4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.1](#openssl-31)
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 3.0 and 3.1 [xx XXX xxxx]
32 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
33 listed here are only a brief description.
34 The migration guide contains more detailed information related to new features,
35 breaking changes, and mappings for the large list of deprecated functions.
37 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
39 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
41 * Due to move of the implementation of cryptographic operations
42 to the providers, validation of various operation parameters can
43 be postponed until the actual operation is executed where previously
44 it happened immediately when an operation parameter was set.
46 For example when setting an unsupported curve with
47 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
48 fail but later keygen operations with the EVP_PKEY_CTX will fail.
50 *OpenSSL team members and many third party contributors*
52 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
53 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
54 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
55 instead to retrieve these algorithms from a provider.
59 * On build targets where the multilib postfix is set in the build
60 configuration the libdir directory was changing based on whether
61 the lib directory with the multilib postfix exists on the system
62 or not. This unpredictable behavior was removed and eventual
63 multilib postfix is now always added to the default libdir. Use
64 `--libdir=lib` to override the libdir if adding the postfix is
69 * The triple DES key wrap functionality now conforms to RFC 3217 but is
70 no longer interoperable with OpenSSL 1.1.1.
74 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
75 function codes, this function can only cause problems for calling
80 * Add a configurable flag to output date formats as ISO 8601. Does not
81 change the default date format.
85 * Version of MSVC earlier than 1300 could get link warnings, which could
86 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
87 Support for this flag has been removed.
91 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
92 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
93 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
94 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
95 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
99 * The signatures of the functions to get and set options on SSL and
100 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
101 Some source code changes may be required.
105 * The public definitions of conf_method_st and conf_st have been
106 deprecated. They will be made opaque in a future release.
108 *Rich Salz and Tomáš Mráz*
110 * Client-initiated renegotiation is disabled by default. To allow it, use
111 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
112 flag, or the "ClientRenegotiation" config parameter as appropriate.
116 * Add "abspath" and "includedir" pragma's to config files, to prevent,
117 or modify relative pathname inclusion.
121 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
122 validated. Please consult the README-FIPS and
123 README-PROVIDERS files, as well as the migration guide.
125 *OpenSSL team members and many third party contributors*
127 * For the key types DH and DHX the allowed settable parameters are now different.
131 * The openssl commands that read keys, certificates, and CRLs now
132 automatically detect the PEM or DER format of the input files.
134 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
136 * Added enhanced PKCS#12 APIs which accept a library context.
140 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
144 * Added support for Kernel TLS (KTLS).
146 *Boris Pismenny, John Baldwin and Andrew Gallatin*
148 * Support for RFC 5746 secure renegotiation is now required by default for
149 SSL or TLS connections to succeed.
153 * The signature of the `copy` functional parameter of the
154 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
155 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
156 the signature of the `pub_decode` functional parameter of the
157 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
158 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
162 * The error return values from some control calls (ctrl) have changed.
166 * A public key check is now performed during EVP_PKEY_derive_set_peer().
170 * Many functions in the EVP_ namespace that are getters of values from
171 implementations or contexts were renamed to include get or get0 in their
172 names. Old names are provided as macro aliases for compatibility and
177 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
178 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
179 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
184 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
187 * The output from the command line applications may have minor
192 * The output from numerous "printing" may have minor changes.
196 * Windows thread synchronization uses read/write primitives (SRWLock) when
197 supported by the OS, otherwise CriticalSection continues to be used.
201 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
202 work on read only BIO source/sinks that do not support these functions.
203 This allows piping or redirection of a file BIO using stdin to be buffered
204 into memory. This is used internally in OSSL_DECODER_from_bio().
208 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
209 this function would return one of the values OSSL_STORE_INFO_NAME,
210 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
211 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
212 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
213 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
214 using this function should be amended to handle the changed return value.
218 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
219 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
220 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
221 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
222 contains more than one certificate identifier: This means that all
223 certificates referenced there MUST be part of the validation chain.
227 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
228 RC5, DESX and DES have been moved to the legacy provider.
232 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
233 RIPEMD-160 have been moved to the legacy provider.
237 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
242 * The deprecated functions EVP_PKEY_get0_RSA(),
243 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
244 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
245 well as the similarly named "get1" functions behave differently in
250 * A number of functions handling low-level keys or engines were deprecated
251 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
252 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
253 EVP_PKEY_get0_siphash().
257 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
258 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
259 will need to load the legacy crypto provider. This includes these PBE
260 algorithms which use this KDF:
261 - NID_pbeWithMD2AndDES_CBC
262 - NID_pbeWithMD5AndDES_CBC
263 - NID_pbeWithSHA1AndRC2_CBC
264 - NID_pbeWithMD2AndRC2_CBC
265 - NID_pbeWithMD5AndRC2_CBC
266 - NID_pbeWithSHA1AndDES_CBC
270 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
271 BIO_debug_callback() functions.
275 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
276 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
280 * The RAND_METHOD APIs have been deprecated.
284 * The SRP APIs have been deprecated.
288 * Add a compile time option to prevent the caching of provider fetched
289 algorithms. This is enabled by including the no-cached-fetch option
290 at configuration time.
294 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
295 count of PKCS12_DEFAULT_ITER.
297 *Tomáš Mráz and Sahana Prasad*
299 * The openssl speed command does not use low-level API calls anymore.
303 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
306 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
308 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
312 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
313 providers may supply their own group implementations (using either the "key
314 exchange" or the "key encapsulation" methods) which will automatically be
315 detected and used by libssl.
317 *Matt Caswell, Nicola Tuveri*
319 * The undocumented function X509_certificate_type() has been deprecated;
323 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
327 * Removed RSA padding mode for SSLv23 (which was only used for
328 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
329 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
334 * Deprecated the obsolete X9.31 RSA key generation related functions.
336 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
337 is not allowed to return a value > 1, this is no more taken as failure.
339 *Viktor Dukhovni and David von Oheimb*
341 * Deprecated the obsolete X9.31 RSA key generation related functions
342 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
343 BN_X931_generate_prime_ex().
347 * The default key generation method for the regular 2-prime RSA keys was
348 changed to the FIPS 186-4 B.3.6 method.
352 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
356 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
360 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
361 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
363 *Rich Salz, Richard Levitte, and David von Oheimb*
365 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
369 * Deprecated `OCSP_parse_url()`.
373 * Validation of SM2 keys has been separated from the validation of regular EC
378 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
379 switches: a validation failure triggers an early exit, returning a failure
380 exit status to the parent process.
384 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
385 to ignore unknown ciphers.
389 * The `-cipher-commands` and `-digest-commands` options
390 of the command line utility `list` have been deprecated.
391 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
395 * Added convenience functions for generating asymmetric key pairs:
396 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
397 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
401 * All of the low level EC_KEY functions have been deprecated.
403 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
405 * Deprecated all the libcrypto and libssl error string loading
410 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
411 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
416 * The `-crypt` option to the `passwd` command line tool has been removed.
420 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
425 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
429 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
430 EVP_PKEY_get1_tls_encodedpoint().
434 * The security callback, which can be customised by application code, supports
435 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
436 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
440 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
441 interface. Their functionality remains unchanged.
445 * Added new option for 'openssl list', '-providers', which will display the
446 list of loaded providers, their names, version and status. It optionally
447 displays their gettable parameters.
451 * Removed EVP_PKEY_set_alias_type().
455 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
456 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
460 * Changed all "STACK" functions to be macros instead of inline functions. Macro
461 parameters are still checked for type safety at compile time via helper
466 * Remove the RAND_DRBG API
468 *Paul Dale and Matthias St. Pierre*
470 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
471 as well as actual hostnames.
475 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
476 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
477 conversely, silently ignore DTLS protocol version bounds when configuring
478 TLS-based contexts. The commands can be repeated to set bounds of both
479 types. The same applies with the corresponding "min_protocol" and
480 "max_protocol" command-line switches, in case some application uses both TLS
483 SSL_CTX instances that are created for a fixed protocol version (e.g.
484 `TLSv1_server_method()`) also silently ignore version bounds. Previously
485 attempts to apply bounds to these protocol versions would result in an
486 error. Now only the "version-flexible" SSL_CTX instances are subject to
487 limits in configuration files in command-line options.
491 * Deprecated the `ENGINE` API. Engines should be replaced with providers
496 * Reworked the recorded ERR codes to make better space for system errors.
497 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
498 given code is a system error (true) or an OpenSSL error (false).
502 * Reworked the test perl framework to better allow parallel testing.
504 *Nicola Tuveri and David von Oheimb*
506 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
507 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
511 * 'Configure' has been changed to figure out the configuration target if
512 none is given on the command line. Consequently, the 'config' script is
513 now only a mere wrapper. All documentation is changed to only mention
516 *Rich Salz and Richard Levitte*
518 * Added a library context `OSSL_LIB_CTX` that applications as well as
519 other libraries can use to form a separate context within which
520 libcrypto operations are performed.
524 * Handshake now fails if Extended Master Secret extension is dropped
529 * Dropped interactive mode from the `openssl` program.
533 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
535 *David von Oheimb and Shane Lontis*
537 * Deprecated `EC_METHOD_get_field_type()`.
541 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
542 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
543 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
547 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
551 * Add CAdES-BES signature verification support, mostly derived
552 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
554 *Filipe Raimundo da Silva*
556 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
560 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
561 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
565 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
569 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
570 EC_KEY_precompute_mult().
574 * Deprecated EC_POINTs_mul().
578 * Removed FIPS_mode() and FIPS_mode_set().
582 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
586 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
587 EC_POINT_get_Jprojective_coordinates_GFp().
591 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
592 arrays to be more easily constructed via a series of utility functions.
593 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
594 the various push functions and finally convert to a passable OSSL_PARAM
595 array using OSSL_PARAM_BLD_to_param().
599 * The security strength of SHA1 and MD5 based signatures in TLS has been
604 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
605 contain a provider side internal key.
609 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
613 * Project text documents not yet having a proper file name extension
614 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
615 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
619 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
620 have been converted to Markdown with the goal to produce documents
621 which not only look pretty when viewed online in the browser, but
622 remain well readable inside a plain text editor.
624 To achieve this goal, a 'minimalistic' Markdown style has been applied
625 which avoids formatting elements that interfere too much with the
626 reading flow in the text file. For example, it
628 * avoids [ATX headings][] and uses [setext headings][] instead
629 (which works for `<h1>` and `<h2>` headings only).
630 * avoids [inline links][] and uses [reference links][] instead.
631 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
633 [ATX headings]: https://github.github.com/gfm/#atx-headings
634 [setext headings]: https://github.github.com/gfm/#setext-headings
635 [inline links]: https://github.github.com/gfm/#inline-link
636 [reference links]: https://github.github.com/gfm/#reference-link
637 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
638 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
640 *Matthias St. Pierre*
642 * The test suite is changed to preserve results of each test recipe.
643 A new directory test-runs/ with subdirectories named like the
644 test recipes are created in the build tree for this purpose.
648 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
649 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
650 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
652 *David von Oheimb, Martin Peylo*
654 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
655 It supports arbitrary request and response content types, GET redirection,
656 TLS, connections via HTTP(S) proxies, connections and exchange via
657 user-defined BIOs (allowing implicit connections), persistent connections,
658 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
659 The legacy OCSP-focused (and only partly documented) API
660 is retained for backward compatibility, while most of it is deprecated.
664 * Added `util/check-format.pl`, a tool for checking adherence to the
665 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
666 The checks performed are incomplete and yield some false positives.
667 Still the tool should be useful for detecting most typical glitches.
671 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
672 If domain name resolution yields multiple IP addresses all of them are tried
673 after `connect()` failures.
677 * All of the low level RSA functions have been deprecated.
681 * X509 certificates signed using SHA1 are no longer allowed at security
686 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
687 modified to use PKEY APIs. These commands are now in maintenance mode
688 and no new features will be added to them.
692 * The command line utility rsautl has been deprecated.
696 * The command line utilities genrsa and rsa have been modified to use PKEY
697 APIs. They now write PKCS#8 keys by default. These commands are now in
698 maintenance mode and no new features will be added to them.
702 * All of the low level DH functions have been deprecated.
704 *Paul Dale and Matt Caswell*
706 * All of the low level DSA functions have been deprecated.
710 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
711 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
715 * Deprecated low level ECDH and ECDSA functions.
719 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
723 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
724 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
725 a new formulation to include all the things it can be used for,
726 as well as words of caution.
730 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
734 * All of the low level HMAC functions have been deprecated.
736 *Paul Dale and David von Oheimb*
738 * Over two thousand fixes were made to the documentation, including:
739 - Common options (such as -rand/-writerand, TLS version control, etc)
740 were refactored and point to newly-enhanced descriptions in openssl.pod.
741 - Added style conformance for all options (with help from Richard Levitte),
742 documented all reported missing options, added a CI build to check
743 that all options are documented and that no unimplemented options
745 - Documented some internals, such as all use of environment variables.
746 - Addressed all internal broken L<> references.
750 * All of the low level CMAC functions have been deprecated.
754 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
755 functions have been deprecated.
757 *Paul Dale and David von Oheimb*
759 * Corrected the documentation of the return values from the `EVP_DigestSign*`
760 set of functions. The documentation mentioned negative values for some
761 errors, but this was never the case, so the mention of negative values
764 Code that followed the documentation and thereby check with something
765 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
769 * All of the low level cipher functions have been deprecated.
771 *Matt Caswell and Paul Dale*
773 * Removed include/openssl/opensslconf.h.in and replaced it with
774 include/openssl/configuration.h.in, which differs in not including
775 <openssl/macros.h>. A short header include/openssl/opensslconf.h
776 was added to include both.
778 This allows internal hacks where one might need to modify the set
779 of configured macros, for example this if deprecated symbols are
780 still supposed to be available internally:
782 #include <openssl/configuration.h>
784 #undef OPENSSL_NO_DEPRECATED
785 #define OPENSSL_SUPPRESS_DEPRECATED
787 #include <openssl/macros.h>
789 This should not be used by applications that use the exported
790 symbols, as that will lead to linking errors.
794 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
795 used in exponentiation with 512-bit moduli. No EC algorithms are
796 affected. Analysis suggests that attacks against 2-prime RSA1024,
797 3-prime RSA1536, and DSA1024 as a result of this defect would be very
798 difficult to perform and are not believed likely. Attacks against DH512
799 are considered just feasible. However, for an attack the target would
800 have to re-use the DH512 private key, which is not recommended anyway.
801 Also applications directly using the low-level API BN_mod_exp may be
802 affected if they use BN_FLG_CONSTTIME.
807 * Most memory-debug features have been deprecated, and the functionality
808 replaced with no-ops.
812 * Added documentation for the STACK API.
816 * Introduced a new method type and API, OSSL_ENCODER, to
817 represent generic encoders.
821 * Added a .pragma directive to the syntax of configuration files, to
822 allow varying behavior in a supported and predictable manner.
823 Currently added pragma:
827 This allows dollar signs to be a keyword character unless it's
828 followed by a opening brace or parenthesis. This is useful for
829 platforms where dollar signs are commonly used in names, such as
830 volume names and system directory names on VMS.
834 * Added functionality to create an EVP_PKEY from user data.
838 * Change the interpretation of the '--api' configuration option to
839 mean that this is a desired API compatibility level with no
840 further meaning. The previous interpretation, that this would
841 also mean to remove all deprecated symbols up to and including
842 the given version, no requires that 'no-deprecated' is also used
843 in the configuration.
845 When building applications, the desired API compatibility level
846 can be set with the OPENSSL_API_COMPAT macro like before. For
847 API compatibility version below 3.0, the old style numerical
848 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
849 For version 3.0 and on, the value is expected to be the decimal
850 value calculated from the major and minor version like this:
852 MAJOR * 10000 + MINOR * 100
856 -DOPENSSL_API_COMPAT=30000 For 3.0
857 -DOPENSSL_API_COMPAT=30200 For 3.2
859 To hide declarations that are deprecated up to and including the
860 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
861 given when building the application as well.
865 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
866 access to certificate and CRL stores via URIs and OSSL_STORE
869 This adds the following functions:
871 - X509_LOOKUP_store()
872 - X509_STORE_load_file()
873 - X509_STORE_load_path()
874 - X509_STORE_load_store()
875 - SSL_add_store_cert_subjects_to_stack()
876 - SSL_CTX_set_default_verify_store()
877 - SSL_CTX_load_verify_file()
878 - SSL_CTX_load_verify_dir()
879 - SSL_CTX_load_verify_store()
883 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
884 The presence of this system service is determined at run-time.
888 * Added functionality to create an EVP_PKEY context based on data
889 for methods from providers. This takes an algorithm name and a
890 property query string and simply stores them, with the intent
891 that any operation that uses this context will use those strings
892 to fetch the needed methods implicitly, thereby making the port
893 of application written for pre-3.0 OpenSSL easier.
897 * The undocumented function NCONF_WIN32() has been deprecated; for
898 conversion details see the HISTORY section of doc/man5/config.pod
902 * Introduced the new functions EVP_DigestSignInit_ex() and
903 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
904 EVP_DigestVerifyUpdate() have been converted to functions. See the man
905 pages for further details.
909 * Over two thousand fixes were made to the documentation, including:
910 adding missing command flags, better style conformance, documentation
913 *Rich Salz, Richard Levitte*
915 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
916 X25519, X448, Ed25519 and Ed448.
920 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
925 * Deprecated the public definition of `ERR_STATE` as well as the function
926 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
931 * Added ERR functionality to give callers access to the stored function
932 names that have replaced the older function code based functions.
934 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
935 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
936 ERR_peek_error_all() and ERR_peek_last_error_all().
938 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
939 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
940 ERR_func_error_string().
944 * Extended testing to be verbose for failing tests only. The make variables
945 VERBOSE_FAILURE or VF can be used to enable this:
947 $ make VF=1 test # Unix
948 $ mms /macro=(VF=1) test ! OpenVMS
949 $ nmake VF=1 test # Windows
953 * Added the `-copy_extensions` option to the `x509` command for use with
954 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
955 all extensions in the request are copied to the certificate or vice versa.
957 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
959 * Added the `-copy_extensions` option to the `req` command for use with
960 `-x509`. When given with the `copy` or `copyall` argument,
961 all extensions in the certification request are copied to the certificate.
965 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
966 they generate are by default RFC 5280 compliant in the following sense:
967 There is a subjectKeyIdentifier extension with a hash value of the public key
968 and for not self-signed certs there is an authorityKeyIdentifier extension
969 with a keyIdentifier field or issuer information identifying the signing key.
970 This is done unless some configuration overrides the new default behavior,
971 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
975 * Added several checks to `X509_verify_cert()` according to requirements in
976 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
977 (which may be done by using the CLI option `-x509_strict`):
978 * The basicConstraints of CA certificates must be marked critical.
979 * CA certificates must explicitly include the keyUsage extension.
980 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
981 * The issuer name of any certificate must not be empty.
982 * The subject name of CA certs, certs with keyUsage crlSign,
983 and certs without subjectAlternativeName must not be empty.
984 * If a subjectAlternativeName extension is given it must not be empty.
985 * The signatureAlgorithm field and the cert signature must be consistent.
986 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
987 must not be marked critical.
988 * The authorityKeyIdentifier must be given for X.509v3 certs
989 unless they are self-signed.
990 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
994 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
995 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
999 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1000 used even when parsing explicit parameters, when loading a encoded key
1001 or calling `EC_GROUP_new_from_ecpkparameters()`/
1002 `EC_GROUP_new_from_ecparameters()`.
1003 This prevents bypass of security hardening and performance gains,
1004 especially for curves with specialized EC_METHODs.
1005 By default, if a key encoded with explicit parameters is loaded and later
1006 encoded, the output is still encoded with explicit parameters, even if
1007 internally a "named" EC_GROUP is used for computation.
1011 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1012 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1013 NULL. After this change, only the cofactor parameter can be NULL. It also
1014 does some minimal sanity checks on the passed order.
1019 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1020 An attack is simple, if the first CMS_recipientInfo is valid but the
1021 second CMS_recipientInfo is chosen ciphertext. If the second
1022 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1023 encryption key will be replaced by garbage, and the message cannot be
1024 decoded, but if the RSA decryption fails, the correct encryption key is
1025 used and the recipient will not notice the attack.
1026 As a work around for this potential attack the length of the decrypted
1027 key must be equal to the cipher default key length, in case the
1028 certifiate is not given and all recipientInfo are tried out.
1029 The old behaviour can be re-enabled in the CMS code by setting the
1030 CMS_DEBUG_DECRYPT flag.
1034 * Early start up entropy quality from the DEVRANDOM seed source has been
1035 improved for older Linux systems. The RAND subsystem will wait for
1036 /dev/random to be producing output before seeding from /dev/urandom.
1037 The seeded state is stored for future library initialisations using
1038 a system global shared memory segment. The shared memory identifier
1039 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1040 the desired value. The default identifier is 114.
1044 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1045 when primes for RSA keys are computed.
1046 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1047 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1048 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1049 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1050 This avoids possible fingerprinting of newly generated RSA modules.
1054 * Correct the extended master secret constant on EBCDIC systems. Without this
1055 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1056 negotiate EMS will fail. Unfortunately this also means that TLS connections
1057 between EBCDIC systems with this fix, and EBCDIC systems without this
1058 fix will fail if they negotiate EMS.
1062 * Changed the library initialisation so that the config file is now loaded
1063 by default. This was already the case for libssl. It now occurs for both
1064 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1065 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1069 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1070 where the former acts as a replacement for `ERR_put_error()`, and the
1071 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1072 `ERR_raise_data()` adds more flexibility by taking a format string and
1073 an arbitrary number of arguments following it, to be processed with
1078 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1079 to check if a named provider is loaded and available. When called, it
1080 will also activate all fallback providers if such are still present.
1084 * Enforce a minimum DH modulus size of 512 bits.
1088 * Changed DH parameters to generate the order q subgroup instead of 2q.
1089 Previously generated DH parameters are still accepted by DH_check
1090 but DH_generate_key works around that by clearing bit 0 of the
1091 private key for those. This avoids leaking bit 0 of the private key.
1095 * Significantly reduce secure memory usage by the randomness pools.
1099 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1104 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1105 algorithms. An implementation of a key exchange algorithm can be obtained
1106 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1107 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1108 the older EVP_PKEY_derive_init() function. See the man pages for the new
1109 functions for further details.
1113 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1117 * Removed the function names from error messages and deprecated the
1122 * Removed NextStep support and the macro OPENSSL_UNISTD
1126 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1127 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1128 Also removed "export var as function" capability; we do not export
1129 variables, only functions.
1133 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1134 an error and 1 indicating success. In previous versions of OpenSSL this
1135 was a void type. If a key was set longer than the maximum possible this
1140 * Support SM2 signing and verification schemes with X509 certificate.
1144 * Use SHA256 as the default digest for TS query in the `ts` app.
1148 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1152 * Default cipher lists/suites are now available via a function, the
1153 #defines are deprecated.
1157 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1158 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1159 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1163 * Join the directories crypto/x509 and crypto/x509v3
1167 * Added command 'openssl kdf' that uses the EVP_KDF API.
1171 * Added command 'openssl mac' that uses the EVP_MAC API.
1175 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1176 as default directories. Also added the command 'openssl info'
1177 for scripting purposes.
1181 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1186 * Add prediction resistance to the DRBG reseeding process.
1190 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1191 mandated by IEEE Std 1619-2018.
1195 * Added newline escaping functionality to a filename when using openssl dgst.
1196 This output format is to replicate the output format found in the `*sum`
1197 checksum programs. This aims to preserve backward compatibility.
1199 *Matt Eaton, Richard Levitte, and Paul Dale*
1201 * Removed the heartbeat message in DTLS feature, as it has very
1202 little usage and doesn't seem to fulfill a valuable purpose.
1203 The configuration option is now deprecated.
1207 * Changed the output of 'openssl {digestname} < file' to display the
1208 digest name in its output.
1212 * Added a new generic trace API which provides support for enabling
1213 instrumentation through trace output.
1215 *Richard Levitte & Matthias St. Pierre*
1217 * Added build tests for C++. These are generated files that only do one
1218 thing, to include one public OpenSSL head file each. This tests that
1219 the public header files can be usefully included in a C++ application.
1221 This test isn't enabled by default. It can be enabled with the option
1222 'enable-buildtest-c++'.
1226 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1230 * Add KMAC to EVP_MAC.
1234 * Added property based algorithm implementation selection framework to
1239 * Added SCA hardening for modular field inversion in EC_GROUP through
1240 a new dedicated field_inv() pointer in EC_METHOD.
1241 This also addresses a leakage affecting conversions from projective
1242 to affine coordinates.
1244 *Billy Bob Brumley, Nicola Tuveri*
1246 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1247 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1248 those algorithms that were already supported through the EVP_PKEY API
1249 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1250 and scrypt are now wrappers that call EVP_KDF.
1254 * Build devcrypto engine as a dynamic engine.
1256 *Eneas U de Queiroz*
1258 * Add keyed BLAKE2 to EVP_MAC.
1262 * Fix a bug in the computation of the endpoint-pair shared secret used
1263 by DTLS over SCTP. This breaks interoperability with older versions
1264 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1265 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1266 interoperability with such broken implementations. However, enabling
1267 this switch breaks interoperability with correct implementations.
1269 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1270 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1274 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1278 * Changed the license to the Apache License v2.0.
1282 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1284 - Major releases (indicated by incrementing the MAJOR release number)
1285 may introduce incompatible API/ABI changes.
1286 - Minor releases (indicated by incrementing the MINOR release number)
1287 may introduce new features but retain API/ABI compatibility.
1288 - Patch releases (indicated by incrementing the PATCH number)
1289 are intended for bug fixes and other improvements of existing
1290 features only (like improving performance or adding documentation)
1291 and retain API/ABI compatibility.
1295 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1299 * Remove the 'dist' target and add a tarball building script. The
1300 'dist' target has fallen out of use, and it shouldn't be
1301 necessary to configure just to create a source distribution.
1305 * Recreate the OS390-Unix config target. It no longer relies on a
1306 special script like it did for OpenSSL pre-1.1.0.
1310 * Instead of having the source directories listed in Configure, add
1311 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1316 * Add GMAC to EVP_MAC.
1320 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1324 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1325 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1326 to facilitate the continued use of MACs through raw private keys in
1327 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1331 * Deprecate ECDH_KDF_X9_62().
1335 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1336 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1337 are retained for backwards compatibility.
1341 * AES-XTS mode now enforces that its two keys are different to mitigate
1342 the attacked described in "Efficient Instantiations of Tweakable
1343 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1344 Details of this attack can be obtained from:
1345 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1349 * Rename the object files, i.e. give them other names than in previous
1350 versions. Their names now include the name of the final product, as
1351 well as its type mnemonic (bin, lib, shlib).
1355 * Added new option for 'openssl list', '-objects', which will display the
1356 list of built in objects, i.e. OIDs with names.
1360 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1361 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1366 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1367 improves application performance by removing data copies and providing
1368 applications with zero-copy system calls such as sendfile and splice.
1372 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1376 * `PKCS12_parse` now maintains the order of the parsed certificates
1377 when outputting them via `*ca` (rather than reversing it).
1381 * Deprecated pthread fork support methods.
1385 * Added support for FFDHE key exchange in TLS 1.3.
1392 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1394 * Fixed an SM2 Decryption Buffer Overflow.
1396 In order to decrypt SM2 encrypted data an application is expected to
1397 call the API function EVP_PKEY_decrypt(). Typically an application will
1398 call this function twice. The first time, on entry, the "out" parameter
1399 can be NULL and, on exit, the "outlen" parameter is populated with the
1400 buffer size required to hold the decrypted plaintext. The application
1401 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1402 again, but this time passing a non-NULL value for the "out" parameter.
1404 A bug in the implementation of the SM2 decryption code means that the
1405 calculation of the buffer size required to hold the plaintext returned
1406 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1407 size required by the second call. This can lead to a buffer overflow
1408 when EVP_PKEY_decrypt() is called by the application a second time with
1409 a buffer that is too small.
1411 A malicious attacker who is able present SM2 content for decryption to
1412 an application could cause attacker chosen data to overflow the buffer
1413 by up to a maximum of 62 bytes altering the contents of other data held
1414 after the buffer, possibly changing application behaviour or causing
1415 the application to crash. The location of the buffer is application
1416 dependent but is typically heap allocated.
1421 * Fixed various read buffer overruns processing ASN.1 strings
1423 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1424 structure which contains a buffer holding the string data and a field
1425 holding the buffer length. This contrasts with normal C strings which
1426 are repesented as a buffer for the string data which is terminated
1427 with a NUL (0) byte.
1429 Although not a strict requirement, ASN.1 strings that are parsed using
1430 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1431 well as any string whose value has been set with the ASN1_STRING_set()
1432 function will additionally NUL terminate the byte array in the
1433 ASN1_STRING structure.
1435 However, it is possible for applications to directly construct valid
1436 ASN1_STRING structures which do not NUL terminate the byte array by
1437 directly setting the "data" and "length" fields in the ASN1_STRING
1438 array. This can also happen by using the ASN1_STRING_set0() function.
1440 Numerous OpenSSL functions that print ASN.1 data have been found to
1441 assume that the ASN1_STRING byte array will be NUL terminated, even
1442 though this is not guaranteed for strings that have been directly
1443 constructed. Where an application requests an ASN.1 structure to be
1444 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1445 been directly constructed by the application without NUL terminating
1446 the "data" field, then a read buffer overrun can occur.
1448 The same thing can also occur during name constraints processing
1449 of certificates (for example if a certificate has been directly
1450 constructed by the application instead of loading it via the OpenSSL
1451 parsing functions, and the certificate contains non NUL terminated
1452 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1453 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1455 If a malicious actor can cause an application to directly construct an
1456 ASN1_STRING and then process it through one of the affected OpenSSL
1457 functions then this issue could be hit. This might result in a crash
1458 (causing a Denial of Service attack). It could also result in the
1459 disclosure of private memory contents (such as private keys, or
1460 sensitive plaintext).
1465 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1467 * Fixed a problem with verifying a certificate chain when using the
1468 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1469 the certificates present in a certificate chain. It is not set by default.
1471 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1472 the chain that have explicitly encoded elliptic curve parameters was added
1473 as an additional strict check.
1475 An error in the implementation of this check meant that the result of a
1476 previous check to confirm that certificates in the chain are valid CA
1477 certificates was overwritten. This effectively bypasses the check
1478 that non-CA certificates must not be able to issue other certificates.
1480 If a "purpose" has been configured then there is a subsequent opportunity
1481 for checks that the certificate is a valid CA. All of the named "purpose"
1482 values implemented in libcrypto perform this check. Therefore, where
1483 a purpose is set the certificate chain will still be rejected even when the
1484 strict flag has been used. A purpose is set by default in libssl client and
1485 server certificate verification routines, but it can be overridden or
1486 removed by an application.
1488 In order to be affected, an application must explicitly set the
1489 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1490 for the certificate verification or, in the case of TLS client or server
1491 applications, override the default purpose.
1496 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1497 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1498 renegotiation ClientHello omits the signature_algorithms extension (where it
1499 was present in the initial ClientHello), but includes a
1500 signature_algorithms_cert extension then a NULL pointer dereference will
1501 result, leading to a crash and a denial of service attack.
1503 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1504 (which is the default configuration). OpenSSL TLS clients are not impacted by
1508 *Peter Kästle and Samuel Sapalski*
1510 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1512 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1513 create a unique hash value based on the issuer and serial number data
1514 contained within an X509 certificate. However it was failing to correctly
1515 handle any errors that may occur while parsing the issuer field (which might
1516 occur if the issuer field is maliciously constructed). This may subsequently
1517 result in a NULL pointer deref and a crash leading to a potential denial of
1523 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1524 padding mode to correctly check for rollback attacks. This is considered a
1525 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1530 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1531 functions. Previously they could overflow the output length argument in some
1532 cases where the input length is close to the maximum permissable length for
1533 an integer on the platform. In such cases the return value from the function
1534 call would be 1 (indicating success), but the output length value would be
1535 negative. This could cause applications to behave incorrectly or crash.
1540 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1541 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1542 could be exploited in a side channel attack to recover the password. Since
1543 the attack is local host only this is outside of the current OpenSSL
1544 threat model and therefore no CVE is assigned.
1546 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1551 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1553 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1554 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1555 If an attacker can control both items being compared then this could lead
1556 to a possible denial of service attack. OpenSSL itself uses the
1557 GENERAL_NAME_cmp function for two purposes:
1558 1) Comparing CRL distribution point names between an available CRL and a
1559 CRL distribution point embedded in an X509 certificate
1560 2) When verifying that a timestamp response token signer matches the
1561 timestamp authority name (exposed via the API functions
1562 TS_RESP_verify_response and TS_RESP_verify_token)
1567 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1569 * Certificates with explicit curve parameters are now disallowed in
1570 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1574 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1575 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1576 conversely, silently ignore DTLS protocol version bounds when configuring
1577 TLS-based contexts. The commands can be repeated to set bounds of both
1578 types. The same applies with the corresponding "min_protocol" and
1579 "max_protocol" command-line switches, in case some application uses both TLS
1582 SSL_CTX instances that are created for a fixed protocol version (e.g.
1583 TLSv1_server_method()) also silently ignore version bounds. Previously
1584 attempts to apply bounds to these protocol versions would result in an
1585 error. Now only the "version-flexible" SSL_CTX instances are subject to
1586 limits in configuration files in command-line options.
1590 * Handshake now fails if Extended Master Secret extension is dropped
1595 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1597 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1599 * Fixed segmentation fault in SSL_check_chain()
1600 Server or client applications that call the SSL_check_chain() function
1601 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1602 dereference as a result of incorrect handling of the
1603 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1604 or unrecognised signature algorithm is received from the peer. This could
1605 be exploited by a malicious peer in a Denial of Service attack.
1610 * Added AES consttime code for no-asm configurations
1611 an optional constant time support for AES was added
1612 when building openssl for no-asm.
1613 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1614 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1615 At this time this feature is by default disabled.
1616 It will be enabled by default in 3.0.
1620 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1622 * Revert the change of EOF detection while reading in libssl to avoid
1623 regressions in applications depending on the current way of reporting
1624 the EOF. As the existing method is not fully accurate the change to
1625 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1626 branch and will be present in the 3.0 release.
1630 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1631 when primes for RSA keys are computed.
1632 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1633 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1634 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1635 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1636 This avoids possible fingerprinting of newly generated RSA modules.
1640 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1642 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1643 while reading in libssl then we would report an error back to the
1644 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1645 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1646 therefore give a hint as to what went wrong.
1650 * Check that ed25519 and ed448 are allowed by the security level. Previously
1651 signature algorithms not using an MD were not being checked that they were
1652 allowed by the security level.
1656 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1657 was not quite right. The behaviour was not consistent between resumption
1658 and normal handshakes, and also not quite consistent with historical
1659 behaviour. The behaviour in various scenarios has been clarified and
1660 it has been updated to make it match historical behaviour as closely as
1665 * *[VMS only]* The header files that the VMS compilers include automatically,
1666 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1667 that the C++ compiler doesn't understand. This is a shortcoming in the
1668 compiler, but can be worked around with `__cplusplus` guards.
1670 C++ applications that use OpenSSL libraries must be compiled using the
1671 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1672 functions. Otherwise, only functions with symbols of less than 31
1673 characters can be used, as the linker will not be able to successfully
1674 resolve symbols with longer names.
1678 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1679 The presence of this system service is determined at run-time.
1683 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1688 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1690 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1691 number generator (RNG). This was intended to include protection in the
1692 event of a fork() system call in order to ensure that the parent and child
1693 processes did not share the same RNG state. However this protection was not
1694 being used in the default case.
1696 A partial mitigation for this issue is that the output from a high
1697 precision timer is mixed into the RNG state so the likelihood of a parent
1698 and child process sharing state is significantly reduced.
1700 If an application already calls OPENSSL_init_crypto() explicitly using
1701 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1704 *Matthias St. Pierre*
1706 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1707 used even when parsing explicit parameters, when loading a encoded key
1708 or calling `EC_GROUP_new_from_ecpkparameters()`/
1709 `EC_GROUP_new_from_ecparameters()`.
1710 This prevents bypass of security hardening and performance gains,
1711 especially for curves with specialized EC_METHODs.
1712 By default, if a key encoded with explicit parameters is loaded and later
1713 encoded, the output is still encoded with explicit parameters, even if
1714 internally a "named" EC_GROUP is used for computation.
1718 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1719 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1720 NULL. After this change, only the cofactor parameter can be NULL. It also
1721 does some minimal sanity checks on the passed order.
1726 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1727 An attack is simple, if the first CMS_recipientInfo is valid but the
1728 second CMS_recipientInfo is chosen ciphertext. If the second
1729 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1730 encryption key will be replaced by garbage, and the message cannot be
1731 decoded, but if the RSA decryption fails, the correct encryption key is
1732 used and the recipient will not notice the attack.
1733 As a work around for this potential attack the length of the decrypted
1734 key must be equal to the cipher default key length, in case the
1735 certifiate is not given and all recipientInfo are tried out.
1736 The old behaviour can be re-enabled in the CMS code by setting the
1737 CMS_DEBUG_DECRYPT flag.
1742 * Early start up entropy quality from the DEVRANDOM seed source has been
1743 improved for older Linux systems. The RAND subsystem will wait for
1744 /dev/random to be producing output before seeding from /dev/urandom.
1745 The seeded state is stored for future library initialisations using
1746 a system global shared memory segment. The shared memory identifier
1747 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1748 the desired value. The default identifier is 114.
1752 * Correct the extended master secret constant on EBCDIC systems. Without this
1753 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1754 negotiate EMS will fail. Unfortunately this also means that TLS connections
1755 between EBCDIC systems with this fix, and EBCDIC systems without this
1756 fix will fail if they negotiate EMS.
1760 * Use Windows installation paths in the mingw builds
1762 Mingw isn't a POSIX environment per se, which means that Windows
1763 paths should be used for installation.
1768 * Changed DH_check to accept parameters with order q and 2q subgroups.
1769 With order 2q subgroups the bit 0 of the private key is not secret
1770 but DH_generate_key works around that by clearing bit 0 of the
1771 private key for those. This avoids leaking bit 0 of the private key.
1775 * Significantly reduce secure memory usage by the randomness pools.
1779 * Revert the DEVRANDOM_WAIT feature for Linux systems
1781 The DEVRANDOM_WAIT feature added a select() call to wait for the
1782 /dev/random device to become readable before reading from the
1783 /dev/urandom device.
1785 It turned out that this change had negative side effects on
1786 performance which were not acceptable. After some discussion it
1787 was decided to revert this feature and leave it up to the OS
1788 resp. the platform maintainer to ensure a proper initialization
1789 during early boot time.
1791 *Matthias St. Pierre*
1793 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1795 * Add build tests for C++. These are generated files that only do one
1796 thing, to include one public OpenSSL head file each. This tests that
1797 the public header files can be usefully included in a C++ application.
1799 This test isn't enabled by default. It can be enabled with the option
1800 'enable-buildtest-c++'.
1804 * Enable SHA3 pre-hashing for ECDSA and DSA.
1808 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1809 This changes the size when using the `genpkey` command when no size is given.
1810 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1811 generation commands to use 2048 bits by default.
1815 * Reorganize the manual pages to consistently have RETURN VALUES,
1816 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1817 util/fix-doc-nits accordingly.
1819 *Paul Yang, Joshua Lock*
1821 * Add the missing accessor EVP_PKEY_get0_engine()
1825 * Have commands like `s_client` and `s_server` output the signature scheme
1826 along with other cipher suite parameters when debugging.
1830 * Make OPENSSL_config() error agnostic again.
1834 * Do the error handling in RSA decryption constant time.
1838 * Prevent over long nonces in ChaCha20-Poly1305.
1840 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1841 for every encryption operation. RFC 7539 specifies that the nonce value
1842 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1843 and front pads the nonce with 0 bytes if it is less than 12
1844 bytes. However it also incorrectly allows a nonce to be set of up to 16
1845 bytes. In this case only the last 12 bytes are significant and any
1846 additional leading bytes are ignored.
1848 It is a requirement of using this cipher that nonce values are
1849 unique. Messages encrypted using a reused nonce value are susceptible to
1850 serious confidentiality and integrity attacks. If an application changes
1851 the default nonce length to be longer than 12 bytes and then makes a
1852 change to the leading bytes of the nonce expecting the new value to be a
1853 new unique nonce then such an application could inadvertently encrypt
1854 messages with a reused nonce.
1856 Additionally the ignored bytes in a long nonce are not covered by the
1857 integrity guarantee of this cipher. Any application that relies on the
1858 integrity of these ignored leading bytes of a long nonce may be further
1859 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1860 is safe because no such use sets such a long nonce value. However user
1861 applications that use this cipher directly and set a non-default nonce
1862 length to be longer than 12 bytes may be vulnerable.
1864 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1870 * Add DEVRANDOM_WAIT feature for Linux systems
1872 On older Linux systems where the getrandom() system call is not available,
1873 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1874 Contrary to getrandom(), the /dev/urandom device will not block during
1875 early boot when the kernel CSPRNG has not been seeded yet.
1877 To mitigate this known weakness, use select() to wait for /dev/random to
1878 become readable before reading from /dev/urandom.
1880 * Ensure that SM2 only uses SM3 as digest algorithm
1884 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1886 * Change the info callback signals for the start and end of a post-handshake
1887 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1888 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1889 confused by this and assume that a TLSv1.2 renegotiation has started. This
1890 can break KeyUpdate handling. Instead we no longer signal the start and end
1891 of a post handshake message exchange (although the messages themselves are
1892 still signalled). This could break some applications that were expecting
1893 the old signals. However without this KeyUpdate is not usable for many
1898 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1900 * Timing vulnerability in DSA signature generation
1902 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1903 timing side channel attack. An attacker could use variations in the signing
1904 algorithm to recover the private key.
1906 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1911 * Timing vulnerability in ECDSA signature generation
1913 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1914 timing side channel attack. An attacker could use variations in the signing
1915 algorithm to recover the private key.
1917 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1922 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1923 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1924 of two gigabytes and the error handling improved.
1926 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1927 categorized as a normal bug, not a security issue, because the DRBG reseeds
1928 automatically and is fully functional even without additional randomness
1929 provided by the application.
1931 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1933 * Add a new ClientHello callback. Provides a callback interface that gives
1934 the application the ability to adjust the nascent SSL object at the
1935 earliest stage of ClientHello processing, immediately after extensions have
1936 been collected but before they have been processed. In particular, this
1937 callback can adjust the supported TLS versions in response to the contents
1942 * Add SM2 base algorithm support.
1946 * s390x assembly pack: add (improved) hardware-support for the following
1947 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1948 aes-cfb/cfb8, aes-ecb.
1952 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1953 parameter is no longer accepted, as it leads to a corrupt table. NULL
1954 pem_str is reserved for alias entries only.
1958 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1959 step for prime curves. The new implementation is based on formulae from
1960 differential addition-and-doubling in homogeneous projective coordinates
1961 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1962 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1963 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1964 to work in projective coordinates.
1966 *Billy Bob Brumley, Nicola Tuveri*
1968 * Change generating and checking of primes so that the error rate of not
1969 being prime depends on the intended use based on the size of the input.
1970 For larger primes this will result in more rounds of Miller-Rabin.
1971 The maximal error rate for primes with more than 1080 bits is lowered
1974 *Kurt Roeckx, Annie Yousar*
1976 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1980 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1981 moving between systems, and to avoid confusion when a Windows build is
1982 done with mingw vs with MSVC. For POSIX installs, there's still a
1983 symlink or copy named 'tsget' to avoid that confusion as well.
1987 * Revert blinding in ECDSA sign and instead make problematic addition
1988 length-invariant. Switch even to fixed-length Montgomery multiplication.
1992 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1993 step for binary curves. The new implementation is based on formulae from
1994 differential addition-and-doubling in mixed Lopez-Dahab projective
1995 coordinates, modified to independently blind the operands.
1997 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1999 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2000 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2001 EC_METHODs to implement their own specialized "ladder step", to take
2002 advantage of more favorable coordinate systems or more efficient
2003 differential addition-and-doubling algorithms.
2005 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2007 * Modified the random device based seed sources to keep the relevant
2008 file descriptors open rather than reopening them on each access.
2009 This allows such sources to operate in a chroot() jail without
2010 the associated device nodes being available. This behaviour can be
2011 controlled using RAND_keep_random_devices_open().
2015 * Numerous side-channel attack mitigations have been applied. This may have
2016 performance impacts for some algorithms for the benefit of improved
2017 security. Specific changes are noted in this change log by their respective
2022 * AIX shared library support overhaul. Switch to AIX "natural" way of
2023 handling shared libraries, which means collecting shared objects of
2024 different versions and bitnesses in one common archive. This allows to
2025 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2026 doesn't affect the way 3rd party applications are linked, only how
2027 multi-version installation is managed.
2031 * Make ec_group_do_inverse_ord() more robust and available to other
2032 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2033 mitigations are applied to the fallback BN_mod_inverse().
2034 When using this function rather than BN_mod_inverse() directly, new
2035 EC cryptosystem implementations are then safer-by-default.
2039 * Add coordinate blinding for EC_POINT and implement projective
2040 coordinate blinding for generic prime curves as a countermeasure to
2041 chosen point SCA attacks.
2043 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2045 * Add blinding to ECDSA and DSA signatures to protect against side channel
2046 attacks discovered by Keegan Ryan (NCC Group).
2050 * Enforce checking in the `pkeyutl` command to ensure that the input
2051 length does not exceed the maximum supported digest length when performing
2052 a sign, verify or verifyrecover operation.
2056 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2057 I/O in combination with something like select() or poll() will hang. This
2058 can be turned off again using SSL_CTX_clear_mode().
2059 Many applications do not properly handle non-application data records, and
2060 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2061 around the problems in those applications, but can also break some.
2062 It's recommended to read the manpages about SSL_read(), SSL_write(),
2063 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2064 SSL_CTX_set_read_ahead() again.
