script
18444 obj_dat.pl. It considered a zero in an object definition to mean
18445 "end of object": none of the objects in objects.h have any zeros
18446 so it wasn't spotted.
18448 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18450 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18451 Masking (CBCM). In the absence of test vectors, the best I have been able
18452 to do is check that the decrypt undoes the encrypt, so far. Send me test
18453 vectors if you have them.
18457 * Correct calculation of key length for export ciphers (too much space was
18458 allocated for null ciphers). This has not been tested!
18462 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18463 message is now correct (it understands "crypto" and "ssl" on its
18464 command line). There is also now an "update" option. This will update
18465 the util/ssleay.num and util/libeay.num files with any new functions.
18467 perl util/mkdef.pl crypto ssl update
18468 it will update them.
18472 * Overhauled the Perl interface:
18473 - ported BN stuff to OpenSSL's different BN library
18474 - made the perl/ source tree CVS-aware
18475 - renamed the package from SSLeay to OpenSSL (the files still contain
18476 their history because I've copied them in the repository)
18477 - removed obsolete files (the test scripts will be replaced
18478 by better Test::Harness variants in the future)
18480 *Ralf S. Engelschall*
18482 * First cut for a very conservative source tree cleanup:
18483 1. merge various obsolete readme texts into doc/ssleay.txt
18484 where we collect the old documents and readme texts.
18485 2. remove the first part of files where I'm already sure that we no
18486 longer need them because of three reasons: either they are just temporary
18487 files which were left by Eric or they are preserved original files where
18488 I've verified that the diff is also available in the CVS via "cvs diff
18489 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18490 the crypto/md/ stuff).
18492 *Ralf S. Engelschall*
18494 * More extension code. Incomplete support for subject and issuer alt
18495 name, issuer and authority key id. Change the i2v function parameters
18496 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18497 what that's for :-) Fix to ASN1 macro which messed up
18498 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18502 * Preliminary support for ENUMERATED type. This is largely copied from the
18507 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18509 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18511 * Make sure `make rehash` target really finds the `openssl` program.
18513 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18515 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18516 like to hear about it if this slows down other processors.
18520 * Add CygWin32 platform information to Configure script.
18522 *Alan Batie <batie@aahz.jf.intel.com>*
18524 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18526 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18528 * New program nseq to manipulate netscape certificate sequences
18532 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18537 * Fixes to BN code. Previously the default was to define BN_RECURSION
18538 but the BN code had some problems that would cause failures when
18539 doing certificate verification and some other functions.
18541 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18543 * Add ASN1 and PEM code to support netscape certificate sequences.
18547 * Add ASN1 and PEM code to support netscape certificate sequences.
18551 * Add several PKIX and private extended key usage OIDs.
18555 * Modify the 'ca' program to handle the new extension code. Modify
18556 openssl.cnf for new extension format, add comments.
18560 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18561 and add a sample to openssl.cnf so req -x509 now adds appropriate
18566 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18567 error code, add initial support to X509_print() and x509 application.
18571 * Takes a deep breath and start adding X509 V3 extension support code. Add
18572 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18573 stuff is currently isolated and isn't even compiled yet.
18577 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18578 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18579 Removed the versions check from X509 routines when loading extensions:
18580 this allows certain broken certificates that don't set the version
18581 properly to be processed.
18585 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18586 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18587 can still be regenerated with "make depend".
18591 * Spelling mistake in C version of CAST-128.
18593 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18595 * Changes to the error generation code. The perl script err-code.pl
18596 now reads in the old error codes and retains the old numbers, only
18597 adding new ones if necessary. It also only changes the .err files if new
18598 codes are added. The makefiles have been modified to only insert errors
18599 when needed (to avoid needlessly modifying header files). This is done
18600 by only inserting errors if the .err file is newer than the auto generated
18601 C file. To rebuild all the error codes from scratch (the old behaviour)
18602 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18603 or delete all the .err files.
18607 * CAST-128 was incorrectly implemented for short keys. The C version has
18608 been fixed, but is untested. The assembler versions are also fixed, but
18609 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18610 to regenerate it if needed.
18611 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18612 Hagino <itojun@kame.net>*
18614 * File was opened incorrectly in randfile.c.
18616 *Ulf Möller <ulf@fitug.de>*
18618 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18619 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18620 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18621 al: it's just almost always a UTCTime. Note this patch adds new error
18622 codes so do a "make errors" if there are problems.
18626 * Correct Linux 1 recognition in config.
18628 *Ulf Möller <ulf@fitug.de>*
18630 * Remove pointless MD5 hash when using DSA keys in ca.