2068 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2069 now allow empty (zero character) pass phrases.
2073 * Apply blinding to binary field modular inversion and remove patent
2074 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2078 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2079 binary and prime elliptic curves.
2083 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2084 constant time fixed point multiplication.
2088 * Revise elliptic curve scalar multiplication with timing attack
2089 defenses: ec_wNAF_mul redirects to a constant time implementation
2090 when computing fixed point and variable point multiplication (which
2091 in OpenSSL are mostly used with secret scalars in keygen, sign,
2092 ECDH derive operations).
2093 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2096 * Updated CONTRIBUTING
2100 * Updated DRBG / RAND to request nonce and additional low entropy
2101 randomness from the system.
2103 *Matthias St. Pierre*
2105 * Updated 'openssl rehash' to use OpenSSL consistent default.
2109 * Moved the load of the ssl_conf module to libcrypto, which helps
2110 loading engines that libssl uses before libssl is initialised.
2114 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2118 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2120 *Ingo Schwarze, Rich Salz*
2122 * Added output of accepting IP address and port for 'openssl s_server'
2126 * Added a new API for TLSv1.3 ciphersuites:
2127 SSL_CTX_set_ciphersuites()
2128 SSL_set_ciphersuites()
2132 * Memory allocation failures consistently add an error to the error
2137 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2138 in libcrypto when run as setuid/setgid.
2142 * Load any config file by default when libssl is used.
2146 * Added new public header file <openssl/rand_drbg.h> and documentation
2147 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2149 *Matthias St. Pierre*
2151 * QNX support removed (cannot find contributors to get their approval
2152 for the license change).
2156 * TLSv1.3 replay protection for early data has been implemented. See the
2157 SSL_read_early_data() man page for further details.
2161 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2162 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2163 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2164 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2165 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2166 configuration has been separated out. See the ciphers man page or the
2167 SSL_CTX_set_ciphersuites() man page for more information.
2171 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2172 in responder mode now supports the new "-multi" option, which
2173 spawns the specified number of child processes to handle OCSP
2174 requests. The "-timeout" option now also limits the OCSP
2175 responder's patience to wait to receive the full client request
2176 on a newly accepted connection. Child processes are respawned
2177 as needed, and the CA index file is automatically reloaded
2178 when changed. This makes it possible to run the "ocsp" responder
2179 as a long-running service, making the OpenSSL CA somewhat more
2180 feature-complete. In this mode, most diagnostic messages logged
2181 after entering the event loop are logged via syslog(3) rather than
2186 * Added support for X448 and Ed448. Heavily based on original work by
2191 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2192 objects loaded. This adds the functions OSSL_STORE_expect() and
2193 OSSL_STORE_find() as well as needed tools to construct searches and
2194 get the search data out of them.
2198 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2199 version of OpenSSL should review their configuration settings to ensure
2200 that they are still appropriate for TLSv1.3. For further information see:
2201 <https://wiki.openssl.org/index.php/TLS1.3>
2205 * Grand redesign of the OpenSSL random generator
2207 The default RAND method now utilizes an AES-CTR DRBG according to
2208 NIST standard SP 800-90Ar1. The new random generator is essentially
2209 a port of the default random generator from the OpenSSL FIPS 2.0
2210 object module. It is a hybrid deterministic random bit generator
2211 using an AES-CTR bit stream and which seeds and reseeds itself
2212 automatically using trusted system entropy sources.
2214 Some of its new features are:
2215 - Support for multiple DRBG instances with seed chaining.
2216 - The default RAND method makes use of a DRBG.
2217 - There is a public and private DRBG instance.
2218 - The DRBG instances are fork-safe.
2219 - Keep all global DRBG instances on the secure heap if it is enabled.
2220 - The public and private DRBG instance are per thread for lock free
2223 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2225 * Changed Configure so it only says what it does and doesn't dump
2226 so much data. Instead, ./configdata.pm should be used as a script
2227 to display all sorts of configuration data.
2231 * Added processing of "make variables" to Configure.
2235 * Added SHA512/224 and SHA512/256 algorithm support.
2239 * The last traces of Netware support, first removed in 1.1.0, have
2244 * Get rid of Makefile.shared, and in the process, make the processing
2245 of certain files (rc.obj, or the .def/.map/.opt files produced from
2246 the ordinal files) more visible and hopefully easier to trace and
2247 debug (or make silent).
2251 * Make it possible to have environment variable assignments as
2252 arguments to config / Configure.
2256 * Add multi-prime RSA (RFC 8017) support.
2260 * Add SM3 implemented according to GB/T 32905-2016
2261 *Jack Lloyd <jack.lloyd@ribose.com>,*
2262 *Ronald Tse <ronald.tse@ribose.com>,*
2263 *Erick Borsboom <erick.borsboom@ribose.com>*
2265 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2266 as documented in RFC6066.
2267 Based on a patch from Tomasz Moń
2269 *Filipe Raimundo da Silva*
2271 * Add SM4 implemented according to GB/T 32907-2016.
2272 *Jack Lloyd <jack.lloyd@ribose.com>,*
2273 *Ronald Tse <ronald.tse@ribose.com>,*
2274 *Erick Borsboom <erick.borsboom@ribose.com>*
2276 * Reimplement -newreq-nodes and ERR_error_string_n; the
2277 original author does not agree with the license change.
2281 * Add ARIA AEAD TLS support.
2285 * Some macro definitions to support VS6 have been removed. Visual
2286 Studio 6 has not worked since 1.1.0
2290 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2291 without clearing the errors.
2295 * Add "atfork" functions. If building on a system that without
2296 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2297 requirements. The RAND facility now uses/requires this.
2305 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2306 not possible to disable entirely. However, it's still possible to
2307 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2310 To disable, configure with 'no-ui-console'. 'no-ui' is still
2311 possible to use as an alias. Check at compile time with the
2312 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2313 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2317 * Add a STORE module, which implements a uniform and URI based reader of
2318 stores that can contain keys, certificates, CRLs and numerous other
2319 objects. The main API is loosely based on a few stdio functions,
2320 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2321 OSSL_STORE_error and OSSL_STORE_close.
2322 The implementation uses backends called "loaders" to implement arbitrary
2323 URI schemes. There is one built in "loader" for the 'file' scheme.
2327 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2328 then adjusted to work on FreeBSD 8.4 as well.
2329 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2330 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2334 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2335 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2336 error code calls like this:
2338 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2340 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2341 that can be encoded in C. For the foreseeable future, this will only
2344 *Richard Levitte and Tim Hudson*
2346 * Removed BSD cryptodev engine.
2350 * Add a build target 'build_all_generated', to build all generated files
2351 and only that. This can be used to prepare everything that requires
2352 things like perl for a system that lacks perl and then move everything
2353 to that system and do the rest of the build there.
2357 * In the UI interface, make it possible to duplicate the user data. This
2358 can be used by engines that need to retain the data for a longer time
2359 than just the call where this user data is passed.
2363 * Ignore the '-named_curve auto' value for compatibility of applications
2366 *Tomáš Mráz <tmraz@fedoraproject.org>*
2368 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2369 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2370 alerts across multiple records (some of which could be empty). In practice
2371 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2372 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2373 support this at all. Supporting it adds significant complexity to the
2374 record layer, and its removal is unlikely to cause interoperability
2379 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2380 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2381 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2386 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2387 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2389 *Richard Levitte, Andy Polyakov*
2391 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2396 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2397 platform rather than 'mingw'.
2401 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2402 success if they are asked to add an object which already exists
2403 in the store. This change cascades to other functions which load
2404 certificates and CRLs.
2408 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2409 facilitate stack unwinding even from assembly subroutines.
2413 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2414 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2418 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2419 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2420 which is the minimum version we support.
2424 * Certificate time validation (X509_cmp_time) enforces stricter
2425 compliance with RFC 5280. Fractional seconds and timezone offsets
2426 are no longer allowed.
2430 * Add support for ARIA
2434 * s_client will now send the Server Name Indication (SNI) extension by
2435 default unless the new "-noservername" option is used. The server name is
2436 based on the host provided to the "-connect" option unless overridden by
2437 using "-servername".
2441 * Add support for SipHash
2445 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2446 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2447 prevent issues where no progress is being made and the peer continually
2448 sends unrecognised record types, using up resources processing them.
2452 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2453 using the algorithm defined in
2454 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2458 * Heartbeat support has been removed; the ABI is changed for now.
2460 *Richard Levitte, Rich Salz*
2462 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2466 * The RSA "null" method, which was partially supported to avoid patent
2467 issues, has been replaced to always returns NULL.
2474 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2476 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2477 used even when parsing explicit parameters, when loading a encoded key
2478 or calling `EC_GROUP_new_from_ecpkparameters()`/
2479 `EC_GROUP_new_from_ecparameters()`.
2480 This prevents bypass of security hardening and performance gains,
2481 especially for curves with specialized EC_METHODs.
2482 By default, if a key encoded with explicit parameters is loaded and later
2483 encoded, the output is still encoded with explicit parameters, even if
2484 internally a "named" EC_GROUP is used for computation.
2488 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2489 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2490 NULL. After this change, only the cofactor parameter can be NULL. It also
2491 does some minimal sanity checks on the passed order.
2496 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2497 An attack is simple, if the first CMS_recipientInfo is valid but the
2498 second CMS_recipientInfo is chosen ciphertext. If the second
2499 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2500 encryption key will be replaced by garbage, and the message cannot be
2501 decoded, but if the RSA decryption fails, the correct encryption key is
2502 used and the recipient will not notice the attack.
2503 As a work around for this potential attack the length of the decrypted
2504 key must be equal to the cipher default key length, in case the
2505 certifiate is not given and all recipientInfo are tried out.
2506 The old behaviour can be re-enabled in the CMS code by setting the
2507 CMS_DEBUG_DECRYPT flag.
2512 * Use Windows installation paths in the mingw builds
2514 Mingw isn't a POSIX environment per se, which means that Windows
2515 paths should be used for installation.
2520 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2522 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2523 This changes the size when using the `genpkey` command when no size is given.
2524 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2525 generation commands to use 2048 bits by default.
2529 * Prevent over long nonces in ChaCha20-Poly1305.
2531 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2532 for every encryption operation. RFC 7539 specifies that the nonce value
2533 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2534 and front pads the nonce with 0 bytes if it is less than 12
2535 bytes. However it also incorrectly allows a nonce to be set of up to 16
2536 bytes. In this case only the last 12 bytes are significant and any
2537 additional leading bytes are ignored.
2539 It is a requirement of using this cipher that nonce values are
2540 unique. Messages encrypted using a reused nonce value are susceptible to
2541 serious confidentiality and integrity attacks. If an application changes
2542 the default nonce length to be longer than 12 bytes and then makes a
2543 change to the leading bytes of the nonce expecting the new value to be a
2544 new unique nonce then such an application could inadvertently encrypt
2545 messages with a reused nonce.
2547 Additionally the ignored bytes in a long nonce are not covered by the
2548 integrity guarantee of this cipher. Any application that relies on the
2549 integrity of these ignored leading bytes of a long nonce may be further
2550 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2551 is safe because no such use sets such a long nonce value. However user
2552 applications that use this cipher directly and set a non-default nonce
2553 length to be longer than 12 bytes may be vulnerable.
2555 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2561 * Added SCA hardening for modular field inversion in EC_GROUP through
2562 a new dedicated field_inv() pointer in EC_METHOD.
2563 This also addresses a leakage affecting conversions from projective
2564 to affine coordinates.
2566 *Billy Bob Brumley, Nicola Tuveri*
2568 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2569 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2573 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2577 * Remove the 'dist' target and add a tarball building script. The
2578 'dist' target has fallen out of use, and it shouldn't be
2579 necessary to configure just to create a source distribution.
2583 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2585 * Timing vulnerability in DSA signature generation
2587 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2588 timing side channel attack. An attacker could use variations in the signing
2589 algorithm to recover the private key.
2591 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2596 * Timing vulnerability in ECDSA signature generation
2598 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2599 timing side channel attack. An attacker could use variations in the signing
2600 algorithm to recover the private key.
2602 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2607 * Add coordinate blinding for EC_POINT and implement projective
2608 coordinate blinding for generic prime curves as a countermeasure to
2609 chosen point SCA attacks.
2611 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2613 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2615 * Client DoS due to large DH parameter
2617 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2618 malicious server can send a very large prime value to the client. This will
2619 cause the client to spend an unreasonably long period of time generating a
2620 key for this prime resulting in a hang until the client has finished. This
2621 could be exploited in a Denial Of Service attack.
2623 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2628 * Cache timing vulnerability in RSA Key Generation
2630 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2631 a cache timing side channel attack. An attacker with sufficient access to
2632 mount cache timing attacks during the RSA key generation process could
2633 recover the private key.
2635 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2636 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2641 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2642 parameter is no longer accepted, as it leads to a corrupt table. NULL
2643 pem_str is reserved for alias entries only.
2647 * Revert blinding in ECDSA sign and instead make problematic addition
2648 length-invariant. Switch even to fixed-length Montgomery multiplication.
2652 * Change generating and checking of primes so that the error rate of not
2653 being prime depends on the intended use based on the size of the input.
2654 For larger primes this will result in more rounds of Miller-Rabin.
2655 The maximal error rate for primes with more than 1080 bits is lowered
2658 *Kurt Roeckx, Annie Yousar*
2660 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2664 * Add blinding to ECDSA and DSA signatures to protect against side channel
2665 attacks discovered by Keegan Ryan (NCC Group).
2669 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2670 now allow empty (zero character) pass phrases.
2674 * Certificate time validation (X509_cmp_time) enforces stricter
2675 compliance with RFC 5280. Fractional seconds and timezone offsets
2676 are no longer allowed.
2680 * Fixed a text canonicalisation bug in CMS
2682 Where a CMS detached signature is used with text content the text goes
2683 through a canonicalisation process first prior to signing or verifying a
2684 signature. This process strips trailing space at the end of lines, converts
2685 line terminators to CRLF and removes additional trailing line terminators
2686 at the end of a file. A bug in the canonicalisation process meant that
2687 some characters, such as form-feed, were incorrectly treated as whitespace
2688 and removed. This is contrary to the specification (RFC5485). This fix
2689 could mean that detached text data signed with an earlier version of
2690 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2691 signed with a fixed OpenSSL may fail to verify with an earlier version of
2692 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2693 and use the "-binary" flag (for the "cms" command line application) or set
2694 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2698 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2700 * Constructed ASN.1 types with a recursive definition could exceed the stack
2702 Constructed ASN.1 types with a recursive definition (such as can be found
2703 in PKCS7) could eventually exceed the stack given malicious input with
2704 excessive recursion. This could result in a Denial Of Service attack. There
2705 are no such structures used within SSL/TLS that come from untrusted sources
2706 so this is considered safe.
2708 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2714 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2716 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2717 effectively reduced to only comparing the least significant bit of each
2718 byte. This allows an attacker to forge messages that would be considered as
2719 authenticated in an amount of tries lower than that guaranteed by the
2720 security claims of the scheme. The module can only be compiled by the
2721 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2723 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2729 * Add a build target 'build_all_generated', to build all generated files
2730 and only that. This can be used to prepare everything that requires
2731 things like perl for a system that lacks perl and then move everything
2732 to that system and do the rest of the build there.
2736 * Backport SSL_OP_NO_RENGOTIATION
2738 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2739 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2740 changes this is no longer possible in 1.1.0. Therefore the new
2741 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2742 1.1.0 to provide equivalent functionality.
2744 Note that if an application built against 1.1.0h headers (or above) is run
2745 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2746 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2750 * Removed the OS390-Unix config target. It relied on a script that doesn't
2755 * rsaz_1024_mul_avx2 overflow bug on x86_64
2757 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2758 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2759 Analysis suggests that attacks against RSA and DSA as a result of this
2760 defect would be very difficult to perform and are not believed likely.
2761 Attacks against DH1024 are considered just feasible, because most of the
2762 work necessary to deduce information about a private key may be performed
2763 offline. The amount of resources required for such an attack would be
2764 significant. However, for an attack on TLS to be meaningful, the server
2765 would have to share the DH1024 private key among multiple clients, which is
2766 no longer an option since CVE-2016-0701.
2768 This only affects processors that support the AVX2 but not ADX extensions
2769 like Intel Haswell (4th generation).
2771 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2772 was originally found via the OSS-Fuzz project.
2777 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2779 * bn_sqrx8x_internal carry bug on x86_64
2781 There is a carry propagating bug in the x86_64 Montgomery squaring
2782 procedure. No EC algorithms are affected. Analysis suggests that attacks
2783 against RSA and DSA as a result of this defect would be very difficult to
2784 perform and are not believed likely. Attacks against DH are considered just
2785 feasible (although very difficult) because most of the work necessary to
2786 deduce information about a private key may be performed offline. The amount
2787 of resources required for such an attack would be very significant and
2788 likely only accessible to a limited number of attackers. An attacker would
2789 additionally need online access to an unpatched system using the target
2790 private key in a scenario with persistent DH parameters and a private
2791 key that is shared between multiple clients.
2793 This only affects processors that support the BMI1, BMI2 and ADX extensions
2794 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2796 This issue was reported to OpenSSL by the OSS-Fuzz project.
2801 * Malformed X.509 IPAddressFamily could cause OOB read
2803 If an X.509 certificate has a malformed IPAddressFamily extension,
2804 OpenSSL could do a one-byte buffer overread. The most likely result
2805 would be an erroneous display of the certificate in text format.
2807 This issue was reported to OpenSSL by the OSS-Fuzz project.
2812 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2814 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2815 platform rather than 'mingw'.
2819 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2820 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2821 which is the minimum version we support.
2825 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2827 * Encrypt-Then-Mac renegotiation crash
2829 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2830 negotiated where it was not in the original handshake (or vice-versa) then
2831 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2832 and servers are affected.
2834 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2839 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2841 * Truncated packet could crash via OOB read
2843 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2844 cipher is being used, then a truncated packet can cause that host to
2845 perform an out-of-bounds read, usually resulting in a crash.
2847 This issue was reported to OpenSSL by Robert Święcki of Google.
2852 * Bad (EC)DHE parameters cause a client crash
2854 If a malicious server supplies bad parameters for a DHE or ECDHE key
2855 exchange then this can result in the client attempting to dereference a
2856 NULL pointer leading to a client crash. This could be exploited in a Denial
2859 This issue was reported to OpenSSL by Guido Vranken.
2864 * BN_mod_exp may produce incorrect results on x86_64
2866 There is a carry propagating bug in the x86_64 Montgomery squaring
2867 procedure. No EC algorithms are affected. Analysis suggests that attacks
2868 against RSA and DSA as a result of this defect would be very difficult to
2869 perform and are not believed likely. Attacks against DH are considered just
2870 feasible (although very difficult) because most of the work necessary to
2871 deduce information about a private key may be performed offline. The amount
2872 of resources required for such an attack would be very significant and
2873 likely only accessible to a limited number of attackers. An attacker would
2874 additionally need online access to an unpatched system using the target
2875 private key in a scenario with persistent DH parameters and a private
2876 key that is shared between multiple clients. For example this can occur by
2877 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2878 similar to CVE-2015-3193 but must be treated as a separate problem.
2880 This issue was reported to OpenSSL by the OSS-Fuzz project.
2885 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2887 * ChaCha20/Poly1305 heap-buffer-overflow
2889 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2890 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2891 crash. This issue is not considered to be exploitable beyond a DoS.
2893 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2898 * CMS Null dereference
2900 Applications parsing invalid CMS structures can crash with a NULL pointer
2901 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2902 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2903 structure callback if an attempt is made to free certain invalid encodings.
2904 Only CHOICE structures using a callback which do not handle NULL value are
2907 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2912 * Montgomery multiplication may produce incorrect results
2914 There is a carry propagating bug in the Broadwell-specific Montgomery
2915 multiplication procedure that handles input lengths divisible by, but
2916 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2917 and DH private keys are impossible. This is because the subroutine in
2918 question is not used in operations with the private key itself and an input
2919 of the attacker's direct choice. Otherwise the bug can manifest itself as
2920 transient authentication and key negotiation failures or reproducible
2921 erroneous outcome of public-key operations with specially crafted input.
2922 Among EC algorithms only Brainpool P-512 curves are affected and one
2923 presumably can attack ECDH key negotiation. Impact was not analyzed in
2924 detail, because pre-requisites for attack are considered unlikely. Namely
2925 multiple clients have to choose the curve in question and the server has to
2926 share the private key among them, neither of which is default behaviour.
2927 Even then only clients that chose the curve will be affected.
2929 This issue was publicly reported as transient failures and was not
2930 initially recognized as a security issue. Thanks to Richard Morgan for
2931 providing reproducible case.
2936 * Removed automatic addition of RPATH in shared libraries and executables,
2937 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2941 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2943 * Fix Use After Free for large message sizes
2945 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2946 message larger than approx 16k is received then the underlying buffer to
2947 store the incoming message is reallocated and moved. Unfortunately a
2948 dangling pointer to the old location is left which results in an attempt to
2949 write to the previously freed location. This is likely to result in a
2950 crash, however it could potentially lead to execution of arbitrary code.
2952 This issue only affects OpenSSL 1.1.0a.
2954 This issue was reported to OpenSSL by Robert Święcki.
2959 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2961 * OCSP Status Request extension unbounded memory growth
2963 A malicious client can send an excessively large OCSP Status Request
2964 extension. If that client continually requests renegotiation, sending a
2965 large OCSP Status Request extension each time, then there will be unbounded
2966 memory growth on the server. This will eventually lead to a Denial Of
2967 Service attack through memory exhaustion. Servers with a default
2968 configuration are vulnerable even if they do not support OCSP. Builds using
2969 the "no-ocsp" build time option are not affected.
2971 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2976 * SSL_peek() hang on empty record
2978 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2979 sends an empty record. This could be exploited by a malicious peer in a
2980 Denial Of Service attack.
2982 This issue was reported to OpenSSL by Alex Gaynor.
2987 * Excessive allocation of memory in tls_get_message_header() and
2988 dtls1_preprocess_fragment()
2990 A (D)TLS message includes 3 bytes for its length in the header for the
2991 message. This would allow for messages up to 16Mb in length. Messages of
2992 this length are excessive and OpenSSL includes a check to ensure that a
2993 peer is sending reasonably sized messages in order to avoid too much memory
2994 being consumed to service a connection. A flaw in the logic of version
2995 1.1.0 means that memory for the message is allocated too early, prior to
2996 the excessive message length check. Due to way memory is allocated in
2997 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2998 to service a connection. This could lead to a Denial of Service through
2999 memory exhaustion. However, the excessive message length check still takes
3000 place, and this would cause the connection to immediately fail. Assuming
3001 that the application calls SSL_free() on the failed connection in a timely
3002 manner then the 21Mb of allocated memory will then be immediately freed
3003 again. Therefore the excessive memory allocation will be transitory in
3004 nature. This then means that there is only a security impact if:
3006 1) The application does not call SSL_free() in a timely manner in the event
3007 that the connection fails
3009 2) The application is working in a constrained environment where there is
3010 very little free memory
3012 3) The attacker initiates multiple connection attempts such that there are
3013 multiple connections in a state where memory has been allocated for the
3014 connection; SSL_free() has not yet been called; and there is insufficient
3015 memory to service the multiple requests.
3017 Except in the instance of (1) above any Denial Of Service is likely to be
3018 transitory because as soon as the connection fails the memory is
3019 subsequently freed again in the SSL_free() call. However there is an
3020 increased risk during this period of application crashes due to the lack of
3021 memory - which would then mean a more serious Denial of Service.
3023 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3024 (CVE-2016-6307 and CVE-2016-6308)
3028 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3029 had to be removed. Primary reason is that vendor assembler can't
3030 assemble our modules with -KPIC flag. As result it, assembly
3031 support, was not even available as option. But its lack means
3032 lack of side-channel resistant code, which is incompatible with
3033 security by todays standards. Fortunately gcc is readily available
3034 prepackaged option, which we firmly point at...
3038 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3040 * Windows command-line tool supports UTF-8 opt-in option for arguments
3041 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3042 (to any value) allows Windows user to access PKCS#12 file generated
3043 with Windows CryptoAPI and protected with non-ASCII password, as well
3044 as files generated under UTF-8 locale on Linux also protected with
3049 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3050 have been disabled by default and removed from DEFAULT, just like RC4.
3051 See the RC4 item below to re-enable both.
3055 * The method for finding the storage location for the Windows RAND seed file
3056 has changed. First we check %RANDFILE%. If that is not set then we check
3057 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3058 all else fails we fall back to C:\.
3062 * The EVP_EncryptUpdate() function has had its return type changed from void
3063 to int. A return of 0 indicates and error while a return of 1 indicates
3068 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3069 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3070 off the constant time implementation for RSA, DSA and DH have been made
3071 no-ops and deprecated.
3075 * Windows RAND implementation was simplified to only get entropy by
3076 calling CryptGenRandom(). Various other RAND-related tickets
3079 *Joseph Wylie Yandle, Rich Salz*
3081 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3082 and `OPENSSL_LH_`, respectively. The old names are available
3083 with API compatibility. They new names are now completely documented.
3087 * Unify TYPE_up_ref(obj) methods signature.
3088 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3089 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3090 int (instead of void) like all others TYPE_up_ref() methods.
3091 So now these methods also check the return value of CRYPTO_atomic_add(),
3092 and the validity of object reference counter.
3094 *fdasilvayy@gmail.com*
3096 * With Windows Visual Studio builds, the .pdb files are installed
3097 alongside the installed libraries and executables. For a static
3098 library installation, ossl_static.pdb is the associate compiler
3099 generated .pdb file to be used when linking programs.
3103 * Remove openssl.spec. Packaging files belong with the packagers.
3107 * Automatic Darwin/OSX configuration has had a refresh, it will now
3108 recognise x86_64 architectures automatically. You can still decide
3109 to build for a different bitness with the environment variable
3110 KERNEL_BITS (can be 32 or 64), for example:
3112 KERNEL_BITS=32 ./config
3116 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3117 256 bit AES and HMAC with SHA256.
3121 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3125 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3129 * To enable users to have their own config files and build file templates,
3130 Configure looks in the directory indicated by the environment variable
3131 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3132 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3133 name and is used as is.
3137 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3138 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3139 X509_CERT_FILE_CTX was removed.
3143 * "shared" builds are now the default. To create only static libraries use
3144 the "no-shared" Configure option.
3148 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3149 All of these option have not worked for some while and are fundamental
3154 * Make various cleanup routines no-ops and mark them as deprecated. Most
3155 global cleanup functions are no longer required because they are handled
3156 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3157 Explicitly de-initing can cause problems (e.g. where a library that uses
3158 OpenSSL de-inits, but an application is still using it). The affected
3159 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3160 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3161 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3162 COMP_zlib_cleanup().
3166 * --strict-warnings no longer enables runtime debugging options
3167 such as REF_DEBUG. Instead, debug options are automatically
3168 enabled with '--debug' builds.
3170 *Andy Polyakov, Emilia Käsper*
3172 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3173 have been moved out of the public header files. New functions for managing
3174 these have been added.
3178 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3179 objects have been moved out of the public header files. New
3180 functions for managing these have been added.
3184 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3185 have been moved out of the public header files. New functions for managing
3186 these have been added.
3190 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3191 moved out of the public header files. New functions for managing these
3196 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3200 * Removed the mk1mf build scripts.
3204 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3205 it is always safe to #include a header now.
3209 * Removed the aged BC-32 config and all its supporting scripts
3213 * Removed support for Ultrix, Netware, and OS/2.
3217 * Add support for HKDF.
3219 *Alessandro Ghedini*
3221 * Add support for blake2b and blake2s
3225 * Added support for "pipelining". Ciphers that have the
3226 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3227 encryptions/decryptions simultaneously. There are currently no built-in
3228 ciphers with this property but the expectation is that engines will be able
3229 to offer it to significantly improve throughput. Support has been extended
3230 into libssl so that multiple records for a single connection can be
3231 processed in one go (for >=TLS 1.1).
3235 * Added the AFALG engine. This is an async capable engine which is able to
3236 offload work to the Linux kernel. In this initial version it only supports
3237 AES128-CBC. The kernel must be version 4.1.0 or greater.
3241 * OpenSSL now uses a new threading API. It is no longer necessary to
3242 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3243 are two supported threading models: pthreads and windows threads. It is
3244 also possible to configure OpenSSL at compile time for "no-threads". The
3245 old threading API should no longer be used. The functions have been
3246 replaced with "no-op" compatibility macros.
3248 *Alessandro Ghedini, Matt Caswell*
3250 * Modify behavior of ALPN to invoke callback after SNI/servername
3251 callback, such that updates to the SSL_CTX affect ALPN.
3255 * Add SSL_CIPHER queries for authentication and key-exchange.
3259 * Changes to the DEFAULT cipherlist:
3260 - Prefer (EC)DHE handshakes over plain RSA.
3261 - Prefer AEAD ciphers over legacy ciphers.
3262 - Prefer ECDSA over RSA when both certificates are available.
3263 - Prefer TLSv1.2 ciphers/PRF.
3264 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3269 * Change the ECC default curve list to be this, in order: x25519,
3270 secp256r1, secp521r1, secp384r1.
3274 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3275 disabled by default. They can be re-enabled using the
3276 enable-weak-ssl-ciphers option to Configure.
3280 * If the server has ALPN configured, but supports no protocols that the
3281 client advertises, send a fatal "no_application_protocol" alert.
3282 This behaviour is SHALL in RFC 7301, though it isn't universally
3283 implemented by other servers.
3287 * Add X25519 support.
3288 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3289 for public and private key encoding using the format documented in
3290 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3291 key generation and key derivation.
3293 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3298 * Deprecate SRP_VBASE_get_by_user.
3299 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3300 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3301 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3302 seed, even if the seed is configured.
3304 Users should use SRP_VBASE_get1_by_user instead. Note that in
3305 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3306 also that even though configuring the SRP seed attempts to hide
3307 invalid usernames by continuing the handshake with fake
3308 credentials, this behaviour is not constant time and no strong
3309 guarantees are made that the handshake is indistinguishable from
3310 that of a valid user.
3314 * Configuration change; it's now possible to build dynamic engines
3315 without having to build shared libraries and vice versa. This
3316 only applies to the engines in `engines/`, those in `crypto/engine/`
3317 will always be built into libcrypto (i.e. "static").
3319 Building dynamic engines is enabled by default; to disable, use
3320 the configuration option "disable-dynamic-engine".
3322 The only requirements for building dynamic engines are the
3323 presence of the DSO module and building with position independent
3324 code, so they will also automatically be disabled if configuring
3325 with "disable-dso" or "disable-pic".
3327 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3328 are also taken away from openssl/opensslconf.h, as they are
3333 * Configuration change; if there is a known flag to compile
3334 position independent code, it will always be applied on the
3335 libcrypto and libssl object files, and never on the application
3336 object files. This means other libraries that use routines from
3337 libcrypto / libssl can be made into shared libraries regardless
3338 of how OpenSSL was configured.
3340 If this isn't desirable, the configuration options "disable-pic"
3341 or "no-pic" can be used to disable the use of PIC. This will
3342 also disable building shared libraries and dynamic engines.
3346 * Removed JPAKE code. It was experimental and has no wide use.
3350 * The INSTALL_PREFIX Makefile variable has been renamed to
3351 DESTDIR. That makes for less confusion on what this variable
3352 is for. Also, the configuration option --install_prefix is
3357 * Heartbeat for TLS has been removed and is disabled by default
3358 for DTLS; configure with enable-heartbeats. Code that uses the
3359 old #define's might need to be updated.
3361 *Emilia Käsper, Rich Salz*
3363 * Rename REF_CHECK to REF_DEBUG.
3367 * New "unified" build system
3369 The "unified" build system is aimed to be a common system for all
3370 platforms we support. With it comes new support for VMS.
3372 This system builds supports building in a different directory tree
3373 than the source tree. It produces one Makefile (for unix family
3374 or lookalikes), or one descrip.mms (for VMS).
3376 The source of information to make the Makefile / descrip.mms is
3377 small files called 'build.info', holding the necessary
3378 information for each directory with source to compile, and a
3379 template in Configurations, like unix-Makefile.tmpl or
3382 With this change, the library names were also renamed on Windows
3383 and on VMS. They now have names that are closer to the standard
3384 on Unix, and include the major version number, and in certain
3385 cases, the architecture they are built for. See "Notes on shared
3386 libraries" in INSTALL.
3388 We rely heavily on the perl module Text::Template.
3392 * Added support for auto-initialisation and de-initialisation of the library.
3393 OpenSSL no longer requires explicit init or deinit routines to be called,
3394 except in certain circumstances. See the OPENSSL_init_crypto() and
3395 OPENSSL_init_ssl() man pages for further information.
3399 * The arguments to the DTLSv1_listen function have changed. Specifically the
3400 "peer" argument is now expected to be a BIO_ADDR object.
3402 * Rewrite of BIO networking library. The BIO library lacked consistent
3403 support of IPv6, and adding it required some more extensive
3404 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3405 which hold all types of addresses and chains of address information.
3406 It also introduces a new API, with functions like BIO_socket,
3407 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3408 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3409 have been adapted accordingly.
3413 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3418 * CRIME protection: disable compression by default, even if OpenSSL is
3419 compiled with zlib enabled. Applications can still enable compression
3420 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3421 using the SSL_CONF library to configure compression.
3425 * The signature of the session callback configured with
3426 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3427 was explicitly marked as `const unsigned char*` instead of
3432 * Always DPURIFY. Remove the use of uninitialized memory in the
3433 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3437 * Removed many obsolete configuration items, including
3438 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3439 MD2_CHAR, MD2_INT, MD2_LONG
3441 IDEA_SHORT, IDEA_LONG
3442 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3444 *Rich Salz, with advice from Andy Polyakov*
3446 * Many BN internals have been moved to an internal header file.
3448 *Rich Salz with help from Andy Polyakov*
3450 * Configuration and writing out the results from it has changed.
3451 Files such as Makefile include/openssl/opensslconf.h and are now
3452 produced through general templates, such as Makefile.in and
3453 crypto/opensslconf.h.in and some help from the perl module
3456 Also, the center of configuration information is no longer
3457 Makefile. Instead, Configure produces a perl module in
3458 configdata.pm which holds most of the config data (in the hash
3459 table %config), the target data that comes from the target
3460 configuration in one of the `Configurations/*.conf` files (in
3465 * To clarify their intended purposes, the Configure options
3466 --prefix and --openssldir change their semantics, and become more
3467 straightforward and less interdependent.
3469 --prefix shall be used exclusively to give the location INSTALLTOP
3470 where programs, scripts, libraries, include files and manuals are
3471 going to be installed. The default is now /usr/local.
3473 --openssldir shall be used exclusively to give the default
3474 location OPENSSLDIR where certificates, private keys, CRLs are
3475 managed. This is also where the default openssl.cnf gets
3477 If the directory given with this option is a relative path, the
3478 values of both the --prefix value and the --openssldir value will
3479 be combined to become OPENSSLDIR.
3480 The default for --openssldir is INSTALLTOP/ssl.
3482 Anyone who uses --openssldir to specify where OpenSSL is to be
3483 installed MUST change to use --prefix instead.
3487 * The GOST engine was out of date and therefore it has been removed. An up
3488 to date GOST engine is now being maintained in an external repository.
3489 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3490 support for GOST ciphersuites (these are only activated if a GOST engine
3495 * EGD is no longer supported by default; use enable-egd when
3498 *Ben Kaduk and Rich Salz*
3500 * The distribution now has Makefile.in files, which are used to
3501 create Makefile's when Configure is run. *Configure must be run
3502 before trying to build now.*
3506 * The return value for SSL_CIPHER_description() for error conditions
3511 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3513 Obtaining and performing DNSSEC validation of TLSA records is
3514 the application's responsibility. The application provides
3515 the TLSA records of its choice to OpenSSL, and these are then
3516 used to authenticate the peer.
3518 The TLSA records need not even come from DNS. They can, for
3519 example, be used to implement local end-entity certificate or
3520 trust-anchor "pinning", where the "pin" data takes the form
3521 of TLSA records, which can augment or replace verification
3522 based on the usual WebPKI public certification authorities.
3526 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3527 continues to support deprecated interfaces in default builds.
3528 However, applications are strongly advised to compile their
3529 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3530 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3531 or the 1.1.0 releases.
3533 In environments in which all applications have been ported to
3534 not use any deprecated interfaces OpenSSL's Configure script
3535 should be used with the --api=1.1.0 option to entirely remove
3536 support for the deprecated features from the library and
3537 unconditionally disable them in the installed headers.
3538 Essentially the same effect can be achieved with the "no-deprecated"
3539 argument to Configure, except that this will always restrict
3540 the build to just the latest API, rather than a fixed API
3543 As applications are ported to future revisions of the API,
3544 they should update their compile-time OPENSSL_API_COMPAT define
3545 accordingly, but in most cases should be able to continue to
3546 compile with later releases.
3548 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3549 0x10000000L and 0x00908000L, respectively. However those
3550 versions did not support the OPENSSL_API_COMPAT feature, and
3551 so applications are not typically tested for explicit support
3552 of just the undeprecated features of either release.
3556 * Add support for setting the minimum and maximum supported protocol.
3557 It can bet set via the SSL_set_min_proto_version() and
3558 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3559 MaxProtocol. It's recommended to use the new APIs to disable
3560 protocols instead of disabling individual protocols using
3561 SSL_set_options() or SSL_CONF's Protocol. This change also
3562 removes support for disabling TLS 1.2 in the OpenSSL TLS
3563 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3567 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3571 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3572 and integrates ECDSA and ECDH functionality into EC. Implementations can
3573 now redirect key generation and no longer need to convert to or from
3576 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3577 include the ec.h header file instead.
3581 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3582 ciphers who are no longer supported and drops support the ephemeral RSA key
3583 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3587 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3588 opaque. For HMAC_CTX, the following constructors and destructors
3591 HMAC_CTX *HMAC_CTX_new(void);
3592 void HMAC_CTX_free(HMAC_CTX *ctx);
3594 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3595 destroy such methods has been added. See EVP_MD_meth_new(3) and
3596 EVP_CIPHER_meth_new(3) for documentation.
3599 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3600 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3601 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3602 an already created structure.
3603 2) For consistency with the majority of our object creators and
3604 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3605 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3606 for deprecated builds.
3610 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3611 cryptographic operations to be performed asynchronously as long as an
3612 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3613 further details. Libssl has also had this capability integrated with the
3614 introduction of the new mode SSL_MODE_ASYNC and associated error
3615 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3616 pages. This work was developed in partnership with Intel Corp.
3620 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3621 always enabled now. If you want to disable the support you should
3622 exclude it using the list of supported ciphers. This also means that the
3623 "-no_ecdhe" option has been removed from s_server.
3627 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3628 SSL_{CTX_}set1_curves() which can set a list.
3632 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3633 curve you want to support using SSL_{CTX_}set1_curves().
3637 * State machine rewrite. The state machine code has been significantly
3638 refactored in order to remove much duplication of code and solve issues
3639 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3640 further details). This change does have some associated API changes.
3641 Notably the SSL_state() function has been removed and replaced by
3642 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3643 SSL_set_state() has been removed altogether. The previous handshake states
3644 defined in ssl.h and ssl3.h have also been removed.
3648 * All instances of the string "ssleay" in the public API were replaced
3649 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3650 Some error codes related to internal RSA_eay API's were renamed.
3654 * The demo files in crypto/threads were moved to demo/threads.
3658 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3661 *Matt Caswell, Rich Salz*
3663 * New ASN.1 embed macro.
3665 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3666 structure is not allocated: it is part of the parent. That is instead of
3674 This reduces memory fragmentation and make it impossible to accidentally
3675 set a mandatory field to NULL.
3677 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3678 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3679 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3684 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3688 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3689 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3690 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3691 DES and RC4 ciphersuites.
3695 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3696 This changes the decoding behaviour for some invalid messages,
3697 though the change is mostly in the more lenient direction, and
3698 legacy behaviour is preserved as much as possible.
3702 * Fix no-stdio build.
3703 *David Woodhouse <David.Woodhouse@intel.com> and also*
3704 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3706 * New testing framework
3707 The testing framework has been largely rewritten and is now using
3708 perl and the perl modules Test::Harness and an extended variant of
3709 Test::More called OpenSSL::Test to do its work. All test scripts in
3710 test/ have been rewritten into test recipes, and all direct calls to
3711 executables in test/Makefile have become individual recipes using the
3712 simplified testing OpenSSL::Test::Simple.
3714 For documentation on our testing modules, do:
3716 perldoc test/testlib/OpenSSL/Test/Simple.pm
3717 perldoc test/testlib/OpenSSL/Test.pm
3721 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3722 are used; the latter aborts on memory leaks (usually checked on exit).
3723 Some undocumented "set malloc, etc., hooks" functions were removed
3724 and others were changed. All are now documented.
3728 * In DSA_generate_parameters_ex, if the provided seed is too short,
3731 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3733 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3734 from RFC4279, RFC4785, RFC5487, RFC5489.
3736 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3737 original RSA_PSK patch.
3741 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3742 era flag was never set throughout the codebase (only read). Also removed
3743 SSL3_FLAGS_POP_BUFFER which was only used if
3744 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3748 * Changed the default name options in the "ca", "crl", "req" and "x509"
3749 to be "oneline" instead of "compat".
3753 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3754 not aware of clients that still exhibit this bug, and the workaround
3755 hasn't been working properly for a while.
3759 * The return type of BIO_number_read() and BIO_number_written() as well as
3760 the corresponding num_read and num_write members in the BIO structure has
3761 changed from unsigned long to uint64_t. On platforms where an unsigned
3762 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3767 * Given the pervasive nature of TLS extensions it is inadvisable to run
3768 OpenSSL without support for them. It also means that maintaining
3769 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3770 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3774 * Removed support for the two export grade static DH ciphersuites
3775 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3776 were newly added (along with a number of other static DH ciphersuites) to
3777 1.0.2. However the two export ones have *never* worked since they were
3778 introduced. It seems strange in any case to be adding new export
3779 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3783 * Version negotiation has been rewritten. In particular SSLv23_method(),
3784 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3785 and turned into macros which simply call the new preferred function names
3786 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3787 should use the new names instead. Also as part of this change the ssl23.h
3788 header file has been removed.
3792 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3793 code and the associated standard is no longer considered fit-for-purpose.
3797 * RT2547 was closed. When generating a private key, try to make the
3798 output file readable only by the owner. This behavior change might
3799 be noticeable when interacting with other software.
3801 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3806 * Added HTTP GET support to the ocsp command.
3810 * Changed default digest for the dgst and enc commands from MD5 to
3815 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3819 * Added support for TLS extended master secret from
3820 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3821 initial patch which was a great help during development.
3825 * All libssl internal structures have been removed from the public header
3826 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3827 now redundant). Users should not attempt to access internal structures
3828 directly. Instead they should use the provided API functions.
3832 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3833 Access to deprecated functions can be re-enabled by running config with
3834 "enable-deprecated". In addition applications wishing to use deprecated
3835 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3836 will, by default, disable some transitive includes that previously existed
3837 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3841 * Added support for OCB mode. OpenSSL has been granted a patent license
3842 compatible with the OpenSSL license for use of OCB. Details are available
3843 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3844 for OCB can be removed by calling config with no-ocb.
3848 * SSLv2 support has been removed. It still supports receiving a SSLv2
3849 compatible client hello.
3853 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3854 done while fixing the error code for the key-too-small case.
3856 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3858 * CA.sh has been removed; use CA.pl instead.
3862 * Removed old DES API.
3866 * Remove various unsupported platforms:
3872 Sinix/ReliantUNIX RM400
3877 16-bit platforms such as WIN16
3881 * Clean up OPENSSL_NO_xxx #define's
3882 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3883 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3884 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3885 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3886 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3887 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3888 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3889 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3890 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3891 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3895 * Cleaned up dead code
3896 Remove all but one '#ifdef undef' which is to be looked at.
3900 * Clean up calling of xxx_free routines.
3901 Just like free(), fix most of the xxx_free routines to accept
3902 NULL. Remove the non-null checks from callers. Save much code.
3906 * Add secure heap for storage of private keys (when possible).
3907 Add BIO_s_secmem(), CBIGNUM, etc.
3908 Contributed by Akamai Technologies under our Corporate CLA.
3912 * Experimental support for a new, fast, unbiased prime candidate generator,
3913 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3915 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3917 * New output format NSS in the sess_id command line tool. This allows
3918 exporting the session id and the master key in NSS keylog format.
3920 *Martin Kaiser <martin@kaiser.cx>*
3922 * Harmonize version and its documentation. -f flag is used to display
3925 *mancha <mancha1@zoho.com>*
3927 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3928 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3930 *mancha <mancha1@zoho.com>*
3932 * Fix some double frees. These are not thought to be exploitable.
3934 *mancha <mancha1@zoho.com>*
3936 * A missing bounds check in the handling of the TLS heartbeat extension
3937 can be used to reveal up to 64k of memory to a connected client or
3940 Thanks for Neel Mehta of Google Security for discovering this bug and to
3941 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3942 preparing the fix ([CVE-2014-0160])
3944 *Adam Langley, Bodo Moeller*
3946 * Fix for the attack described in the paper "Recovering OpenSSL
3947 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3948 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3949 <http://eprint.iacr.org/2014/140>
3951 Thanks to Yuval Yarom and Naomi Benger for discovering this
3952 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3954 *Yuval Yarom and Naomi Benger*
3956 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3957 this fixes a limitation in previous versions of OpenSSL.
3961 * Experimental encrypt-then-mac support.
3963 Experimental support for encrypt then mac from
3964 draft-gutmann-tls-encrypt-then-mac-02.txt
3966 To enable it set the appropriate extension number (0x42 for the test
3967 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3969 For non-compliant peers (i.e. just about everything) this should have no
3972 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3976 * Add EVP support for key wrapping algorithms, to avoid problems with
3977 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3978 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3979 algorithms and include tests cases.
3983 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3988 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3989 MGF1 digest and OAEP label.
3993 * Make openssl verify return errors.
3995 *Chris Palmer <palmer@google.com> and Ben Laurie*
3997 * New function ASN1_TIME_diff to calculate the difference between two
3998 ASN1_TIME structures or one structure and the current time.
4002 * Update fips_test_suite to support multiple command line options. New
4003 test to induce all self test errors in sequence and check expected
4008 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4009 sign or verify all in one operation.
4013 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4014 test programs and fips_test_suite. Includes functionality to parse
4015 the minimal script output of fipsalgest.pl directly.
4019 * Add authorisation parameter to FIPS_module_mode_set().
4023 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4027 * Use separate DRBG fields for internal and external flags. New function
4028 FIPS_drbg_health_check() to perform on demand health checking. Add
4029 generation tests to fips_test_suite with reduced health check interval to
4030 demonstrate periodic health checking. Add "nodh" option to
4031 fips_test_suite to skip very slow DH test.
4035 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4040 * More extensive health check for DRBG checking many more failure modes.
4041 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4042 combination: call this in fips_test_suite.
4046 * Add support for canonical generation of DSA parameter 'g'. See
4049 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4050 POST to handle HMAC cases.
4054 * Add functions FIPS_module_version() and FIPS_module_version_text()
4055 to return numerical and string versions of the FIPS module number.
4059 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4060 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4061 outside the validated module in the FIPS capable OpenSSL.
4065 * Minor change to DRBG entropy callback semantics. In some cases
4066 there is no multiple of the block length between min_len and
4067 max_len. Allow the callback to return more than max_len bytes
4068 of entropy but discard any extra: it is the callback's responsibility
4069 to ensure that the extra data discarded does not impact the
4070 requested amount of entropy.
4074 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4075 information in FIPS186-3, SP800-57 and SP800-131A.
4079 * CCM support via EVP. Interface is very similar to GCM case except we
4080 must supply all data in one chunk (i.e. no update, final) and the
4081 message length must be supplied if AAD is used. Add algorithm test
4086 * Initial version of POST overhaul. Add POST callback to allow the status
4087 of POST to be monitored and/or failures induced. Modify fips_test_suite
4088 to use callback. Always run all selftests even if one fails.
4092 * XTS support including algorithm test driver in the fips_gcmtest program.
4093 Note: this does increase the maximum key length from 32 to 64 bytes but
4094 there should be no binary compatibility issues as existing applications
4095 will never use XTS mode.
4099 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4100 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4101 performs algorithm blocking for unapproved PRNG types. Also do not
4102 set PRNG type in FIPS_mode_set(): leave this to the application.
4103 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4104 the standard OpenSSL PRNG: set additional data to a date time vector.
4108 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4109 This shouldn't present any incompatibility problems because applications
4110 shouldn't be using these directly and any that are will need to rethink
4111 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4115 * Extensive self tests and health checking required by SP800-90 DRBG.
4116 Remove strength parameter from FIPS_drbg_instantiate and always
4117 instantiate at maximum supported strength.
4121 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4125 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4129 * New function DH_compute_key_padded() to compute a DH key and pad with
4130 leading zeroes if needed: this complies with SP800-56A et al.
4134 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4135 anything, incomplete, subject to change and largely untested at present.
4139 * Modify fipscanisteronly build option to only build the necessary object
4140 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4144 * Add experimental option FIPSSYMS to give all symbols in
4145 fipscanister.o and FIPS or fips prefix. This will avoid
4146 conflicts with future versions of OpenSSL. Add perl script
4147 util/fipsas.pl to preprocess assembly language source files
4148 and rename any affected symbols.
4152 * Add selftest checks and algorithm block of non-fips algorithms in
4153 FIPS mode. Remove DES2 from selftests.
4157 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4158 return internal method without any ENGINE dependencies. Add new
4159 tiny fips sign and verify functions.
4163 * New build option no-ec2m to disable characteristic 2 code.
4167 * New build option "fipscanisteronly". This only builds fipscanister.o
4168 and (currently) associated fips utilities. Uses the file Makefile.fips
4169 instead of Makefile.org as the prototype.
4173 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4174 Update fips_gcmtest to use IV generator.
4178 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4179 setting output buffer to NULL. The `*Final` function must be
4180 called although it will not retrieve any additional data. The tag
4181 can be set or retrieved with a ctrl. The IV length is by default 12
4182 bytes (96 bits) but can be set to an alternative value. If the IV
4183 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4188 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4189 underlying do_cipher function handles all cipher semantics itself
4190 including padding and finalisation. This is useful if (for example)
4191 an ENGINE cipher handles block padding itself. The behaviour of
4192 do_cipher is subtly changed if this flag is set: the return value
4193 is the number of characters written to the output buffer (zero is
4194 no longer an error code) or a negative error code. Also if the
4195 input buffer is NULL and length 0 finalisation should be performed.
4199 * If a candidate issuer certificate is already part of the constructed
4200 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4204 * Improve forward-security support: add functions
4206 void SSL_CTX_set_not_resumable_session_callback(
4207 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4208 void SSL_set_not_resumable_session_callback(
4209 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4211 for use by SSL/TLS servers; the callback function will be called whenever a
4212 new session is created, and gets to decide whether the session may be
4213 cached to make it resumable (return 0) or not (return 1). (As by the
4214 SSL/TLS protocol specifications, the session_id sent by the server will be
4215 empty to indicate that the session is not resumable; also, the server will
4216 not generate RFC 4507 (RFC 5077) session tickets.)
4218 A simple reasonable callback implementation is to return is_forward_secure.
4219 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4220 by the SSL/TLS server library, indicating whether it can provide forward
4223 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4225 * New -verify_name option in command line utilities to set verification
4230 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4231 Add CMAC pkey methods.
4235 * Experimental renegotiation in s_server -www mode. If the client
4236 browses /reneg connection is renegotiated. If /renegcert it is
4237 renegotiated requesting a certificate.
4241 * Add an "external" session cache for debugging purposes to s_server. This
4242 should help trace issues which normally are only apparent in deployed
4243 multi-process servers.
4247 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4248 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4249 BIO_set_cipher() and some obscure PEM functions were changed so they
4250 can now return an error. The RAND changes required a change to the
4251 RAND_METHOD structure.
4255 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4256 a gcc attribute to warn if the result of a function is ignored. This
4257 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4258 whose return value is often ignored.
4262 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4263 These allow SCTs (signed certificate timestamps) to be requested and
4264 validated when establishing a connection.
4266 *Rob Percival <robpercival@google.com>*
4271 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4273 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4274 used even when parsing explicit parameters, when loading a encoded key
4275 or calling `EC_GROUP_new_from_ecpkparameters()`/
4276 `EC_GROUP_new_from_ecparameters()`.
4277 This prevents bypass of security hardening and performance gains,
4278 especially for curves with specialized EC_METHODs.
4279 By default, if a key encoded with explicit parameters is loaded and later
4280 encoded, the output is still encoded with explicit parameters, even if
4281 internally a "named" EC_GROUP is used for computation.
4285 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4286 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4287 NULL. After this change, only the cofactor parameter can be NULL. It also
4288 does some minimal sanity checks on the passed order.
4293 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4294 An attack is simple, if the first CMS_recipientInfo is valid but the
4295 second CMS_recipientInfo is chosen ciphertext. If the second
4296 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4297 encryption key will be replaced by garbage, and the message cannot be
4298 decoded, but if the RSA decryption fails, the correct encryption key is
4299 used and the recipient will not notice the attack.
4300 As a work around for this potential attack the length of the decrypted
4301 key must be equal to the cipher default key length, in case the
4302 certifiate is not given and all recipientInfo are tried out.
4303 The old behaviour can be re-enabled in the CMS code by setting the
4304 CMS_DEBUG_DECRYPT flag.
4309 * Document issue with installation paths in diverse Windows builds
4311 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4312 binaries and run-time config file.
4317 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4319 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4320 This changes the size when using the `genpkey` command when no size is given.
4321 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4322 generation commands to use 2048 bits by default.
4326 * Add FIPS support for Android Arm 64-bit
4328 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4329 Module in Version 2.0.10. For some reason, the corresponding target
4330 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4331 built with FIPS support on Android Arm 64-bit. This omission has been
4334 *Matthias St. Pierre*
4336 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4338 * 0-byte record padding oracle
4340 If an application encounters a fatal protocol error and then calls
4341 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4342 then OpenSSL can respond differently to the calling application if a 0 byte
4343 record is received with invalid padding compared to if a 0 byte record is
4344 received with an invalid MAC. If the application then behaves differently
4345 based on that in a way that is detectable to the remote peer, then this
4346 amounts to a padding oracle that could be used to decrypt data.
4348 In order for this to be exploitable "non-stitched" ciphersuites must be in
4349 use. Stitched ciphersuites are optimised implementations of certain
4350 commonly used ciphersuites. Also the application must call SSL_shutdown()
4351 twice even if a protocol error has occurred (applications should not do
4352 this but some do anyway).
4354 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4355 Aviram, with additional investigation by Steven Collison and Andrew
4356 Hourselt. It was reported to OpenSSL on 10th December 2018.
4361 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4365 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4367 * Microarchitecture timing vulnerability in ECC scalar multiplication
4369 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4370 shown to be vulnerable to a microarchitecture timing side channel attack.
4371 An attacker with sufficient access to mount local timing attacks during
4372 ECDSA signature generation could recover the private key.
4374 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4375 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4381 * Timing vulnerability in DSA signature generation
4383 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4384 timing side channel attack. An attacker could use variations in the signing
4385 algorithm to recover the private key.
4387 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4392 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4393 Module, accidentally introduced while backporting security fixes from the
4394 development branch and hindering the use of ECC in FIPS mode.
4398 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4400 * Client DoS due to large DH parameter
4402 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4403 malicious server can send a very large prime value to the client. This will
4404 cause the client to spend an unreasonably long period of time generating a
4405 key for this prime resulting in a hang until the client has finished. This
4406 could be exploited in a Denial Of Service attack.
4408 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4413 * Cache timing vulnerability in RSA Key Generation
4415 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4416 a cache timing side channel attack. An attacker with sufficient access to
4417 mount cache timing attacks during the RSA key generation process could
4418 recover the private key.
4420 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4421 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4426 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4427 parameter is no longer accepted, as it leads to a corrupt table. NULL
4428 pem_str is reserved for alias entries only.
4432 * Revert blinding in ECDSA sign and instead make problematic addition
4433 length-invariant. Switch even to fixed-length Montgomery multiplication.
4437 * Change generating and checking of primes so that the error rate of not
4438 being prime depends on the intended use based on the size of the input.
4439 For larger primes this will result in more rounds of Miller-Rabin.
4440 The maximal error rate for primes with more than 1080 bits is lowered
4443 *Kurt Roeckx, Annie Yousar*
4445 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4449 * Add blinding to ECDSA and DSA signatures to protect against side channel
4450 attacks discovered by Keegan Ryan (NCC Group).
4454 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4455 now allow empty (zero character) pass phrases.
4459 * Certificate time validation (X509_cmp_time) enforces stricter
4460 compliance with RFC 5280. Fractional seconds and timezone offsets
4461 are no longer allowed.
4465 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4467 * Constructed ASN.1 types with a recursive definition could exceed the stack
4469 Constructed ASN.1 types with a recursive definition (such as can be found
4470 in PKCS7) could eventually exceed the stack given malicious input with
4471 excessive recursion. This could result in a Denial Of Service attack. There
4472 are no such structures used within SSL/TLS that come from untrusted sources
4473 so this is considered safe.
4475 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4481 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4483 * Read/write after SSL object in error state
4485 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4486 mechanism. The intent was that if a fatal error occurred during a handshake
4487 then OpenSSL would move into the error state and would immediately fail if
4488 you attempted to continue the handshake. This works as designed for the
4489 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4490 SSL_connect()), however due to a bug it does not work correctly if
4491 SSL_read() or SSL_write() is called directly. In that scenario, if the
4492 handshake fails then a fatal error will be returned in the initial function
4493 call. If SSL_read()/SSL_write() is subsequently called by the application
4494 for the same SSL object then it will succeed and the data is passed without
4495 being decrypted/encrypted directly from the SSL/TLS record layer.
4497 In order to exploit this issue an application bug would have to be present
4498 that resulted in a call to SSL_read()/SSL_write() being issued after having
4499 already received a fatal error.
4501 This issue was reported to OpenSSL by David Benjamin (Google).
4506 * rsaz_1024_mul_avx2 overflow bug on x86_64
4508 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4509 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4510 Analysis suggests that attacks against RSA and DSA as a result of this
4511 defect would be very difficult to perform and are not believed likely.
4512 Attacks against DH1024 are considered just feasible, because most of the
4513 work necessary to deduce information about a private key may be performed
4514 offline. The amount of resources required for such an attack would be
4515 significant. However, for an attack on TLS to be meaningful, the server
4516 would have to share the DH1024 private key among multiple clients, which is
4517 no longer an option since CVE-2016-0701.
4519 This only affects processors that support the AVX2 but not ADX extensions
4520 like Intel Haswell (4th generation).
4522 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4523 was originally found via the OSS-Fuzz project.
4528 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4530 * bn_sqrx8x_internal carry bug on x86_64
4532 There is a carry propagating bug in the x86_64 Montgomery squaring
4533 procedure. No EC algorithms are affected. Analysis suggests that attacks
4534 against RSA and DSA as a result of this defect would be very difficult to
4535 perform and are not believed likely. Attacks against DH are considered just
4536 feasible (although very difficult) because most of the work necessary to
4537 deduce information about a private key may be performed offline. The amount
4538 of resources required for such an attack would be very significant and
4539 likely only accessible to a limited number of attackers. An attacker would
4540 additionally need online access to an unpatched system using the target
4541 private key in a scenario with persistent DH parameters and a private
4542 key that is shared between multiple clients.
4544 This only affects processors that support the BMI1, BMI2 and ADX extensions
4545 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4547 This issue was reported to OpenSSL by the OSS-Fuzz project.
4552 * Malformed X.509 IPAddressFamily could cause OOB read
4554 If an X.509 certificate has a malformed IPAddressFamily extension,
4555 OpenSSL could do a one-byte buffer overread. The most likely result
4556 would be an erroneous display of the certificate in text format.
4558 This issue was reported to OpenSSL by the OSS-Fuzz project.
4562 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4564 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4565 platform rather than 'mingw'.
4569 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4571 * Truncated packet could crash via OOB read
4573 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4574 cipher is being used, then a truncated packet can cause that host to
4575 perform an out-of-bounds read, usually resulting in a crash.
4577 This issue was reported to OpenSSL by Robert Święcki of Google.
4582 * BN_mod_exp may produce incorrect results on x86_64
4584 There is a carry propagating bug in the x86_64 Montgomery squaring
4585 procedure. No EC algorithms are affected. Analysis suggests that attacks
4586 against RSA and DSA as a result of this defect would be very difficult to
4587 perform and are not believed likely. Attacks against DH are considered just
4588 feasible (although very difficult) because most of the work necessary to
4589 deduce information about a private key may be performed offline. The amount
4590 of resources required for such an attack would be very significant and
4591 likely only accessible to a limited number of attackers. An attacker would
4592 additionally need online access to an unpatched system using the target
4593 private key in a scenario with persistent DH parameters and a private
4594 key that is shared between multiple clients. For example this can occur by
4595 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4596 similar to CVE-2015-3193 but must be treated as a separate problem.
4598 This issue was reported to OpenSSL by the OSS-Fuzz project.
4603 * Montgomery multiplication may produce incorrect results
4605 There is a carry propagating bug in the Broadwell-specific Montgomery
4606 multiplication procedure that handles input lengths divisible by, but
4607 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4608 and DH private keys are impossible. This is because the subroutine in
4609 question is not used in operations with the private key itself and an input
4610 of the attacker's direct choice. Otherwise the bug can manifest itself as
4611 transient authentication and key negotiation failures or reproducible
4612 erroneous outcome of public-key operations with specially crafted input.
4613 Among EC algorithms only Brainpool P-512 curves are affected and one
4614 presumably can attack ECDH key negotiation. Impact was not analyzed in
4615 detail, because pre-requisites for attack are considered unlikely. Namely
4616 multiple clients have to choose the curve in question and the server has to
4617 share the private key among them, neither of which is default behaviour.
4618 Even then only clients that chose the curve will be affected.
4620 This issue was publicly reported as transient failures and was not
4621 initially recognized as a security issue. Thanks to Richard Morgan for
4622 providing reproducible case.
4627 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4628 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4629 prevent issues where no progress is being made and the peer continually
4630 sends unrecognised record types, using up resources processing them.
4634 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4636 * Missing CRL sanity check
4638 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4639 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4640 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4642 This issue only affects the OpenSSL 1.0.2i
4647 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4649 * OCSP Status Request extension unbounded memory growth
4651 A malicious client can send an excessively large OCSP Status Request
4652 extension. If that client continually requests renegotiation, sending a
4653 large OCSP Status Request extension each time, then there will be unbounded
4654 memory growth on the server. This will eventually lead to a Denial Of
4655 Service attack through memory exhaustion. Servers with a default
4656 configuration are vulnerable even if they do not support OCSP. Builds using
4657 the "no-ocsp" build time option are not affected.
4659 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4664 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4667 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4673 * OOB write in MDC2_Update()
4675 An overflow can occur in MDC2_Update() either if called directly or
4676 through the EVP_DigestUpdate() function using MDC2. If an attacker
4677 is able to supply very large amounts of input data after a previous
4678 call to EVP_EncryptUpdate() with a partial block then a length check
4679 can overflow resulting in a heap corruption.
4681 The amount of data needed is comparable to SIZE_MAX which is impractical
4684 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4689 * Malformed SHA512 ticket DoS
4691 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4692 DoS attack where a malformed ticket will result in an OOB read which will
4695 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4696 a custom server callback and ticket lookup mechanism.
4698 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4703 * OOB write in BN_bn2dec()
4705 The function BN_bn2dec() does not check the return value of BN_div_word().
4706 This can cause an OOB write if an application uses this function with an
4707 overly large BIGNUM. This could be a problem if an overly large certificate
4708 or CRL is printed out from an untrusted source. TLS is not affected because
4709 record limits will reject an oversized certificate before it is parsed.
4711 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4716 * OOB read in TS_OBJ_print_bio()
4718 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4719 the total length the OID text representation would use and not the amount
4720 of data written. This will result in OOB reads when large OIDs are
4723 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4728 * Pointer arithmetic undefined behaviour
4730 Avoid some undefined pointer arithmetic
4732 A common idiom in the codebase is to check limits in the following manner:
4735 Where "p" points to some malloc'd data of SIZE bytes and
4738 "len" here could be from some externally supplied data (e.g. from a TLS
4741 The rules of C pointer arithmetic are such that "p + len" is only well
4742 defined where len <= SIZE. Therefore the above idiom is actually
4743 undefined behaviour.
4745 For example this could cause problems if some malloc implementation
4746 provides an address for "p" such that "p + len" actually overflows for
4747 values of len that are too big and therefore p + len < limit.
4749 This issue was reported to OpenSSL by Guido Vranken
4754 * Constant time flag not preserved in DSA signing
4756 Operations in the DSA signing algorithm should run in constant time in
4757 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4758 implementation means that a non-constant time codepath is followed for
4759 certain operations. This has been demonstrated through a cache-timing
4760 attack to be sufficient for an attacker to recover the private DSA key.
4762 This issue was reported by César Pereida (Aalto University), Billy Brumley
4763 (Tampere University of Technology), and Yuval Yarom (The University of
4764 Adelaide and NICTA).
4769 * DTLS buffered message DoS
4771 In a DTLS connection where handshake messages are delivered out-of-order
4772 those messages that OpenSSL is not yet ready to process will be buffered
4773 for later use. Under certain circumstances, a flaw in the logic means that
4774 those messages do not get removed from the buffer even though the handshake
4775 has been completed. An attacker could force up to approx. 15 messages to
4776 remain in the buffer when they are no longer required. These messages will
4777 be cleared when the DTLS connection is closed. The default maximum size for
4778 a message is 100k. Therefore the attacker could force an additional 1500k
4779 to be consumed per connection. By opening many simulataneous connections an
4780 attacker could cause a DoS attack through memory exhaustion.
4782 This issue was reported to OpenSSL by Quan Luo.
4787 * DTLS replay protection DoS
4789 A flaw in the DTLS replay attack protection mechanism means that records
4790 that arrive for future epochs update the replay protection "window" before
4791 the MAC for the record has been validated. This could be exploited by an
4792 attacker by sending a record for the next epoch (which does not have to
4793 decrypt or have a valid MAC), with a very large sequence number. This means
4794 that all subsequent legitimate packets are dropped causing a denial of
4795 service for a specific DTLS connection.
4797 This issue was reported to OpenSSL by the OCAP audit team.
4802 * Certificate message OOB reads
4804 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4805 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4806 theoretical DoS risk but this has not been observed in practice on common
4809 The messages affected are client certificate, client certificate request
4810 and server certificate. As a result the attack can only be performed
4811 against a client or a server which enables client authentication.
4813 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4818 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4820 * Prevent padding oracle in AES-NI CBC MAC check
4822 A MITM attacker can use a padding oracle attack to decrypt traffic
4823 when the connection uses an AES CBC cipher and the server support
4826 This issue was introduced as part of the fix for Lucky 13 padding
4827 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4828 constant time by making sure that always the same bytes are read and
4829 compared against either the MAC or padding bytes. But it no longer
4830 checked that there was enough data to have both the MAC and padding
4833 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4837 * Fix EVP_EncodeUpdate overflow
4839 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4840 Base64 encoding of binary data. If an attacker is able to supply very large
4841 amounts of input data then a length check can overflow resulting in a heap
4844 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4845 the `PEM_write_bio*` family of functions. These are mainly used within the
4846 OpenSSL command line applications, so any application which processes data
4847 from an untrusted source and outputs it as a PEM file should be considered
4848 vulnerable to this issue. User applications that call these APIs directly
4849 with large amounts of untrusted data may also be vulnerable.
4851 This issue was reported by Guido Vranken.
4856 * Fix EVP_EncryptUpdate overflow
4858 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4859 is able to supply very large amounts of input data after a previous call to
4860 EVP_EncryptUpdate() with a partial block then a length check can overflow
4861 resulting in a heap corruption. Following an analysis of all OpenSSL
4862 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4863 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4864 the first called function after an EVP_EncryptInit(), and therefore that
4865 specific call must be safe. The second form is where the length passed to
4866 EVP_EncryptUpdate() can be seen from the code to be some small value and
4867 therefore there is no possibility of an overflow. Since all instances are
4868 one of these two forms, it is believed that there can be no overflows in
4869 internal code due to this problem. It should be noted that
4870 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4871 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4872 of these calls have also been analysed too and it is believed there are no
4873 instances in internal usage where an overflow could occur.
4875 This issue was reported by Guido Vranken.
4880 * Prevent ASN.1 BIO excessive memory allocation
4882 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4883 a short invalid encoding can cause allocation of large amounts of memory
4884 potentially consuming excessive resources or exhausting memory.
4886 Any application parsing untrusted data through d2i BIO functions is
4887 affected. The memory based functions such as d2i_X509() are *not* affected.
4888 Since the memory based functions are used by the TLS library, TLS
4889 applications are not affected.
4891 This issue was reported by Brian Carpenter.
4898 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4899 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4900 in arbitrary stack data being returned in the buffer.
4902 This issue was reported by Guido Vranken.
4907 * Modify behavior of ALPN to invoke callback after SNI/servername
4908 callback, such that updates to the SSL_CTX affect ALPN.
4912 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4917 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4918 methods are enabled and ssl2 is disabled the methods return NULL.
4922 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4924 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4925 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4926 provide any "EXPORT" or "LOW" strength ciphers.
4930 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4931 is by default disabled at build-time. Builds that are not configured with
4932 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4933 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4934 will need to explicitly call either of:
4936 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4938 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4940 as appropriate. Even if either of those is used, or the application
4941 explicitly uses the version-specific SSLv2_method() or its client and
4942 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4943 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4944 ciphers, and SSLv2 56-bit DES are no longer available.
4949 * Fix a double-free in DSA code
4951 A double free bug was discovered when OpenSSL parses malformed DSA private
4952 keys and could lead to a DoS attack or memory corruption for applications
4953 that receive DSA private keys from untrusted sources. This scenario is
4956 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4962 * Disable SRP fake user seed to address a server memory leak.
4964 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4966 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4967 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4968 was changed to ignore the "fake user" SRP seed, even if the seed
4971 Users should use SRP_VBASE_get1_by_user instead. Note that in
4972 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4973 also that even though configuring the SRP seed attempts to hide
4974 invalid usernames by continuing the handshake with fake
4975 credentials, this behaviour is not constant time and no strong
4976 guarantees are made that the handshake is indistinguishable from
4977 that of a valid user.
4982 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4984 In the BN_hex2bn function the number of hex digits is calculated using an
4985 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4986 large values of `i` this can result in `bn_expand` not allocating any
4987 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4988 field as NULL leading to a subsequent NULL ptr deref. For very large values
4989 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4990 In this case memory is allocated to the internal BIGNUM data field, but it
4991 is insufficiently sized leading to heap corruption. A similar issue exists
4992 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4993 is ever called by user applications with very large untrusted hex/dec data.
4994 This is anticipated to be a rare occurrence.
4996 All OpenSSL internal usage of these functions use data that is not expected
4997 to be untrusted, e.g. config file data or application command line
4998 arguments. If user developed applications generate config file data based
4999 on untrusted data then it is possible that this could also lead to security
5000 consequences. This is also anticipated to be rare.
5002 This issue was reported to OpenSSL by Guido Vranken.
5007 * Fix memory issues in `BIO_*printf` functions
5009 The internal `fmtstr` function used in processing a "%s" format string in
5010 the `BIO_*printf` functions could overflow while calculating the length of a
5011 string and cause an OOB read when printing very long strings.
5013 Additionally the internal `doapr_outch` function can attempt to write to an
5014 OOB memory location (at an offset from the NULL pointer) in the event of a
5015 memory allocation failure. In 1.0.2 and below this could be caused where
5016 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5017 could be in processing a very long "%s" format string. Memory leaks can
5020 The first issue may mask the second issue dependent on compiler behaviour.
5021 These problems could enable attacks where large amounts of untrusted data
5022 is passed to the `BIO_*printf` functions. If applications use these functions
5023 in this way then they could be vulnerable. OpenSSL itself uses these
5024 functions when printing out human-readable dumps of ASN.1 data. Therefore
5025 applications that print this data could be vulnerable if the data is from
5026 untrusted sources. OpenSSL command line applications could also be
5027 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5028 as command line arguments.
5030 Libssl is not considered directly vulnerable. Additionally certificates etc
5031 received via remote connections via libssl are also unlikely to be able to
5032 trigger these issues because of message size limits enforced within libssl.
5034 This issue was reported to OpenSSL Guido Vranken.
5039 * Side channel attack on modular exponentiation
5041 A side-channel attack was found which makes use of cache-bank conflicts on
5042 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5043 of RSA keys. The ability to exploit this issue is limited as it relies on
5044 an attacker who has control of code in a thread running on the same
5045 hyper-threaded core as the victim thread which is performing decryptions.
5047 This issue was reported to OpenSSL by Yuval Yarom, The University of
5048 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5049 Nadia Heninger, University of Pennsylvania with more information at
5050 <http://cachebleed.info>.
5055 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5056 if no keysize is specified with default_bits. This fixes an
5057 omission in an earlier change that changed all RSA/DSA key generation
5058 commands to use 2048 bits by default.
5062 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5064 * DH small subgroups
5066 Historically OpenSSL only ever generated DH parameters based on "safe"
5067 primes. More recently (in version 1.0.2) support was provided for
5068 generating X9.42 style parameter files such as those required for RFC 5114
5069 support. The primes used in such files may not be "safe". Where an
5070 application is using DH configured with parameters based on primes that are
5071 not "safe" then an attacker could use this fact to find a peer's private
5072 DH exponent. This attack requires that the attacker complete multiple
5073 handshakes in which the peer uses the same private DH exponent. For example
5074 this could be used to discover a TLS server's private DH exponent if it's
5075 reusing the private DH exponent or it's using a static DH ciphersuite.
5077 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5078 TLS. It is not on by default. If the option is not set then the server
5079 reuses the same private DH exponent for the life of the server process and
5080 would be vulnerable to this attack. It is believed that many popular
5081 applications do set this option and would therefore not be at risk.
5083 The fix for this issue adds an additional check where a "q" parameter is
5084 available (as is the case in X9.42 based parameters). This detects the
5085 only known attack, and is the only possible defense for static DH
5086 ciphersuites. This could have some performance impact.
5088 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5089 default and cannot be disabled. This could have some performance impact.
5091 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5096 * SSLv2 doesn't block disabled ciphers
5098 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5099 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5100 been disabled, provided that the SSLv2 protocol was not also disabled via
5103 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5104 and Sebastian Schinzel.
5109 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5111 * BN_mod_exp may produce incorrect results on x86_64
5113 There is a carry propagating bug in the x86_64 Montgomery squaring
5114 procedure. No EC algorithms are affected. Analysis suggests that attacks
5115 against RSA and DSA as a result of this defect would be very difficult to
5116 perform and are not believed likely. Attacks against DH are considered just
5117 feasible (although very difficult) because most of the work necessary to
5118 deduce information about a private key may be performed offline. The amount
5119 of resources required for such an attack would be very significant and
5120 likely only accessible to a limited number of attackers. An attacker would
5121 additionally need online access to an unpatched system using the target
5122 private key in a scenario with persistent DH parameters and a private
5123 key that is shared between multiple clients. For example this can occur by
5124 default in OpenSSL DHE based SSL/TLS ciphersuites.
5126 This issue was reported to OpenSSL by Hanno Böck.
5131 * Certificate verify crash with missing PSS parameter
5133 The signature verification routines will crash with a NULL pointer
5134 dereference if presented with an ASN.1 signature using the RSA PSS
5135 algorithm and absent mask generation function parameter. Since these
5136 routines are used to verify certificate signature algorithms this can be
5137 used to crash any certificate verification operation and exploited in a
5138 DoS attack. Any application which performs certificate verification is
5139 vulnerable including OpenSSL clients and servers which enable client
5142 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5147 * X509_ATTRIBUTE memory leak
5149 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5150 memory. This structure is used by the PKCS#7 and CMS routines so any
5151 application which reads PKCS#7 or CMS data from untrusted sources is
5152 affected. SSL/TLS is not affected.
5154 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5160 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5161 This changes the decoding behaviour for some invalid messages,
5162 though the change is mostly in the more lenient direction, and
5163 legacy behaviour is preserved as much as possible.
5167 * In DSA_generate_parameters_ex, if the provided seed is too short,
5170 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5172 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5174 * Alternate chains certificate forgery
5176 During certificate verification, OpenSSL will attempt to find an
5177 alternative certificate chain if the first attempt to build such a chain
5178 fails. An error in the implementation of this logic can mean that an
5179 attacker could cause certain checks on untrusted certificates to be
5180 bypassed, such as the CA flag, enabling them to use a valid leaf
5181 certificate to act as a CA and "issue" an invalid certificate.
5183 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5188 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5190 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5191 incompatibility in the handling of HMAC. The previous ABI has now been
5196 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5198 * Malformed ECParameters causes infinite loop
5200 When processing an ECParameters structure OpenSSL enters an infinite loop
5201 if the curve specified is over a specially malformed binary polynomial
5204 This can be used to perform denial of service against any
5205 system which processes public keys, certificate requests or
5206 certificates. This includes TLS clients and TLS servers with
5207 client authentication enabled.
5209 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5214 * Exploitable out-of-bounds read in X509_cmp_time
5216 X509_cmp_time does not properly check the length of the ASN1_TIME
5217 string and can read a few bytes out of bounds. In addition,
5218 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5221 An attacker can use this to craft malformed certificates and CRLs of
5222 various sizes and potentially cause a segmentation fault, resulting in
5223 a DoS on applications that verify certificates or CRLs. TLS clients
5224 that verify CRLs are affected. TLS clients and servers with client
5225 authentication enabled may be affected if they use custom verification
5228 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5229 independently by Hanno Böck.
5234 * PKCS7 crash with missing EnvelopedContent
5236 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5237 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5238 with missing content and trigger a NULL pointer dereference on parsing.
5240 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5241 structures from untrusted sources are affected. OpenSSL clients and
5242 servers are not affected.
5244 This issue was reported to OpenSSL by Michal Zalewski (Google).
5249 * CMS verify infinite loop with unknown hash function
5251 When verifying a signedData message the CMS code can enter an infinite loop
5252 if presented with an unknown hash function OID. This can be used to perform
5253 denial of service against any system which verifies signedData messages using
5255 This issue was reported to OpenSSL by Johannes Bauer.
5260 * Race condition handling NewSessionTicket
5262 If a NewSessionTicket is received by a multi-threaded client when attempting to
5263 reuse a previous ticket then a race condition can occur potentially leading to
5264 a double free of the ticket data.
5269 * Only support 256-bit or stronger elliptic curves with the
5270 'ecdh_auto' setting (server) or by default (client). Of supported
5271 curves, prefer P-256 (both).
5275 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5277 * ClientHello sigalgs DoS fix
5279 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5280 invalid signature algorithms extension a NULL pointer dereference will
5281 occur. This can be exploited in a DoS attack against the server.
5283 This issue was was reported to OpenSSL by David Ramos of Stanford
5287 *Stephen Henson and Matt Caswell*
5289 * Multiblock corrupted pointer fix
5291 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5292 feature only applies on 64 bit x86 architecture platforms that support AES
5293 NI instructions. A defect in the implementation of "multiblock" can cause
5294 OpenSSL's internal write buffer to become incorrectly set to NULL when
5295 using non-blocking IO. Typically, when the user application is using a
5296 socket BIO for writing, this will only result in a failed connection.
5297 However if some other BIO is used then it is likely that a segmentation
5298 fault will be triggered, thus enabling a potential DoS attack.
5300 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5305 * Segmentation fault in DTLSv1_listen fix
5307 The DTLSv1_listen function is intended to be stateless and processes the
5308 initial ClientHello from many peers. It is common for user code to loop
5309 over the call to DTLSv1_listen until a valid ClientHello is received with
5310 an associated cookie. A defect in the implementation of DTLSv1_listen means
5311 that state is preserved in the SSL object from one invocation to the next
5312 that can lead to a segmentation fault. Errors processing the initial
5313 ClientHello can trigger this scenario. An example of such an error could be
5314 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5317 This issue was reported to OpenSSL by Per Allansson.
5322 * Segmentation fault in ASN1_TYPE_cmp fix
5324 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5325 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5326 certificate signature algorithm consistency this can be used to crash any
5327 certificate verification operation and exploited in a DoS attack. Any
5328 application which performs certificate verification is vulnerable including
5329 OpenSSL clients and servers which enable client authentication.
5334 * Segmentation fault for invalid PSS parameters fix
5336 The signature verification routines will crash with a NULL pointer
5337 dereference if presented with an ASN.1 signature using the RSA PSS
5338 algorithm and invalid parameters. Since these routines are used to verify
5339 certificate signature algorithms this can be used to crash any
5340 certificate verification operation and exploited in a DoS attack. Any
5341 application which performs certificate verification is vulnerable including
5342 OpenSSL clients and servers which enable client authentication.
5344 This issue was was reported to OpenSSL by Brian Carpenter.
5349 * ASN.1 structure reuse memory corruption fix
5351 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5352 memory corruption via an invalid write. Such reuse is and has been
5353 strongly discouraged and is believed to be rare.
5355 Applications that parse structures containing CHOICE or ANY DEFINED BY
5356 components may be affected. Certificate parsing (d2i_X509 and related
5357 functions) are however not affected. OpenSSL clients and servers are
5363 * PKCS7 NULL pointer dereferences fix
5365 The PKCS#7 parsing code does not handle missing outer ContentInfo
5366 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5367 missing content and trigger a NULL pointer dereference on parsing.
5369 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5370 otherwise parse PKCS#7 structures from untrusted sources are
5371 affected. OpenSSL clients and servers are not affected.
5373 This issue was reported to OpenSSL by Michal Zalewski (Google).
5378 * DoS via reachable assert in SSLv2 servers fix
5380 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5381 servers that both support SSLv2 and enable export cipher suites by sending
5382 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5384 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5385 (OpenSSL development team).
5390 * Empty CKE with client auth and DHE fix
5392 If client auth is used then a server can seg fault in the event of a DHE
5393 ciphersuite being selected and a zero length ClientKeyExchange message
5394 being sent by the client. This could be exploited in a DoS attack.
5399 * Handshake with unseeded PRNG fix
5401 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5402 with an unseeded PRNG. The conditions are:
5403 - The client is on a platform where the PRNG has not been seeded
5404 automatically, and the user has not seeded manually
5405 - A protocol specific client method version has been used (i.e. not
5406 SSL_client_methodv23)
5407 - A ciphersuite is used that does not require additional random data from
5408 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5410 If the handshake succeeds then the client random that has been used will
5411 have been generated from a PRNG with insufficient entropy and therefore the
5412 output may be predictable.
5414 For example using the following command with an unseeded openssl will
5415 succeed on an unpatched platform:
5417 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5422 * Use After Free following d2i_ECPrivatekey error fix
5424 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5425 could cause a use after free condition. This, in turn, could cause a double
5426 free in several private key parsing functions (such as d2i_PrivateKey
5427 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5428 for applications that receive EC private keys from untrusted
5429 sources. This scenario is considered rare.
5431 This issue was discovered by the BoringSSL project and fixed in their
5437 * X509_to_X509_REQ NULL pointer deref fix
5439 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5440 the certificate key is invalid. This function is rarely used in practice.
5442 This issue was discovered by Brian Carpenter.
5447 * Removed the export ciphers from the DEFAULT ciphers
5451 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5453 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5454 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5455 So far those who have to target multiple platforms would compromise
5456 and argue that binary targeting say ARMv5 would still execute on
5457 ARMv8. "Universal" build resolves this compromise by providing
5458 near-optimal performance even on newer platforms.
5462 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5463 (other platforms pending).
5465 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5467 * Add support for the SignedCertificateTimestampList certificate and
5468 OCSP response extensions from RFC6962.
5472 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5473 for corner cases. (Certain input points at infinity could lead to
5474 bogus results, with non-infinity inputs mapped to infinity too.)
5478 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5479 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5480 common cases are optimized and there still is room for further
5481 improvements. Vector Permutation AES for Altivec is also added.
5485 * Add support for little-endian ppc64 Linux target.
5487 *Marcelo Cerri (IBM)*
5489 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5490 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5491 are optimized and there still is room for further improvements.
5492 Both 32- and 64-bit modes are supported.
5494 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5496 * Improved ARMv7 NEON support.
5500 * Support for SPARC Architecture 2011 crypto extensions, first
5501 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5502 SHA256/512, MD5, GHASH and modular exponentiation.
5504 *Andy Polyakov, David Miller*
5506 * Accelerated modular exponentiation for Intel processors, a.k.a.
5509 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5511 * Support for new and upcoming Intel processors, including AVX2,
5512 BMI and SHA ISA extensions. This includes additional "stitched"
5513 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5516 This work was sponsored by Intel Corp.
5520 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5521 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5522 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5526 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5527 this fixes a limitation in previous versions of OpenSSL.
5531 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5532 MGF1 digest and OAEP label.
5536 * Add EVP support for key wrapping algorithms, to avoid problems with
5537 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5538 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5539 algorithms and include tests cases.
5543 * Add functions to allocate and set the fields of an ECDSA_METHOD
5546 *Douglas E. Engert, Steve Henson*
5548 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5549 difference in days and seconds between two tm or ASN1_TIME structures.
5553 * Add -rev test option to s_server to just reverse order of characters
5554 received by client and send back to server. Also prints an abbreviated
5555 summary of the connection parameters.
5559 * New option -brief for s_client and s_server to print out a brief summary
5560 of connection parameters.
5564 * Add callbacks for arbitrary TLS extensions.
5566 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5568 * New option -crl_download in several openssl utilities to download CRLs
5569 from CRLDP extension in certificates.
5573 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5577 * New function X509_CRL_diff to generate a delta CRL from the difference
5578 of two full CRLs. Add support to "crl" utility.
5582 * New functions to set lookup_crls function and to retrieve
5583 X509_STORE from X509_STORE_CTX.
5587 * Print out deprecated issuer and subject unique ID fields in
5592 * Extend OCSP I/O functions so they can be used for simple general purpose
5593 HTTP as well as OCSP. New wrapper function which can be used to download
5594 CRLs using the OCSP API.
5598 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5602 * `SSL_CONF*` functions. These provide a common framework for application
5603 configuration using configuration files or command lines.
5607 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5608 message callback and prints the results. Needs compile time option
5609 "enable-ssl-trace". New options to s_client and s_server to enable
5614 * New ctrl and macro to retrieve supported points extensions.
5615 Print out extension in s_server and s_client.
5619 * New functions to retrieve certificate signature and signature
5624 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5629 * New Suite B modes for TLS code. These use and enforce the requirements
5630 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5631 only use Suite B curves. The Suite B modes can be set by using the
5632 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5636 * New chain verification flags for Suite B levels of security. Check
5637 algorithms are acceptable when flags are set in X509_verify_cert.
5641 * Make tls1_check_chain return a set of flags indicating checks passed
5642 by a certificate chain. Add additional tests to handle client
5643 certificates: checks for matching certificate type and issuer name
5648 * If an attempt is made to use a signature algorithm not in the peer
5649 preference list abort the handshake. If client has no suitable
5650 signature algorithms in response to a certificate request do not
5651 use the certificate.
5655 * If server EC tmp key is not in client preference list abort handshake.
5659 * Add support for certificate stores in CERT structure. This makes it
5660 possible to have different stores per SSL structure or one store in
5661 the parent SSL_CTX. Include distinct stores for certificate chain
5662 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5663 to build and store a certificate chain in CERT structure: returning
5664 an error if the chain cannot be built: this will allow applications
5665 to test if a chain is correctly configured.
5667 Note: if the CERT based stores are not set then the parent SSL_CTX
5668 store is used to retain compatibility with existing behaviour.
5672 * New function ssl_set_client_disabled to set a ciphersuite disabled
5673 mask based on the current session, check mask when sending client
5674 hello and checking the requested ciphersuite.
5678 * New ctrls to retrieve and set certificate types in a certificate
5679 request message. Print out received values in s_client. If certificate
5680 types is not set with custom values set sensible values based on
5681 supported signature algorithms.
5685 * Support for distinct client and server supported signature algorithms.
5689 * Add certificate callback. If set this is called whenever a certificate
5690 is required by client or server. An application can decide which
5691 certificate chain to present based on arbitrary criteria: for example
5692 supported signature algorithms. Add very simple example to s_server.
5693 This fixes many of the problems and restrictions of the existing client
5694 certificate callback: for example you can now clear an existing
5695 certificate and specify the whole chain.
5699 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5700 the certificate can be used for (if anything). Set valid_flags field
5701 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5702 to have similar checks in it.
5704 Add new "cert_flags" field to CERT structure and include a "strict mode".
5705 This enforces some TLS certificate requirements (such as only permitting
5706 certificate signature algorithms contained in the supported algorithms
5707 extension) which some implementations ignore: this option should be used
5708 with caution as it could cause interoperability issues.
5712 * Update and tidy signature algorithm extension processing. Work out
5713 shared signature algorithms based on preferences and peer algorithms
5714 and print them out in s_client and s_server. Abort handshake if no
5715 shared signature algorithms.
5719 * Add new functions to allow customised supported signature algorithms
5720 for SSL and SSL_CTX structures. Add options to s_client and s_server
5725 * New function SSL_certs_clear() to delete all references to certificates
5726 from an SSL structure. Before this once a certificate had been added
5727 it couldn't be removed.
5731 * Integrate hostname, email address and IP address checking with certificate
5732 verification. New verify options supporting checking in openssl utility.
5736 * Fixes and wildcard matching support to hostname and email checking
5737 functions. Add manual page.
5739 *Florian Weimer (Red Hat Product Security Team)*
5741 * New functions to check a hostname email or IP address against a
5742 certificate. Add options x509 utility to print results of checks against
5747 * Fix OCSP checking.
5749 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5751 * Initial experimental support for explicitly trusted non-root CAs.
5752 OpenSSL still tries to build a complete chain to a root but if an
5753 intermediate CA has a trust setting included that is used. The first
5754 setting is used: whether to trust (e.g., -addtrust option to the x509
5759 * Add -trusted_first option which attempts to find certificates in the
5760 trusted store even if an untrusted chain is also supplied.
5764 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5765 platform support for Linux and Android.
5769 * Support for linux-x32, ILP32 environment in x86_64 framework.
5773 * Experimental multi-implementation support for FIPS capable OpenSSL.
5774 When in FIPS mode the approved implementations are used as normal,
5775 when not in FIPS mode the internal unapproved versions are used instead.
5776 This means that the FIPS capable OpenSSL isn't forced to use the
5777 (often lower performance) FIPS implementations outside FIPS mode.
5781 * Transparently support X9.42 DH parameters when calling
5782 PEM_read_bio_DHparameters. This means existing applications can handle
5783 the new parameter format automatically.
5787 * Initial experimental support for X9.42 DH parameter format: mainly
5788 to support use of 'q' parameter for RFC5114 parameters.
5792 * Add DH parameters from RFC5114 including test data to dhtest.
5796 * Support for automatic EC temporary key parameter selection. If enabled
5797 the most preferred EC parameters are automatically used instead of
5798 hardcoded fixed parameters. Now a server just has to call:
5799 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5800 support ECDH and use the most appropriate parameters.
5804 * Enhance and tidy EC curve and point format TLS extension code. Use
5805 static structures instead of allocation if default values are used.
5806 New ctrls to set curves we wish to support and to retrieve shared curves.
5807 Print out shared curves in s_server. New options to s_server and s_client
5808 to set list of supported curves.
5812 * New ctrls to retrieve supported signature algorithms and
5813 supported curve values as an array of NIDs. Extend openssl utility
5814 to print out received values.