18632 *Anonymous <nobody@replay.com>*
18634 * Generate an error if given an empty string as a cert directory. Also
18635 generate an error if handed NULL (previously returned 0 to indicate an
18636 error, but didn't set one).
18638 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18640 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18644 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18645 parameters. This was causing a warning which killed off the Win32 compile.
18649 * Remove C++ style comments from crypto/bn/bn_local.h.
18651 *Neil Costigan <neil.costigan@celocom.com>*
18653 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18654 based on a text string, looking up short and long names and finally
18655 "dot" format. The "dot" format stuff didn't work. Added new function
18656 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18657 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18658 OID is not part of the table.
18662 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18663 X509_LOOKUP_by_alias().
18667 * Sort openssl functions by name.
18671 * Get the `gendsa` command working and add it to the `list` command. Remove
18672 encryption from sample DSA keys (in case anyone is interested the password
18677 * Make *all* `*_free` functions accept a NULL pointer.
18679 *Frans Heymans <fheymans@isaserver.be>*
18681 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18684 *Anonymous <nobody@replay.com>*
18686 * s_server should send the CAfile as acceptable CAs, not its own cert.
18688 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18690 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18692 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18694 * Temp key "for export" tests were wrong in s3_srvr.c.
18696 *Anonymous <nobody@replay.com>*
18698 * Add prototype for temp key callback functions
18699 SSL_CTX_set_tmp_{rsa,dh}_callback().
18703 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18704 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18708 * X509_name_add_entry() freed the wrong thing after an error.
18710 *Arne Ansper <arne@ats.cyber.ee>*
18712 * rsa_eay.c would attempt to free a NULL context.
18714 *Arne Ansper <arne@ats.cyber.ee>*
18716 * BIO_s_socket() had a broken should_retry() on Windoze.
18718 *Arne Ansper <arne@ats.cyber.ee>*
18720 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18722 *Arne Ansper <arne@ats.cyber.ee>*
18724 * Make sure the already existing X509_STORE->depth variable is initialized
18725 in X509_STORE_new(), but document the fact that this variable is still
18726 unused in the certificate verification process.
18728 *Ralf S. Engelschall*
18730 * Fix the various library and `apps/` files to free up pkeys obtained from
18731 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18735 * Fix reference counting in X509_PUBKEY_get(). This makes
18736 demos/maurice/example2.c work, amongst others, probably.
18738 *Steve Henson and Ben Laurie*
18740 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18741 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18742 are no longer created. This way we have a single and consistent command
18743 line interface `openssl <command>`, similar to `cvs <command>`.
18745 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18747 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18748 BIT STRING wrapper always have zero unused bits.
18752 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18756 * Make the top-level INSTALL documentation easier to understand.
18760 * Makefiles updated to exit if an error occurs in a sub-directory
18761 make (including if user presses ^C) [Paul Sutton]
18763 * Make Montgomery context stuff explicit in RSA data structure.
18767 * Fix build order of pem and err to allow for generated pem.h.
18771 * Fix renumbering bug in X509_NAME_delete_entry().
18775 * Enhanced the err-ins.pl script so it makes the error library number
18776 global and can add a library name. This is needed for external ASN1 and
18777 other error libraries.
18781 * Fixed sk_insert which never worked properly.
18785 * Fix ASN1 macros so they can handle indefinite length constructed
18786 EXPLICIT tags. Some non standard certificates use these: they can now
18791 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18792 into a single doc/ssleay.txt bundle. This way the information is still
18793 preserved but no longer messes up this directory. Now it's new room for
18794 the new set of documentation files.
18796 *Ralf S. Engelschall*
18798 * SETs were incorrectly DER encoded. This was a major pain, because they
18799 shared code with SEQUENCEs, which aren't coded the same. This means that
18800 almost everything to do with SETs or SEQUENCEs has either changed name or
18801 number of arguments.
18803 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18805 * Fix test data to work with the above.
18809 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18810 was already fixed by Eric for 0.9.1 it seems.
18812 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18814 * Autodetect FreeBSD3.
18818 * Fix various bugs in Configure. This affects the following platforms:
18822 unixware-2.0-pentium
18827 * Eliminate generated files from CVS. Reorder tests to regenerate files
18828 before they are needed.
18832 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18836 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18838 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18839 changed SSLeay to OpenSSL in version strings.
18841 *Ralf S. Engelschall*
18843 * Some fixups to the top-level documents.
18847 * Fixed the nasty bug where rsaref.h was not found under compile-time
18848 because the symlink to include/ was missing.
18850 *Ralf S. Engelschall*
18852 * Incorporated the popular no-RSA/DSA-only patches
18853 which allow to compile a RSA-free SSLeay.