5818 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5819 between NIDs and the more common NIST names such as "P-256". Enhance
5820 ecparam utility and ECC method to recognise the NIST names for curves.
5824 * Enhance SSL/TLS certificate chain handling to support different
5825 chains for each certificate instead of one chain in the parent SSL_CTX.
5829 * Support for fixed DH ciphersuite client authentication: where both
5830 server and client use DH certificates with common parameters.
5834 * Support for fixed DH ciphersuites: those requiring DH server
5839 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5841 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5842 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5843 X509_CINF_get_signature were reverted post internal team review.
5848 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5850 * OCSP Status Request extension unbounded memory growth
5852 A malicious client can send an excessively large OCSP Status Request
5853 extension. If that client continually requests renegotiation, sending a
5854 large OCSP Status Request extension each time, then there will be unbounded
5855 memory growth on the server. This will eventually lead to a Denial Of
5856 Service attack through memory exhaustion. Servers with a default
5857 configuration are vulnerable even if they do not support OCSP. Builds using
5858 the "no-ocsp" build time option are not affected.
5860 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5865 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5868 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5874 * OOB write in MDC2_Update()
5876 An overflow can occur in MDC2_Update() either if called directly or
5877 through the EVP_DigestUpdate() function using MDC2. If an attacker
5878 is able to supply very large amounts of input data after a previous
5879 call to EVP_EncryptUpdate() with a partial block then a length check
5880 can overflow resulting in a heap corruption.
5882 The amount of data needed is comparable to SIZE_MAX which is impractical
5885 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5890 * Malformed SHA512 ticket DoS
5892 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5893 DoS attack where a malformed ticket will result in an OOB read which will
5896 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5897 a custom server callback and ticket lookup mechanism.
5899 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5904 * OOB write in BN_bn2dec()
5906 The function BN_bn2dec() does not check the return value of BN_div_word().
5907 This can cause an OOB write if an application uses this function with an
5908 overly large BIGNUM. This could be a problem if an overly large certificate
5909 or CRL is printed out from an untrusted source. TLS is not affected because
5910 record limits will reject an oversized certificate before it is parsed.
5912 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5917 * OOB read in TS_OBJ_print_bio()
5919 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5920 the total length the OID text representation would use and not the amount
5921 of data written. This will result in OOB reads when large OIDs are
5924 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5929 * Pointer arithmetic undefined behaviour
5931 Avoid some undefined pointer arithmetic
5933 A common idiom in the codebase is to check limits in the following manner:
5936 Where "p" points to some malloc'd data of SIZE bytes and
5939 "len" here could be from some externally supplied data (e.g. from a TLS
5942 The rules of C pointer arithmetic are such that "p + len" is only well
5943 defined where len <= SIZE. Therefore the above idiom is actually
5944 undefined behaviour.
5946 For example this could cause problems if some malloc implementation
5947 provides an address for "p" such that "p + len" actually overflows for
5948 values of len that are too big and therefore p + len < limit.
5950 This issue was reported to OpenSSL by Guido Vranken
5955 * Constant time flag not preserved in DSA signing
5957 Operations in the DSA signing algorithm should run in constant time in
5958 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5959 implementation means that a non-constant time codepath is followed for
5960 certain operations. This has been demonstrated through a cache-timing
5961 attack to be sufficient for an attacker to recover the private DSA key.
5963 This issue was reported by César Pereida (Aalto University), Billy Brumley
5964 (Tampere University of Technology), and Yuval Yarom (The University of
5965 Adelaide and NICTA).
5970 * DTLS buffered message DoS
5972 In a DTLS connection where handshake messages are delivered out-of-order
5973 those messages that OpenSSL is not yet ready to process will be buffered
5974 for later use. Under certain circumstances, a flaw in the logic means that
5975 those messages do not get removed from the buffer even though the handshake
5976 has been completed. An attacker could force up to approx. 15 messages to
5977 remain in the buffer when they are no longer required. These messages will
5978 be cleared when the DTLS connection is closed. The default maximum size for
5979 a message is 100k. Therefore the attacker could force an additional 1500k
5980 to be consumed per connection. By opening many simulataneous connections an
5981 attacker could cause a DoS attack through memory exhaustion.
5983 This issue was reported to OpenSSL by Quan Luo.
5988 * DTLS replay protection DoS
5990 A flaw in the DTLS replay attack protection mechanism means that records
5991 that arrive for future epochs update the replay protection "window" before
5992 the MAC for the record has been validated. This could be exploited by an
5993 attacker by sending a record for the next epoch (which does not have to
5994 decrypt or have a valid MAC), with a very large sequence number. This means
5995 that all subsequent legitimate packets are dropped causing a denial of
5996 service for a specific DTLS connection.
5998 This issue was reported to OpenSSL by the OCAP audit team.
6003 * Certificate message OOB reads
6005 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6006 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6007 theoretical DoS risk but this has not been observed in practice on common
6010 The messages affected are client certificate, client certificate request
6011 and server certificate. As a result the attack can only be performed
6012 against a client or a server which enables client authentication.
6014 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6019 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6021 * Prevent padding oracle in AES-NI CBC MAC check
6023 A MITM attacker can use a padding oracle attack to decrypt traffic
6024 when the connection uses an AES CBC cipher and the server support
6027 This issue was introduced as part of the fix for Lucky 13 padding
6028 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6029 constant time by making sure that always the same bytes are read and
6030 compared against either the MAC or padding bytes. But it no longer
6031 checked that there was enough data to have both the MAC and padding
6034 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6039 * Fix EVP_EncodeUpdate overflow
6041 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6042 Base64 encoding of binary data. If an attacker is able to supply very large
6043 amounts of input data then a length check can overflow resulting in a heap
6046 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6047 the `PEM_write_bio*` family of functions. These are mainly used within the
6048 OpenSSL command line applications, so any application which processes data
6049 from an untrusted source and outputs it as a PEM file should be considered
6050 vulnerable to this issue. User applications that call these APIs directly
6051 with large amounts of untrusted data may also be vulnerable.
6053 This issue was reported by Guido Vranken.
6058 * Fix EVP_EncryptUpdate overflow
6060 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6061 is able to supply very large amounts of input data after a previous call to
6062 EVP_EncryptUpdate() with a partial block then a length check can overflow
6063 resulting in a heap corruption. Following an analysis of all OpenSSL
6064 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6065 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6066 the first called function after an EVP_EncryptInit(), and therefore that
6067 specific call must be safe. The second form is where the length passed to
6068 EVP_EncryptUpdate() can be seen from the code to be some small value and
6069 therefore there is no possibility of an overflow. Since all instances are
6070 one of these two forms, it is believed that there can be no overflows in
6071 internal code due to this problem. It should be noted that
6072 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6073 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6074 of these calls have also been analysed too and it is believed there are no
6075 instances in internal usage where an overflow could occur.
6077 This issue was reported by Guido Vranken.
6082 * Prevent ASN.1 BIO excessive memory allocation
6084 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6085 a short invalid encoding can casuse allocation of large amounts of memory
6086 potentially consuming excessive resources or exhausting memory.
6088 Any application parsing untrusted data through d2i BIO functions is
6089 affected. The memory based functions such as d2i_X509() are *not* affected.
6090 Since the memory based functions are used by the TLS library, TLS
6091 applications are not affected.
6093 This issue was reported by Brian Carpenter.
6100 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6101 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6102 in arbitrary stack data being returned in the buffer.
6104 This issue was reported by Guido Vranken.
6109 * Modify behavior of ALPN to invoke callback after SNI/servername
6110 callback, such that updates to the SSL_CTX affect ALPN.
6114 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6119 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6120 methods are enabled and ssl2 is disabled the methods return NULL.
6124 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6126 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6127 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6128 provide any "EXPORT" or "LOW" strength ciphers.
6132 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6133 is by default disabled at build-time. Builds that are not configured with
6134 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6135 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6136 will need to explicitly call either of:
6138 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6140 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6142 as appropriate. Even if either of those is used, or the application
6143 explicitly uses the version-specific SSLv2_method() or its client and
6144 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6145 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6146 ciphers, and SSLv2 56-bit DES are no longer available.
6151 * Fix a double-free in DSA code
6153 A double free bug was discovered when OpenSSL parses malformed DSA private
6154 keys and could lead to a DoS attack or memory corruption for applications
6155 that receive DSA private keys from untrusted sources. This scenario is
6158 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6164 * Disable SRP fake user seed to address a server memory leak.
6166 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6168 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6169 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6170 was changed to ignore the "fake user" SRP seed, even if the seed
6173 Users should use SRP_VBASE_get1_by_user instead. Note that in
6174 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6175 also that even though configuring the SRP seed attempts to hide
6176 invalid usernames by continuing the handshake with fake
6177 credentials, this behaviour is not constant time and no strong
6178 guarantees are made that the handshake is indistinguishable from
6179 that of a valid user.
6184 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6186 In the BN_hex2bn function the number of hex digits is calculated using an
6187 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6188 large values of `i` this can result in `bn_expand` not allocating any
6189 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6190 field as NULL leading to a subsequent NULL ptr deref. For very large values
6191 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6192 In this case memory is allocated to the internal BIGNUM data field, but it
6193 is insufficiently sized leading to heap corruption. A similar issue exists
6194 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6195 is ever called by user applications with very large untrusted hex/dec data.
6196 This is anticipated to be a rare occurrence.
6198 All OpenSSL internal usage of these functions use data that is not expected
6199 to be untrusted, e.g. config file data or application command line
6200 arguments. If user developed applications generate config file data based
6201 on untrusted data then it is possible that this could also lead to security
6202 consequences. This is also anticipated to be rare.
6204 This issue was reported to OpenSSL by Guido Vranken.
6209 * Fix memory issues in `BIO_*printf` functions
6211 The internal `fmtstr` function used in processing a "%s" format string in
6212 the `BIO_*printf` functions could overflow while calculating the length of a
6213 string and cause an OOB read when printing very long strings.
6215 Additionally the internal `doapr_outch` function can attempt to write to an
6216 OOB memory location (at an offset from the NULL pointer) in the event of a
6217 memory allocation failure. In 1.0.2 and below this could be caused where
6218 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6219 could be in processing a very long "%s" format string. Memory leaks can
6222 The first issue may mask the second issue dependent on compiler behaviour.
6223 These problems could enable attacks where large amounts of untrusted data
6224 is passed to the `BIO_*printf` functions. If applications use these functions
6225 in this way then they could be vulnerable. OpenSSL itself uses these
6226 functions when printing out human-readable dumps of ASN.1 data. Therefore
6227 applications that print this data could be vulnerable if the data is from
6228 untrusted sources. OpenSSL command line applications could also be
6229 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6230 as command line arguments.
6232 Libssl is not considered directly vulnerable. Additionally certificates etc
6233 received via remote connections via libssl are also unlikely to be able to
6234 trigger these issues because of message size limits enforced within libssl.
6236 This issue was reported to OpenSSL Guido Vranken.
6241 * Side channel attack on modular exponentiation
6243 A side-channel attack was found which makes use of cache-bank conflicts on
6244 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6245 of RSA keys. The ability to exploit this issue is limited as it relies on
6246 an attacker who has control of code in a thread running on the same
6247 hyper-threaded core as the victim thread which is performing decryptions.
6249 This issue was reported to OpenSSL by Yuval Yarom, The University of
6250 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6251 Nadia Heninger, University of Pennsylvania with more information at
6252 <http://cachebleed.info>.
6257 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6258 if no keysize is specified with default_bits. This fixes an
6259 omission in an earlier change that changed all RSA/DSA key generation
6260 commands to use 2048 bits by default.
6264 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6266 * Protection for DH small subgroup attacks
6268 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6269 switched on by default and cannot be disabled. This could have some
6274 * SSLv2 doesn't block disabled ciphers
6276 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6277 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6278 been disabled, provided that the SSLv2 protocol was not also disabled via
6281 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6282 and Sebastian Schinzel.
6287 * Reject DH handshakes with parameters shorter than 1024 bits.
6291 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6293 * Certificate verify crash with missing PSS parameter
6295 The signature verification routines will crash with a NULL pointer
6296 dereference if presented with an ASN.1 signature using the RSA PSS
6297 algorithm and absent mask generation function parameter. Since these
6298 routines are used to verify certificate signature algorithms this can be
6299 used to crash any certificate verification operation and exploited in a
6300 DoS attack. Any application which performs certificate verification is
6301 vulnerable including OpenSSL clients and servers which enable client
6304 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6309 * X509_ATTRIBUTE memory leak
6311 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6312 memory. This structure is used by the PKCS#7 and CMS routines so any
6313 application which reads PKCS#7 or CMS data from untrusted sources is
6314 affected. SSL/TLS is not affected.
6316 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6322 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6323 This changes the decoding behaviour for some invalid messages,
6324 though the change is mostly in the more lenient direction, and
6325 legacy behaviour is preserved as much as possible.
6329 * In DSA_generate_parameters_ex, if the provided seed is too short,
6330 use a random seed, as already documented.
6332 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6334 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6336 * Alternate chains certificate forgery
6338 During certificate verfification, OpenSSL will attempt to find an
6339 alternative certificate chain if the first attempt to build such a chain
6340 fails. An error in the implementation of this logic can mean that an
6341 attacker could cause certain checks on untrusted certificates to be
6342 bypassed, such as the CA flag, enabling them to use a valid leaf
6343 certificate to act as a CA and "issue" an invalid certificate.
6345 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6351 * Race condition handling PSK identify hint
6353 If PSK identity hints are received by a multi-threaded client then
6354 the values are wrongly updated in the parent SSL_CTX structure. This can
6355 result in a race condition potentially leading to a double free of the
6361 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6363 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6364 incompatibility in the handling of HMAC. The previous ABI has now been
6367 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6369 * Malformed ECParameters causes infinite loop
6371 When processing an ECParameters structure OpenSSL enters an infinite loop
6372 if the curve specified is over a specially malformed binary polynomial
6375 This can be used to perform denial of service against any
6376 system which processes public keys, certificate requests or
6377 certificates. This includes TLS clients and TLS servers with
6378 client authentication enabled.
6380 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6385 * Exploitable out-of-bounds read in X509_cmp_time
6387 X509_cmp_time does not properly check the length of the ASN1_TIME
6388 string and can read a few bytes out of bounds. In addition,
6389 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6392 An attacker can use this to craft malformed certificates and CRLs of
6393 various sizes and potentially cause a segmentation fault, resulting in
6394 a DoS on applications that verify certificates or CRLs. TLS clients
6395 that verify CRLs are affected. TLS clients and servers with client
6396 authentication enabled may be affected if they use custom verification
6399 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6400 independently by Hanno Böck.
6405 * PKCS7 crash with missing EnvelopedContent
6407 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6408 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6409 with missing content and trigger a NULL pointer dereference on parsing.
6411 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6412 structures from untrusted sources are affected. OpenSSL clients and
6413 servers are not affected.
6415 This issue was reported to OpenSSL by Michal Zalewski (Google).
6420 * CMS verify infinite loop with unknown hash function
6422 When verifying a signedData message the CMS code can enter an infinite loop
6423 if presented with an unknown hash function OID. This can be used to perform
6424 denial of service against any system which verifies signedData messages using
6426 This issue was reported to OpenSSL by Johannes Bauer.
6431 * Race condition handling NewSessionTicket
6433 If a NewSessionTicket is received by a multi-threaded client when attempting to
6434 reuse a previous ticket then a race condition can occur potentially leading to
6435 a double free of the ticket data.
6440 * Reject DH handshakes with parameters shorter than 768 bits.
6442 *Kurt Roeckx and Emilia Kasper*
6444 * dhparam: generate 2048-bit parameters by default.
6446 *Kurt Roeckx and Emilia Kasper*
6448 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6450 * Segmentation fault in ASN1_TYPE_cmp fix
6452 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6453 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6454 certificate signature algorithm consistency this can be used to crash any
6455 certificate verification operation and exploited in a DoS attack. Any
6456 application which performs certificate verification is vulnerable including
6457 OpenSSL clients and servers which enable client authentication.
6462 * ASN.1 structure reuse memory corruption fix
6464 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6465 memory corruption via an invalid write. Such reuse is and has been
6466 strongly discouraged and is believed to be rare.
6468 Applications that parse structures containing CHOICE or ANY DEFINED BY
6469 components may be affected. Certificate parsing (d2i_X509 and related
6470 functions) are however not affected. OpenSSL clients and servers are
6476 * PKCS7 NULL pointer dereferences fix
6478 The PKCS#7 parsing code does not handle missing outer ContentInfo
6479 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6480 missing content and trigger a NULL pointer dereference on parsing.
6482 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6483 otherwise parse PKCS#7 structures from untrusted sources are
6484 affected. OpenSSL clients and servers are not affected.
6486 This issue was reported to OpenSSL by Michal Zalewski (Google).
6491 * DoS via reachable assert in SSLv2 servers fix
6493 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6494 servers that both support SSLv2 and enable export cipher suites by sending
6495 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6497 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6498 (OpenSSL development team).
6503 * Use After Free following d2i_ECPrivatekey error fix
6505 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6506 could cause a use after free condition. This, in turn, could cause a double
6507 free in several private key parsing functions (such as d2i_PrivateKey
6508 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6509 for applications that receive EC private keys from untrusted
6510 sources. This scenario is considered rare.
6512 This issue was discovered by the BoringSSL project and fixed in their
6518 * X509_to_X509_REQ NULL pointer deref fix
6520 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6521 the certificate key is invalid. This function is rarely used in practice.
6523 This issue was discovered by Brian Carpenter.
6528 * Removed the export ciphers from the DEFAULT ciphers
6532 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6534 * Build fixes for the Windows and OpenVMS platforms
6536 *Matt Caswell and Richard Levitte*
6538 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6540 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6541 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6542 dereference. This could lead to a Denial Of Service attack. Thanks to
6543 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6548 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6549 dtls1_buffer_record function under certain conditions. In particular this
6550 could occur if an attacker sent repeated DTLS records with the same
6551 sequence number but for the next epoch. The memory leak could be exploited
6552 by an attacker in a Denial of Service attack through memory exhaustion.
6553 Thanks to Chris Mueller for reporting this issue.
6558 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6559 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6560 method would be set to NULL which could later result in a NULL pointer
6561 dereference. Thanks to Frank Schmirler for reporting this issue.
6566 * Abort handshake if server key exchange message is omitted for ephemeral
6569 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6570 reporting this issue.
6575 * Remove non-export ephemeral RSA code on client and server. This code
6576 violated the TLS standard by allowing the use of temporary RSA keys in
6577 non-export ciphersuites and could be used by a server to effectively
6578 downgrade the RSA key length used to a value smaller than the server
6579 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6580 INRIA or reporting this issue.
6585 * Fixed issue where DH client certificates are accepted without verification.
6586 An OpenSSL server will accept a DH certificate for client authentication
6587 without the certificate verify message. This effectively allows a client to
6588 authenticate without the use of a private key. This only affects servers
6589 which trust a client certificate authority which issues certificates
6590 containing DH keys: these are extremely rare and hardly ever encountered.
6591 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6597 * Ensure that the session ID context of an SSL is updated when its
6598 SSL_CTX is updated via SSL_set_SSL_CTX.
6600 The session ID context is typically set from the parent SSL_CTX,
6601 and can vary with the CTX.
6605 * Fix various certificate fingerprint issues.
6607 By using non-DER or invalid encodings outside the signed portion of a
6608 certificate the fingerprint can be changed without breaking the signature.
6609 Although no details of the signed portion of the certificate can be changed
6610 this can cause problems with some applications: e.g. those using the
6611 certificate fingerprint for blacklists.
6613 1. Reject signatures with non zero unused bits.
6615 If the BIT STRING containing the signature has non zero unused bits reject
6616 the signature. All current signature algorithms require zero unused bits.
6618 2. Check certificate algorithm consistency.
6620 Check the AlgorithmIdentifier inside TBS matches the one in the
6621 certificate signature. NB: this will result in signature failure
6622 errors for some broken certificates.
6624 Thanks to Konrad Kraszewski from Google for reporting this issue.
6626 3. Check DSA/ECDSA signatures use DER.
6628 Re-encode DSA/ECDSA signatures and compare with the original received
6629 signature. Return an error if there is a mismatch.
6631 This will reject various cases including garbage after signature
6632 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6633 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6634 (negative or with leading zeroes).
6636 Further analysis was conducted and fixes were developed by Stephen Henson
6637 of the OpenSSL core team.
6643 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6644 results on some platforms, including x86_64. This bug occurs at random
6645 with a very low probability, and is not known to be exploitable in any
6646 way, though its exact impact is difficult to determine. Thanks to Pieter
6647 Wuille (Blockstream) who reported this issue and also suggested an initial
6648 fix. Further analysis was conducted by the OpenSSL development team and
6649 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6650 the OpenSSL core team.
6655 * Do not resume sessions on the server if the negotiated protocol
6656 version does not match the session's version. Resuming with a different
6657 version, while not strictly forbidden by the RFC, is of questionable
6658 sanity and breaks all known clients.
6660 *David Benjamin, Emilia Käsper*
6662 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6663 early CCS messages during renegotiation. (Note that because
6664 renegotiation is encrypted, this early CCS was not exploitable.)
6668 * Tighten client-side session ticket handling during renegotiation:
6669 ensure that the client only accepts a session ticket if the server sends
6670 the extension anew in the ServerHello. Previously, a TLS client would
6671 reuse the old extension state and thus accept a session ticket if one was
6672 announced in the initial ServerHello.
6674 Similarly, ensure that the client requires a session ticket if one
6675 was advertised in the ServerHello. Previously, a TLS client would
6676 ignore a missing NewSessionTicket message.
6680 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6684 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6685 sends a carefully crafted handshake message, to cause OpenSSL to fail
6686 to free up to 64k of memory causing a memory leak. This could be
6687 exploited in a Denial Of Service attack. This issue affects OpenSSL
6688 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6689 whether SRTP is used or configured. Implementations of OpenSSL that
6690 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6692 The fix was developed by the OpenSSL team.
6697 * Session Ticket Memory Leak.
6699 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6700 integrity of that ticket is first verified. In the event of a session
6701 ticket integrity check failing, OpenSSL will fail to free memory
6702 causing a memory leak. By sending a large number of invalid session
6703 tickets an attacker could exploit this issue in a Denial Of Service
6709 * Build option no-ssl3 is incomplete.
6711 When OpenSSL is configured with "no-ssl3" as a build option, servers
6712 could accept and complete a SSL 3.0 handshake, and clients could be
6713 configured to send them.
6716 *Akamai and the OpenSSL team*
6718 * Add support for TLS_FALLBACK_SCSV.
6719 Client applications doing fallback retries should call
6720 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6723 *Adam Langley, Bodo Moeller*
6725 * Add additional DigestInfo checks.
6727 Re-encode DigestInto in DER and check against the original when
6728 verifying RSA signature: this will reject any improperly encoded
6729 DigestInfo structures.
6731 Note: this is a precautionary measure and no attacks are currently known.
6735 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6737 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6738 SRP code can be overrun an internal buffer. Add sanity check that
6739 g, A, B < N to SRP code.
6741 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6742 Group for discovering this issue.
6747 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6748 TLS 1.0 instead of higher protocol versions when the ClientHello message
6749 is badly fragmented. This allows a man-in-the-middle attacker to force a
6750 downgrade to TLS 1.0 even if both the server and the client support a
6751 higher protocol version, by modifying the client's TLS records.
6753 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6754 researching this issue.
6759 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6760 to a denial of service attack. A malicious server can crash the client
6761 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6762 ciphersuite and sending carefully crafted handshake messages.
6764 Thanks to Felix Gröbert (Google) for discovering and researching this
6770 * By sending carefully crafted DTLS packets an attacker could cause openssl
6771 to leak memory. This can be exploited through a Denial of Service attack.
6772 Thanks to Adam Langley for discovering and researching this issue.
6777 * An attacker can force openssl to consume large amounts of memory whilst
6778 processing DTLS handshake messages. This can be exploited through a
6779 Denial of Service attack.
6780 Thanks to Adam Langley for discovering and researching this issue.
6785 * An attacker can force an error condition which causes openssl to crash
6786 whilst processing DTLS packets due to memory being freed twice. This
6787 can be exploited through a Denial of Service attack.
6788 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6794 * If a multithreaded client connects to a malicious server using a resumed
6795 session and the server sends an ec point format extension it could write
6796 up to 255 bytes to freed memory.
6798 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6804 * A malicious server can crash an OpenSSL client with a null pointer
6805 dereference (read) by specifying an SRP ciphersuite even though it was not
6806 properly negotiated with the client. This can be exploited through a
6807 Denial of Service attack.
6809 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6810 discovering and researching this issue.
6815 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6816 X509_name_oneline, X509_name_print_ex et al. to leak some information
6817 from the stack. Applications may be affected if they echo pretty printing
6818 output to the attacker.
6820 Thanks to Ivan Fratric (Google) for discovering this issue.
6823 *Emilia Käsper, and Steve Henson*
6825 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6826 for corner cases. (Certain input points at infinity could lead to
6827 bogus results, with non-infinity inputs mapped to infinity too.)
6831 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6833 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6834 handshake can force the use of weak keying material in OpenSSL
6835 SSL/TLS clients and servers.
6837 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6838 researching this issue. ([CVE-2014-0224])
6840 *KIKUCHI Masashi, Steve Henson*
6842 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6843 OpenSSL DTLS client the code can be made to recurse eventually crashing
6846 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6849 *Imre Rad, Steve Henson*
6851 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6852 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6853 client or server. This is potentially exploitable to run arbitrary
6854 code on a vulnerable client or server.
6856 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6858 *Jüri Aedla, Steve Henson*
6860 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6861 are subject to a denial of service attack.
6863 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6864 this issue. ([CVE-2014-3470])
6866 *Felix Gröbert, Ivan Fratric, Steve Henson*
6868 * Harmonize version and its documentation. -f flag is used to display
6871 *mancha <mancha1@zoho.com>*
6873 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6874 in i2d_ECPrivateKey.
6876 *mancha <mancha1@zoho.com>*
6878 * Fix some double frees. These are not thought to be exploitable.
6880 *mancha <mancha1@zoho.com>*
6882 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6884 * A missing bounds check in the handling of the TLS heartbeat extension
6885 can be used to reveal up to 64k of memory to a connected client or
6888 Thanks for Neel Mehta of Google Security for discovering this bug and to
6889 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6890 preparing the fix ([CVE-2014-0160])
6892 *Adam Langley, Bodo Moeller*
6894 * Fix for the attack described in the paper "Recovering OpenSSL
6895 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6896 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6897 <http://eprint.iacr.org/2014/140>
6899 Thanks to Yuval Yarom and Naomi Benger for discovering this
6900 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6902 *Yuval Yarom and Naomi Benger*
6904 * TLS pad extension: draft-agl-tls-padding-03
6906 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6907 TLS client Hello record length value would otherwise be > 255 and
6908 less that 512 pad with a dummy extension containing zeroes so it
6909 is at least 512 bytes long.
6911 *Adam Langley, Steve Henson*
6913 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6915 * Fix for TLS record tampering bug. A carefully crafted invalid
6916 handshake could crash OpenSSL with a NULL pointer exception.
6917 Thanks to Anton Johansson for reporting this issues.
6920 * Keep original DTLS digest and encryption contexts in retransmission
6921 structures so we can use the previous session parameters if they need
6922 to be resent. ([CVE-2013-6450])
6926 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6927 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6928 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6929 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6930 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6931 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6933 *Rob Stradling, Adam Langley*
6935 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6937 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6938 supporting platforms or when small records were transferred.
6940 *Andy Polyakov, Steve Henson*
6942 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6944 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6946 This addresses the flaw in CBC record processing discovered by
6947 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6948 at: <http://www.isg.rhul.ac.uk/tls/>
6950 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6951 Security Group at Royal Holloway, University of London
6952 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6953 Emilia Käsper for the initial patch.
6956 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6958 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6959 ciphersuites which can be exploited in a denial of service attack.
6960 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6961 and detecting this bug and to Wolfgang Ettlinger
6962 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6967 * Return an error when checking OCSP signatures when key is NULL.
6968 This fixes a DoS attack. ([CVE-2013-0166])
6972 * Make openssl verify return errors.
6974 *Chris Palmer <palmer@google.com> and Ben Laurie*
6976 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6977 the right response is stapled. Also change SSL_get_certificate()
6978 so it returns the certificate actually sent.
6979 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6981 *Rob Stradling <rob.stradling@comodo.com>*
6983 * Fix possible deadlock when decoding public keys.
6987 * Don't use TLS 1.0 record version number in initial client hello
6992 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6994 * Sanity check record length before skipping explicit IV in TLS
6995 1.2, 1.1 and DTLS to fix DoS attack.
6997 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6998 fuzzing as a service testing platform.
7003 * Initialise tkeylen properly when encrypting CMS messages.
7004 Thanks to Solar Designer of Openwall for reporting this issue.
7008 * In FIPS mode don't try to use composite ciphers as they are not
7013 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7015 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7016 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7017 mean any application compiled against OpenSSL 1.0.0 headers setting
7018 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7019 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7020 0x10000000L Any application which was previously compiled against
7021 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7022 will need to be recompiled as a result. Letting be results in
7023 inability to disable specifically TLS 1.1 and in client context,
7024 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7028 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7029 disable just protocol X, but all protocols above X *if* there are
7030 protocols *below* X still enabled. In more practical terms it means
7031 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7032 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7033 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7038 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7040 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7041 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7042 in CRYPTO_realloc_clean.
7044 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7045 issue and to Adam Langley <agl@chromium.org> for fixing it.
7048 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7050 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7054 * Workarounds for some broken servers that "hang" if a client hello
7055 record length exceeds 255 bytes.
7057 1. Do not use record version number > TLS 1.0 in initial client
7058 hello: some (but not all) hanging servers will now work.
7059 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7060 the number of ciphers sent in the client hello. This should be
7061 set to an even number, such as 50, for example by passing:
7062 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7063 Most broken servers should now work.
7064 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7065 TLS 1.2 client support entirely.
7069 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7073 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7075 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7076 STRING form instead of a DigestInfo.
7080 * The format used for MDC2 RSA signatures is inconsistent between EVP
7081 and the RSA_sign/RSA_verify functions. This was made more apparent when
7082 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7083 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7084 the correct format in RSA_verify so both forms transparently work.
7088 * Some servers which support TLS 1.0 can choke if we initially indicate
7089 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7090 encrypted premaster secret. As a workaround use the maximum permitted
7091 client version in client hello, this should keep such servers happy
7092 and still work with previous versions of OpenSSL.
7096 * Add support for TLS/DTLS heartbeats.
7098 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7100 * Add support for SCTP.
7102 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7104 * Improved PRNG seeding for VOS.
7106 *Paul Green <Paul.Green@stratus.com>*
7108 * Extensive assembler packs updates, most notably:
7110 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7111 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7112 - x86_64: bit-sliced AES implementation;
7113 - ARM: NEON support, contemporary platforms optimizations;
7114 - s390x: z196 support;
7115 - `*`: GHASH and GF(2^m) multiplication implementations;
7119 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7120 (removal of unnecessary code)
7122 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7124 * Add TLS key material exporter from RFC 5705.
7128 * Add DTLS-SRTP negotiation from RFC 5764.
7132 * Add Next Protocol Negotiation,
7133 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7134 disabled with a no-npn flag to config or Configure. Code donated
7137 *Adam Langley <agl@google.com> and Ben Laurie*
7139 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7140 NIST-P256, NIST-P521, with constant-time single point multiplication on
7141 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7142 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7143 Code made available under Apache License version 2.0.
7145 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7146 line to include this in your build of OpenSSL, and run "make depend" (or
7147 "make update"). This enables the following EC_METHODs:
7149 EC_GFp_nistp224_method()
7150 EC_GFp_nistp256_method()
7151 EC_GFp_nistp521_method()
7153 EC_GROUP_new_by_curve_name() will automatically use these (while
7154 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7157 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7159 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7160 all platforms. Move ssize_t definition from e_os.h to the public
7161 header file e_os2.h as it now appears in public header file cms.h
7165 * New -sigopt option to the ca, req and x509 utilities. Additional
7166 signature parameters can be passed using this option and in
7171 * Add RSA PSS signing function. This will generate and set the
7172 appropriate AlgorithmIdentifiers for PSS based on those in the
7173 corresponding EVP_MD_CTX structure. No application support yet.
7177 * Support for companion algorithm specific ASN1 signing routines.
7178 New function ASN1_item_sign_ctx() signs a pre-initialised
7179 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7180 the appropriate parameters.
7184 * Add new algorithm specific ASN1 verification initialisation function
7185 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7186 handling will be the same no matter what EVP_PKEY_METHOD is used.
7187 Add a PSS handler to support verification of PSS signatures: checked
7188 against a number of sample certificates.
7192 * Add signature printing for PSS. Add PSS OIDs.
7194 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7196 * Add algorithm specific signature printing. An individual ASN1 method
7197 can now print out signatures instead of the standard hex dump.
7199 More complex signatures (e.g. PSS) can print out more meaningful
7200 information. Include DSA version that prints out the signature
7205 * Password based recipient info support for CMS library: implementing
7210 * Split password based encryption into PBES2 and PBKDF2 functions. This
7211 neatly separates the code into cipher and PBE sections and is required
7212 for some algorithms that split PBES2 into separate pieces (such as
7213 password based CMS).
7217 * Session-handling fixes:
7218 - Fix handling of connections that are resuming with a session ID,
7219 but also support Session Tickets.
7220 - Fix a bug that suppressed issuing of a new ticket if the client
7221 presented a ticket with an expired session.
7222 - Try to set the ticket lifetime hint to something reasonable.
7223 - Make tickets shorter by excluding irrelevant information.
7224 - On the client side, don't ignore renewed tickets.
7226 *Adam Langley, Bodo Moeller (Google)*
7228 * Fix PSK session representation.
7232 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7234 This work was sponsored by Intel.
7238 * Add GCM support to TLS library. Some custom code is needed to split
7239 the IV between the fixed (from PRF) and explicit (from TLS record)
7240 portions. This adds all GCM ciphersuites supported by RFC5288 and
7241 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7242 add a special AESGCM string for GCM only.
7246 * Expand range of ctrls for AES GCM. Permit setting invocation
7247 field on decrypt and retrieval of invocation field only on encrypt.
7251 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7252 As required by RFC5289 these ciphersuites cannot be used if for
7253 versions of TLS earlier than 1.2.
7257 * For FIPS capable OpenSSL interpret a NULL default public key method
7258 as unset and return the appropriate default but do *not* set the default.
7259 This means we can return the appropriate method in applications that
7260 switch between FIPS and non-FIPS modes.
7264 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7265 ENGINE is used then we cannot handle that in the FIPS module so we
7266 keep original code iff non-FIPS operations are allowed.
7270 * Add -attime option to openssl utilities.
7272 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7274 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7278 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7279 FIPS EC methods unconditionally for now.
7283 * New build option no-ec2m to disable characteristic 2 code.
7287 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7288 all cases can be covered as some introduce binary incompatibilities.
7292 * Redirect RSA operations to FIPS module including keygen,
7293 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7297 * Add similar low-level API blocking to ciphers.
7301 * low-level digest APIs are not approved in FIPS mode: any attempt
7302 to use these will cause a fatal error. Applications that *really* want
7303 to use them can use the `private_*` version instead.
7307 * Redirect cipher operations to FIPS module for FIPS builds.
7311 * Redirect digest operations to FIPS module for FIPS builds.
7315 * Update build system to add "fips" flag which will link in fipscanister.o
7316 for static and shared library builds embedding a signature if needed.
7320 * Output TLS supported curves in preference order instead of numerical
7321 order. This is currently hardcoded for the highest order curves first.
7322 This should be configurable so applications can judge speed vs strength.
7326 * Add TLS v1.2 server support for client authentication.
7330 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7335 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7336 FIPS modules versions.
7340 * Add TLS v1.2 client side support for client authentication. Keep cache
7341 of handshake records longer as we don't know the hash algorithm to use
7342 until after the certificate request message is received.
7346 * Initial TLS v1.2 client support. Add a default signature algorithms
7347 extension including all the algorithms we support. Parse new signature
7348 format in client key exchange. Relax some ECC signing restrictions for
7349 TLS v1.2 as indicated in RFC5246.
7353 * Add server support for TLS v1.2 signature algorithms extension. Switch
7354 to new signature format when needed using client digest preference.
7355 All server ciphersuites should now work correctly in TLS v1.2. No client
7356 support yet and no support for client certificates.
7360 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7361 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7362 ciphersuites. At present only RSA key exchange ciphersuites work with
7363 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7364 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7365 and version checking.
7369 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7370 with this defined it will not be affected by any changes to ssl internal
7371 structures. Add several utility functions to allow openssl application
7372 to work with OPENSSL_NO_SSL_INTERN defined.
7376 * A long standing patch to add support for SRP from EdelWeb (Peter
7377 Sylvester and Christophe Renou) was integrated.
7378 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7379 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7382 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7386 * Permit abbreviated handshakes when renegotiating using the function
7387 SSL_renegotiate_abbreviated().
7389 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7391 * Add call to ENGINE_register_all_complete() to
7392 ENGINE_load_builtin_engines(), so some implementations get used
7393 automatically instead of needing explicit application support.
7397 * Add support for TLS key exporter as described in RFC5705.
7399 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7401 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7402 a few changes are required:
7404 Add SSL_OP_NO_TLSv1_1 flag.
7405 Add TLSv1_1 methods.
7406 Update version checking logic to handle version 1.1.
7407 Add explicit IV handling (ported from DTLS code).
7408 Add command line options to s_client/s_server.
7415 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7417 * X509_ATTRIBUTE memory leak
7419 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7420 memory. This structure is used by the PKCS#7 and CMS routines so any
7421 application which reads PKCS#7 or CMS data from untrusted sources is
7422 affected. SSL/TLS is not affected.
7424 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7430 * Race condition handling PSK identify hint
7432 If PSK identity hints are received by a multi-threaded client then
7433 the values are wrongly updated in the parent SSL_CTX structure. This can
7434 result in a race condition potentially leading to a double free of the
7440 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7442 * Malformed ECParameters causes infinite loop
7444 When processing an ECParameters structure OpenSSL enters an infinite loop
7445 if the curve specified is over a specially malformed binary polynomial
7448 This can be used to perform denial of service against any
7449 system which processes public keys, certificate requests or
7450 certificates. This includes TLS clients and TLS servers with
7451 client authentication enabled.
7453 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7458 * Exploitable out-of-bounds read in X509_cmp_time
7460 X509_cmp_time does not properly check the length of the ASN1_TIME
7461 string and can read a few bytes out of bounds. In addition,
7462 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7465 An attacker can use this to craft malformed certificates and CRLs of
7466 various sizes and potentially cause a segmentation fault, resulting in
7467 a DoS on applications that verify certificates or CRLs. TLS clients
7468 that verify CRLs are affected. TLS clients and servers with client
7469 authentication enabled may be affected if they use custom verification
7472 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7473 independently by Hanno Böck.
7478 * PKCS7 crash with missing EnvelopedContent
7480 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7481 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7482 with missing content and trigger a NULL pointer dereference on parsing.
7484 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7485 structures from untrusted sources are affected. OpenSSL clients and
7486 servers are not affected.
7488 This issue was reported to OpenSSL by Michal Zalewski (Google).
7493 * CMS verify infinite loop with unknown hash function
7495 When verifying a signedData message the CMS code can enter an infinite loop
7496 if presented with an unknown hash function OID. This can be used to perform
7497 denial of service against any system which verifies signedData messages using
7499 This issue was reported to OpenSSL by Johannes Bauer.
7504 * Race condition handling NewSessionTicket
7506 If a NewSessionTicket is received by a multi-threaded client when attempting to
7507 reuse a previous ticket then a race condition can occur potentially leading to
7508 a double free of the ticket data.
7513 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7515 * Segmentation fault in ASN1_TYPE_cmp fix
7517 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7518 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7519 certificate signature algorithm consistency this can be used to crash any
7520 certificate verification operation and exploited in a DoS attack. Any
7521 application which performs certificate verification is vulnerable including
7522 OpenSSL clients and servers which enable client authentication.
7527 * ASN.1 structure reuse memory corruption fix
7529 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7530 memory corruption via an invalid write. Such reuse is and has been
7531 strongly discouraged and is believed to be rare.
7533 Applications that parse structures containing CHOICE or ANY DEFINED BY
7534 components may be affected. Certificate parsing (d2i_X509 and related
7535 functions) are however not affected. OpenSSL clients and servers are
7541 * PKCS7 NULL pointer dereferences fix
7543 The PKCS#7 parsing code does not handle missing outer ContentInfo
7544 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7545 missing content and trigger a NULL pointer dereference on parsing.
7547 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7548 otherwise parse PKCS#7 structures from untrusted sources are
7549 affected. OpenSSL clients and servers are not affected.
7551 This issue was reported to OpenSSL by Michal Zalewski (Google).
7556 * DoS via reachable assert in SSLv2 servers fix
7558 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7559 servers that both support SSLv2 and enable export cipher suites by sending
7560 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7562 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7563 (OpenSSL development team).
7568 * Use After Free following d2i_ECPrivatekey error fix
7570 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7571 could cause a use after free condition. This, in turn, could cause a double
7572 free in several private key parsing functions (such as d2i_PrivateKey
7573 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7574 for applications that receive EC private keys from untrusted
7575 sources. This scenario is considered rare.
7577 This issue was discovered by the BoringSSL project and fixed in their
7583 * X509_to_X509_REQ NULL pointer deref fix
7585 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7586 the certificate key is invalid. This function is rarely used in practice.
7588 This issue was discovered by Brian Carpenter.
7593 * Removed the export ciphers from the DEFAULT ciphers
7597 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7599 * Build fixes for the Windows and OpenVMS platforms
7601 *Matt Caswell and Richard Levitte*
7603 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7605 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7606 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7607 dereference. This could lead to a Denial Of Service attack. Thanks to
7608 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7613 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7614 dtls1_buffer_record function under certain conditions. In particular this
7615 could occur if an attacker sent repeated DTLS records with the same
7616 sequence number but for the next epoch. The memory leak could be exploited
7617 by an attacker in a Denial of Service attack through memory exhaustion.
7618 Thanks to Chris Mueller for reporting this issue.
7623 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7624 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7625 method would be set to NULL which could later result in a NULL pointer
7626 dereference. Thanks to Frank Schmirler for reporting this issue.
7631 * Abort handshake if server key exchange message is omitted for ephemeral
7634 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7635 reporting this issue.
7640 * Remove non-export ephemeral RSA code on client and server. This code
7641 violated the TLS standard by allowing the use of temporary RSA keys in
7642 non-export ciphersuites and could be used by a server to effectively
7643 downgrade the RSA key length used to a value smaller than the server
7644 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7645 INRIA or reporting this issue.
7650 * Fixed issue where DH client certificates are accepted without verification.
7651 An OpenSSL server will accept a DH certificate for client authentication
7652 without the certificate verify message. This effectively allows a client to
7653 authenticate without the use of a private key. This only affects servers
7654 which trust a client certificate authority which issues certificates
7655 containing DH keys: these are extremely rare and hardly ever encountered.
7656 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7662 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7663 results on some platforms, including x86_64. This bug occurs at random
7664 with a very low probability, and is not known to be exploitable in any
7665 way, though its exact impact is difficult to determine. Thanks to Pieter
7666 Wuille (Blockstream) who reported this issue and also suggested an initial
7667 fix. Further analysis was conducted by the OpenSSL development team and
7668 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7669 the OpenSSL core team.
7674 * Fix various certificate fingerprint issues.
7676 By using non-DER or invalid encodings outside the signed portion of a
7677 certificate the fingerprint can be changed without breaking the signature.
7678 Although no details of the signed portion of the certificate can be changed
7679 this can cause problems with some applications: e.g. those using the
7680 certificate fingerprint for blacklists.
7682 1. Reject signatures with non zero unused bits.
7684 If the BIT STRING containing the signature has non zero unused bits reject
7685 the signature. All current signature algorithms require zero unused bits.
7687 2. Check certificate algorithm consistency.
7689 Check the AlgorithmIdentifier inside TBS matches the one in the
7690 certificate signature. NB: this will result in signature failure
7691 errors for some broken certificates.
7693 Thanks to Konrad Kraszewski from Google for reporting this issue.
7695 3. Check DSA/ECDSA signatures use DER.
7697 Reencode DSA/ECDSA signatures and compare with the original received
7698 signature. Return an error if there is a mismatch.
7700 This will reject various cases including garbage after signature
7701 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7702 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7703 (negative or with leading zeroes).
7705 Further analysis was conducted and fixes were developed by Stephen Henson
7706 of the OpenSSL core team.
7712 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7714 * Session Ticket Memory Leak.
7716 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7717 integrity of that ticket is first verified. In the event of a session
7718 ticket integrity check failing, OpenSSL will fail to free memory
7719 causing a memory leak. By sending a large number of invalid session
7720 tickets an attacker could exploit this issue in a Denial Of Service
7726 * Build option no-ssl3 is incomplete.
7728 When OpenSSL is configured with "no-ssl3" as a build option, servers
7729 could accept and complete a SSL 3.0 handshake, and clients could be
7730 configured to send them.
7733 *Akamai and the OpenSSL team*
7735 * Add support for TLS_FALLBACK_SCSV.
7736 Client applications doing fallback retries should call
7737 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7740 *Adam Langley, Bodo Moeller*
7742 * Add additional DigestInfo checks.
7744 Reencode DigestInto in DER and check against the original when
7745 verifying RSA signature: this will reject any improperly encoded
7746 DigestInfo structures.
7748 Note: this is a precautionary measure and no attacks are currently known.
7752 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7754 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7755 to a denial of service attack. A malicious server can crash the client
7756 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7757 ciphersuite and sending carefully crafted handshake messages.
7759 Thanks to Felix Gröbert (Google) for discovering and researching this
7765 * By sending carefully crafted DTLS packets an attacker could cause openssl
7766 to leak memory. This can be exploited through a Denial of Service attack.
7767 Thanks to Adam Langley for discovering and researching this issue.
7772 * An attacker can force openssl to consume large amounts of memory whilst
7773 processing DTLS handshake messages. This can be exploited through a
7774 Denial of Service attack.
7775 Thanks to Adam Langley for discovering and researching this issue.
7780 * An attacker can force an error condition which causes openssl to crash
7781 whilst processing DTLS packets due to memory being freed twice. This
7782 can be exploited through a Denial of Service attack.
7783 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7789 * If a multithreaded client connects to a malicious server using a resumed
7790 session and the server sends an ec point format extension it could write
7791 up to 255 bytes to freed memory.
7793 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7799 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7800 X509_name_oneline, X509_name_print_ex et al. to leak some information
7801 from the stack. Applications may be affected if they echo pretty printing
7802 output to the attacker.
7804 Thanks to Ivan Fratric (Google) for discovering this issue.
7807 *Emilia Käsper, and Steve Henson*
7809 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7810 for corner cases. (Certain input points at infinity could lead to
7811 bogus results, with non-infinity inputs mapped to infinity too.)
7815 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7817 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7818 handshake can force the use of weak keying material in OpenSSL
7819 SSL/TLS clients and servers.
7821 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7822 researching this issue. ([CVE-2014-0224])
7824 *KIKUCHI Masashi, Steve Henson*
7826 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7827 OpenSSL DTLS client the code can be made to recurse eventually crashing
7830 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7833 *Imre Rad, Steve Henson*
7835 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7836 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7837 client or server. This is potentially exploitable to run arbitrary
7838 code on a vulnerable client or server.
7840 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7842 *Jüri Aedla, Steve Henson*
7844 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7845 are subject to a denial of service attack.
7847 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7848 this issue. ([CVE-2014-3470])
7850 *Felix Gröbert, Ivan Fratric, Steve Henson*
7852 * Harmonize version and its documentation. -f flag is used to display
7855 *mancha <mancha1@zoho.com>*
7857 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7858 in i2d_ECPrivateKey.
7860 *mancha <mancha1@zoho.com>*
7862 * Fix some double frees. These are not thought to be exploitable.
7864 *mancha <mancha1@zoho.com>*
7866 * Fix for the attack described in the paper "Recovering OpenSSL
7867 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7868 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7869 <http://eprint.iacr.org/2014/140>
7871 Thanks to Yuval Yarom and Naomi Benger for discovering this
7872 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7874 *Yuval Yarom and Naomi Benger*
7876 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7878 * Keep original DTLS digest and encryption contexts in retransmission
7879 structures so we can use the previous session parameters if they need
7880 to be resent. ([CVE-2013-6450])
7884 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7885 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7886 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7887 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7888 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7889 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7891 *Rob Stradling, Adam Langley*
7893 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7895 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7897 This addresses the flaw in CBC record processing discovered by
7898 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7899 at: <http://www.isg.rhul.ac.uk/tls/>
7901 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7902 Security Group at Royal Holloway, University of London
7903 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7904 Emilia Käsper for the initial patch.
7907 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7909 * Return an error when checking OCSP signatures when key is NULL.
7910 This fixes a DoS attack. ([CVE-2013-0166])
7914 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7915 the right response is stapled. Also change SSL_get_certificate()
7916 so it returns the certificate actually sent.
7917 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7918 (This is a backport)
7920 *Rob Stradling <rob.stradling@comodo.com>*
7922 * Fix possible deadlock when decoding public keys.
7926 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7928 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7931 * Sanity check record length before skipping explicit IV in DTLS
7934 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7935 fuzzing as a service testing platform.
7940 * Initialise tkeylen properly when encrypting CMS messages.
7941 Thanks to Solar Designer of Openwall for reporting this issue.
7945 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7947 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7948 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7949 in CRYPTO_realloc_clean.
7951 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7952 issue and to Adam Langley <agl@chromium.org> for fixing it.
7955 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7957 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7959 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7960 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7961 content decryption and always return the same error. Note: this attack
7962 needs on average 2^20 messages so it only affects automated senders. The
7963 old behaviour can be re-enabled in the CMS code by setting the
7964 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7965 an MMA defence is not necessary.
7966 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7967 this issue. ([CVE-2012-0884])
7971 * Fix CVE-2011-4619: make sure we really are receiving a
7972 client hello before rejecting multiple SGC restarts. Thanks to
7973 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7977 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7979 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7980 Thanks to Antonio Martin, Enterprise Secure Access Research and
7981 Development, Cisco Systems, Inc. for discovering this bug and
7982 preparing a fix. ([CVE-2012-0050])
7986 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7988 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7989 of the Vaudenay padding oracle attack on CBC mode encryption
7990 which enables an efficient plaintext recovery attack against
7991 the OpenSSL implementation of DTLS. Their attack exploits timing
7992 differences arising during decryption processing. A research
7993 paper describing this attack can be found at:
7994 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7995 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7996 Security Group at Royal Holloway, University of London
7997 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7998 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7999 for preparing the fix. ([CVE-2011-4108])
8001 *Robin Seggelmann, Michael Tuexen*
8003 * Clear bytes used for block padding of SSL 3.0 records.
8006 *Adam Langley (Google)*
8008 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8009 Kadianakis <desnacked@gmail.com> for discovering this issue and
8010 Adam Langley for preparing the fix. ([CVE-2011-4619])
8012 *Adam Langley (Google)*
8014 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8016 *Andrey Kulikov <amdeich@gmail.com>*
8018 * Prevent malformed RFC3779 data triggering an assertion failure.
8019 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8020 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8022 *Rob Austein <sra@hactrn.net>*
8024 * Improved PRNG seeding for VOS.
8026 *Paul Green <Paul.Green@stratus.com>*
8028 * Fix ssl_ciph.c set-up race.
8030 *Adam Langley (Google)*
8032 * Fix spurious failures in ecdsatest.c.
8034 *Emilia Käsper (Google)*
8036 * Fix the BIO_f_buffer() implementation (which was mixing different
8037 interpretations of the `..._len` fields).
8039 *Adam Langley (Google)*
8041 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8042 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8043 threads won't reuse the same blinding coefficients.
8045 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8046 lock to call BN_BLINDING_invert_ex, and avoids one use of
8047 BN_BLINDING_update for each BN_BLINDING structure (previously,
8048 the last update always remained unused).
8050 *Emilia Käsper (Google)*
8052 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8054 *Bob Buckholz (Google)*
8056 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8058 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8059 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8061 *Kaspar Brand <ossl@velox.ch>*
8063 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8064 for multi-threaded use of ECDH. ([CVE-2011-3210])
8066 *Adam Langley (Google)*
8068 * Fix x509_name_ex_d2i memory leak on bad inputs.
8072 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8073 signature public key algorithm by using OID xref utilities instead.
8074 Before this you could only use some ECC ciphersuites with SHA1 only.
8078 * Add protection against ECDSA timing attacks as mentioned in the paper
8079 by Billy Bob Brumley and Nicola Tuveri, see:
8080 <http://eprint.iacr.org/2011/232.pdf>
8082 *Billy Bob Brumley and Nicola Tuveri*
8084 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8086 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8088 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8090 * Fix bug in string printing code: if *any* escaping is enabled we must
8091 escape the escape character (backslash) or the resulting string is
8096 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8098 * Disable code workaround for ancient and obsolete Netscape browsers
8099 and servers: an attacker can use it in a ciphersuite downgrade attack.
8100 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8104 * Fixed J-PAKE implementation error, originally discovered by
8105 Sebastien Martini, further info and confirmation from Stefan
8106 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8110 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8112 * Fix extension code to avoid race conditions which can result in a buffer
8113 overrun vulnerability: resumed sessions must not be modified as they can
8114 be shared by multiple threads. CVE-2010-3864
8118 * Fix WIN32 build system to correctly link an ENGINE directory into
8123 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8125 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8128 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8130 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8132 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8133 context. The operation can be customised via the ctrl mechanism in
8134 case ENGINEs want to include additional functionality.
8138 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8142 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8143 output hashes compatible with older versions of OpenSSL.
8145 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8147 * Fix compression algorithm handling: if resuming a session use the
8148 compression algorithm of the resumed session instead of determining
8149 it from client hello again. Don't allow server to change algorithm.
8153 * Add load_crls() function to commands tidying load_certs() too. Add option
8154 to verify utility to allow additional CRLs to be included.
8158 * Update OCSP request code to permit adding custom headers to the request:
8159 some responders need this.
8163 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8166 *Julia Lawall <julia@diku.dk>*
8168 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8169 needlessly dereferenced structures, used obsolete functions and
8170 didn't handle all updated verify codes correctly.
8174 * Disable MD2 in the default configuration.
8178 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8179 indicate the initial BIO being pushed or popped. This makes it possible
8180 to determine whether the BIO is the one explicitly called or as a result
8181 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8182 it handles reference counts correctly and doesn't zero out the I/O bio
8183 when it is not being explicitly popped. WARNING: applications which
8184 included workarounds for the old buggy behaviour will need to be modified
8185 or they could free up already freed BIOs.
8189 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8190 renaming to all platforms (within the 0.9.8 branch, this was
8191 done conditionally on Netware platforms to avoid a name clash).
8193 *Guenter <lists@gknw.net>*
8195 * Add ECDHE and PSK support to DTLS.
8197 *Michael Tuexen <tuexen@fh-muenster.de>*
8199 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8204 * Add "missing" function EVP_MD_flags() (without this the only way to
8205 retrieve a digest flags is by accessing the structure directly. Update
8206 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8207 or cipher is registered as in the "from" argument. Print out all
8208 registered digests in the dgst usage message instead of manually
8209 attempting to work them out.
8213 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8214 this allows the use of compression and extensions. Change default cipher
8215 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8216 by default unless an application cipher string requests it.
8220 * Alter match criteria in PKCS12_parse(). It used to try to use local
8221 key ids to find matching certificates and keys but some PKCS#12 files
8222 don't follow the (somewhat unwritten) rules and this strategy fails.
8223 Now just gather all certificates together and the first private key
8224 then look for the first certificate that matches the key.
8228 * Support use of registered digest and cipher names for dgst and cipher
8229 commands instead of having to add each one as a special case. So now
8236 openssl dgst -sha256 foo
8238 and this works for ENGINE based algorithms too.
8242 * Update Gost ENGINE to support parameter files.
8244 *Victor B. Wagner <vitus@cryptocom.ru>*
8246 * Support GeneralizedTime in ca utility.
8248 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8250 * Enhance the hash format used for certificate directory links. The new
8251 form uses the canonical encoding (meaning equivalent names will work
8252 even if they aren't identical) and uses SHA1 instead of MD5. This form
8253 is incompatible with the older format and as a result c_rehash should
8254 be used to rebuild symbolic links.
8258 * Make PKCS#8 the default write format for private keys, replacing the
8259 traditional format. This form is standardised, more secure and doesn't
8260 include an implicit MD5 dependency.
8264 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8265 committed to OpenSSL should pass this lot as a minimum.
8269 * Add session ticket override functionality for use by EAP-FAST.
8271 *Jouni Malinen <j@w1.fi>*
8273 * Modify HMAC functions to return a value. Since these can be implemented
8274 in an ENGINE errors can occur.
8278 * Type-checked OBJ_bsearch_ex.
8282 * Type-checked OBJ_bsearch. Also some constification necessitated
8283 by type-checking. Still to come: TXT_DB, bsearch(?),
8284 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8289 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8290 seconds to a tm structure directly, instead of going through OS
8291 specific date routines. This avoids any issues with OS routines such
8292 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8293 and X509_time_adj_ex() to cover the extended range. The existing
8294 X509_time_adj() is still usable and will no longer have any date issues.
8298 * Delta CRL support. New use deltas option which will attempt to locate
8299 and search any appropriate delta CRLs available.
8301 This work was sponsored by Google.
8305 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8306 code and add additional score elements. Validate alternate CRL paths
8307 as part of the CRL checking and indicate a new error "CRL path validation
8308 error" in this case. Applications wanting additional details can use
8309 the verify callback and check the new "parent" field. If this is not
8310 NULL CRL path validation is taking place. Existing applications won't
8311 see this because it requires extended CRL support which is off by
8314 This work was sponsored by Google.
8318 * Support for freshest CRL extension.
8320 This work was sponsored by Google.
8324 * Initial indirect CRL support. Currently only supported in the CRLs
8325 passed directly and not via lookup. Process certificate issuer
8326 CRL entry extension and lookup CRL entries by bother issuer name
8327 and serial number. Check and process CRL issuer entry in IDP extension.
8329 This work was sponsored by Google.
8333 * Add support for distinct certificate and CRL paths. The CRL issuer
8334 certificate is validated separately in this case. Only enabled if
8335 an extended CRL support flag is set: this flag will enable additional
8336 CRL functionality in future.
8338 This work was sponsored by Google.
8342 * Add support for policy mappings extension.
8344 This work was sponsored by Google.
8348 * Fixes to pathlength constraint, self issued certificate handling,
8349 policy processing to align with RFC3280 and PKITS tests.
8351 This work was sponsored by Google.
8355 * Support for name constraints certificate extension. DN, email, DNS
8356 and URI types are currently supported.
8358 This work was sponsored by Google.
8362 * To cater for systems that provide a pointer-based thread ID rather
8363 than numeric, deprecate the current numeric thread ID mechanism and
8364 replace it with a structure and associated callback type. This
8365 mechanism allows a numeric "hash" to be extracted from a thread ID in
8366 either case, and on platforms where pointers are larger than 'long',
8367 mixing is done to help ensure the numeric 'hash' is usable even if it
8368 can't be guaranteed unique. The default mechanism is to use "&errno"
8369 as a pointer-based thread ID to distinguish between threads.
8371 Applications that want to provide their own thread IDs should now use
8372 CRYPTO_THREADID_set_callback() to register a callback that will call
8373 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8375 Note that ERR_remove_state() is now deprecated, because it is tied
8376 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8377 to free the current thread's error state should be replaced by
8378 ERR_remove_thread_state(NULL).
8380 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8381 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8382 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8383 application was previously providing a numeric thread callback that
8384 was inappropriate for distinguishing threads, then uniqueness might
8385 have been obtained with &errno that happened immediately in the
8386 intermediate development versions of OpenSSL; this is no longer the
8387 case, the numeric thread callback will now override the automatic use
8390 *Geoff Thorpe, with help from Bodo Moeller*
8392 * Initial support for different CRL issuing certificates. This covers a
8393 simple case where the self issued certificates in the chain exist and
8394 the real CRL issuer is higher in the existing chain.
8396 This work was sponsored by Google.
8400 * Removed effectively defunct crypto/store from the build.
8404 * Revamp of STACK to provide stronger type-checking. Still to come:
8405 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8406 ASN1_STRING, CONF_VALUE.
8410 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8411 RAM on SSL connections. This option can save about 34k per idle SSL.
8415 * Revamp of LHASH to provide stronger type-checking. Still to come:
8416 STACK, TXT_DB, bsearch, qsort.
8420 * Initial support for Cryptographic Message Syntax (aka CMS) based
8421 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8422 support for data, signedData, compressedData, digestedData and
8423 encryptedData, envelopedData types included. Scripts to check against
8424 RFC4134 examples draft and interop and consistency checks of many
8425 content types and variants.
8429 * Add options to enc utility to support use of zlib compression BIO.
8433 * Extend mk1mf to support importing of options and assembly language
8434 files from Configure script, currently only included in VC-WIN32.
8435 The assembly language rules can now optionally generate the source
8436 files from the associated perl scripts.
8440 * Implement remaining functionality needed to support GOST ciphersuites.
8441 Interop testing has been performed using CryptoPro implementations.
8443 *Victor B. Wagner <vitus@cryptocom.ru>*
8445 * s390x assembler pack.
8449 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8454 * Implement Opaque PRF Input TLS extension as specified in
8455 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8456 official specification yet and no extension type assignment by
8457 IANA exists, this extension (for now) will have to be explicitly
8458 enabled when building OpenSSL by providing the extension number
8459 to use. For example, specify an option
8461 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8463 to the "config" or "Configure" script to enable the extension,
8464 assuming extension number 0x9527 (which is a completely arbitrary
8465 and unofficial assignment based on the MD5 hash of the Internet
8466 Draft). Note that by doing so, you potentially lose
8467 interoperability with other TLS implementations since these might
8468 be using the same extension number for other purposes.
8470 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8471 opaque PRF input value to use in the handshake. This will create
8472 an internal copy of the length-'len' string at 'src', and will
8473 return non-zero for success.
8475 To get more control and flexibility, provide a callback function
8478 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8479 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8483 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8486 Callback function 'cb' will be called in handshakes, and is
8487 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8488 Argument 'arg' is for application purposes (the value as given to
8489 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8490 be provided to the callback function). The callback function
8491 has to return non-zero to report success: usually 1 to use opaque
8492 PRF input just if possible, or 2 to enforce use of the opaque PRF
8493 input. In the latter case, the library will abort the handshake
8494 if opaque PRF input is not successfully negotiated.
8496 Arguments 'peerinput' and 'len' given to the callback function
8497 will always be NULL and 0 in the case of a client. A server will
8498 see the client's opaque PRF input through these variables if
8499 available (NULL and 0 otherwise). Note that if the server
8500 provides an opaque PRF input, the length must be the same as the
8501 length of the client's opaque PRF input.
8503 Note that the callback function will only be called when creating
8504 a new session (session resumption can resume whatever was
8505 previously negotiated), and will not be called in SSL 2.0
8506 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8507 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8508 for applications that need to enforce opaque PRF input.
8512 * Update ssl code to support digests other than SHA1+MD5 for handshake
8515 *Victor B. Wagner <vitus@cryptocom.ru>*
8517 * Add RFC4507 support to OpenSSL. This includes the corrections in
8518 RFC4507bis. The encrypted ticket format is an encrypted encoded
8519 SSL_SESSION structure, that way new session features are automatically
8522 If a client application caches session in an SSL_SESSION structure
8523 support is transparent because tickets are now stored in the encoded
8526 The SSL_CTX structure automatically generates keys for ticket
8527 protection in servers so again support should be possible
8528 with no application modification.
8530 If a client or server wishes to disable RFC4507 support then the option
8531 SSL_OP_NO_TICKET can be set.
8533 Add a TLS extension debugging callback to allow the contents of any client
8534 or server extensions to be examined.
8536 This work was sponsored by Google.
8540 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8541 OpenSSL should now compile cleanly on gcc 4.2
8543 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8545 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8546 support including streaming MAC support: this is required for GOST
8547 ciphersuite support.
8549 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8551 * Add option -stream to use PKCS#7 streaming in smime utility. New
8552 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8553 to output in BER and PEM format.
8557 * Experimental support for use of HMAC via EVP_PKEY interface. This
8558 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8559 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8560 ENGINE support for HMAC keys which are unextractable. New -mac and
8561 -macopt options to dgst utility.
8565 * New option -sigopt to dgst utility. Update dgst to use
8566 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8567 alternative signing parameters such as X9.31 or PSS in the dgst
8572 * Change ssl_cipher_apply_rule(), the internal function that does
8573 the work each time a ciphersuite string requests enabling
8574 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8575 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8576 the order of disabled ciphersuites such that those ciphersuites
8577 that most recently went from enabled to disabled not only stay
8578 in order with respect to each other, but also have higher priority
8579 than other disabled ciphersuites the next time ciphersuites are
8582 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8583 the same ciphersuites as with "HIGH" alone, but in a specific
8584 order where the PSK ciphersuites come first (since they are the
8585 most recently disabled ciphersuites when "HIGH" is parsed).
8587 Also, change ssl_create_cipher_list() (using this new
8588 functionality) such that between otherwise identical
8589 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8594 * Change ssl_create_cipher_list() so that it automatically
8595 arranges the ciphersuites in reasonable order before starting
8596 to process the rule string. Thus, the definition for "DEFAULT"
8597 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8598 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8599 This makes it much easier to arrive at a reasonable default order
8600 in applications for which anonymous ciphers are OK (meaning
8601 that you can't actually use DEFAULT).
8603 *Bodo Moeller; suggested by Victor Duchovni*
8605 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8606 processing) into multiple integers instead of setting
8607 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8608 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8609 (These masks as well as the individual bit definitions are hidden
8610 away into the non-exported interface ssl/ssl_locl.h, so this
8611 change to the definition of the SSL_CIPHER structure shouldn't
8612 affect applications.) This give us more bits for each of these
8613 categories, so there is no longer a need to coagulate AES128 and
8614 AES256 into a single algorithm bit, and to coagulate Camellia128
8615 and Camellia256 into a single algorithm bit, which has led to all
8618 Thus, among other things, the kludge introduced in 0.9.7m and
8619 0.9.8e for masking out AES256 independently of AES128 or masking
8620 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8622 With the change, we also introduce new ciphersuite aliases that
8623 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8628 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8629 Use the leftmost N bytes of the signature input if the input is
8630 larger than the prime q (with N being the size in bytes of q).
8634 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8635 it yet and it is largely untested.
8639 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8643 * Initial incomplete changes to avoid need for function casts in OpenSSL
8644 some compilers (gcc 4.2 and later) reject their use. Safestack is
8645 reimplemented. Update ASN1 to avoid use of legacy functions.
8649 * Win32/64 targets are linked with Winsock2.
8653 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8654 to external functions. This can be used to increase CRL handling
8655 efficiency especially when CRLs are very large by (for example) storing
8656 the CRL revoked certificates in a database.
8660 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8661 new CRLs added to a directory can be used. New command line option
8662 -verify_return_error to s_client and s_server. This causes real errors
8663 to be returned by the verify callback instead of carrying on no matter
8664 what. This reflects the way a "real world" verify callback would behave.
8668 * GOST engine, supporting several GOST algorithms and public key formats.
8669 Kindly donated by Cryptocom.
8673 * Partial support for Issuing Distribution Point CRL extension. CRLs
8674 partitioned by DP are handled but no indirect CRL or reason partitioning
8675 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8676 selected via a scoring technique which handles IDP and AKID in CRLs.
8680 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8681 will ultimately be used for all verify operations: this will remove the
8682 X509_STORE dependency on certificate verification and allow alternative
8683 lookup methods. X509_STORE based implementations of these two callbacks.
8687 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8688 Modify get_crl() to find a valid (unexpired) CRL if possible.
8692 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8693 this would be called X509_CRL_cmp() but that name is already used by
8694 a function that just compares CRL issuer names. Cache several CRL
8695 extensions in X509_CRL structure and cache CRLDP in X509.
8699 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8700 this maps equivalent X509_NAME structures into a consistent structure.
8701 Name comparison can then be performed rapidly using memcmp().
8705 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8710 * Allow digests to supply their own micalg string for S/MIME type using
8711 the ctrl EVP_MD_CTRL_MICALG.
8715 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8716 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8717 ctrl. It can then customise the structure before and/or after signing
8722 * New function OBJ_add_sigid() to allow application defined signature OIDs
8723 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8724 to free up any added signature OIDs.
8728 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8729 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8730 digest and cipher tables. New options added to openssl utility:
8731 list-message-digest-algorithms and list-cipher-algorithms.
8735 * Change the array representation of binary polynomials: the list
8736 of degrees of non-zero coefficients is now terminated with -1.
8737 Previously it was terminated with 0, which was also part of the
8738 value; thus, the array representation was not applicable to
8739 polynomials where t^0 has coefficient zero. This change makes
8740 the array representation useful in a more general context.
8744 * Various modifications and fixes to SSL/TLS cipher string
8745 handling. For ECC, the code now distinguishes between fixed ECDH
8746 with RSA certificates on the one hand and with ECDSA certificates
8747 on the other hand, since these are separate ciphersuites. The
8748 unused code for Fortezza ciphersuites has been removed.
8750 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8751 (not "ECDHE"). For consistency with the code for DH
8752 certificates, use of ECDH certificates is now considered ECDH
8753 authentication, not RSA or ECDSA authentication (the latter is
8754 merely the CA's signing algorithm and not actively used in the
8757 The temporary ciphersuite alias "ECCdraft" is no longer
8758 available, and ECC ciphersuites are no longer excluded from "ALL"
8759 and "DEFAULT". The following aliases now exist for RFC 4492
8760 ciphersuites, most of these by analogy with the DH case:
8762 kECDHr - ECDH cert, signed with RSA
8763 kECDHe - ECDH cert, signed with ECDSA
8764 kECDH - ECDH cert (signed with either RSA or ECDSA)
8765 kEECDH - ephemeral ECDH
8766 ECDH - ECDH cert or ephemeral ECDH
8772 AECDH - anonymous ECDH
8773 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8777 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8778 Use correct micalg parameters depending on digest(s) in signed message.
8782 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8783 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8787 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8788 an engine to register a method. Add ENGINE lookups for methods and
8789 functional reference processing.
8793 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8794 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8799 * New -resign option to smime utility. This adds one or more signers
8800 to an existing PKCS#7 signedData structure. Also -md option to use an
8801 alternative message digest algorithm for signing.
8805 * Tidy up PKCS#7 routines and add new functions to make it easier to
8806 create PKCS7 structures containing multiple signers. Update smime
8807 application to support multiple signers.
8811 * New -macalg option to pkcs12 utility to allow setting of an alternative
8816 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8817 Reorganize PBE internals to lookup from a static table using NIDs,
8818 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8819 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8820 PRF which will be automatically used with PBES2.
8824 * Replace the algorithm specific calls to generate keys in "req" with the
8829 * Update PKCS#7 enveloped data routines to use new API. This is now
8830 supported by any public key method supporting the encrypt operation. A
8831 ctrl is added to allow the public key algorithm to examine or modify
8832 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8837 * Add a ctrl to asn1 method to allow a public key algorithm to express
8838 a default digest type to use. In most cases this will be SHA1 but some
8839 algorithms (such as GOST) need to specify an alternative digest. The
8840 return value indicates how strong the preference is 1 means optional and
8841 2 is mandatory (that is it is the only supported type). Modify
8842 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8843 use the default md. Update openssl utilities to use the default digest
8844 type for signing if it is not explicitly indicated.
8848 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8849 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8850 signing method from the key type. This effectively removes the link
8851 between digests and public key types.
8855 * Add an OID cross reference table and utility functions. Its purpose is to
8856 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8857 rsaEncryption. This will allow some of the algorithm specific hackery
8858 needed to use the correct OID to be removed.
8862 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8863 structures for PKCS7_sign(). They are now set up by the relevant public
8868 * Add provisional EC pkey method with support for ECDSA and ECDH.
8872 * Add support for key derivation (agreement) in the API, DH method and
8877 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8878 public and private key formats. As a side effect these add additional
8879 command line functionality not previously available: DSA signatures can be
8880 generated and verified using pkeyutl and DH key support and generation in
8887 *Oliver Tappe <zooey@hirschkaefer.de>*
8889 * New make target "install_html_docs" installs HTML renditions of the
8892 *Oliver Tappe <zooey@hirschkaefer.de>*
8894 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8895 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8896 support key and parameter generation and add initial key generation
8897 functionality for RSA.
8901 * Add functions for main EVP_PKEY_method operations. The undocumented
8902 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8903 `EVP_PKEY_{encrypt,decrypt}_old`.
8907 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8908 key API, doesn't do much yet.
8912 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8913 public key algorithms. New option to openssl utility:
8914 "list-public-key-algorithms" to print out info.
8918 * Implement the Supported Elliptic Curves Extension for
8919 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8923 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8924 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8928 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8929 utilities such as rsa, dsa, dsaparam etc except they process any key
8934 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8935 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8936 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8941 * Initial support for pluggable public key ASN1.
8942 De-spaghettify the public key ASN1 handling. Move public and private
8943 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8944 algorithm specific handling to a single module within the relevant
8945 algorithm directory. Add functions to allow (near) opaque processing
8946 of public and private key structures.
8950 * Implement the Supported Point Formats Extension for
8951 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8955 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8956 for the psk identity [hint] and the psk callback functions to the
8957 SSL_SESSION, SSL and SSL_CTX structure.
8960 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8964 SSL_CTX_use_psk_identity_hint
8965 SSL_get_psk_identity_hint
8966 SSL_get_psk_identity
8967 SSL_use_psk_identity_hint
8969 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8971 * Add RFC 3161 compliant time stamp request creation, response generation
8972 and response verification functionality.
8974 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8976 * Add initial support for TLS extensions, specifically for the server_name
8977 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8978 have new members for a host name. The SSL data structure has an
8979 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8980 stored in that context to allow for session resumption, even after the
8981 SSL has been switched to a new SSL_CTX in reaction to a client's
8982 server_name extension.
8984 New functions (subject to change):
8986 SSL_get_servername()
8987 SSL_get_servername_type()
8990 New CTRL codes and macros (subject to change):
8992 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8993 - SSL_CTX_set_tlsext_servername_callback()
8994 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8995 - SSL_CTX_set_tlsext_servername_arg()
8996 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8998 openssl s_client has a new '-servername ...' option.
9000 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9001 '-key2 ...', '-servername_fatal' (subject to change). This allows
9002 testing the HostName extension for a specific single host name ('-cert'
9003 and '-key' remain fallbacks for handshakes without HostName
9004 negotiation). If the unrecognized_name alert has to be sent, this by
9005 default is a warning; it becomes fatal with the '-servername_fatal'
9008 *Peter Sylvester, Remy Allais, Christophe Renou*
9010 * Whirlpool hash implementation is added.
9014 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9015 bn(64,32). Because of instruction set limitations it doesn't have
9016 any negative impact on performance. This was done mostly in order
9017 to make it possible to share assembler modules, such as bn_mul_mont
9018 implementations, between 32- and 64-bit builds without hassle.
9022 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9023 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9028 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9029 dedicated Montgomery multiplication procedure, is introduced.
9030 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9031 "64-bit" performance on certain 32-bit targets.
9035 * New option SSL_OP_NO_COMP to disable use of compression selectively
9036 in SSL structures. New SSL ctrl to set maximum send fragment size.
9037 Save memory by setting the I/O buffer sizes dynamically instead of
9038 using the maximum available value.
9042 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9043 in addition to the text details.
9047 * Very, very preliminary EXPERIMENTAL support for printing of general
9048 ASN1 structures. This currently produces rather ugly output and doesn't
9049 handle several customised structures at all.
9053 * Integrated support for PVK file format and some related formats such
9054 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9055 these in the 'rsa' and 'dsa' utilities.
9059 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9063 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9064 place for the (very old) "NETSCAPE" format certificates which are now
9065 handled using new ASN1 code equivalents.
9069 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9070 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9071 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9075 * Modify CRL distribution points extension code to print out previously
9076 unsupported fields. Enhance extension setting code to allow setting of
9081 * Add print and set support for Issuing Distribution Point CRL extension.
9085 * Change 'Configure' script to enable Camellia by default.
9092 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9094 * When rejecting SSL/TLS records due to an incorrect version number, never
9095 update s->server with a new major version number. As of
9096 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9097 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9098 the previous behavior could result in a read attempt at NULL when
9099 receiving specific incorrect SSL/TLS records once record payload
9100 protection is active. ([CVE-2010-0740])
9102 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9104 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9105 could be crashed if the relevant tables were not present (e.g. chrooted).
9107 *Tomas Hoger <thoger@redhat.com>*
9109 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9111 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9113 *Martin Olsson, Neel Mehta*
9115 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9116 accommodate for stack sorting, always a write lock!).
9120 * On some versions of WIN32 Heap32Next is very slow. This can cause
9121 excessive delays in the RAND_poll(): over a minute. As a workaround
9122 include a time check in the inner Heap32Next loop too.
9126 * The code that handled flushing of data in SSL/TLS originally used the
9127 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9128 the problem outlined in PR#1949. The fix suggested there however can
9129 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9130 of Apache). So instead simplify the code to flush unconditionally.
9131 This should be fine since flushing with no data to flush is a no op.
9135 * Handle TLS versions 2.0 and later properly and correctly use the
9136 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9137 off ancient servers have a habit of sticking around for a while...
9141 * Modify compression code so it frees up structures without using the
9142 ex_data callbacks. This works around a problem where some applications
9143 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9144 restarting) then use compression (e.g. SSL with compression) later.
9145 This results in significant per-connection memory leaks and
9146 has caused some security issues including CVE-2008-1678 and
9151 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9152 change when encrypting or decrypting.
9156 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9157 connect and renegotiate with servers which do not support RI.
9158 Until RI is more widely deployed this option is enabled by default.
9162 * Add "missing" ssl ctrls to clear options and mode.
9166 * If client attempts to renegotiate and doesn't support RI respond with
9167 a no_renegotiation alert as required by RFC5746. Some renegotiating
9168 TLS clients will continue a connection gracefully when they receive
9169 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9170 waiting for a server hello which it will never receive. Now we treat a
9171 received no_renegotiation alert as a fatal error. This is because
9172 applications requesting a renegotiation might well expect it to succeed
9173 and would have no code in place to handle the server denying it so the
9174 only safe thing to do is to terminate the connection.
9178 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9179 peer supports secure renegotiation and 0 otherwise. Print out peer
9180 renegotiation support in s_client/s_server.
9184 * Replace the highly broken and deprecated SPKAC certification method with
9185 the updated NID creation version. This should correctly handle UTF8.
9189 * Implement RFC5746. Re-enable renegotiation but require the extension
9190 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9191 turns out to be a bad idea. It has been replaced by
9192 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9193 SSL_CTX_set_options(). This is really not recommended unless you
9194 know what you are doing.
9196 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9198 * Fixes to stateless session resumption handling. Use initial_ctx when
9199 issuing and attempting to decrypt tickets in case it has changed during
9200 servername handling. Use a non-zero length session ID when attempting
9201 stateless session resumption: this makes it possible to determine if
9202 a resumption has occurred immediately after receiving server hello
9203 (several places in OpenSSL subtly assume this) instead of later in
9208 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9209 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9210 fixes for a few places where the return code is not checked
9213 *Julia Lawall <julia@diku.dk>*
9215 * Add --strict-warnings option to Configure script to include devteam
9216 warnings in other configurations.
9220 * Add support for --libdir option and LIBDIR variable in makefiles. This
9221 makes it possible to install openssl libraries in locations which
9222 have names other than "lib", for example "/usr/lib64" which some
9225 *Steve Henson, based on patch from Jeremy Utley*
9227 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9228 X690 8.9.12 and can produce some misleading textual output of OIDs.
9230 *Steve Henson, reported by Dan Kaminsky*
9232 * Delete MD2 from algorithm tables. This follows the recommendation in
9233 several standards that it is not used in new applications due to
9234 several cryptographic weaknesses. For binary compatibility reasons
9235 the MD2 API is still compiled in by default.
9239 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9244 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9245 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9248 *Guenter <lists@gknw.net>*
9250 * Fix the server certificate chain building code to use X509_verify_cert(),
9251 it used to have an ad-hoc builder which was unable to cope with anything
9252 other than a simple chain.
9254 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9256 * Don't check self signed certificate signatures in X509_verify_cert()
9257 by default (a flag can override this): it just wastes time without
9258 adding any security. As a useful side effect self signed root CAs
9259 with non-FIPS digests are now usable in FIPS mode.
9263 * In dtls1_process_out_of_seq_message() the check if the current message
9264 is already buffered was missing. For every new message was memory
9265 allocated, allowing an attacker to perform an denial of service attack
9266 with sending out of seq handshake messages until there is no memory
9267 left. Additionally every future message was buffered, even if the
9268 sequence number made no sense and would be part of another handshake.
9269 So only messages with sequence numbers less than 10 in advance will be
9270 buffered. ([CVE-2009-1378])
9272 *Robin Seggelmann, discovered by Daniel Mentz*
9274 * Records are buffered if they arrive with a future epoch to be
9275 processed after finishing the corresponding handshake. There is
9276 currently no limitation to this buffer allowing an attacker to perform
9277 a DOS attack with sending records with future epochs until there is no
9278 memory left. This patch adds the pqueue_size() function to determine
9279 the size of a buffer and limits the record buffer to 100 entries.
9282 *Robin Seggelmann, discovered by Daniel Mentz*
9284 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9285 parent structure is freed. ([CVE-2009-1379])
9289 * Handle non-blocking I/O properly in SSL_shutdown() call.
9291 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9293 * Add `2.5.4.*` OIDs
9295 *Ilya O. <vrghost@gmail.com>*
9297 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9299 * Disable renegotiation completely - this fixes a severe security
9300 problem ([CVE-2009-3555]) at the cost of breaking all
9301 renegotiation. Renegotiation can be re-enabled by setting
9302 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9303 run-time. This is really not recommended unless you know what
9308 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9310 * Don't set val to NULL when freeing up structures, it is freed up by
9311 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9312 zeroing past the valid field. ([CVE-2009-0789])
9314 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9316 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9317 checked correctly. This would allow some invalid signed attributes to
9318 appear to verify correctly. ([CVE-2009-0591])
9320 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9322 * Reject UniversalString and BMPString types with invalid lengths. This
9323 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9324 a legal length. ([CVE-2009-0590])
9328 * Set S/MIME signing as the default purpose rather than setting it
9329 unconditionally. This allows applications to override it at the store
9334 * Permit restricted recursion of ASN1 strings. This is needed in practice
9335 to handle some structures.
9339 * Improve efficiency of mem_gets: don't search whole buffer each time
9342 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9344 * New -hex option for openssl rand.
9348 * Print out UTF8String and NumericString when parsing ASN1.
9352 * Support NumericString type for name components.
9356 * Allow CC in the environment to override the automatically chosen
9357 compiler. Note that nothing is done to ensure flags work with the
9362 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9364 * Properly check EVP_VerifyFinal() and similar return values
9367 *Ben Laurie, Bodo Moeller, Google Security Team*
9369 * Enable TLS extensions by default.
9373 * Allow the CHIL engine to be loaded, whether the application is
9374 multithreaded or not. (This does not release the developer from the
9375 obligation to set up the dynamic locking callbacks.)
9377 *Sander Temme <sander@temme.net>*
9379 * Use correct exit code if there is an error in dgst command.
9381 *Steve Henson; problem pointed out by Roland Dirlewanger*
9383 * Tweak Configure so that you need to say "experimental-jpake" to enable
9384 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9388 * Add experimental JPAKE support, including demo authentication in
9389 s_client and s_server.
9393 * Set the comparison function in v3_addr_canonize().
9395 *Rob Austein <sra@hactrn.net>*
9397 * Add support for XMPP STARTTLS in s_client.
9399 *Philip Paeps <philip@freebsd.org>*
9401 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9402 to ensure that even with this option, only ciphersuites in the
9403 server's preference list will be accepted. (Note that the option
9404 applies only when resuming a session, so the earlier behavior was
9405 just about the algorithm choice for symmetric cryptography.)
9409 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9411 * Fix NULL pointer dereference if a DTLS server received
9412 ChangeCipherSpec as first record ([CVE-2009-1386]).
9416 * Fix a state transition in s3_srvr.c and d1_srvr.c
9417 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9421 * The fix in 0.9.8c that supposedly got rid of unsafe
9422 double-checked locking was incomplete for RSA blinding,
9423 addressing just one layer of what turns out to have been
9424 doubly unsafe triple-checked locking.
9426 So now fix this for real by retiring the MONT_HELPER macro
9427 in crypto/rsa/rsa_eay.c.
9429 *Bodo Moeller; problem pointed out by Marius Schilder*
9431 * Various precautionary measures:
9433 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9435 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9436 (NB: This would require knowledge of the secret session ticket key
9437 to exploit, in which case you'd be SOL either way.)
9439 - Change bn_nist.c so that it will properly handle input BIGNUMs
9440 outside the expected range.
9442 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9445 *Neel Mehta, Bodo Moeller*
9447 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9448 the load fails. Useful for distros.
9450 *Ben Laurie and the FreeBSD team*
9452 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9456 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9460 * Expand ENGINE to support engine supplied SSL client certificate functions.
9462 This work was sponsored by Logica.
9466 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9467 keystores. Support for SSL/TLS client authentication too.
9468 Not compiled unless enable-capieng specified to Configure.
9470 This work was sponsored by Logica.
9474 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9475 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9476 attribute creation routines such as certificate requests and PKCS#12
9481 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9483 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9484 handshake which could lead to a client crash as found using the
9485 Codenomicon TLS test suite ([CVE-2008-1672])
9487 *Steve Henson, Mark Cox*
9489 * Fix double free in TLS server name extensions which could lead to
9490 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9494 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9496 Clear the error queue to ensure that error entries left from
9497 older function calls do not interfere with the correct operation.
9499 *Lutz Jaenicke, Erik de Castro Lopo*
9501 * Remove root CA certificates of commercial CAs:
9503 The OpenSSL project does not recommend any specific CA and does not
9504 have any policy with respect to including or excluding any CA.
9505 Therefore it does not make any sense to ship an arbitrary selection
9506 of root CA certificates with the OpenSSL software.
9510 * RSA OAEP patches to fix two separate invalid memory reads.
9511 The first one involves inputs when 'lzero' is greater than
9512 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9513 before the beginning of from). The second one involves inputs where
9514 the 'db' section contains nothing but zeroes (there is a one-byte
9515 invalid read after the end of 'db').
9517 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9519 * Partial backport from 0.9.9-dev:
9521 Introduce bn_mul_mont (dedicated Montgomery multiplication
9522 procedure) as a candidate for BIGNUM assembler implementation.
9523 While 0.9.9-dev uses assembler for various architectures, only
9524 x86_64 is available by default here in the 0.9.8 branch, and
9525 32-bit x86 is available through a compile-time setting.
9527 To try the 32-bit x86 assembler implementation, use Configure
9528 option "enable-montasm" (which exists only for this backport).
9530 As "enable-montasm" for 32-bit x86 disclaims code stability
9531 anyway, in this constellation we activate additional code
9532 backported from 0.9.9-dev for further performance improvements,
9533 namely BN_from_montgomery_word. (To enable this otherwise,
9534 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9536 *Andy Polyakov (backport partially by Bodo Moeller)*
9538 * Add TLS session ticket callback. This allows an application to set
9539 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9540 values. This is useful for key rollover for example where several key
9541 sets may exist with different names.
9545 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9546 This was broken until now in 0.9.8 releases, such that the only way
9547 a registered ENGINE could be used (assuming it initialises
9548 successfully on the host) was to explicitly set it as the default
9549 for the relevant algorithms. This is in contradiction with 0.9.7
9550 behaviour and the documentation. With this fix, when an ENGINE is
9551 registered into a given algorithm's table of implementations, the
9552 'uptodate' flag is reset so that auto-discovery will be used next
9553 time a new context for that algorithm attempts to select an
9556 *Ian Lister (tweaked by Geoff Thorpe)*
9558 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9559 implementation in the following ways:
9561 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9564 Lack of BER streaming support means one pass streaming processing is
9565 only supported if data is detached: setting the streaming flag is
9566 ignored for embedded content.
9568 CMS support is disabled by default and must be explicitly enabled
9569 with the enable-cms configuration option.
9573 * Update the GMP engine glue to do direct copies between BIGNUM and
9574 mpz_t when openssl and GMP use the same limb size. Otherwise the
9575 existing "conversion via a text string export" trick is still used.
9577 *Paul Sheer <paulsheer@gmail.com>*
9579 * Zlib compression BIO. This is a filter BIO which compressed and
9580 uncompresses any data passed through it.
9584 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9585 RFC3394 compatible AES key wrapping.
9589 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9590 sets string data without copying. X509_ALGOR_set0() and
9591 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9592 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9593 from an X509_ATTRIBUTE structure optionally checking it occurs only
9594 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9599 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9600 to get the expected BN_FLG_CONSTTIME behavior.
9602 *Bodo Moeller (Google)*
9606 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9607 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9608 - added some more tests to do_tests.pl
9609 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9610 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9611 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9612 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9613 - various changes to netware.pl to enable gcc-cross builds on Win32
9615 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9616 - various changes to fix missing prototype warnings
9617 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9618 - added AES, WHIRLPOOL and CPUID assembler code to build files
9619 - added missing AES assembler make rules to mk1mf.pl
9620 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9622 *Guenter Knauf <eflash@gmx.net>*
9624 * Implement certificate status request TLS extension defined in RFC3546.
9625 A client can set the appropriate parameters and receive the encoded
9626 OCSP response via a callback. A server can query the supplied parameters
9627 and set the encoded OCSP response in the callback. Add simplified examples
9628 to s_client and s_server.
9632 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9635 + Binary incompatibility of ssl_ctx_st structure
9636 + DTLS interoperation with non-compliant servers
9637 + Don't call get_session_cb() without proposed session
9638 + Fix ia64 assembler code
9640 *Andy Polyakov, Steve Henson*
9642 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9644 * DTLS Handshake overhaul. There were longstanding issues with
9645 OpenSSL DTLS implementation, which were making it impossible for
9646 RFC 4347 compliant client to communicate with OpenSSL server.
9647 Unfortunately just fixing these incompatibilities would "cut off"
9648 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9649 server keeps tolerating non RFC compliant syntax. The opposite is
9650 not true, 0.9.8f client can not communicate with earlier server.
9651 This update even addresses CVE-2007-4995.
9655 * Changes to avoid need for function casts in OpenSSL: some compilers
9656 (gcc 4.2 and later) reject their use.
9657 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9660 * Add RFC4507 support to OpenSSL. This includes the corrections in
9661 RFC4507bis. The encrypted ticket format is an encrypted encoded
9662 SSL_SESSION structure, that way new session features are automatically
9665 If a client application caches session in an SSL_SESSION structure
9666 support is transparent because tickets are now stored in the encoded
9669 The SSL_CTX structure automatically generates keys for ticket
9670 protection in servers so again support should be possible
9671 with no application modification.
9673 If a client or server wishes to disable RFC4507 support then the option
9674 SSL_OP_NO_TICKET can be set.
9676 Add a TLS extension debugging callback to allow the contents of any client
9677 or server extensions to be examined.
9679 This work was sponsored by Google.
9683 * Add initial support for TLS extensions, specifically for the server_name
9684 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9685 have new members for a host name. The SSL data structure has an
9686 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9687 stored in that context to allow for session resumption, even after the
9688 SSL has been switched to a new SSL_CTX in reaction to a client's
9689 server_name extension.