18855 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18857 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18858 when "ssleay" is still not found.
18860 *Ralf S. Engelschall*
18862 * Added more platforms to Configure: Cray T3E, HPUX 11,
18864 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18866 * Updated the README file.
18868 *Ralf S. Engelschall*
18870 * Added various .cvsignore files in the CVS repository subdirs
18871 to make a "cvs update" really silent.
18873 *Ralf S. Engelschall*
18875 * Recompiled the error-definition header files and added
18876 missing symbols to the Win32 linker tables.
18878 *Ralf S. Engelschall*
18880 * Cleaned up the top-level documents;
18881 o new files: CHANGES and LICENSE
18882 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18883 o merged COPYRIGHT into LICENSE
18884 o removed obsolete TODO file
18885 o renamed MICROSOFT to INSTALL.W32
18887 *Ralf S. Engelschall*
18889 * Removed dummy files from the 0.9.1b source tree:
18890 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18891 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18892 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18893 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18894 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18896 *Ralf S. Engelschall*
18898 * Added various platform portability fixes.
18902 * The Genesis of the OpenSSL rpject:
18903 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18904 Young and Tim J. Hudson created while they were working for C2Net until
18907 *The OpenSSL Project*
18909 ### Changes between 0.9.0b and 0.9.1b [not released]
18911 * Updated a few CA certificates under certs/
18915 * Changed some BIGNUM api stuff.
18919 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18920 DGUX x86, Linux Alpha, etc.
18924 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18925 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18930 * Add -strparse option to asn1pars program which parses nested
18933 *Dr Stephen Henson <shenson@bigfoot.com>*
18935 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18939 * DSA fix for "ca" program.
18943 * Added "-genkey" option to "dsaparam" program.
18947 * Added RIPE MD160 (rmd160) message digest.
18951 * Added -a (all) option to "ssleay version" command.
18955 * Added PLATFORM define which is the id given to Configure.
18959 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18963 * Extended the ASN.1 parser routines.
18967 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18971 * Added a BN_CTX to the BN library.
18975 * Fixed the weak key values in DES library
18979 * Changed API in EVP library for cipher aliases.
18983 * Added support for RC2/64bit cipher.
18987 * Converted the lhash library to the crypto/mem.c functions.
18991 * Added more recognized ASN.1 object ids.
18995 * Added more RSA padding checks for SSL/TLS.
18999 * Added BIO proxy/filter functionality.
19003 * Added extra_certs to SSL_CTX which can be used
19004 send extra CA certificates to the client in the CA cert chain sending
19005 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19009 * Now Fortezza is denied in the authentication phase because
19010 this is key exchange mechanism is not supported by SSLeay at all.
19014 * Additional PKCS1 checks.
19018 * Support the string "TLSv1" for all TLS v1 ciphers.
19022 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19023 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19027 * Fixed a few memory leaks.
19031 * Fixed various code and comment typos.
19035 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19036 bytes sent in the client random.
19038 *Edward Bishop <ebishop@spyglass.com>*
19042 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19043 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19044 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19045 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19046 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19047 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19048 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19049 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19050 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19051 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19052 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19053 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19054 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19055 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19056 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19057 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19058 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19059 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19060 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19061 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19062 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19063 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19064 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19065 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19066 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19067 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19068 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19069 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19070 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19071 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19072 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19073 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19074 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19075 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19076 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19077 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19078 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19079 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19080 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19081 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19082 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19083 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19084 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19085 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19086 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19087 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19088 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19089 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19090 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19091 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19092 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19093 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19094 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19095 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19096 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19097 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19098 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19099 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19100 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19101 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19102 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19103 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19104 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19105 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19106 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19107 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19108 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19109 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19110 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19111 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19112 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19113 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19114 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19115 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19116 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19117 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19118 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19119 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19120 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19121 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19122 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19123 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19124 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19125 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19126 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19127 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19128 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19129 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19130 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19131 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19132 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19133 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19134 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19135 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19136 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19137 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19138 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19139 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19140 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19141 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19142 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19143 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19144 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19145 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19146 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19147 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19148 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19149 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19150 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19151 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19152 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19153 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19154 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19155 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19156 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19157 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19158 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19159 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19160 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19161 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19162 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19163 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19164 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19165 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19166 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19167 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19168 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19169 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19170 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19171 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19172 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19173 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19174 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19175 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19176 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19177 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19178 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19179 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19180 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19181 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19182 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19183 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19184 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19185 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19186 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19187 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19188 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19189 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19190 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19191 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19192 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19193 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19194 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19195 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19196 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19197 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19198 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19199 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19200 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19201 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19202 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19203 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655