9691 New functions (subject to change):
9693 SSL_get_servername()
9694 SSL_get_servername_type()
9697 New CTRL codes and macros (subject to change):
9699 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9700 - SSL_CTX_set_tlsext_servername_callback()
9701 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9702 - SSL_CTX_set_tlsext_servername_arg()
9703 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9705 openssl s_client has a new '-servername ...' option.
9707 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9708 '-key2 ...', '-servername_fatal' (subject to change). This allows
9709 testing the HostName extension for a specific single host name ('-cert'
9710 and '-key' remain fallbacks for handshakes without HostName
9711 negotiation). If the unrecognized_name alert has to be sent, this by
9712 default is a warning; it becomes fatal with the '-servername_fatal'
9715 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9717 * Add AES and SSE2 assembly language support to VC++ build.
9721 * Mitigate attack on final subtraction in Montgomery reduction.
9725 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9726 (which previously caused an internal error).
9730 * Squeeze another 10% out of IGE mode when in != out.
9734 * AES IGE mode speedup.
9736 *Dean Gaudet (Google)*
9738 * Add the Korean symmetric 128-bit cipher SEED (see
9739 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9740 add SEED ciphersuites from RFC 4162:
9742 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9743 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9744 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9745 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9747 To minimize changes between patchlevels in the OpenSSL 0.9.8
9748 series, SEED remains excluded from compilation unless OpenSSL
9749 is configured with 'enable-seed'.
9751 *KISA, Bodo Moeller*
9753 * Mitigate branch prediction attacks, which can be practical if a
9754 single processor is shared, allowing a spy process to extract
9755 information. For detailed background information, see
9756 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9757 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9758 and Necessary Software Countermeasures"). The core of the change
9759 are new versions BN_div_no_branch() and
9760 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9761 respectively, which are slower, but avoid the security-relevant
9762 conditional branches. These are automatically called by BN_div()
9763 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9764 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9765 remove a conditional branch.
9767 BN_FLG_CONSTTIME is the new name for the previous
9768 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9769 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9770 in the exponent causes BN_mod_exp_mont() to use the alternative
9771 implementation in BN_mod_exp_mont_consttime().) The old name
9772 remains as a deprecated alias.
9774 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9775 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9776 constant-time implementations for more than just exponentiation.
9777 Here too the old name is kept as a deprecated alias.
9779 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9780 the BN_BLINDING structure gets an independent copy of the
9781 modulus. This means that the previous `BIGNUM *m` argument to
9782 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9783 essentially becomes `const BIGNUM *m`, although we can't actually
9784 change this in the header file before 0.9.9. It allows
9785 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9786 enable BN_FLG_CONSTTIME.
9788 *Matthew D Wood (Intel Corp)*
9790 * In the SSL/TLS server implementation, be strict about session ID
9791 context matching (which matters if an application uses a single
9792 external cache for different purposes). Previously,
9793 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9794 set. This did ensure strict client verification, but meant that,
9795 with applications using a single external cache for quite
9796 different requirements, clients could circumvent ciphersuite
9797 restrictions for a given session ID context by starting a session
9798 in a different context.
9802 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9803 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9804 authentication-only ciphersuites.
9808 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9809 not complete and could lead to a possible single byte overflow
9810 ([CVE-2007-5135]) [Ben Laurie]
9812 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9814 * Since AES128 and AES256 (and similarly Camellia128 and
9815 Camellia256) share a single mask bit in the logic of
9816 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9817 kludge to work properly if AES128 is available and AES256 isn't
9818 (or if Camellia128 is available and Camellia256 isn't).
9822 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9823 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9824 When a point or a seed is encoded in a BIT STRING, we need to
9825 prevent the removal of trailing zero bits to get the proper DER
9826 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9827 of a NamedBitList, for which trailing 0 bits need to be removed.)
9831 * Have SSL/TLS server implementation tolerate "mismatched" record
9832 protocol version while receiving ClientHello even if the
9833 ClientHello is fragmented. (The server can't insist on the
9834 particular protocol version it has chosen before the ServerHello
9835 message has informed the client about his choice.)
9839 * Add RFC 3779 support.
9841 *Rob Austein for ARIN, Ben Laurie*
9843 * Load error codes if they are not already present instead of using a
9844 static variable. This allows them to be cleanly unloaded and reloaded.
9845 Improve header file function name parsing.
9849 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9850 or CAPABILITY handshake as required by RFCs.
9854 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9856 * Introduce limits to prevent malicious keys being able to
9857 cause a denial of service. ([CVE-2006-2940])
9859 *Steve Henson, Bodo Moeller*
9861 * Fix ASN.1 parsing of certain invalid structures that can result
9862 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9864 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9865 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9867 * Fix SSL client code which could crash if connecting to a
9868 malicious SSLv2 server. ([CVE-2006-4343])
9870 *Tavis Ormandy and Will Drewry, Google Security Team*
9872 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9873 match only those. Before that, "AES256-SHA" would be interpreted
9874 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9875 the same strength classification in 0.9.7h) as we currently only
9876 have a single AES bit in the ciphersuite description bitmap.
9877 That change, however, also applied to ciphersuite strings such as
9878 "RC4-MD5" that intentionally matched multiple ciphersuites --
9879 namely, SSL 2.0 ciphersuites in addition to the more common ones
9880 from SSL 3.0/TLS 1.0.
9882 So we change the selection algorithm again: Naming an explicit
9883 ciphersuite selects this one ciphersuite, and any other similar
9884 ciphersuite (same bitmap) from *other* protocol versions.
9885 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9886 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9888 Since SSL 2.0 does not have any ciphersuites for which the
9889 128/256 bit distinction would be relevant, this works for now.
9890 The proper fix will be to use different bits for AES128 and
9891 AES256, which would have avoided the problems from the beginning;
9892 however, bits are scarce, so we can only do this in a new release
9893 (not just a patchlevel) when we can change the SSL_CIPHER
9894 definition to split the single 'unsigned long mask' bitmap into
9895 multiple values to extend the available space.
9899 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9901 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9902 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9904 * Add AES IGE and biIGE modes.
9908 * Change the Unix randomness entropy gathering to use poll() when
9909 possible instead of select(), since the latter has some
9910 undesirable limitations.
9912 *Darryl Miles via Richard Levitte and Bodo Moeller*
9914 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9915 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9916 cannot be implicitly activated as part of, e.g., the "AES" alias.
9917 However, please upgrade to OpenSSL 0.9.9[-dev] for
9918 non-experimental use of the ECC ciphersuites to get TLS extension
9919 support, which is required for curve and point format negotiation
9920 to avoid potential handshake problems.
9924 * Disable rogue ciphersuites:
9926 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9927 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9928 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9930 The latter two were purportedly from
9931 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9934 Also deactivate the remaining ciphersuites from
9935 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9936 unofficial, and the ID has long expired.
9940 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9941 dual-core machines) and other potential thread-safety issues.
9945 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9946 versions), which is now available for royalty-free use
9947 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9948 Also, add Camellia TLS ciphersuites from RFC 4132.
9950 To minimize changes between patchlevels in the OpenSSL 0.9.8
9951 series, Camellia remains excluded from compilation unless OpenSSL
9952 is configured with 'enable-camellia'.
9956 * Disable the padding bug check when compression is in use. The padding
9957 bug check assumes the first packet is of even length, this is not
9958 necessarily true if compression is enabled and can result in false
9959 positives causing handshake failure. The actual bug test is ancient
9960 code so it is hoped that implementations will either have fixed it by
9961 now or any which still have the bug do not support compression.
9965 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9967 * When applying a cipher rule check to see if string match is an explicit
9968 cipher suite and only match that one cipher suite if it is.
9972 * Link in manifests for VC++ if needed.
9974 *Austin Ziegler <halostatue@gmail.com>*
9976 * Update support for ECC-based TLS ciphersuites according to
9977 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9978 TLS extensions, which are supported starting with the 0.9.9
9979 branch, not in the OpenSSL 0.9.8 branch).
9983 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9984 opaque EVP_CIPHER_CTX handling.
9988 * Fixes and enhancements to zlib compression code. We now only use
9989 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9990 to conform with the standards mentioned here:
9991 <http://www.zlib.net/DLL_FAQ.txt>
9992 Static zlib linking now works on Windows and the new --with-zlib-include
9993 --with-zlib-lib options to Configure can be used to supply the location
9994 of the headers and library. Gracefully handle case where zlib library
9999 * Several fixes and enhancements to the OID generation code. The old code
10000 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10001 handle numbers larger than ULONG_MAX, truncated printing and had a
10002 non standard OBJ_obj2txt() behaviour.
10006 * Add support for building of engines under engine/ as shared libraries
10007 under VC++ build system.
10011 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10012 Hopefully, we will not see any false combination of paths any more.
10016 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10018 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10019 (part of SSL_OP_ALL). This option used to disable the
10020 countermeasure against man-in-the-middle protocol-version
10021 rollback in the SSL 2.0 server implementation, which is a bad
10022 idea. ([CVE-2005-2969])
10024 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10025 for Information Security, National Institute of Advanced Industrial
10026 Science and Technology [AIST], Japan)*
10028 * Add two function to clear and return the verify parameter flags.
10032 * Keep cipherlists sorted in the source instead of sorting them at
10033 runtime, thus removing the need for a lock.
10037 * Avoid some small subgroup attacks in Diffie-Hellman.
10039 *Nick Mathewson and Ben Laurie*
10041 * Add functions for well-known primes.
10045 * Extended Windows CE support.
10047 *Satoshi Nakamura and Andy Polyakov*
10049 * Initialize SSL_METHOD structures at compile time instead of during
10050 runtime, thus removing the need for a lock.
10054 * Make PKCS7_decrypt() work even if no certificate is supplied by
10055 attempting to decrypt each encrypted key in turn. Add support to
10060 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10062 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10065 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10069 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10070 key into the same file any more.
10074 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10078 * Add -utf8 command line and config file option to 'ca'.
10080 *Stefan <stf@udoma.org*
10082 * Removed the macro des_crypt(), as it seems to conflict with some
10083 libraries. Use DES_crypt().
10087 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10088 involves renaming the source and generated shared-libs for
10089 both. The engines will accept the corrected or legacy ids
10090 ('ncipher' and '4758_cca' respectively) when binding. NB,
10091 this only applies when building 'shared'.
10093 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10095 * Add attribute functions to EVP_PKEY structure. Modify
10096 PKCS12_create() to recognize a CSP name attribute and
10097 use it. Make -CSP option work again in pkcs12 utility.
10101 * Add new functionality to the bn blinding code:
10102 - automatic re-creation of the BN_BLINDING parameters after
10103 a fixed number of uses (currently 32)
10104 - add new function for parameter creation
10105 - introduce flags to control the update behaviour of the
10106 BN_BLINDING parameters
10107 - hide BN_BLINDING structure
10108 Add a second BN_BLINDING slot to the RSA structure to improve
10109 performance when a single RSA object is shared among several
10114 * Add support for DTLS.
10116 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10118 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10119 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10123 * Remove buggy and incomplete DH cert support from
10124 ssl/ssl_rsa.c and ssl/s3_both.c
10128 * Use SHA-1 instead of MD5 as the default digest algorithm for
10129 the `apps/openssl` commands.
10133 * Compile clean with "-Wall -Wmissing-prototypes
10134 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10135 DEBUG_SAFESTACK must also be set.
10139 * Change ./Configure so that certain algorithms can be disabled by default.
10140 The new counterpiece to "no-xxx" is "enable-xxx".
10142 The patented RC5 and MDC2 algorithms will now be disabled unless
10143 "enable-rc5" and "enable-mdc2", respectively, are specified.
10145 (IDEA remains enabled despite being patented. This is because IDEA
10146 is frequently required for interoperability, and there is no license
10147 fee for non-commercial use. As before, "no-idea" can be used to
10148 avoid this algorithm.)
10152 * Add processing of proxy certificates (see RFC 3820). This work was
10153 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10154 EGEE (Enabling Grids for E-science in Europe).
10158 * RC4 performance overhaul on modern architectures/implementations, such
10159 as Intel P4, IA-64 and AMD64.
10163 * New utility extract-section.pl. This can be used specify an alternative
10164 section number in a pod file instead of having to treat each file as
10165 a separate case in Makefile. This can be done by adding two lines to the
10168 =for comment openssl_section:XXX
10170 The blank line is mandatory.
10174 * New arguments -certform, -keyform and -pass for s_client and s_server
10175 to allow alternative format key and certificate files and passphrase
10180 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10181 update associated structures and add various utility functions.
10183 Add new policy related verify parameters, include policy checking in
10184 standard verify code. Enhance 'smime' application with extra parameters
10185 to support policy checking and print out.
10189 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10190 Nehemiah processors. These extensions support AES encryption in hardware
10191 as well as RNG (though RNG support is currently disabled).
10193 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10195 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10199 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10201 *Andy Polyakov and a number of other people*
10203 * Improved PowerPC platform support. Most notably BIGNUM assembler
10204 implementation contributed by IBM.
10206 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10208 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10209 exponent rather than 'unsigned long'. There is a corresponding change to
10210 the new 'rsa_keygen' element of the RSA_METHOD structure.
10212 *Jelte Jansen, Geoff Thorpe*
10214 * Functionality for creating the initial serial number file is now
10215 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10217 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10218 number file to 1, which is bound to cause problems. To avoid
10219 the problems while respecting compatibility between different 0.9.7
10220 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10221 CA.pl for serial number initialization. With the new release 0.9.8,
10222 we can fix the problem directly in the 'ca' utility.)
10226 * Reduced header interdependencies by declaring more opaque objects in
10227 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10228 give fewer recursive includes, which could break lazy source code - so
10229 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10230 developers should define this symbol when building and using openssl to
10231 ensure they track the recommended behaviour, interfaces, [etc], but
10232 backwards-compatible behaviour prevails when this isn't defined.
10236 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10240 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10241 This will generate a random key of the appropriate length based on the
10242 cipher context. The EVP_CIPHER can provide its own random key generation
10243 routine to support keys of a specific form. This is used in the des and
10244 3des routines to generate a key of the correct parity. Update S/MIME
10245 code to use new functions and hence generate correct parity DES keys.
10246 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10247 valid (weak or incorrect parity).
10251 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10252 as looking them up. This is useful when the verified structure may contain
10253 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10254 present unless the new PKCS7_NO_CRL flag is asserted.
10258 * Extend ASN1 oid configuration module. It now additionally accepts the
10261 shortName = some long name, 1.2.3.4
10265 * Reimplemented the BN_CTX implementation. There is now no more static
10266 limitation on the number of variables it can handle nor the depth of the
10267 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10268 information can now expand as required, and rather than having a single
10269 static array of bignums, BN_CTX now uses a linked-list of such arrays
10270 allowing it to expand on demand whilst maintaining the usefulness of
10271 BN_CTX's "bundling".
10275 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10276 to allow all RSA operations to function using a single BN_CTX.
10280 * Preliminary support for certificate policy evaluation and checking. This
10281 is initially intended to pass the tests outlined in "Conformance Testing
10282 of Relying Party Client Certificate Path Processing Logic" v1.07.
10286 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10287 remained unused and not that useful. A variety of other little bignum
10288 tweaks and fixes have also been made continuing on from the audit (see
10293 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10294 associated ASN1, EVP and SSL functions and old ASN1 macros.
10298 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10299 and this should never fail. So the return value from the use of
10300 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10301 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10305 * BN_CTX_get() should return zero-valued bignums, providing the same
10306 initialised value as BN_new().
10308 *Geoff Thorpe, suggested by Ulf Möller*
10310 * Support for inhibitAnyPolicy certificate extension.
10314 * An audit of the BIGNUM code is underway, for which debugging code is
10315 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10316 is considered valid when processing BIGNUMs, and causes execution to
10317 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10318 further steps are taken to deliberately pollute unused data in BIGNUM
10319 structures to try and expose faulty code further on. For now, openssl will
10320 (in its default mode of operation) continue to tolerate the inconsistent
10321 forms that it has tolerated in the past, but authors and packagers should
10322 consider trying openssl and their own applications when compiled with
10323 these debugging symbols defined. It will help highlight potential bugs in
10324 their own code, and will improve the test coverage for OpenSSL itself. At
10325 some point, these tighter rules will become openssl's default to improve
10326 maintainability, though the assert()s and other overheads will remain only
10327 in debugging configurations. See bn.h for more details.
10329 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10331 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10332 that can only be obtained through BN_CTX_new() (which implicitly
10333 initialises it). The presence of this function only made it possible
10334 to overwrite an existing structure (and cause memory leaks).
10338 * Because of the callback-based approach for implementing LHASH as a
10339 template type, lh_insert() adds opaque objects to hash-tables and
10340 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10341 to clean up those corresponding objects before destroying the hash table
10342 (and losing the object pointers). So some over-zealous constifications in
10343 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10344 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10345 prototyped to have "const" restrictions on the object pointers they are
10346 given (and so aren't required to cast them away any more).
10350 * The tmdiff.h API was so ugly and minimal that our own timing utility
10351 (speed) prefers to use its own implementation. The two implementations
10352 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10353 its object type properly exposed (MS_TM) instead of casting to/from
10354 `char *`. This may still change yet if someone realises MS_TM and
10356 aren't necessarily the greatest nomenclatures - but this is what was used
10357 internally to the implementation so I've used that for now.
10361 * Ensure that deprecated functions do not get compiled when
10362 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10363 the self-tests were still using deprecated key-generation functions so
10364 these have been updated also.
10368 * Reorganise PKCS#7 code to separate the digest location functionality
10369 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10370 New function PKCS7_set_digest() to set the digest type for PKCS#7
10371 digestedData type. Add additional code to correctly generate the
10372 digestedData type and add support for this type in PKCS7 initialization
10377 * New function PKCS7_set0_type_other() this initializes a PKCS7
10378 structure of type "other".
10382 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10383 sure the loop does correctly stop and breaking ("division by zero")
10384 modulus operations are not performed. The (pre-generated) prime
10385 table crypto/bn/bn_prime.h was already correct, but it could not be
10386 re-generated on some platforms because of the "division by zero"
10387 situation in the script.
10389 *Ralf S. Engelschall*
10391 * Update support for ECC-based TLS ciphersuites according to
10392 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10393 SHA-1 now is only used for "small" curves (where the
10394 representation of a field element takes up to 24 bytes); for
10395 larger curves, the field element resulting from ECDH is directly
10396 used as premaster secret.
10398 *Douglas Stebila (Sun Microsystems Laboratories)*
10400 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10401 curve secp160r1 to the tests.
10403 *Douglas Stebila (Sun Microsystems Laboratories)*
10405 * Add the possibility to load symbols globally with DSO.
10407 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10409 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10410 control of the error stack.
10414 * Add support for STORE in ENGINE.
10418 * Add the STORE type. The intention is to provide a common interface
10419 to certificate and key stores, be they simple file-based stores, or
10420 HSM-type store, or LDAP stores, or...
10421 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10425 * Add a generic structure called OPENSSL_ITEM. This can be used to
10426 pass a list of arguments to any function as well as provide a way
10427 for a function to pass data back to the caller.
10431 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10432 works like BUF_strdup() but can be used to duplicate a portion of
10433 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10438 * Add the function sk_find_ex() which works like sk_find(), but will
10439 return an index to an element even if an exact match couldn't be
10440 found. The index is guaranteed to point at the element where the
10441 searched-for key would be inserted to preserve sorting order.
10445 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10446 takes an extra flags argument for optional functionality. Currently,
10447 the following flags are defined:
10449 OBJ_BSEARCH_VALUE_ON_NOMATCH
10450 This one gets OBJ_bsearch_ex() to return a pointer to the first
10451 element where the comparing function returns a negative or zero
10454 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10455 This one gets OBJ_bsearch_ex() to return a pointer to the first
10456 element where the comparing function returns zero. This is useful
10457 if there are more than one element where the comparing function
10462 * Make it possible to create self-signed certificates with 'openssl ca'
10463 in such a way that the self-signed certificate becomes part of the
10464 CA database and uses the same mechanisms for serial number generation
10465 as all other certificate signing. The new flag '-selfsign' enables
10466 this functionality. Adapt CA.sh and CA.pl.in.
10470 * Add functionality to check the public key of a certificate request
10471 against a given private. This is useful to check that a certificate
10472 request can be signed by that key (self-signing).
10476 * Make it possible to have multiple active certificates with the same
10477 subject in the CA index file. This is done only if the keyword
10478 'unique_subject' is set to 'no' in the main CA section (default
10479 if 'CA_default') of the configuration file. The value is saved
10480 with the database itself in a separate index attribute file,
10481 named like the index file with '.attr' appended to the name.
10485 * Generate multi-valued AVAs using '+' notation in config files for
10490 * Support for nameConstraints certificate extension.
10494 * Support for policyConstraints certificate extension.
10498 * Support for policyMappings certificate extension.
10502 * Make sure the default DSA_METHOD implementation only uses its
10503 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10504 and change its own handlers to be NULL so as to remove unnecessary
10505 indirection. This lets alternative implementations fallback to the
10506 default implementation more easily.
10510 * Support for directoryName in GeneralName related extensions
10515 * Make it possible to link applications using Makefile.shared.
10516 Make that possible even when linking against static libraries!
10520 * Support for single pass processing for S/MIME signing. This now
10521 means that S/MIME signing can be done from a pipe, in addition
10522 cleartext signing (multipart/signed type) is effectively streaming
10523 and the signed data does not need to be all held in memory.
10525 This is done with a new flag PKCS7_STREAM. When this flag is set
10526 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10527 is done after the data is output (and digests calculated) in
10528 SMIME_write_PKCS7().
10532 * Add full support for -rpath/-R, both in shared libraries and
10533 applications, at least on the platforms where it's known how
10538 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10539 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10540 will now compute a table of multiples of the generator that
10541 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10542 faster (notably in the case of a single point multiplication,
10543 scalar * generator).
10545 *Nils Larsch, Bodo Moeller*
10547 * IPv6 support for certificate extensions. The various extensions
10548 which use the IP:a.b.c.d can now take IPv6 addresses using the
10549 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10554 * Added an ENGINE that implements RSA by performing private key
10555 exponentiations with the GMP library. The conversions to and from
10556 GMP's mpz_t format aren't optimised nor are any montgomery forms
10557 cached, and on x86 it appears OpenSSL's own performance has caught up.
10558 However there are likely to be other architectures where GMP could
10559 provide a boost. This ENGINE is not built in by default, but it can be
10560 specified at Configure time and should be accompanied by the necessary
10561 linker additions, eg;
10562 ./config -DOPENSSL_USE_GMP -lgmp
10566 * "openssl engine" will not display ENGINE/DSO load failure errors when
10567 testing availability of engines with "-t" - the old behaviour is
10568 produced by increasing the feature's verbosity with "-tt".
10572 * ECDSA routines: under certain error conditions uninitialized BN objects
10573 could be freed. Solution: make sure initialization is performed early
10574 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10579 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10580 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10581 software implementations. For DSA and DH, parameter generation can
10582 also be overridden by providing the appropriate method callbacks.
10586 * Change the "progress" mechanism used in key-generation and
10587 primality testing to functions that take a new BN_GENCB pointer in
10588 place of callback/argument pairs. The new API functions have `_ex`
10589 postfixes and the older functions are reimplemented as wrappers for
10590 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10591 declarations of the old functions to help (graceful) attempts to
10592 migrate to the new functions. Also, the new key-generation API
10593 functions operate on a caller-supplied key-structure and return
10594 success/failure rather than returning a key or NULL - this is to
10595 help make "keygen" another member function of RSA_METHOD etc.
10597 Example for using the new callback interface:
10599 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10600 void *my_arg = ...;
10603 BN_GENCB_set(&my_cb, my_callback, my_arg);
10605 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10606 /* For the meaning of a, b in calls to my_callback(), see the
10607 * documentation of the function that calls the callback.
10608 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10609 * my_callback should return 1 if it wants BN_is_prime_ex()
10610 * to continue, or 0 to stop.
10615 * Change the ZLIB compression method to be stateful, and make it
10616 available to TLS with the number defined in
10617 draft-ietf-tls-compression-04.txt.
10621 * Add the ASN.1 structures and functions for CertificatePair, which
10622 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10624 CertificatePair ::= SEQUENCE {
10625 forward [0] Certificate OPTIONAL,
10626 reverse [1] Certificate OPTIONAL,
10627 -- at least one of the pair shall be present -- }
10629 Also implement the PEM functions to read and write certificate
10630 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10632 This needed to be defined, mostly for the sake of the LDAP
10633 attribute crossCertificatePair, but may prove useful elsewhere as
10638 * Make it possible to inhibit symlinking of shared libraries in
10639 Makefile.shared, for Cygwin's sake.
10643 * Extend the BIGNUM API by creating a function
10644 void BN_set_negative(BIGNUM *a, int neg);
10645 and a macro that behave like
10646 int BN_is_negative(const BIGNUM *a);
10648 to avoid the need to access 'a->neg' directly in applications.
10652 * Implement fast modular reduction for pseudo-Mersenne primes
10653 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10654 EC_GROUP_new_curve_GFp() will now automatically use this
10657 *Nils Larsch <nla@trustcenter.de>*
10659 * Add new lock type (CRYPTO_LOCK_BN).
10663 * Change the ENGINE framework to automatically load engines
10664 dynamically from specific directories unless they could be
10665 found to already be built in or loaded. Move all the
10666 current engines except for the cryptodev one to a new
10667 directory engines/.
10668 The engines in engines/ are built as shared libraries if
10669 the "shared" options was given to ./Configure or ./config.
10670 Otherwise, they are inserted in libcrypto.a.
10671 /usr/local/ssl/engines is the default directory for dynamic
10672 engines, but that can be overridden at configure time through
10673 the usual use of --prefix and/or --openssldir, and at run
10674 time with the environment variable OPENSSL_ENGINES.
10676 *Geoff Thorpe and Richard Levitte*
10678 * Add Makefile.shared, a helper makefile to build shared
10679 libraries. Adapt Makefile.org.
10683 * Add version info to Win32 DLLs.
10685 *Peter 'Luna' Runestig" <peter@runestig.com>*
10687 * Add new 'medium level' PKCS#12 API. Certificates and keys
10688 can be added using this API to created arbitrary PKCS#12
10689 files while avoiding the low-level API.
10691 New options to PKCS12_create(), key or cert can be NULL and
10692 will then be omitted from the output file. The encryption
10693 algorithm NIDs can be set to -1 for no encryption, the mac
10694 iteration count can be set to 0 to omit the mac.
10696 Enhance pkcs12 utility by making the -nokeys and -nocerts
10697 options work when creating a PKCS#12 file. New option -nomac
10698 to omit the mac, NONE can be set for an encryption algorithm.
10699 New code is modified to use the enhanced PKCS12_create()
10700 instead of the low-level API.
10704 * Extend ASN1 encoder to support indefinite length constructed
10705 encoding. This can output sequences tags and octet strings in
10706 this form. Modify pk7_asn1.c to support indefinite length
10707 encoding. This is experimental and needs additional code to
10708 be useful, such as an ASN1 bio and some enhanced streaming
10711 Extend template encode functionality so that tagging is passed
10712 down to the template encoder.
10716 * Let 'openssl req' fail if an argument to '-newkey' is not
10717 recognized instead of using RSA as a default.
10721 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10722 As these are not official, they are not included in "ALL";
10723 the "ECCdraft" ciphersuite group alias can be used to select them.
10725 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10727 * Add ECDH engine support.
10729 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10731 * Add ECDH in new directory crypto/ecdh/.
10733 *Douglas Stebila (Sun Microsystems Laboratories)*
10735 * Let BN_rand_range() abort with an error after 100 iterations
10736 without success (which indicates a broken PRNG).
10740 * Change BN_mod_sqrt() so that it verifies that the input value
10741 is really the square of the return value. (Previously,
10742 BN_mod_sqrt would show GIGO behaviour.)
10746 * Add named elliptic curves over binary fields from X9.62, SECG,
10747 and WAP/WTLS; add OIDs that were still missing.
10749 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10751 * Extend the EC library for elliptic curves over binary fields
10752 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10755 EC_GF2m_simple_method
10759 EC_GROUP_new_curve_GF2m
10760 EC_GROUP_set_curve_GF2m
10761 EC_GROUP_get_curve_GF2m
10762 EC_POINT_set_affine_coordinates_GF2m
10763 EC_POINT_get_affine_coordinates_GF2m
10764 EC_POINT_set_compressed_coordinates_GF2m
10766 Point compression for binary fields is disabled by default for
10767 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10770 As binary polynomials are represented as BIGNUMs, various members
10771 of the EC_GROUP and EC_POINT data structures can be shared
10772 between the implementations for prime fields and binary fields;
10773 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10774 are essentially identical to their `..._GFp` counterparts.
10775 (For simplicity, the `..._GFp` prefix has been dropped from
10776 various internal method names.)
10778 An internal 'field_div' method (similar to 'field_mul' and
10779 'field_sqr') has been added; this is used only for binary fields.
10781 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10783 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10784 through methods ('mul', 'precompute_mult').
10786 The generic implementations (now internally called 'ec_wNAF_mul'
10787 and 'ec_wNAF_precomputed_mult') remain the default if these
10788 methods are undefined.
10790 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10792 * New function EC_GROUP_get_degree, which is defined through
10793 EC_METHOD. For curves over prime fields, this returns the bit
10794 length of the modulus.
10796 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10798 * New functions EC_GROUP_dup, EC_POINT_dup.
10799 (These simply call ..._new and ..._copy).
10801 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10803 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10804 Polynomials are represented as BIGNUMs (where the sign bit is not
10805 used) in the following functions [macros]:
10808 BN_GF2m_sub [= BN_GF2m_add]
10809 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10810 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10811 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10813 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10814 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10815 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10816 BN_GF2m_cmp [= BN_ucmp]
10818 (Note that only the 'mod' functions are actually for fields GF(2^m).
10819 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10821 For some functions, an the irreducible polynomial defining a
10822 field can be given as an 'unsigned int[]' with strictly
10823 decreasing elements giving the indices of those bits that are set;
10824 i.e., p[] represents the polynomial
10825 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10827 p[0] > p[1] > ... > p[k] = 0.
10828 This applies to the following functions:
10831 BN_GF2m_mod_mul_arr
10832 BN_GF2m_mod_sqr_arr
10833 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10834 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10835 BN_GF2m_mod_exp_arr
10836 BN_GF2m_mod_sqrt_arr
10837 BN_GF2m_mod_solve_quad_arr
10841 Conversion can be performed by the following functions:
10846 bntest.c has additional tests for binary polynomial arithmetic.
10848 Two implementations for BN_GF2m_mod_div() are available.
10849 The default algorithm simply uses BN_GF2m_mod_inv() and
10850 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10851 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10852 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10854 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10856 * Add new error code 'ERR_R_DISABLED' that can be used when some
10857 functionality is disabled at compile-time.
10859 *Douglas Stebila <douglas.stebila@sun.com>*
10861 * Change default behaviour of 'openssl asn1parse' so that more
10862 information is visible when viewing, e.g., a certificate:
10864 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10865 mode the content of non-printable OCTET STRINGs is output in a
10866 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10867 avoid the appearance of a printable string.
10869 *Nils Larsch <nla@trustcenter.de>*
10871 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10873 EC_GROUP_set_asn1_flag()
10874 EC_GROUP_get_asn1_flag()
10875 EC_GROUP_set_point_conversion_form()
10876 EC_GROUP_get_point_conversion_form()
10877 These control ASN1 encoding details:
10878 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10879 has been set to OPENSSL_EC_NAMED_CURVE.
10880 - Points are encoded in uncompressed form by default; options for
10881 asn1_for are as for point2oct, namely
10882 POINT_CONVERSION_COMPRESSED
10883 POINT_CONVERSION_UNCOMPRESSED
10884 POINT_CONVERSION_HYBRID
10886 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10888 EC_GROUP_set_seed()
10889 EC_GROUP_get0_seed()
10890 EC_GROUP_get_seed_len()
10891 This is used only for ASN1 purposes (so far).
10893 *Nils Larsch <nla@trustcenter.de>*
10895 * Add 'field_type' member to EC_METHOD, which holds the NID
10896 of the appropriate field type OID. The new function
10897 EC_METHOD_get_field_type() returns this value.
10899 *Nils Larsch <nla@trustcenter.de>*
10902 EC_POINT_point2bn()
10903 EC_POINT_bn2point()
10904 EC_POINT_point2hex()
10905 EC_POINT_hex2point()
10906 providing useful interfaces to EC_POINT_point2oct() and
10907 EC_POINT_oct2point().
10909 *Nils Larsch <nla@trustcenter.de>*
10911 * Change internals of the EC library so that the functions
10912 EC_GROUP_set_generator()
10913 EC_GROUP_get_generator()
10914 EC_GROUP_get_order()
10915 EC_GROUP_get_cofactor()
10916 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10917 to methods, which would lead to unnecessary code duplication when
10918 adding different types of curves.
10920 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10922 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10923 arithmetic, and such that modified wNAFs are generated
10924 (which avoid length expansion in many cases).
10928 * Add a function EC_GROUP_check_discriminant() (defined via
10929 EC_METHOD) that verifies that the curve discriminant is non-zero.
10931 Add a function EC_GROUP_check() that makes some sanity tests
10932 on a EC_GROUP, its generator and order. This includes
10933 EC_GROUP_check_discriminant().
10935 *Nils Larsch <nla@trustcenter.de>*
10937 * Add ECDSA in new directory crypto/ecdsa/.
10939 Add applications 'openssl ecparam' and 'openssl ecdsa'
10940 (these are based on 'openssl dsaparam' and 'openssl dsa').
10942 ECDSA support is also included in various other files across the
10943 library. Most notably,
10944 - 'openssl req' now has a '-newkey ecdsa:file' option;
10945 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10946 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10947 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10948 them suitable for ECDSA where domain parameters must be
10949 extracted before the specific public key;
10950 - ECDSA engine support has been added.
10952 *Nils Larsch <nla@trustcenter.de>*
10954 * Include some named elliptic curves, and add OIDs from X9.62,
10955 SECG, and WAP/WTLS. Each curve can be obtained from the new
10957 EC_GROUP_new_by_curve_name(),
10958 and the list of available named curves can be obtained with
10959 EC_get_builtin_curves().
10960 Also add a 'curve_name' member to EC_GROUP objects, which can be
10962 EC_GROUP_set_curve_name()
10963 EC_GROUP_get_curve_name()
10965 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10967 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10968 was actually never needed) and in BN_mul(). The removal in BN_mul()
10969 required a small change in bn_mul_part_recursive() and the addition
10970 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10971 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10972 bn_sub_words() and bn_add_words() except they take arrays with
10977 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10979 * Cleanse PEM buffers before freeing them since they may contain
10982 *Benjamin Bennett <ben@psc.edu>*
10984 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10985 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10986 authentication-only ciphersuites.
10990 * Since AES128 and AES256 share a single mask bit in the logic of
10991 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10992 kludge to work properly if AES128 is available and AES256 isn't.
10996 * Expand security boundary to match 1.1.1 module.
11000 * Remove redundant features: hash file source, editing of test vectors
11001 modify fipsld to use external fips_premain.c signature.
11005 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11006 run algorithm test programs.
11010 * Make algorithm test programs more tolerant of whitespace.
11014 * Have SSL/TLS server implementation tolerate "mismatched" record
11015 protocol version while receiving ClientHello even if the
11016 ClientHello is fragmented. (The server can't insist on the
11017 particular protocol version it has chosen before the ServerHello
11018 message has informed the client about his choice.)
11022 * Load error codes if they are not already present instead of using a
11023 static variable. This allows them to be cleanly unloaded and reloaded.
11027 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11029 * Introduce limits to prevent malicious keys being able to
11030 cause a denial of service. ([CVE-2006-2940])
11032 *Steve Henson, Bodo Moeller*
11034 * Fix ASN.1 parsing of certain invalid structures that can result
11035 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11037 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11038 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11040 * Fix SSL client code which could crash if connecting to a
11041 malicious SSLv2 server. ([CVE-2006-4343])
11043 *Tavis Ormandy and Will Drewry, Google Security Team*
11045 * Change ciphersuite string processing so that an explicit
11046 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11047 will no longer include "AES128-SHA"), and any other similar
11048 ciphersuite (same bitmap) from *other* protocol versions (so that
11049 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11050 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11051 changes from 0.9.8b and 0.9.8d.
11055 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11057 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11058 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11060 * Change the Unix randomness entropy gathering to use poll() when
11061 possible instead of select(), since the latter has some
11062 undesirable limitations.
11064 *Darryl Miles via Richard Levitte and Bodo Moeller*
11066 * Disable rogue ciphersuites:
11068 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11069 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11070 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11072 The latter two were purportedly from
11073 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11076 Also deactivate the remaining ciphersuites from
11077 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11078 unofficial, and the ID has long expired.
11082 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11083 dual-core machines) and other potential thread-safety issues.
11087 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11089 * Adapt fipsld and the build system to link against the validated FIPS
11090 module in FIPS mode.
11094 * Fixes for VC++ 2005 build under Windows.
11098 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11099 from a Windows bash shell such as MSYS. It is autodetected from the
11100 "config" script when run from a VC++ environment. Modify standard VC++
11101 build to use fipscanister.o from the GNU make build.
11105 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11107 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11108 The value now differs depending on if you build for FIPS or not.
11109 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11110 safely run with a non-FIPSed libcrypto, as it may crash because of
11111 the difference induced by this change.
11115 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11117 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11118 (part of SSL_OP_ALL). This option used to disable the
11119 countermeasure against man-in-the-middle protocol-version
11120 rollback in the SSL 2.0 server implementation, which is a bad
11121 idea. ([CVE-2005-2969])
11123 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11124 for Information Security, National Institute of Advanced Industrial
11125 Science and Technology [AIST, Japan)]*
11127 * Minimal support for X9.31 signatures and PSS padding modes. This is
11128 mainly for FIPS compliance and not fully integrated at this stage.
11132 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11133 the exponentiation using a fixed-length exponent. (Otherwise,
11134 the information leaked through timing could expose the secret key
11135 after many signatures; cf. Bleichenbacher's attack on DSA with
11140 * Make a new fixed-window mod_exp implementation the default for
11141 RSA, DSA, and DH private-key operations so that the sequence of
11142 squares and multiplies and the memory access pattern are
11143 independent of the particular secret key. This will mitigate
11144 cache-timing and potential related attacks.
11146 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11147 and this is automatically used by BN_mod_exp_mont() if the new flag
11148 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11149 will use this BN flag for private exponents unless the flag
11150 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11151 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11153 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11155 * Change the client implementation for SSLv23_method() and
11156 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11157 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11158 (Previously, the SSL 2.0 backwards compatible Client Hello
11159 message format would be used even with SSL_OP_NO_SSLv2.)
11163 * Add support for smime-type MIME parameter in S/MIME messages which some
11168 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11169 a threadsafe manner. Modify rsa code to use new function and add calls
11170 to dsa and dh code (which had race conditions before).
11174 * Include the fixed error library code in the C error file definitions
11175 instead of fixing them up at runtime. This keeps the error code
11176 structures constant.
11180 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11182 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11185 * Fixes for newer kerberos headers. NB: the casts are needed because
11186 the 'length' field is signed on one version and unsigned on another
11187 with no (?) obvious way to tell the difference, without these VC++
11188 complains. Also the "definition" of FAR (blank) is no longer included
11189 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11190 some needed definitions.
11194 * Undo Cygwin change.
11198 * Added support for proxy certificates according to RFC 3820.
11199 Because they may be a security thread to unaware applications,
11200 they must be explicitly allowed in run-time. See
11201 docs/HOWTO/proxy_certificates.txt for further information.
11205 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11207 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11208 server and client random values. Previously
11209 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11210 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11212 This change has negligible security impact because:
11214 1. Server and client random values still have 24 bytes of pseudo random
11217 2. Server and client random values are sent in the clear in the initial
11220 3. The master secret is derived using the premaster secret (48 bytes in
11221 size for static RSA ciphersuites) as well as client server and random
11224 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11227 *Stephen Henson, reported by UK NISCC*
11229 * Use Windows randomness collection on Cygwin.
11233 * Fix hang in EGD/PRNGD query when communication socket is closed
11234 prematurely by EGD/PRNGD.
11236 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11238 * Prompt for pass phrases when appropriate for PKCS12 input format.
11242 * Back-port of selected performance improvements from development
11243 branch, as well as improved support for PowerPC platforms.
11247 * Add lots of checks for memory allocation failure, error codes to indicate
11248 failure and freeing up memory if a failure occurs.
11250 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11252 * Add new -passin argument to dgst.
11256 * Perform some character comparisons of different types in X509_NAME_cmp:
11257 this is needed for some certificates that re-encode DNs into UTF8Strings
11258 (in violation of RFC3280) and can't or won't issue name rollover
11263 * Make an explicit check during certificate validation to see that
11264 the CA setting in each certificate on the chain is correct. As a
11265 side effect always do the following basic checks on extensions,
11266 not just when there's an associated purpose to the check:
11268 - if there is an unhandled critical extension (unless the user
11269 has chosen to ignore this fault)
11270 - if the path length has been exceeded (if one is set at all)
11271 - that certain extensions fit the associated purpose (if one has
11276 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11278 * Avoid a race condition when CRLs are checked in a multi threaded
11279 environment. This would happen due to the reordering of the revoked
11280 entries during signature checking and serial number lookup. Now the
11281 encoding is cached and the serial number sort performed under a lock.
11282 Add new STACK function sk_is_sorted().
11286 * Add Delta CRL to the extension code.
11290 * Various fixes to s3_pkt.c so alerts are sent properly.
11292 *David Holmes <d.holmes@f5.com>*
11294 * Reduce the chances of duplicate issuer name and serial numbers (in
11295 violation of RFC3280) using the OpenSSL certificate creation utilities.
11296 This is done by creating a random 64 bit value for the initial serial
11297 number when a serial number file is created or when a self signed
11298 certificate is created using 'openssl req -x509'. The initial serial
11299 number file is created using 'openssl x509 -next_serial' in CA.pl
11300 rather than being initialized to 1.
11304 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11306 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11307 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11309 *Joe Orton, Steve Henson*
11311 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11314 *Joe Orton, Steve Henson*
11316 * Make it possible to have multiple active certificates with the same
11317 subject in the CA index file. This is done only if the keyword
11318 'unique_subject' is set to 'no' in the main CA section (default
11319 if 'CA_default') of the configuration file. The value is saved
11320 with the database itself in a separate index attribute file,
11321 named like the index file with '.attr' appended to the name.
11325 * X509 verify fixes. Disable broken certificate workarounds when
11326 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11327 keyUsage extension present. Don't accept CRLs with unhandled critical
11328 extensions: since verify currently doesn't process CRL extensions this
11329 rejects a CRL with *any* critical extensions. Add new verify error codes
11334 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11335 A clarification of RFC2560 will require the use of OCTET STRINGs and
11336 some implementations cannot handle the current raw format. Since OpenSSL
11337 copies and compares OCSP nonces as opaque blobs without any attempt at
11338 parsing them this should not create any compatibility issues.
11342 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11343 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11344 this HMAC (and other) operations are several times slower than OpenSSL
11349 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11351 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11353 * Use the correct content when signing type "other".
11357 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11359 * Fix various bugs revealed by running the NISCC test suite:
11361 Stop out of bounds reads in the ASN1 code when presented with
11362 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11364 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11366 If verify callback ignores invalid public key errors don't try to check
11367 certificate signature with the NULL public key.
11371 * New -ignore_err option in ocsp application to stop the server
11372 exiting on the first error in a request.
11376 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11377 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11382 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11383 extra data after the compression methods not only for TLS 1.0
11384 but also for SSL 3.0 (as required by the specification).
11386 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11388 * Change X509_certificate_type() to mark the key as exported/exportable
11389 when it's 512 *bits* long, not 512 bytes.
11393 * Change AES_cbc_encrypt() so it outputs exact multiple of
11394 blocks during encryption.
11398 * Various fixes to base64 BIO and non blocking I/O. On write
11399 flushes were not handled properly if the BIO retried. On read
11400 data was not being buffered properly and had various logic bugs.
11401 This also affects blocking I/O when the data being decoded is a
11406 * Various S/MIME bugfixes and compatibility changes:
11407 output correct application/pkcs7 MIME type if
11408 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11409 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11410 of files as .eml work). Correctly handle very long lines in MIME
11415 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11417 * Countermeasure against the Klima-Pokorny-Rosa extension of
11418 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11419 a protocol version number mismatch like a decryption error
11420 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11424 * Turn on RSA blinding by default in the default implementation
11425 to avoid a timing attack. Applications that don't want it can call
11426 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11427 They would be ill-advised to do so in most cases.
11429 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11431 * Change RSA blinding code so that it works when the PRNG is not
11432 seeded (in this case, the secret RSA exponent is abused as
11433 an unpredictable seed -- if it is not unpredictable, there
11434 is no point in blinding anyway). Make RSA blinding thread-safe
11435 by remembering the creator's thread ID in rsa->blinding and
11436 having all other threads use local one-time blinding factors
11437 (this requires more computation than sharing rsa->blinding, but
11438 avoids excessive locking; and if an RSA object is not shared
11439 between threads, blinding will still be very fast).
11443 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11444 ENGINE as defaults for all supported algorithms irrespective of
11445 the 'flags' parameter. 'flags' is now honoured, so applications
11446 should make sure they are passing it correctly.
11450 * Target "mingw" now allows native Windows code to be generated in
11451 the Cygwin environment as well as with the MinGW compiler.
11455 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11457 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11458 via timing by performing a MAC computation even if incorrect
11459 block cipher padding has been found. This is a countermeasure
11460 against active attacks where the attacker has to distinguish
11461 between bad padding and a MAC verification error. ([CVE-2003-0078])
11463 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11464 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11465 Martin Vuagnoux (EPFL, Ilion)*
11467 * Make the no-err option work as intended. The intention with no-err
11468 is not to have the whole error stack handling routines removed from
11469 libcrypto, it's only intended to remove all the function name and
11470 reason texts, thereby removing some of the footprint that may not
11471 be interesting if those errors aren't displayed anyway.
11473 NOTE: it's still possible for any application or module to have its
11474 own set of error texts inserted. The routines are there, just not
11475 used by default when no-err is given.
11479 * Add support for FreeBSD on IA64.
11481 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11483 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11484 Kerberos function mit_des_cbc_cksum(). Before this change,
11485 the value returned by DES_cbc_cksum() was like the one from
11486 mit_des_cbc_cksum(), except the bytes were swapped.
11488 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11490 * Allow an application to disable the automatic SSL chain building.
11491 Before this a rather primitive chain build was always performed in
11492 ssl3_output_cert_chain(): an application had no way to send the
11493 correct chain if the automatic operation produced an incorrect result.
11495 Now the chain builder is disabled if either:
11497 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11499 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11501 The reasoning behind this is that an application would not want the
11502 auto chain building to take place if extra chain certificates are
11503 present and it might also want a means of sending no additional
11504 certificates (for example the chain has two certificates and the
11509 * Add the possibility to build without the ENGINE framework.
11511 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11513 * Under Win32 gmtime() can return NULL: check return value in
11514 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11518 * DSA routines: under certain error conditions uninitialized BN objects
11519 could be freed. Solution: make sure initialization is performed early
11520 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11521 Nils Larsch <nla@trustcenter.de> via PR#459)
11525 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11526 checked on reconnect on the client side, therefore session resumption
11527 could still fail with a "ssl session id is different" error. This
11528 behaviour is masked when SSL_OP_ALL is used due to
11529 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11530 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11531 followup to PR #377.
11535 * IA-32 assembler support enhancements: unified ELF targets, support
11536 for SCO/Caldera platforms, fix for Cygwin shared build.
11540 * Add support for FreeBSD on sparc64. As a consequence, support for
11541 FreeBSD on non-x86 processors is separate from x86 processors on
11542 the config script, much like the NetBSD support.
11544 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11546 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11548 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11551 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11552 code (06) was taken as the first octet of the session ID and the last
11553 octet was ignored consequently. As a result SSLv2 client side session
11554 caching could not have worked due to the session ID mismatch between
11556 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11561 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11562 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11567 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11568 seems that in spite of existing for more than a year, many application
11569 author have done nothing to provide the necessary callbacks, which
11570 means that this particular engine will not work properly anywhere.
11571 This is a very unfortunate situation which forces us, in the name
11572 of usability, to give the hw_ncipher.c a static lock, which is part
11574 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11575 appear in 0.9.8 or later. We EXPECT application authors to have
11576 dealt properly with this when 0.9.8 is released (unless we actually
11577 make such changes in the libcrypto locking code that changes will
11578 have to be made anyway).
11582 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11583 octets have been read, EOF or an error occurs. Without this change
11584 some truncated ASN1 structures will not produce an error.
11588 * Disable Heimdal support, since it hasn't been fully implemented.
11589 Still give the possibility to force the use of Heimdal, but with
11590 warnings and a request that patches get sent to openssl-dev.
11594 * Add the VC-CE target, introduce the WINCE sysname, and add
11595 INSTALL.WCE and appropriate conditionals to make it build.
11597 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11599 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11600 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11601 edit numbers of the version.
11603 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11605 * Introduce safe string copy and catenation functions
11606 (BUF_strlcpy() and BUF_strlcat()).
11608 *Ben Laurie (CHATS) and Richard Levitte*
11610 * Avoid using fixed-size buffers for one-line DNs.
11612 *Ben Laurie (CHATS)*
11614 * Add BUF_MEM_grow_clean() to avoid information leakage when
11615 resizing buffers containing secrets, and use where appropriate.
11617 *Ben Laurie (CHATS)*
11619 * Avoid using fixed size buffers for configuration file location.
11621 *Ben Laurie (CHATS)*
11623 * Avoid filename truncation for various CA files.
11625 *Ben Laurie (CHATS)*
11627 * Use sizeof in preference to magic numbers.
11629 *Ben Laurie (CHATS)*
11631 * Avoid filename truncation in cert requests.
11633 *Ben Laurie (CHATS)*
11635 * Add assertions to check for (supposedly impossible) buffer
11638 *Ben Laurie (CHATS)*
11640 * Don't cache truncated DNS entries in the local cache (this could
11641 potentially lead to a spoofing attack).
11643 *Ben Laurie (CHATS)*
11645 * Fix various buffers to be large enough for hex/decimal
11646 representations in a platform independent manner.
11648 *Ben Laurie (CHATS)*
11650 * Add CRYPTO_realloc_clean() to avoid information leakage when
11651 resizing buffers containing secrets, and use where appropriate.
11653 *Ben Laurie (CHATS)*
11655 * Add BIO_indent() to avoid much slightly worrying code to do
11658 *Ben Laurie (CHATS)*
11660 * Convert sprintf()/BIO_puts() to BIO_printf().
11662 *Ben Laurie (CHATS)*
11664 * buffer_gets() could terminate with the buffer only half
11667 *Ben Laurie (CHATS)*
11669 * Add assertions to prevent user-supplied crypto functions from
11670 overflowing internal buffers by having large block sizes, etc.
11672 *Ben Laurie (CHATS)*
11674 * New OPENSSL_assert() macro (similar to assert(), but enabled
11677 *Ben Laurie (CHATS)*
11679 * Eliminate unused copy of key in RC4.
11681 *Ben Laurie (CHATS)*
11683 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11685 *Ben Laurie (CHATS)*
11687 * Fix off-by-one error in EGD path.
11689 *Ben Laurie (CHATS)*
11691 * If RANDFILE path is too long, ignore instead of truncating.
11693 *Ben Laurie (CHATS)*
11695 * Eliminate unused and incorrectly sized X.509 structure
11698 *Ben Laurie (CHATS)*
11700 * Eliminate unused and dangerous function knumber().
11702 *Ben Laurie (CHATS)*
11704 * Eliminate unused and dangerous structure, KSSL_ERR.
11706 *Ben Laurie (CHATS)*
11708 * Protect against overlong session ID context length in an encoded
11709 session object. Since these are local, this does not appear to be
11712 *Ben Laurie (CHATS)*
11714 * Change from security patch (see 0.9.6e below) that did not affect
11715 the 0.9.6 release series:
11717 Remote buffer overflow in SSL3 protocol - an attacker could
11718 supply an oversized master key in Kerberos-enabled versions.
11721 *Ben Laurie (CHATS)*
11723 * Change the SSL kerb5 codes to match RFC 2712.
11727 * Make -nameopt work fully for req and add -reqopt switch.
11729 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11731 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11733 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11735 * Make sure tests can be performed even if the corresponding algorithms
11736 have been removed entirely. This was also the last step to make
11737 OpenSSL compilable with DJGPP under all reasonable conditions.
11739 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11741 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11742 to allow version independent disabling of normally unselected ciphers,
11743 which may be activated as a side-effect of selecting a single cipher.
11745 (E.g., cipher list string "RSA" enables ciphersuites that are left
11746 out of "ALL" because they do not provide symmetric encryption.
11747 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11749 *Lutz Jaenicke, Bodo Moeller*
11751 * Add appropriate support for separate platform-dependent build
11752 directories. The recommended way to make a platform-dependent
11753 build directory is the following (tested on Linux), maybe with
11756 # Place yourself outside of the OpenSSL source tree. In
11757 # this example, the environment variable OPENSSL_SOURCE
11758 # is assumed to contain the absolute OpenSSL source directory.
11759 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11760 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11761 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11762 mkdir -p `dirname $F`
11763 ln -s $OPENSSL_SOURCE/$F $F
11766 To be absolutely sure not to disturb the source tree, a "make clean"
11767 is a good thing. If it isn't successful, don't worry about it,
11768 it probably means the source directory is very clean.
11772 * Make sure any ENGINE control commands make local copies of string
11773 pointers passed to them whenever necessary. Otherwise it is possible
11774 the caller may have overwritten (or deallocated) the original string
11775 data when a later ENGINE operation tries to use the stored values.
11777 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11779 * Improve diagnostics in file reading and command-line digests.
11781 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11783 * Add AES modes CFB and OFB to the object database. Correct an
11784 error in AES-CFB decryption.
11788 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11789 allows existing EVP_CIPHER_CTX structures to be reused after
11790 calling `EVP_*Final()`. This behaviour is used by encryption
11791 BIOs and some applications. This has the side effect that
11792 applications must explicitly clean up cipher contexts with
11793 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11797 * Check the values of dna and dnb in bn_mul_recursive before calling
11798 bn_mul_comba (a non zero value means the a or b arrays do not contain
11799 n2 elements) and fallback to bn_mul_normal if either is not zero.
11803 * Fix escaping of non-ASCII characters when using the -subj option
11804 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11808 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11809 form for "surname", serialNumber has no short form.
11810 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11811 therefore remove "mail" short name for "internet 7".
11812 The OID for unique identifiers in X509 certificates is
11813 x500UniqueIdentifier, not uniqueIdentifier.
11814 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11818 * Add an "init" command to the ENGINE config module and auto initialize
11819 ENGINEs. Without any "init" command the ENGINE will be initialized
11820 after all ctrl commands have been executed on it. If init=1 the
11821 ENGINE is initialized at that point (ctrls before that point are run
11822 on the uninitialized ENGINE and after on the initialized one). If
11823 init=0 then the ENGINE will not be initialized at all.
11827 * Fix the 'app_verify_callback' interface so that the user-defined
11828 argument is actually passed to the callback: In the
11829 SSL_CTX_set_cert_verify_callback() prototype, the callback
11830 declaration has been changed from
11833 int (*cb)(X509_STORE_CTX *,void *);
11834 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11835 i=s->ctx->app_verify_callback(&ctx)
11836 has been changed into
11837 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11839 To update applications using SSL_CTX_set_cert_verify_callback(),
11840 a dummy argument can be added to their callback functions.
11842 *D. K. Smetters <smetters@parc.xerox.com>*
11844 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11846 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11848 * Add and OPENSSL_LOAD_CONF define which will cause
11849 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11850 This allows older applications to transparently support certain
11851 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11852 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11853 load the config file and OPENSSL_add_all_algorithms_conf() which will
11854 always load it have also been added.
11858 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11859 Adjust NIDs and EVP layer.
11861 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11863 * Config modules support in openssl utility.
11865 Most commands now load modules from the config file,
11866 though in a few (such as version) this isn't done
11867 because it couldn't be used for anything.
11869 In the case of ca and req the config file used is
11870 the same as the utility itself: that is the -config
11871 command line option can be used to specify an
11876 * Move default behaviour from OPENSSL_config(). If appname is NULL
11877 use "openssl_conf" if filename is NULL use default openssl config file.
11881 * Add an argument to OPENSSL_config() to allow the use of an alternative
11882 config section name. Add a new flag to tolerate a missing config file
11883 and move code to CONF_modules_load_file().
11887 * Support for crypto accelerator cards from Accelerated Encryption
11888 Processing, www.aep.ie. (Use engine 'aep')
11889 The support was copied from 0.9.6c [engine] and adapted/corrected
11890 to work with the new engine framework.
11892 *AEP Inc. and Richard Levitte*
11894 * Support for SureWare crypto accelerator cards from Baltimore
11895 Technologies. (Use engine 'sureware')
11896 The support was copied from 0.9.6c [engine] and adapted
11897 to work with the new engine framework.
11901 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11902 make the newer ENGINE framework commands for the CHIL engine work.
11904 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11906 * Make it possible to produce shared libraries on ReliantUNIX.
11908 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11910 * Add the configuration target debug-linux-ppro.
11911 Make 'openssl rsa' use the general key loading routines
11912 implemented in `apps.c`, and make those routines able to
11913 handle the key format FORMAT_NETSCAPE and the variant
11916 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11918 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11920 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11922 * Add -keyform to rsautl, and document -engine.
11924 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11926 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11927 BIO_R_NO_SUCH_FILE error code rather than the generic
11928 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11932 * Add new functions
11933 ERR_peek_last_error
11934 ERR_peek_last_error_line
11935 ERR_peek_last_error_line_data.
11936 These are similar to
11938 ERR_peek_error_line
11939 ERR_peek_error_line_data,
11940 but report on the latest error recorded rather than the first one
11941 still in the error queue.
11943 *Ben Laurie, Bodo Moeller*
11945 * default_algorithms option in ENGINE config module. This allows things
11947 default_algorithms = ALL
11948 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11952 * Preliminary ENGINE config module.
11956 * New experimental application configuration code.
11960 * Change the AES code to follow the same name structure as all other
11961 symmetric ciphers, and behave the same way. Move everything to
11962 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11964 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11966 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11968 *Ben Laurie and Theo de Raadt*
11970 * Add option to output public keys in req command.
11972 *Massimiliano Pala madwolf@openca.org*
11974 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11975 (up to about 10% better than before for P-192 and P-224).
11979 * New functions/macros
11981 SSL_CTX_set_msg_callback(ctx, cb)
11982 SSL_CTX_set_msg_callback_arg(ctx, arg)
11983 SSL_set_msg_callback(ssl, cb)
11984 SSL_set_msg_callback_arg(ssl, arg)
11986 to request calling a callback function
11988 void cb(int write_p, int version, int content_type,
11989 const void *buf, size_t len, SSL *ssl, void *arg)
11991 whenever a protocol message has been completely received
11992 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11993 protocol version according to which the SSL library interprets
11994 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11995 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11996 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11997 specification (change_cipher_spec(20), alert(21), handshake(22)).
11998 'buf' and 'len' point to the actual message, 'ssl' to the
11999 SSL object, and 'arg' is the application-defined value set by
12000 SSL[_CTX]_set_msg_callback_arg().
12002 'openssl s_client' and 'openssl s_server' have new '-msg' options
12003 to enable a callback that displays all protocol messages.
12007 * Change the shared library support so shared libraries are built as
12008 soon as the corresponding static library is finished, and thereby get
12009 openssl and the test programs linked against the shared library.
12010 This still only happens when the keyword "shard" has been given to
12011 the configuration scripts.
12013 NOTE: shared library support is still an experimental thing, and
12014 backward binary compatibility is still not guaranteed.
12016 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12018 * Add support for Subject Information Access extension.
12020 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12022 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12023 additional bytes when new memory had to be allocated, not just
12024 when reusing an existing buffer.
12028 * New command line and configuration option 'utf8' for the req command.
12029 This allows field values to be specified as UTF8 strings.
12033 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12034 runs for the former and machine-readable output for the latter.
12038 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12039 of the e-mail address in the DN (i.e., it will go into a certificate
12040 extension only). The new configuration file option 'email_in_dn = no'
12041 has the same effect.
12043 *Massimiliano Pala madwolf@openca.org*
12045 * Change all functions with names starting with `des_` to be starting
12046 with `DES_` instead. Add wrappers that are compatible with libdes,
12047 but are named `_ossl_old_des_*`. Finally, add macros that map the
12048 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12049 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12050 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12053 Since we provide two compatibility mappings, the user needs to
12054 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12055 compatibility is desired. The default (i.e., when that macro
12056 isn't defined) is OpenSSL 0.9.6c compatibility.
12058 There are also macros that enable and disable the support of old
12059 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12060 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12061 are defined, the default will apply: to support the old des routines.
12063 In either case, one must include openssl/des.h to get the correct
12064 definitions. Do not try to just include openssl/des_old.h, that
12067 NOTE: This is a major break of an old API into a new one. Software
12068 authors are encouraged to switch to the `DES_` style functions. Some
12069 time in the future, des_old.h and the libdes compatibility functions
12070 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12071 default), and then completely removed.
12075 * Test for certificates which contain unsupported critical extensions.
12076 If such a certificate is found during a verify operation it is
12077 rejected by default: this behaviour can be overridden by either
12078 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12079 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12080 X509_supported_extension() has also been added which returns 1 if a
12081 particular extension is supported.
12085 * Modify the behaviour of EVP cipher functions in similar way to digests
12086 to retain compatibility with existing code.
12090 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12091 compatibility with existing code. In particular the 'ctx' parameter does
12092 not have to be to be initialized before the call to EVP_DigestInit() and
12093 it is tidied up after a call to EVP_DigestFinal(). New function
12094 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12095 EVP_MD_CTX_copy() changed to not require the destination to be
12096 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12097 requires the destination to be valid.
12099 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12100 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12104 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12105 so that complete 'Handshake' protocol structures are kept in memory
12106 instead of overwriting 'msg_type' and 'length' with 'body' data.
12110 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12112 *Massimo Santin via Richard Levitte*
12114 * Major restructuring to the underlying ENGINE code. This includes
12115 reduction of linker bloat, separation of pure "ENGINE" manipulation
12116 (initialisation, etc) from functionality dealing with implementations
12117 of specific crypto interfaces. This change also introduces integrated
12118 support for symmetric ciphers and digest implementations - so ENGINEs
12119 can now accelerate these by providing EVP_CIPHER and EVP_MD
12120 implementations of their own. This is detailed in
12121 [crypto/engine/README.md](crypto/engine/README.md)
12122 as it couldn't be adequately described here. However, there are a few
12123 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12124 were changed in the original introduction of ENGINE code have now
12125 reverted back - the hooking from this code to ENGINE is now a good
12126 deal more passive and at run-time, operations deal directly with
12127 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12128 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12129 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12130 they were not being used by the framework as there is no concept of a
12131 BIGNUM_METHOD and they could not be generalised to the new
12132 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12133 ENGINE_cpy() has been removed as it cannot be consistently defined in
12138 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12142 * Change mkdef.pl to sort symbols that get the same entry number,
12143 and make sure the automatically generated functions `ERR_load_*`
12144 become part of libeay.num as well.
12148 * New function SSL_renegotiate_pending(). This returns true once
12149 renegotiation has been requested (either SSL_renegotiate() call
12150 or HelloRequest/ClientHello received from the peer) and becomes
12151 false once a handshake has been completed.
12152 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12153 sends a HelloRequest, but does not ensure that a handshake takes
12154 place. SSL_renegotiate_pending() is useful for checking if the
12155 client has followed the request.)
12159 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12160 By default, clients may request session resumption even during
12161 renegotiation (if session ID contexts permit); with this option,
12162 session resumption is possible only in the first handshake.
12164 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12165 more bits available for options that should not be part of
12166 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12170 * Add some demos for certificate and certificate request creation.
12174 * Make maximum certificate chain size accepted from the peer application
12175 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12176 "Douglas E. Engert" <deengert@anl.gov>.
12180 * Add support for shared libraries for Unixware-7
12181 (Boyd Lynn Gerber <gerberb@zenez.com>).
12185 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12186 be done prior to destruction. Use this to unload error strings from
12187 ENGINEs that load their own error strings. NB: This adds two new API
12188 functions to "get" and "set" this destroy handler in an ENGINE.
12192 * Alter all existing ENGINE implementations (except "openssl" and
12193 "openbsd") to dynamically instantiate their own error strings. This
12194 makes them more flexible to be built both as statically-linked ENGINEs
12195 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12196 Also, add stub code to each that makes building them as self-contained
12197 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12201 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12202 implementations into applications that are completely implemented in
12203 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12204 commands that can be used to configure what shared-library to load and
12205 to control aspects of the way it is handled. Also, made an update to
12206 the [README-Engine.md](README-Engine.md) file
12207 that brings its information up-to-date and
12208 provides some information and instructions on the "dynamic" ENGINE
12209 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12213 * Make it possible to unload ranges of ERR strings with a new
12214 "ERR_unload_strings" function.
12218 * Add a copy() function to EVP_MD.
12222 * Make EVP_MD routines take a context pointer instead of just the
12223 md_data void pointer.
12227 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12228 that the digest can only process a single chunk of data
12229 (typically because it is provided by a piece of
12230 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12231 is only going to provide a single chunk of data, and hence the
12232 framework needn't accumulate the data for oneshot drivers.
12236 * As with "ERR", make it possible to replace the underlying "ex_data"
12237 functions. This change also alters the storage and management of global
12238 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12239 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12240 index counters. The API functions that use this state have been changed
12241 to take a "class_index" rather than pointers to the class's local STACK
12242 and counter, and there is now an API function to dynamically create new
12243 classes. This centralisation allows us to (a) plug a lot of the
12244 thread-safety problems that existed, and (b) makes it possible to clean
12245 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12246 such data would previously have always leaked in application code and
12247 workarounds were in place to make the memory debugging turn a blind eye
12248 to it. Application code that doesn't use this new function will still
12249 leak as before, but their memory debugging output will announce it now
12250 rather than letting it slide.
12252 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12253 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12254 has a return value to indicate success or failure.
12258 * Make it possible to replace the underlying "ERR" functions such that the
12259 global state (2 LHASH tables and 2 locks) is only used by the "default"
12260 implementation. This change also adds two functions to "get" and "set"
12261 the implementation prior to it being automatically set the first time
12262 any other ERR function takes place. Ie. an application can call "get",
12263 pass the return value to a module it has just loaded, and that module
12264 can call its own "set" function using that value. This means the
12265 module's "ERR" operations will use (and modify) the error state in the
12266 application and not in its own statically linked copy of OpenSSL code.
12270 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12271 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12272 the operation, and provides a more encapsulated way for external code
12273 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12274 to use these functions rather than manually incrementing the counts.
12276 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12280 * Add EVP test program.
12284 * Add symmetric cipher support to ENGINE. Expect the API to change!
12288 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12289 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12290 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12291 These allow a CRL to be built without having to access X509_CRL fields
12292 directly. Modify 'ca' application to use new functions.
12296 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12297 bug workarounds. Rollback attack detection is a security feature.
12298 The problem will only arise on OpenSSL servers when TLSv1 is not
12299 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12300 Software authors not wanting to support TLSv1 will have special reasons
12301 for their choice and can explicitly enable this option.
12303 *Bodo Moeller, Lutz Jaenicke*
12305 * Rationalise EVP so it can be extended: don't include a union of
12306 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12307 (similar to those existing for EVP_CIPHER_CTX).
12312 EVP_MD_CTX_init(&md); /* new function call */
12313 EVP_DigestInit(&md, EVP_sha1());
12314 EVP_DigestUpdate(&md, in, len);
12315 EVP_DigestFinal(&md, out, NULL);
12316 EVP_MD_CTX_cleanup(&md); /* new function call */
12320 * Make DES key schedule conform to the usual scheme, as well as
12321 correcting its structure. This means that calls to DES functions
12322 now have to pass a pointer to a des_key_schedule instead of a
12323 plain des_key_schedule (which was actually always a pointer
12326 des_key_schedule ks;
12328 des_set_key_checked(..., &ks);
12329 des_ncbc_encrypt(..., &ks, ...);
12331 (Note that a later change renames 'des_...' into 'DES_...'.)
12335 * Initial reduction of linker bloat: the use of some functions, such as
12336 PEM causes large amounts of unused functions to be linked in due to
12337 poor organisation. For example pem_all.c contains every PEM function
12338 which has a knock on effect of linking in large amounts of (unused)
12339 ASN1 code. Grouping together similar functions and splitting unrelated
12340 functions prevents this.
12344 * Cleanup of EVP macros.
12348 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12349 correct `_ecb suffix`.
12353 * Add initial OCSP responder support to ocsp application. The
12354 revocation information is handled using the text based index
12355 use by the ca application. The responder can either handle
12356 requests generated internally, supplied in files (for example
12357 via a CGI script) or using an internal minimal server.
12361 * Add configuration choices to get zlib compression for TLS.
12365 * Changes to Kerberos SSL for RFC 2712 compliance:
12366 1. Implemented real KerberosWrapper, instead of just using
12367 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12368 2. Implemented optional authenticator field of KerberosWrapper.
12370 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12371 and authenticator structs; see crypto/krb5/.
12373 Generalized Kerberos calls to support multiple Kerberos libraries.
12374 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12375 via Richard Levitte*
12377 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12378 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12379 values for each of the key sizes rather than having just
12380 parameters (and 'speed' generating keys each time).
12384 * Speed up EVP routines.
12387 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12388 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12389 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12390 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12392 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12393 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12394 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12397 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12399 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12403 * Added the OS2-EMX target.
12405 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12407 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12408 New functions to support `NCONF` routines in extension code.
12409 New function `CONF_set_nconf()`
12410 to allow functions which take an `NCONF` to also handle the old `LHASH`
12411 structure: this means that the old `CONF` compatible routines can be
12412 retained (in particular w.rt. extensions) without having to duplicate the
12413 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12417 * Enhance the general user interface with mechanisms for inner control
12418 and with possibilities to have yes/no kind of prompts.
12422 * Change all calls to low-level digest routines in the library and
12423 applications to use EVP. Add missing calls to HMAC_cleanup() and
12424 don't assume HMAC_CTX can be copied using memcpy().
12426 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12428 * Add the possibility to control engines through control names but with
12429 arbitrary arguments instead of just a string.
12430 Change the key loaders to take a UI_METHOD instead of a callback
12431 function pointer. NOTE: this breaks binary compatibility with earlier
12432 versions of OpenSSL [engine].
12433 Adapt the nCipher code for these new conditions and add a card insertion
12438 * Enhance the general user interface with mechanisms to better support
12439 dialog box interfaces, application-defined prompts, the possibility
12440 to use defaults (for example default passwords from somewhere else)
12441 and interrupts/cancellations.
12445 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12446 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12450 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12451 tidy up some unnecessarily weird code in 'sk_new()').
12453 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12455 * Change the key loading routines for ENGINEs to use the same kind
12456 callback (pem_password_cb) as all other routines that need this
12461 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12462 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12463 than this minimum value is recommended.
12467 * New random seeder for OpenVMS, using the system process statistics
12468 that are easily reachable.
12472 * Windows apparently can't transparently handle global
12473 variables defined in DLLs. Initialisations such as:
12475 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12477 won't compile. This is used by the any applications that need to
12478 declare their own ASN1 modules. This was fixed by adding the option
12479 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12480 needed for static libraries under Win32.
12484 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12485 setting of purpose and trust fields. New X509_STORE trust and
12486 purpose functions and tidy up setting in other SSL functions.
12490 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12491 structure. These are inherited by X509_STORE_CTX when it is
12492 initialised. This allows various defaults to be set in the
12493 X509_STORE structure (such as flags for CRL checking and custom
12494 purpose or trust settings) for functions which only use X509_STORE_CTX
12495 internally such as S/MIME.
12497 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12498 trust settings if they are not set in X509_STORE. This allows X509_STORE
12499 purposes and trust (in S/MIME for example) to override any set by default.
12501 Add command line options for CRL checking to smime, s_client and s_server
12506 * Initial CRL based revocation checking. If the CRL checking flag(s)
12507 are set then the CRL is looked up in the X509_STORE structure and
12508 its validity and signature checked, then if the certificate is found
12509 in the CRL the verify fails with a revoked error.
12511 Various new CRL related callbacks added to X509_STORE_CTX structure.
12513 Command line options added to 'verify' application to support this.
12515 This needs some additional work, such as being able to handle multiple
12516 CRLs with different times, extension based lookup (rather than just
12517 by subject name) and ultimately more complete V2 CRL extension
12522 * Add a general user interface API (crypto/ui/). This is designed
12523 to replace things like des_read_password and friends (backward
12524 compatibility functions using this new API are provided).
12525 The purpose is to remove prompting functions from the DES code
12526 section as well as provide for prompting through dialog boxes in
12527 a window system and the like.
12531 * Add "ex_data" support to ENGINE so implementations can add state at a
12532 per-structure level rather than having to store it globally.
12536 * Make it possible for ENGINE structures to be copied when retrieved by
12537 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12538 This causes the "original" ENGINE structure to act like a template,
12539 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12540 operational state can be localised to each ENGINE structure, despite the
12541 fact they all share the same "methods". New ENGINE structures returned in
12542 this case have no functional references and the return value is the single
12543 structural reference. This matches the single structural reference returned
12544 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12549 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12550 needs to match any other type at all we need to manually clear the
12555 * Changes to the "openssl engine" utility to include;
12556 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12557 about an ENGINE's available control commands.
12558 - executing control commands from command line arguments using the
12559 '-pre' and '-post' switches. '-post' is only used if '-t' is
12560 specified and the ENGINE is successfully initialised. The syntax for
12561 the individual commands are colon-separated, for example;
12562 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12566 * New dynamic control command support for ENGINEs. ENGINEs can now
12567 declare their own commands (numbers), names (strings), descriptions,
12568 and input types for run-time discovery by calling applications. A
12569 subset of these commands are implicitly classed as "executable"
12570 depending on their input type, and only these can be invoked through
12571 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12572 can be based on user input, config files, etc). The distinction is
12573 that "executable" commands cannot return anything other than a boolean
12574 result and can only support numeric or string input, whereas some
12575 discoverable commands may only be for direct use through
12576 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12577 pointers, or other custom uses. The "executable" commands are to
12578 support parameterisations of ENGINE behaviour that can be
12579 unambiguously defined by ENGINEs and used consistently across any
12580 OpenSSL-based application. Commands have been added to all the
12581 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12582 control over shared-library paths without source code alterations.
12586 * Changed all ENGINE implementations to dynamically allocate their
12587 ENGINEs rather than declaring them statically. Apart from this being
12588 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12589 this also allows the implementations to compile without using the
12590 internal engine_int.h header.
12594 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12595 'const' value. Any code that should be able to modify a RAND_METHOD
12596 should already have non-const pointers to it (ie. they should only
12597 modify their own ones).
12601 * Made a variety of little tweaks to the ENGINE code.
12602 - "atalla" and "ubsec" string definitions were moved from header files
12603 to C code. "nuron" string definitions were placed in variables
12604 rather than hard-coded - allowing parameterisation of these values
12605 later on via ctrl() commands.
12606 - Removed unused "#if 0"'d code.
12607 - Fixed engine list iteration code so it uses ENGINE_free() to release
12608 structural references.
12609 - Constified the RAND_METHOD element of ENGINE structures.
12610 - Constified various get/set functions as appropriate and added
12611 missing functions (including a catch-all ENGINE_cpy that duplicates
12612 all ENGINE values onto a new ENGINE except reference counts/state).
12613 - Removed NULL parameter checks in get/set functions. Setting a method
12614 or function to NULL is a way of cancelling out a previously set
12615 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12616 and doesn't justify the extra error symbols and code.
12617 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12618 flags from engine_int.h to engine.h.
12619 - Changed prototypes for ENGINE handler functions (init(), finish(),
12620 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12624 * Implement binary inversion algorithm for BN_mod_inverse in addition
12625 to the algorithm using long division. The binary algorithm can be
12626 used only if the modulus is odd. On 32-bit systems, it is faster
12627 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12628 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12629 up to 450 bits. In 64-bit environments, the binary algorithm
12630 appears to be advantageous for much longer moduli; here we use it
12631 for moduli up to 2048 bits.
12635 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12636 could not support the combine flag in choice fields.
12640 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12641 extensions from a certificate request to the certificate.
12645 * Allow multiple 'certopt' and 'nameopt' options to be separated
12646 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12647 file: this allows the display of the certificate about to be
12648 signed to be customised, to allow certain fields to be included
12649 or excluded and extension details. The old system didn't display
12650 multicharacter strings properly, omitted fields not in the policy
12651 and couldn't display additional details such as extensions.
12655 * Function EC_POINTs_mul for multiple scalar multiplication
12656 of an arbitrary number of elliptic curve points
12657 \sum scalars[i]*points[i],
12658 optionally including the generator defined for the EC_GROUP:
12659 scalar*generator + \sum scalars[i]*points[i].
12661 EC_POINT_mul is a simple wrapper function for the typical case
12662 that the point list has just one item (besides the optional
12667 * First EC_METHODs for curves over GF(p):
12669 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12670 operations and provides various method functions that can also
12671 operate with faster implementations of modular arithmetic.
12673 EC_GFp_mont_method() reuses most functions that are part of
12674 EC_GFp_simple_method, but uses Montgomery arithmetic.
12676 *Bodo Moeller; point addition and point doubling
12677 implementation directly derived from source code provided by
12678 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12680 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12681 crypto/ec/ec_lib.c):
12683 Curves are EC_GROUP objects (with an optional group generator)
12684 based on EC_METHODs that are built into the library.
12686 Points are EC_POINT objects based on EC_GROUP objects.
12688 Most of the framework would be able to handle curves over arbitrary
12689 finite fields, but as there are no obvious types for fields other
12690 than GF(p), some functions are limited to that for now.
12694 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12695 that the file contains a complete HTTP response.
12699 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12700 change the def and num file printf format specifier from "%-40sXXX"
12701 to "%-39s XXX". The latter will always guarantee a space after the
12702 field while the former will cause them to run together if the field
12703 is 40 of more characters long.
12707 * Constify the cipher and digest 'method' functions and structures
12708 and modify related functions to take constant EVP_MD and EVP_CIPHER
12713 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12714 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12718 * Modify `EVP_Digest*()` routines so they now return values. Although the
12719 internal software routines can never fail additional hardware versions
12724 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12726 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12727 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12730 ERR_R_NESTED_ASN1_ERROR
12732 ERR_R_MISSING_ASN1_EOS
12733 were 4 .. 9, conflicting with
12734 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12736 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12737 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12739 Add new error code 'ERR_R_INTERNAL_ERROR'.
12743 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12748 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12749 sets the subject name for a new request or supersedes the
12750 subject name in a given request. Formats that can be parsed are
12751 'CN=Some Name, OU=myOU, C=IT'
12753 'CN=Some Name/OU=myOU/C=IT'.
12755 Add options '-batch' and '-verbose' to 'openssl req'.
12757 *Massimiliano Pala <madwolf@hackmasters.net>*
12759 * Introduce the possibility to access global variables through
12760 functions on platform were that's the best way to handle exporting
12761 global variables in shared libraries. To enable this functionality,
12762 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12763 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12764 is normally done by Configure or something similar).
12766 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12767 in the source file (foo.c) like this:
12769 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12770 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12772 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12773 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12775 OPENSSL_DECLARE_GLOBAL(int,foo);
12776 #define foo OPENSSL_GLOBAL_REF(foo)
12777 OPENSSL_DECLARE_GLOBAL(double,bar);
12778 #define bar OPENSSL_GLOBAL_REF(bar)
12780 The #defines are very important, and therefore so is including the
12781 header file everywhere where the defined globals are used.
12783 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12784 of ASN.1 items, but that structure is a bit different.
12786 The largest change is in util/mkdef.pl which has been enhanced with
12787 better and easier to understand logic to choose which symbols should
12788 go into the Windows .def files as well as a number of fixes and code
12789 cleanup (among others, algorithm keywords are now sorted
12790 lexicographically to avoid constant rewrites).
12794 * In BN_div() keep a copy of the sign of 'num' before writing the
12795 result to 'rm' because if rm==num the value will be overwritten
12796 and produce the wrong result if 'num' is negative: this caused
12797 problems with BN_mod() and BN_nnmod().
12801 * Function OCSP_request_verify(). This checks the signature on an
12802 OCSP request and verifies the signer certificate. The signer
12803 certificate is just checked for a generic purpose and OCSP request
12808 * Add OCSP_check_validity() function to check the validity of OCSP
12809 responses. OCSP responses are prepared in real time and may only
12810 be a few seconds old. Simply checking that the current time lies
12811 between thisUpdate and nextUpdate max reject otherwise valid responses
12812 caused by either OCSP responder or client clock inaccuracy. Instead
12813 we allow thisUpdate and nextUpdate to fall within a certain period of
12814 the current time. The age of the response can also optionally be
12815 checked. Two new options -validity_period and -status_age added to
12820 * If signature or public key algorithm is unrecognized print out its
12821 OID rather that just UNKNOWN.
12825 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12826 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12827 ID to be generated from the issuer certificate alone which can then be
12828 passed to OCSP_id_issuer_cmp().
12832 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12833 ASN1 modules to export functions returning ASN1_ITEM pointers
12834 instead of the ASN1_ITEM structures themselves. This adds several
12835 new macros which allow the underlying ASN1 function/structure to
12836 be accessed transparently. As a result code should not use ASN1_ITEM
12837 references directly (such as &X509_it) but instead use the relevant
12838 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12839 use of the new ASN1 code on platforms where exporting structures
12840 is problematical (for example in shared libraries) but exporting
12841 functions returning pointers to structures is not.
12845 * Add support for overriding the generation of SSL/TLS session IDs.
12846 These callbacks can be registered either in an SSL_CTX or per SSL.
12847 The purpose of this is to allow applications to control, if they wish,
12848 the arbitrary values chosen for use as session IDs, particularly as it
12849 can be useful for session caching in multiple-server environments. A
12850 command-line switch for testing this (and any client code that wishes
12851 to use such a feature) has been added to "s_server".
12853 *Geoff Thorpe, Lutz Jaenicke*
12855 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12856 of the form `#if defined(...) || defined(...) || ...` and
12857 `#if !defined(...) && !defined(...) && ...`. This also avoids
12858 the growing number of special cases it was previously handling.
12862 * Make all configuration macros available for application by making
12863 sure they are available in opensslconf.h, by giving them names starting
12864 with `OPENSSL_` to avoid conflicts with other packages and by making
12865 sure e_os2.h will cover all platform-specific cases together with
12867 Additionally, it is now possible to define configuration/platform-
12868 specific names (called "system identities"). In the C code, these
12869 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12870 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12871 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12876 * New option -set_serial to 'req' and 'x509' this allows the serial
12877 number to use to be specified on the command line. Previously self
12878 signed certificates were hard coded with serial number 0 and the
12879 CA options of 'x509' had to use a serial number in a file which was
12884 * New options to 'ca' utility to support V2 CRL entry extensions.
12885 Currently CRL reason, invalidity date and hold instruction are
12886 supported. Add new CRL extensions to V3 code and some new objects.
12890 * New function EVP_CIPHER_CTX_set_padding() this is used to
12891 disable standard block padding (aka PKCS#5 padding) in the EVP
12892 API, which was previously mandatory. This means that the data is
12893 not padded in any way and so the total length much be a multiple
12894 of the block size, otherwise an error occurs.
12898 * Initial (incomplete) OCSP SSL support.
12902 * New function OCSP_parse_url(). This splits up a URL into its host,
12903 port and path components: primarily to parse OCSP URLs. New -url
12904 option to ocsp utility.
12908 * New nonce behavior. The return value of OCSP_check_nonce() now
12909 reflects the various checks performed. Applications can decide
12910 whether to tolerate certain situations such as an absent nonce
12911 in a response when one was present in a request: the ocsp application
12912 just prints out a warning. New function OCSP_add1_basic_nonce()
12913 this is to allow responders to include a nonce in a response even if
12914 the request is nonce-less.
12918 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12919 skipped when using openssl x509 multiple times on a single input file,
12920 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12924 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12925 set string type: to handle setting ASN1_TIME structures. Fix ca
12926 utility to correctly initialize revocation date of CRLs.
12930 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12931 the clients preferred ciphersuites and rather use its own preferences.
12932 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12933 Internet Explorer by ensuring unchanged hash method during stepup.
12934 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12938 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12939 to aes and add a new 'exist' option to print out symbols that don't
12944 * Additional options to ocsp utility to allow flags to be set and
12945 additional certificates supplied.
12949 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12950 OCSP client a number of certificate to only verify the response
12955 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12956 handle the new API. Currently only ECB, CBC modes supported. Add new
12959 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12960 Encryption Standard (AES) Ciphersuites for Transport Layer
12961 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12962 not enabled by default and were not part of the "ALL" ciphersuite
12963 alias because they were not yet official; they could be
12964 explicitly requested by specifying the "AESdraft" ciphersuite
12965 group alias. In the final release of OpenSSL 0.9.7, the group
12966 alias is called "AES" and is part of "ALL".)
12968 *Ben Laurie, Steve Henson, Bodo Moeller*
12970 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12971 request to response.
12975 * Functions for OCSP responders. OCSP_request_onereq_count(),
12976 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12977 extract information from a certificate request. OCSP_response_create()
12978 creates a response and optionally adds a basic response structure.
12979 OCSP_basic_add1_status() adds a complete single response to a basic
12980 response and returns the OCSP_SINGLERESP structure just added (to allow
12981 extensions to be included for example). OCSP_basic_add1_cert() adds a
12982 certificate to a basic response and OCSP_basic_sign() signs a basic
12983 response with various flags. New helper functions ASN1_TIME_check()
12984 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12985 (converts ASN1_TIME to GeneralizedTime).
12989 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12990 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12991 structure from a certificate. X509_pubkey_digest() digests the public_key
12992 contents: this is used in various key identifiers.
12996 * Make sk_sort() tolerate a NULL argument.
12998 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13000 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13001 passed by the function are trusted implicitly. If any of them signed the
13002 response then it is assumed to be valid and is not verified.
13006 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13007 to data. This was previously part of the PKCS7 ASN1 code. This
13008 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13009 *Steve Henson, reported by Kenneth R. Robinette
13010 <support@securenetterm.com>*
13012 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13013 routines: without these tracing memory leaks is very painful.
13014 Fix leaks in PKCS12 and PKCS7 routines.
13018 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13019 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13020 effectively meant GeneralizedTime would never be used. Now it
13021 is initialised to -1 but X509_time_adj() now has to check the value
13022 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13023 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13024 *Steve Henson, reported by Kenneth R. Robinette
13025 <support@securenetterm.com>*
13027 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13028 result in a zero length in the ASN1_INTEGER structure which was
13029 not consistent with the structure when d2i_ASN1_INTEGER() was used
13030 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13031 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13032 where it did not print out a minus for negative ASN1_INTEGER.
13036 * Add summary printout to ocsp utility. The various functions which
13037 convert status values to strings have been renamed to:
13038 OCSP_response_status_str(), OCSP_cert_status_str() and
13039 OCSP_crl_reason_str() and are no longer static. New options
13040 to verify nonce values and to disable verification. OCSP response
13041 printout format cleaned up.
13045 * Add additional OCSP certificate checks. These are those specified
13046 in RFC2560. This consists of two separate checks: the CA of the
13047 certificate being checked must either be the OCSP signer certificate
13048 or the issuer of the OCSP signer certificate. In the latter case the
13049 OCSP signer certificate must contain the OCSP signing extended key
13050 usage. This check is performed by attempting to match the OCSP
13051 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13052 in the OCSP_CERTID structures of the response.
13056 * Initial OCSP certificate verification added to OCSP_basic_verify()
13057 and related routines. This uses the standard OpenSSL certificate
13058 verify routines to perform initial checks (just CA validity) and
13059 to obtain the certificate chain. Then additional checks will be
13060 performed on the chain. Currently the root CA is checked to see
13061 if it is explicitly trusted for OCSP signing. This is used to set
13062 a root CA as a global signing root: that is any certificate that
13063 chains to that CA is an acceptable OCSP signing certificate.
13067 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13068 extensions from a separate configuration file.
13069 As when reading extensions from the main configuration file,
13070 the '-extensions ...' option may be used for specifying the
13073 *Massimiliano Pala <madwolf@comune.modena.it>*
13075 * New OCSP utility. Allows OCSP requests to be generated or
13076 read. The request can be sent to a responder and the output
13077 parsed, outputted or printed in text form. Not complete yet:
13078 still needs to check the OCSP response validity.
13082 * New subcommands for 'openssl ca':
13083 `openssl ca -status <serial>` prints the status of the cert with
13084 the given serial number (according to the index file).
13085 `openssl ca -updatedb` updates the expiry status of certificates
13088 *Massimiliano Pala <madwolf@comune.modena.it>*
13090 * New '-newreq-nodes' command option to CA.pl. This is like
13091 '-newreq', but calls 'openssl req' with the '-nodes' option
13092 so that the resulting key is not encrypted.
13094 *Damien Miller <djm@mindrot.org>*
13096 * New configuration for the GNU Hurd.
13098 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13100 * Initial code to implement OCSP basic response verify. This
13101 is currently incomplete. Currently just finds the signer's
13102 certificate and verifies the signature on the response.
13106 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13107 value of OPENSSLDIR. This is available via the new '-d' option
13108 to 'openssl version', and is also included in 'openssl version -a'.
13112 * Allowing defining memory allocation callbacks that will be given
13113 file name and line number information in additional arguments
13114 (a `const char*` and an int). The basic functionality remains, as
13115 well as the original possibility to just replace malloc(),
13116 realloc() and free() by functions that do not know about these
13117 additional arguments. To register and find out the current
13118 settings for extended allocation functions, the following
13119 functions are provided:
13121 CRYPTO_set_mem_ex_functions
13122 CRYPTO_set_locked_mem_ex_functions
13123 CRYPTO_get_mem_ex_functions
13124 CRYPTO_get_locked_mem_ex_functions
13126 These work the same way as CRYPTO_set_mem_functions and friends.
13127 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13128 extended allocation function is enabled.
13129 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13130 a conventional allocation function is enabled.
13132 *Richard Levitte, Bodo Moeller*
13134 * Finish off removing the remaining LHASH function pointer casts.
13135 There should no longer be any prototype-casting required when using
13136 the LHASH abstraction, and any casts that remain are "bugs". See
13137 the callback types and macros at the head of lhash.h for details
13138 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13142 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13143 If /dev/[u]random devices are not available or do not return enough
13144 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13146 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13147 /etc/entropy will be queried once each in this sequence, querying stops
13148 when enough entropy was collected without querying more sockets.
13152 * Change the Unix RAND_poll() variant to be able to poll several
13153 random devices, as specified by DEVRANDOM, until a sufficient amount
13154 of data has been collected. We spend at most 10 ms on each file
13155 (select timeout) and read in non-blocking mode. DEVRANDOM now
13156 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13157 (previously it was just the string "/dev/urandom"), so on typical
13158 platforms the 10 ms delay will never occur.
13159 Also separate out the Unix variant to its own file, rand_unix.c.
13160 For VMS, there's a currently-empty rand_vms.c.
13164 * Move OCSP client related routines to ocsp_cl.c. These
13165 provide utility functions which an application needing
13166 to issue a request to an OCSP responder and analyse the
13167 response will typically need: as opposed to those which an
13168 OCSP responder itself would need which will be added later.
13170 OCSP_request_sign() signs an OCSP request with an API similar
13171 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13172 response. OCSP_response_get1_basic() extracts basic response
13173 from response. OCSP_resp_find_status(): finds and extracts status
13174 information from an OCSP_CERTID structure (which will be created
13175 when the request structure is built). These are built from lower
13176 level functions which work on OCSP_SINGLERESP structures but
13177 won't normally be used unless the application wishes to examine
13178 extensions in the OCSP response for example.
13180 Replace nonce routines with a pair of functions.
13181 OCSP_request_add1_nonce() adds a nonce value and optionally
13182 generates a random value. OCSP_check_nonce() checks the
13183 validity of the nonce in an OCSP response.
13187 * Change function OCSP_request_add() to OCSP_request_add0_id().
13188 This doesn't copy the supplied OCSP_CERTID and avoids the
13189 need to free up the newly created id. Change return type
13190 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13191 This can then be used to add extensions to the request.
13192 Deleted OCSP_request_new(), since most of its functionality
13193 is now in OCSP_REQUEST_new() (and the case insensitive name
13194 clash) apart from the ability to set the request name which
13195 will be added elsewhere.
13199 * Update OCSP API. Remove obsolete extensions argument from
13200 various functions. Extensions are now handled using the new
13201 OCSP extension code. New simple OCSP HTTP function which
13202 can be used to send requests and parse the response.
13206 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13207 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13208 uses the special reorder version of SET OF to sort the attributes
13209 and reorder them to match the encoded order. This resolves a long
13210 standing problem: a verify on a PKCS7 structure just after signing
13211 it used to fail because the attribute order did not match the
13212 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13213 it uses the received order. This is necessary to tolerate some broken
13214 software that does not order SET OF. This is handled by encoding
13215 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13216 to produce the required SET OF.
13220 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13221 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13222 files to get correct declarations of the ASN.1 item variables.
13226 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13227 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13228 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13229 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13230 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13231 ASN1_ITEM and no wrapper functions.
13235 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13236 replace the old function pointer based I/O routines. Change most of
13237 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13241 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13242 lines, recognize more "algorithms" that can be deselected, and make
13243 it complain about algorithm deselection that isn't recognised.
13247 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13248 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13249 to use new functions. Add NO_ASN1_OLD which can be set to remove
13250 some old style ASN1 functions: this can be used to determine if old
13251 code will still work when these eventually go away.
13255 * New extension functions for OCSP structures, these follow the
13256 same conventions as certificates and CRLs.
13260 * New function X509V3_add1_i2d(). This automatically encodes and
13261 adds an extension. Its behaviour can be customised with various
13262 flags to append, replace or delete. Various wrappers added for
13263 certificates and CRLs.
13267 * Fix to avoid calling the underlying ASN1 print routine when
13268 an extension cannot be parsed. Correct a typo in the
13269 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13273 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13274 entries for variables.
13278 * Add functionality to `apps/openssl.c` for detecting locking
13279 problems: As the program is single-threaded, all we have
13280 to do is register a locking callback using an array for
13281 storing which locks are currently held by the program.
13285 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13286 SSL_get_ex_data_X509_STORE_idx(), which is used in
13287 ssl_verify_cert_chain() and thus can be called at any time
13288 during TLS/SSL handshakes so that thread-safety is essential.
13289 Unfortunately, the ex_data design is not at all suited
13290 for multi-threaded use, so it probably should be abolished.
13294 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13296 *Broadcom, tweaked and integrated by Geoff Thorpe*
13298 * Move common extension printing code to new function
13299 X509V3_print_extensions(). Reorganise OCSP print routines and
13300 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13304 * New function X509_signature_print() to remove duplication in some
13309 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13310 set (this was treated exactly the same as SET OF previously). This
13311 is used to reorder the STACK representing the structure to match the
13312 encoding. This will be used to get round a problem where a PKCS7
13313 structure which was signed could not be verified because the STACK
13314 order did not reflect the encoded order.
13318 * Reimplement the OCSP ASN1 module using the new code.
13322 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13323 for its ASN1 operations. The old style function pointers still exist
13324 for now but they will eventually go away.
13328 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13329 completely replaces the old ASN1 functionality with a table driven
13330 encoder and decoder which interprets an ASN1_ITEM structure describing
13331 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13332 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13333 has also been converted to the new form.
13337 * Change BN_mod_exp_recp so that negative moduli are tolerated
13338 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13339 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13340 for negative moduli.
13344 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13345 of not touching the result's sign bit.
13349 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13354 * Changed the LHASH code to use prototypes for callbacks, and created
13355 macros to declare and implement thin (optionally static) functions
13356 that provide type-safety and avoid function pointer casting for the
13357 type-specific callbacks.
13361 * Added Kerberos Cipher Suites to be used with TLS, as written in
13363 *Veers Staats <staatsvr@asc.hpc.mil>,
13364 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13366 * Reformat the FAQ so the different questions and answers can be divided
13367 in sections depending on the subject.
13371 * Have the zlib compression code load ZLIB.DLL dynamically under
13376 * New function BN_mod_sqrt for computing square roots modulo a prime
13377 (using the probabilistic Tonelli-Shanks algorithm unless
13378 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13379 be handled deterministically).
13381 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13383 * Make BN_mod_inverse faster by explicitly handling small quotients
13384 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13385 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13389 * New function BN_kronecker.
13393 * Fix BN_gcd so that it works on negative inputs; the result is
13394 positive unless both parameters are zero.
13395 Previously something reasonably close to an infinite loop was
13396 possible because numbers could be growing instead of shrinking
13397 in the implementation of Euclid's algorithm.
13401 * Fix BN_is_word() and BN_is_one() macros to take into account the
13402 sign of the number in question.
13404 Fix BN_is_word(a,w) to work correctly for w == 0.
13406 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13407 because its test if the absolute value of 'a' equals 'w'.
13408 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13409 it exists mostly for use in the implementations of BN_is_zero(),
13410 BN_is_one(), and BN_is_word().
13414 * New function BN_swap.
13418 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13419 the exponentiation functions are more likely to produce reasonable
13420 results on negative inputs.
13424 * Change BN_mod_mul so that the result is always non-negative.
13425 Previously, it could be negative if one of the factors was negative;
13426 I don't think anyone really wanted that behaviour.
13430 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13431 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13432 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13433 and add new functions:
13442 BN_mod_lshift1_quick
13444 BN_mod_lshift_quick
13446 These functions always generate non-negative results.
13448 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13449 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13451 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13452 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13453 be reduced modulo `m`.
13455 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13458 The following entry accidentally appeared in the CHANGES file
13459 distributed with OpenSSL 0.9.7. The modifications described in
13460 it do *not* apply to OpenSSL 0.9.7.
13462 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13463 was actually never needed) and in BN_mul(). The removal in BN_mul()
13464 required a small change in bn_mul_part_recursive() and the addition
13465 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13466 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13467 bn_sub_words() and bn_add_words() except they take arrays with
13473 * In 'openssl passwd', verify passwords read from the terminal
13474 unless the '-salt' option is used (which usually means that
13475 verification would just waste user's time since the resulting
13476 hash is going to be compared with some given password hash)
13477 or the new '-noverify' option is used.
13479 This is an incompatible change, but it does not affect
13480 non-interactive use of 'openssl passwd' (passwords on the command
13481 line, '-stdin' option, '-in ...' option) and thus should not
13482 cause any problems.
13486 * Remove all references to RSAref, since there's no more need for it.
13490 * Make DSO load along a path given through an environment variable
13491 (SHLIB_PATH) with shl_load().
13495 * Constify the ENGINE code as a result of BIGNUM constification.
13496 Also constify the RSA code and most things related to it. In a
13497 few places, most notable in the depth of the ASN.1 code, ugly
13498 casts back to non-const were required (to be solved at a later
13503 * Make it so the openssl application has all engines loaded by default.
13507 * Constify the BIGNUM routines a little more.
13511 * Add the following functions:
13513 ENGINE_load_cswift()
13515 ENGINE_load_atalla()
13516 ENGINE_load_nuron()
13517 ENGINE_load_builtin_engines()
13519 That way, an application can itself choose if external engines that
13520 are built-in in OpenSSL shall ever be used or not. The benefit is
13521 that applications won't have to be linked with libdl or other dso
13522 libraries unless it's really needed.
13524 Changed 'openssl engine' to load all engines on demand.
13525 Changed the engine header files to avoid the duplication of some
13526 declarations (they differed!).
13530 * 'openssl engine' can now list capabilities.
13534 * Better error reporting in 'openssl engine'.
13538 * Never call load_dh_param(NULL) in s_server.
13542 * Add engine application. It can currently list engines by name and
13543 identity, and test if they are actually available.
13547 * Improve RPM specification file by forcing symbolic linking and making
13548 sure the installed documentation is also owned by root.root.
13550 *Damien Miller <djm@mindrot.org>*
13552 * Give the OpenSSL applications more possibilities to make use of
13553 keys (public as well as private) handled by engines.
13557 * Add OCSP code that comes from CertCo.
13561 * Add VMS support for the Rijndael code.
13565 * Added untested support for Nuron crypto accelerator.
13569 * Add support for external cryptographic devices. This code was
13570 previously distributed separately as the "engine" branch.
13572 *Geoff Thorpe, Richard Levitte*
13574 * Rework the filename-translation in the DSO code. It is now possible to
13575 have far greater control over how a "name" is turned into a filename
13576 depending on the operating environment and any oddities about the
13577 different shared library filenames on each system.
13581 * Support threads on FreeBSD-elf in Configure.
13585 * Fix for SHA1 assembly problem with MASM: it produces
13586 warnings about corrupt line number information when assembling
13587 with debugging information. This is caused by the overlapping
13590 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13593 NCONF_get_number() has no error checking at all. As a replacement,
13594 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13595 promoted strongly. The old NCONF_get_number is kept around for
13596 binary backward compatibility.
13597 Make it possible for methods to load from something other than a BIO,
13598 by providing a function pointer that is given a name instead of a BIO.
13599 For example, this could be used to load configuration data from an
13604 * Fix for non blocking accept BIOs. Added new I/O special reason
13605 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13606 with non blocking I/O was not possible because no retry code was
13607 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13612 * Added the beginnings of Rijndael support.
13616 * Fix for bug in DirectoryString mask setting. Add support for
13617 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13618 to allow certificate printing to more controllable, additional
13619 'certopt' option to 'x509' to allow new printing options to be
13624 * Clean old EAY MD5 hack from e_os.h.
13628 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13630 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13631 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13633 *Joe Orton, Steve Henson*
13635 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13637 * Fix additional bug revealed by the NISCC test suite:
13639 Stop bug triggering large recursion when presented with
13640 certain ASN.1 tags ([CVE-2003-0851])
13644 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13646 * Fix various bugs revealed by running the NISCC test suite:
13648 Stop out of bounds reads in the ASN1 code when presented with
13649 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13651 If verify callback ignores invalid public key errors don't try to check
13652 certificate signature with the NULL public key.
13656 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13657 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13662 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13663 extra data after the compression methods not only for TLS 1.0
13664 but also for SSL 3.0 (as required by the specification).
13666 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13668 * Change X509_certificate_type() to mark the key as exported/exportable
13669 when it's 512 *bits* long, not 512 bytes.
13673 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13675 * Countermeasure against the Klima-Pokorny-Rosa extension of
13676 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13677 a protocol version number mismatch like a decryption error
13678 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13682 * Turn on RSA blinding by default in the default implementation
13683 to avoid a timing attack. Applications that don't want it can call
13684 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13685 They would be ill-advised to do so in most cases.
13687 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13689 * Change RSA blinding code so that it works when the PRNG is not
13690 seeded (in this case, the secret RSA exponent is abused as
13691 an unpredictable seed -- if it is not unpredictable, there
13692 is no point in blinding anyway). Make RSA blinding thread-safe
13693 by remembering the creator's thread ID in rsa->blinding and
13694 having all other threads use local one-time blinding factors
13695 (this requires more computation than sharing rsa->blinding, but
13696 avoids excessive locking; and if an RSA object is not shared
13697 between threads, blinding will still be very fast).
13701 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13703 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13704 via timing by performing a MAC computation even if incorrect
13705 block cipher padding has been found. This is a countermeasure
13706 against active attacks where the attacker has to distinguish
13707 between bad padding and a MAC verification error. ([CVE-2003-0078])
13709 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13710 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13711 Martin Vuagnoux (EPFL, Ilion)*
13713 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13715 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13716 memory from its contents. This is done with a counter that will
13717 place alternating values in each byte. This can be used to solve
13718 two issues: 1) the removal of calls to memset() by highly optimizing
13719 compilers, and 2) cleansing with other values than 0, since those can
13720 be read through on certain media, for example a swap space on disk.
13724 * Bugfix: client side session caching did not work with external caching,
13725 because the session->cipher setting was not restored when reloading
13726 from the external cache. This problem was masked, when
13727 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13728 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13732 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13733 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13735 *Zeev Lieber <zeev-l@yahoo.com>*
13737 * Undo an undocumented change introduced in 0.9.6e which caused
13738 repeated calls to OpenSSL_add_all_ciphers() and
13739 OpenSSL_add_all_digests() to be ignored, even after calling
13744 * Change the default configuration reader to deal with last line not
13745 being properly terminated.
13749 * Change X509_NAME_cmp() so it applies the special rules on handling
13750 DN values that are of type PrintableString, as well as RDNs of type
13751 emailAddress where the value has the type ia5String.
13753 *stefank@valicert.com via Richard Levitte*
13755 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13756 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13757 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13758 the bitwise-OR of the two for use by the majority of applications
13759 wanting this behaviour, and update the docs. The documented
13760 behaviour and actual behaviour were inconsistent and had been
13761 changing anyway, so this is more a bug-fix than a behavioural
13764 *Geoff Thorpe, diagnosed by Nadav Har'El*
13766 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13767 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13771 * Fix initialization code race conditions in
13772 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13773 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13774 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13775 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13776 ssl2_get_cipher_by_char(),
13777 ssl3_get_cipher_by_char().
13779 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13781 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13782 the cached sessions are flushed, as the remove_cb() might use ex_data
13783 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13784 (see [openssl.org #212]).
13786 *Geoff Thorpe, Lutz Jaenicke*
13788 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13789 length, instead of the encoding length to d2i_ASN1_OBJECT.
13793 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13795 * [In 0.9.6g-engine release:]
13796 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13798 *Lynn Gazis <lgazis@rainbow.com>*
13800 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13802 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13803 and get fix the header length calculation.
13804 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13805 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13807 * Use proper error handling instead of 'assertions' in buffer
13808 overflow checks added in 0.9.6e. This prevents DoS (the
13809 assertions could call abort()).
13811 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13813 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13815 * Add various sanity checks to asn1_get_length() to reject
13816 the ASN1 length bytes if they exceed sizeof(long), will appear
13817 negative or the content length exceeds the length of the
13820 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13822 * Fix cipher selection routines: ciphers without encryption had no flags
13823 for the cipher strength set and where therefore not handled correctly
13824 by the selection routines (PR #130).
13828 * Fix EVP_dsa_sha macro.
13833 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13834 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13835 that was added in OpenSSL 0.9.6d.
13837 As the countermeasure turned out to be incompatible with some
13838 broken SSL implementations, the new option is part of SSL_OP_ALL.
13839 SSL_OP_ALL is usually employed when compatibility with weird SSL
13840 implementations is desired (e.g. '-bugs' option to 's_client' and
13841 's_server'), so the new option is automatically set in many
13846 * Changes in security patch:
13848 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13849 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13850 Air Force Materiel Command, USAF, under agreement number
13853 * Add various sanity checks to asn1_get_length() to reject
13854 the ASN1 length bytes if they exceed sizeof(long), will appear
13855 negative or the content length exceeds the length of the
13856 supplied buffer. ([CVE-2002-0659])
13858 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13860 * Assertions for various potential buffer overflows, not known to
13861 happen in practice.
13863 *Ben Laurie (CHATS)*
13865 * Various temporary buffers to hold ASCII versions of integers were
13866 too small for 64 bit platforms. ([CVE-2002-0655])
13867 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13869 * Remote buffer overflow in SSL3 protocol - an attacker could
13870 supply an oversized session ID to a client. ([CVE-2002-0656])
13872 *Ben Laurie (CHATS)*
13874 * Remote buffer overflow in SSL2 protocol - an attacker could
13875 supply an oversized client master key. ([CVE-2002-0656])
13877 *Ben Laurie (CHATS)*
13879 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13881 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13882 encoded as NULL) with id-dsa-with-sha1.
13884 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13886 * Check various `X509_...()` return values in `apps/req.c`.
13888 *Nils Larsch <nla@trustcenter.de>*
13890 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13891 an end-of-file condition would erroneously be flagged, when the CRLF
13892 was just at the end of a processed block. The bug was discovered when
13893 processing data through a buffering memory BIO handing the data to a
13894 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13895 <ptsekov@syntrex.com> and Nedelcho Stanev.
13899 * Implement a countermeasure against a vulnerability recently found
13900 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13901 before application data chunks to avoid the use of known IVs
13902 with data potentially chosen by the attacker.
13906 * Fix length checks in ssl3_get_client_hello().
13910 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13911 to prevent ssl3_read_internal() from incorrectly assuming that
13912 ssl3_read_bytes() found application data while handshake
13913 processing was enabled when in fact s->s3->in_read_app_data was
13914 merely automatically cleared during the initial handshake.
13916 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13918 * Fix object definitions for Private and Enterprise: they were not
13919 recognized in their shortname (=lowercase) representation. Extend
13920 obj_dat.pl to issue an error when using undefined keywords instead
13921 of silently ignoring the problem (Svenning Sorensen
13922 <sss@sss.dnsalias.net>).
13926 * Fix DH_generate_parameters() so that it works for 'non-standard'
13927 generators, i.e. generators other than 2 and 5. (Previously, the
13928 code did not properly initialise the 'add' and 'rem' values to
13929 BN_generate_prime().)
13931 In the new general case, we do not insist that 'generator' is
13932 actually a primitive root: This requirement is rather pointless;
13933 a generator of the order-q subgroup is just as good, if not
13938 * Map new X509 verification errors to alerts. Discovered and submitted by
13939 Tom Wu <tom@arcot.com>.
13943 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13944 returning non-zero before the data has been completely received
13945 when using non-blocking I/O.
13947 *Bodo Moeller; problem pointed out by John Hughes*
13949 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13951 *Ben Laurie, Lutz Jaenicke*
13953 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13954 Yoram Zahavi <YoramZ@gilian.com>).
13958 * Add information about CygWin 1.3 and on, and preserve proper
13959 configuration for the versions before that.
13961 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13963 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13964 check whether we deal with a copy of a session and do not delete from
13965 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13966 <izhar@checkpoint.com>.
13970 * Do not store session data into the internal session cache, if it
13971 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13972 flag is set). Proposed by Aslam <aslam@funk.com>.
13976 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13981 * [In 0.9.6d-engine release:]
13982 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13984 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13986 * Add the configuration target linux-s390x.
13988 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13990 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13991 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13992 variable as an indication that a ClientHello message has been
13993 received. As the flag value will be lost between multiple
13994 invocations of ssl3_accept when using non-blocking I/O, the
13995 function may not be aware that a handshake has actually taken
13996 place, thus preventing a new session from being added to the
13999 To avoid this problem, we now set s->new_session to 2 instead of
14000 using a local variable.
14002 *Lutz Jaenicke, Bodo Moeller*
14004 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14005 if the SSL_R_LENGTH_MISMATCH error is detected.
14007 *Geoff Thorpe, Bodo Moeller*
14009 * New 'shared_ldflag' column in Configure platform table.
14013 * Fix EVP_CIPHER_mode macro.
14015 *"Dan S. Camper" <dan@bti.net>*
14017 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14018 type, we must throw them away by setting rr->length to 0.
14020 *D P Chang <dpc@qualys.com>*
14022 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14024 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14025 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14026 worked incorrectly for those cases where range = `10..._2` and
14027 `3*range` is two bits longer than range.)
14031 * Only add signing time to PKCS7 structures if it is not already
14036 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14037 OBJ_ld_ce should be OBJ_id_ce.
14038 Also some ip-pda OIDs in crypto/objects/objects.txt were
14039 incorrect (cf. RFC 3039).
14041 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14043 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14044 returns early because it has nothing to do.
14046 *Andy Schneider <andy.schneider@bjss.co.uk>*
14048 * [In 0.9.6c-engine release:]
14049 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14051 *Andy Schneider <andy.schneider@bjss.co.uk>*
14053 * [In 0.9.6c-engine release:]
14054 Add support for Cryptographic Appliance's keyserver technology.
14055 (Use engine 'keyclient')
14057 *Cryptographic Appliances and Geoff Thorpe*
14059 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14060 is called via tools/c89.sh because arguments have to be
14061 rearranged (all '-L' options must appear before the first object
14064 *Richard Shapiro <rshapiro@abinitio.com>*
14066 * [In 0.9.6c-engine release:]
14067 Add support for Broadcom crypto accelerator cards, backported
14070 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14072 * [In 0.9.6c-engine release:]
14073 Add support for SureWare crypto accelerator cards from
14074 Baltimore Technologies. (Use engine 'sureware')
14076 *Baltimore Technologies and Mark Cox*
14078 * [In 0.9.6c-engine release:]
14079 Add support for crypto accelerator cards from Accelerated
14080 Encryption Processing, www.aep.ie. (Use engine 'aep')
14082 *AEP Inc. and Mark Cox*
14084 * Add a configuration entry for gcc on UnixWare.
14086 *Gary Benson <gbenson@redhat.com>*
14088 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14089 messages are stored in a single piece (fixed-length part and
14090 variable-length part combined) and fix various bugs found on the way.
14094 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14095 instead. BIO_gethostbyname() does not know what timeouts are
14096 appropriate, so entries would stay in cache even when they have
14098 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14100 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14101 faced with a pathologically small ClientHello fragment that does
14102 not contain client_version: Instead of aborting with an error,
14103 simply choose the highest available protocol version (i.e.,
14104 TLS 1.0 unless it is disabled). In practice, ClientHello
14105 messages are never sent like this, but this change gives us
14106 strictly correct behaviour at least for TLS.
14110 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14111 never resets s->method to s->ctx->method when called from within
14112 one of the SSL handshake functions.
14114 *Bodo Moeller; problem pointed out by Niko Baric*
14116 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14117 (sent using the client's version number) if client_version is
14118 smaller than the protocol version in use. Also change
14119 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14120 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14121 the client will at least see that alert.
14125 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14130 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14131 client receives HelloRequest while in a handshake.
14133 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14135 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14136 should end in 'break', not 'goto end' which circumvents various
14137 cleanups done in state SSL_ST_OK. But session related stuff
14138 must be disabled for SSL_ST_OK in the case that we just sent a
14141 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14142 before just sending a HelloRequest.
14144 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14146 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14147 reveal whether illegal block cipher padding was found or a MAC
14148 verification error occurred. (Neither SSLerr() codes nor alerts
14149 are directly visible to potential attackers, but the information
14150 may leak via logfiles.)
14152 Similar changes are not required for the SSL 2.0 implementation
14153 because the number of padding bytes is sent in clear for SSL 2.0,
14154 and the extra bytes are just ignored. However ssl/s2_pkt.c
14155 failed to verify that the purported number of padding bytes is in
14160 * Add OpenUNIX-8 support including shared libraries
14161 (Boyd Lynn Gerber <gerberb@zenez.com>).
14165 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14166 'wristwatch attack' using huge encoding parameters (cf.
14167 James H. Manger's CRYPTO 2001 paper). Note that the
14168 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14169 encoding parameters and hence was not vulnerable.
14173 * BN_sqr() bug fix.
14175 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14177 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14178 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14179 followed by modular reduction.
14181 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14183 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14184 equivalent based on BN_pseudo_rand() instead of BN_rand().
14188 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14189 This function was broken, as the check for a new client hello message
14190 to handle SGC did not allow these large messages.
14191 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14195 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14199 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14200 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14204 * Rework the configuration and shared library support for Tru64 Unix.
14205 The configuration part makes use of modern compiler features and
14206 still retains old compiler behavior for those that run older versions
14207 of the OS. The shared library support part includes a variant that
14208 uses the RPATH feature, and is available through the special
14209 configuration target "alpha-cc-rpath", which will never be selected
14212 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14214 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14215 with the same message size as in ssl3_get_certificate_request().
14216 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14217 messages might inadvertently be reject as too long.
14219 *Petr Lampa <lampa@fee.vutbr.cz>*
14221 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14225 * Modified SSL library such that the verify_callback that has been set
14226 specifically for an SSL object with SSL_set_verify() is actually being
14227 used. Before the change, a verify_callback set with this function was
14228 ignored and the verify_callback() set in the SSL_CTX at the time of
14229 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14230 to allow the necessary settings.
14234 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14235 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14236 done automatically (in contradiction to the requirements of the C
14237 standard). This made problems when used from OpenSSH.
14241 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14242 dh->length and always used
14244 BN_rand_range(priv_key, dh->p).
14246 BN_rand_range() is not necessary for Diffie-Hellman, and this
14247 specific range makes Diffie-Hellman unnecessarily inefficient if
14248 dh->length (recommended exponent length) is much smaller than the
14249 length of dh->p. We could use BN_rand_range() if the order of
14250 the subgroup was stored in the DH structure, but we only have
14255 BN_rand(priv_key, l, ...)
14257 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14264 RSA_eay_public_encrypt
14265 RSA_eay_private_decrypt
14266 RSA_eay_private_encrypt (signing)
14267 RSA_eay_public_decrypt (signature verification)
14269 (default implementations for RSA_public_encrypt,
14270 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14271 always reject numbers >= n.
14275 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14276 to synchronize access to 'locking_thread'. This is necessary on
14277 systems where access to 'locking_thread' (an 'unsigned long'
14278 variable) is not atomic.
14282 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14283 *before* setting the 'crypto_lock_rand' flag. The previous code had
14284 a race condition if 0 is a valid thread ID.
14286 *Travis Vitek <vitek@roguewave.com>*
14288 * Add support for shared libraries under Irix.
14290 *Albert Chin-A-Young <china@thewrittenword.com>*
14292 * Add configuration option to build on Linux on both big-endian and
14293 little-endian MIPS.
14295 *Ralf Baechle <ralf@uni-koblenz.de>*
14297 * Add the possibility to create shared libraries on HP-UX.
14301 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14303 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14304 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14305 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14306 PRNG state recovery was possible based on the output of
14307 one PRNG request appropriately sized to gain knowledge on
14308 'md' followed by enough consecutive 1-byte PRNG requests
14309 to traverse all of 'state'.
14311 1. When updating 'md_local' (the current thread's copy of 'md')
14312 during PRNG output generation, hash all of the previous
14313 'md_local' value, not just the half used for PRNG output.
14315 2. Make the number of bytes from 'state' included into the hash
14316 independent from the number of PRNG bytes requested.
14318 The first measure alone would be sufficient to avoid
14319 Markku-Juhani's attack. (Actually it had never occurred
14320 to me that the half of 'md_local' used for chaining was the
14321 half from which PRNG output bytes were taken -- I had always
14322 assumed that the secret half would be used.) The second
14323 measure makes sure that additional data from 'state' is never
14324 mixed into 'md_local' in small portions; this heuristically
14325 further strengthens the PRNG.
14329 * Fix crypto/bn/asm/mips3.s.
14333 * When only the key is given to "enc", the IV is undefined. Print out
14334 an error message in this case.
14338 * Handle special case when X509_NAME is empty in X509 printing routines.
14342 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14343 positive and less than q.
14347 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14348 used: it isn't thread safe and the add_lock_callback should handle
14351 *Paul Rose <Paul.Rose@bridge.com>*
14353 * Verify that incoming data obeys the block size in
14354 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14360 *Ulf Möller, Bodo Möller*
14362 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14363 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14364 when fixing the server behaviour for backwards-compatible 'client
14365 hello' messages. (Note that the attack is impractical against
14366 SSL 3.0 and TLS 1.0 anyway because length and version checking
14367 means that the probability of guessing a valid ciphertext is
14368 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14371 Before 0.9.5, the countermeasure (hide the error by generating a
14372 random 'decryption result') did not work properly because
14373 ERR_clear_error() was missing, meaning that SSL_get_error() would
14374 detect the supposedly ignored error.
14376 Both problems are now fixed.
14380 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14381 (previously it was 1024).
14385 * Fix for compatibility mode trust settings: ignore trust settings
14386 unless some valid trust or reject settings are present.
14390 * Fix for blowfish EVP: its a variable length cipher.
14394 * Fix various bugs related to DSA S/MIME verification. Handle missing
14395 parameters in DSA public key structures and return an error in the
14396 DSA routines if parameters are absent.
14400 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14401 in the current directory if neither $RANDFILE nor $HOME was set.
14402 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14403 caused some confusion to Windows users who haven't defined $HOME.
14404 Thus RAND_file_name() is changed again: e_os.h can define a
14405 DEFAULT_HOME, which will be used if $HOME is not set.
14406 For Windows, we use "C:"; on other platforms, we still require
14407 environment variables.
14409 * Move 'if (!initialized) RAND_poll()' into regions protected by
14410 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14411 having multiple threads call RAND_poll() concurrently.
14415 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14416 combination of a flag and a thread ID variable.
14417 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14418 flag), *other* threads can enter ssleay_add_bytes without obeying
14419 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14420 that they do not hold after the first thread unsets add_do_not_lock).
14424 * Change bctest again: '-x' expressions are not available in all
14425 versions of 'test'.
14429 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14431 * Fix a couple of memory leaks in PKCS7_dataDecode()
14433 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14435 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14436 the default extension for executables, if any. Also, make the perl
14437 scripts that use symlink() to test if it really exists and use "cp"
14438 if it doesn't. All this made OpenSSL compilable and installable in
14443 * Fix for asn1_GetSequence() for indefinite length constructed data.
14444 If SEQUENCE is length is indefinite just set c->slen to the total
14445 amount of data available.
14447 *Steve Henson, reported by shige@FreeBSD.org*
14449 *This change does not apply to 0.9.7.*
14451 * Change bctest to avoid here-documents inside command substitution
14452 (workaround for FreeBSD /bin/sh bug).
14453 For compatibility with Ultrix, avoid shell functions (introduced
14454 in the bctest version that searches along $PATH).
14458 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14459 with des_encrypt() defined on some operating systems, like Solaris
14464 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14465 On the Importance of Eliminating Errors in Cryptographic
14466 Computations, J. Cryptology 14 (2001) 2, 101-119,
14467 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14471 * MIPS assembler BIGNUM division bug fix.
14475 * Disabled incorrect Alpha assembler code.
14479 * Fix PKCS#7 decode routines so they correctly update the length
14480 after reading an EOC for the EXPLICIT tag.
14484 *This change does not apply to 0.9.7.*
14486 * Fix bug in PKCS#12 key generation routines. This was triggered
14487 if a 3DES key was generated with a 0 initial byte. Include
14488 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14489 (but broken) behaviour.
14493 * Enhance bctest to search for a working bc along $PATH and print
14496 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14498 * Fix memory leaks in err.c: free err_data string if necessary;
14499 don't write to the wrong index in ERR_set_error_data.
14503 * Implement ssl23_peek (analogous to ssl23_read), which previously
14508 * Replace rdtsc with `_emit` statements for VC++ version 5.
14510 *Jeremy Cooper <jeremy@baymoo.org>*
14512 * Make it possible to reuse SSLv2 sessions.
14516 * In copy_email() check for >= 0 as a return value for
14517 X509_NAME_get_index_by_NID() since 0 is a valid index.
14519 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14521 * Avoid coredump with unsupported or invalid public keys by checking if
14522 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14523 PKCS7_verify() fails with non detached data.
14527 * Don't use getenv in library functions when run as setuid/setgid.
14528 New function OPENSSL_issetugid().
14532 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14533 due to incorrect handling of multi-threading:
14535 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14537 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14539 3. Count how many times MemCheck_off() has been called so that
14540 nested use can be treated correctly. This also avoids
14541 inband-signalling in the previous code (which relied on the
14542 assumption that thread ID 0 is impossible).
14546 * Add "-rand" option also to s_client and s_server.
14550 * Fix CPU detection on Irix 6.x.
14551 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14552 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14554 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14559 *This change does not apply to 0.9.7.*
14561 * Use the cached encoding of an X509_NAME structure rather than
14562 copying it. This is apparently the reason for the libsafe "errors"
14563 but the code is actually correct.
14567 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14568 Bleichenbacher's DSA attack.
14569 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14570 to be set and top=0 forces the highest bit to be set; top=-1 is new
14571 and leaves the highest bit random.
14573 *Ulf Moeller, Bodo Moeller*
14575 * In the `NCONF_...`-based implementations for `CONF_...` queries
14576 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14577 a temporary CONF structure with the data component set to NULL
14578 (which gives segmentation faults in lh_retrieve).
14579 Instead, use NULL for the CONF pointer in CONF_get_string and
14580 CONF_get_number (which may use environment variables) and directly
14581 return NULL from CONF_get_section.
14585 * Fix potential buffer overrun for EBCDIC.
14589 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14590 keyUsage if basicConstraints absent for a CA.
14594 * Make SMIME_write_PKCS7() write mail header values with a format that
14595 is more generally accepted (no spaces before the semicolon), since
14596 some programs can't parse those values properly otherwise. Also make
14597 sure BIO's that break lines after each write do not create invalid
14602 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14603 macros previously used would not encode an empty SEQUENCE OF
14604 and break the signature.
14608 *This change does not apply to 0.9.7.*
14610 * Zero the premaster secret after deriving the master secret in
14615 * Add some EVP_add_digest_alias registrations (as found in
14616 OpenSSL_add_all_digests()) to SSL_library_init()
14617 aka OpenSSL_add_ssl_algorithms(). This provides improved
14618 compatibility with peers using X.509 certificates
14619 with unconventional AlgorithmIdentifier OIDs.
14623 * Fix for Irix with NO_ASM.
14625 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14627 * ./config script fixes.
14629 *Ulf Moeller, Richard Levitte*
14631 * Fix 'openssl passwd -1'.
14635 * Change PKCS12_key_gen_asc() so it can cope with non null
14636 terminated strings whose length is passed in the passlen
14637 parameter, for example from PEM callbacks. This was done
14638 by adding an extra length parameter to asc2uni().
14640 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14642 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14643 call failed, free the DSA structure.
14647 * Fix to uni2asc() to cope with zero length Unicode strings.
14648 These are present in some PKCS#12 files.
14652 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14653 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14654 when writing a 32767 byte record.
14656 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14658 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14659 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14661 (RSA objects have a reference count access to which is protected
14662 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14663 so they are meant to be shared between threads.)
14664 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14665 "Reddie, Steven" <Steven.Reddie@ca.com>*
14667 * Fix a deadlock in CRYPTO_mem_leaks().
14671 * Use better test patterns in bntest.
14675 * rand_win.c fix for Borland C.
14679 * BN_rshift bugfix for n == 0.
14683 * Add a 'bctest' script that checks for some known 'bc' bugs
14684 so that 'make test' does not abort just because 'bc' is broken.
14688 * Store verify_result within SSL_SESSION also for client side to
14689 avoid potential security hole. (Re-used sessions on the client side
14690 always resulted in verify_result==X509_V_OK, not using the original
14691 result of the server certificate verification.)
14695 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14696 SSL3_RT_APPLICATION_DATA, return 0.
14697 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14702 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14703 releases, have been re-implemented by renaming the previous
14704 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14705 and ssl3_read_internal, respectively, and adding 'peek' parameters
14706 to them. The new ssl[23]_{read,peek} functions are calls to
14707 ssl[23]_read_internal with the 'peek' flag set appropriately.
14708 A 'peek' parameter has also been added to ssl3_read_bytes, which
14709 does the actual work for ssl3_read_internal.
14713 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14714 the method-specific "init()" handler. Also clean up ex_data after
14715 calling the method-specific "finish()" handler. Previously, this was
14716 happening the other way round.
14720 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14721 The previous value, 12, was not always sufficient for BN_mod_exp().
14725 * Make sure that shared libraries get the internal name engine with
14726 the full version number and not just 0. This should mark the
14727 shared libraries as not backward compatible. Of course, this should
14728 be changed again when we can guarantee backward binary compatibility.
14732 * Fix typo in get_cert_by_subject() in by_dir.c
14734 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14736 * Rework the system to generate shared libraries:
14738 - Make note of the expected extension for the shared libraries and
14739 if there is a need for symbolic links from for example libcrypto.so.0
14740 to libcrypto.so.0.9.7. There is extended info in Configure for
14743 - Make as few rebuilds of the shared libraries as possible.
14745 - Still avoid linking the OpenSSL programs with the shared libraries.
14747 - When installing, install the shared libraries separately from the
14752 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14754 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14755 and not in SSL_clear because the latter is also used by the
14756 accept/connect functions; previously, the settings made by
14757 SSL_set_read_ahead would be lost during the handshake.
14759 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14761 * Correct util/mkdef.pl to be selective about disabled algorithms.
14762 Previously, it would create entries for disabled algorithms no
14767 * Added several new manual pages for SSL_* function.
14771 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14773 * In ssl23_get_client_hello, generate an error message when faced
14774 with an initial SSL 3.0/TLS record that is too small to contain the
14775 first two bytes of the ClientHello message, i.e. client_version.
14776 (Note that this is a pathologic case that probably has never happened
14777 in real life.) The previous approach was to use the version number
14778 from the record header as a substitute; but our protocol choice
14779 should not depend on that one because it is not authenticated
14780 by the Finished messages.
14784 * More robust randomness gathering functions for Windows.
14786 *Jeffrey Altman <jaltman@columbia.edu>*
14788 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14789 not set then we don't setup the error code for issuer check errors
14790 to avoid possibly overwriting other errors which the callback does
14791 handle. If an application does set the flag then we assume it knows
14792 what it is doing and can handle the new informational codes
14797 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14798 a general "ANY" type, as such it should be able to decode anything
14799 including tagged types. However it didn't check the class so it would
14800 wrongly interpret tagged types in the same way as their universal
14801 counterpart and unknown types were just rejected. Changed so that the
14802 tagged and unknown types are handled in the same way as a SEQUENCE:
14803 that is the encoding is stored intact. There is also a new type
14804 "V_ASN1_OTHER" which is used when the class is not universal, in this
14805 case we have no idea what the actual type is so we just lump them all
14810 * On VMS, stdout may very well lead to a file that is written to
14811 in a record-oriented fashion. That means that every write() will
14812 write a separate record, which will be read separately by the
14813 programs trying to read from it. This can be very confusing.
14815 The solution is to put a BIO filter in the way that will buffer
14816 text until a linefeed is reached, and then write everything a
14817 line at a time, so every record written will be an actual line,
14818 not chunks of lines and not (usually doesn't happen, but I've
14819 seen it once) several lines in one record. BIO_f_linebuffer() is
14822 Currently, it's a VMS-only method, because that's where it has
14823 been tested well enough.
14827 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14828 it can return incorrect results.
14829 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14830 but it was in 0.9.6-beta[12].)
14834 * Disable the check for content being present when verifying detached
14835 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14836 include zero length content when signing messages.
14840 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14841 BIO_ctrl (for BIO pairs).
14845 * Add DSO method for VMS.
14849 * Bug fix: Montgomery multiplication could produce results with the
14854 * Add RPM specification openssl.spec and modify it to build three
14855 packages. The default package contains applications, application
14856 documentation and run-time libraries. The devel package contains
14857 include files, static libraries and function documentation. The
14858 doc package contains the contents of the doc directory. The original
14859 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14863 * Add a large number of documentation files for many SSL routines.
14865 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14867 * Add a configuration entry for Sony News 4.
14869 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14871 * Don't set the two most significant bits to one when generating a
14872 random number < q in the DSA library.
14876 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14877 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14878 the underlying transport is blocking) if a handshake took place.
14879 (The default behaviour is needed by applications such as s_client
14880 and s_server that use select() to determine when to use SSL_read;
14881 but for applications that know in advance when to expect data, it
14882 just makes things more complicated.)
14886 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14891 * Add a few more EBCDIC conditionals that make `req` and `x509`
14892 work better on such systems.
14894 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14896 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14897 Update PKCS12_parse() so it copies the friendlyName and the
14898 keyid to the certificates aux info.
14902 * Fix bug in PKCS7_verify() which caused an infinite loop
14903 if there was more than one signature.
14905 *Sven Uszpelkat <su@celocom.de>*
14907 * Major change in util/mkdef.pl to include extra information
14908 about each symbol, as well as presenting variables as well
14909 as functions. This change means that there's n more need
14910 to rebuild the .num files when some algorithms are excluded.
14914 * Allow the verify time to be set by an application,
14915 rather than always using the current time.
14919 * Phase 2 verify code reorganisation. The certificate
14920 verify code now looks up an issuer certificate by a
14921 number of criteria: subject name, authority key id
14922 and key usage. It also verifies self signed certificates
14923 by the same criteria. The main comparison function is
14924 X509_check_issued() which performs these checks.
14926 Lot of changes were necessary in order to support this
14927 without completely rewriting the lookup code.
14929 Authority and subject key identifier are now cached.
14931 The LHASH 'certs' is X509_STORE has now been replaced
14932 by a STACK_OF(X509_OBJECT). This is mainly because an
14933 LHASH can't store or retrieve multiple objects with
14934 the same hash value.
14936 As a result various functions (which were all internal
14937 use only) have changed to handle the new X509_STORE
14938 structure. This will break anything that messed round
14939 with X509_STORE internally.
14941 The functions X509_STORE_add_cert() now checks for an
14942 exact match, rather than just subject name.
14944 The X509_STORE API doesn't directly support the retrieval
14945 of multiple certificates matching a given criteria, however
14946 this can be worked round by performing a lookup first
14947 (which will fill the cache with candidate certificates)
14948 and then examining the cache for matches. This is probably
14949 the best we can do without throwing out X509_LOOKUP
14950 entirely (maybe later...).
14952 The X509_VERIFY_CTX structure has been enhanced considerably.
14954 All certificate lookup operations now go via a get_issuer()
14955 callback. Although this currently uses an X509_STORE it
14956 can be replaced by custom lookups. This is a simple way
14957 to bypass the X509_STORE hackery necessary to make this
14958 work and makes it possible to use more efficient techniques
14959 in future. A very simple version which uses a simple
14960 STACK for its trusted certificate store is also provided
14961 using X509_STORE_CTX_trusted_stack().
14963 The verify_cb() and verify() callbacks now have equivalents
14964 in the X509_STORE_CTX structure.
14966 X509_STORE_CTX also has a 'flags' field which can be used
14967 to customise the verify behaviour.
14971 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14972 excludes S/MIME capabilities.
14976 * When a certificate request is read in keep a copy of the
14977 original encoding of the signed data and use it when outputting
14978 again. Signatures then use the original encoding rather than
14979 a decoded, encoded version which may cause problems if the
14980 request is improperly encoded.
14984 * For consistency with other BIO_puts implementations, call
14985 buffer_write(b, ...) directly in buffer_puts instead of calling
14988 In BIO_puts, increment b->num_write as in BIO_write.
14990 *Peter.Sylvester@EdelWeb.fr*
14992 * Fix BN_mul_word for the case where the word is 0. (We have to use
14993 BN_zero, we may not return a BIGNUM with an array consisting of
14994 words set to zero.)
14998 * Avoid calling abort() from within the library when problems are
14999 detected, except if preprocessor symbols have been defined
15000 (such as REF_CHECK, BN_DEBUG etc.).
15004 * New openssl application 'rsautl'. This utility can be
15005 used for low-level RSA operations. DER public key
15006 BIO/fp routines also added.
15010 * New Configure entry and patches for compiling on QNX 4.
15012 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15014 * A demo state-machine implementation was sponsored by
15015 Nuron (<http://www.nuron.com/>) and is now available in
15016 demos/state_machine.
15020 * New options added to the 'dgst' utility for signature
15021 generation and verification.
15025 * Unrecognized PKCS#7 content types are now handled via a
15026 catch all ASN1_TYPE structure. This allows unsupported
15027 types to be stored as a "blob" and an application can
15028 encode and decode it manually.
15032 * Fix various signed/unsigned issues to make a_strex.c
15033 compile under VC++.
15035 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15037 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15038 length if passed a buffer. ASN1_INTEGER_to_BN failed
15039 if passed a NULL BN and its argument was negative.
15041 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15043 * Modification to PKCS#7 encoding routines to output definite
15044 length encoding. Since currently the whole structures are in
15045 memory there's not real point in using indefinite length
15046 constructed encoding. However if OpenSSL is compiled with
15047 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15051 * Added BIO_vprintf() and BIO_vsnprintf().
15055 * Added more prefixes to parse for in the strings written
15056 through a logging bio, to cover all the levels that are available
15057 through syslog. The prefixes are now:
15059 PANIC, EMERG, EMR => LOG_EMERG
15060 ALERT, ALR => LOG_ALERT
15061 CRIT, CRI => LOG_CRIT
15062 ERROR, ERR => LOG_ERR
15063 WARNING, WARN, WAR => LOG_WARNING
15064 NOTICE, NOTE, NOT => LOG_NOTICE
15065 INFO, INF => LOG_INFO
15066 DEBUG, DBG => LOG_DEBUG
15068 and as before, if none of those prefixes are present at the
15069 beginning of the string, LOG_ERR is chosen.
15071 On Win32, the `LOG_*` levels are mapped according to this:
15073 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15074 LOG_WARNING => EVENTLOG_WARNING_TYPE
15075 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15079 * Made it possible to reconfigure with just the configuration
15080 argument "reconf" or "reconfigure". The command line arguments
15081 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15082 and are retrieved from there when reconfiguring.
15088 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15090 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15094 * The obj_dat.pl script was messing up the sorting of object
15095 names. The reason was that it compared the quoted version
15096 of strings as a result "OCSP" > "OCSP Signing" because
15097 " > SPACE. Changed script to store unquoted versions of
15098 names and add quotes on output. It was also omitting some
15099 names from the lookup table if they were given a default
15100 value (that is if SN is missing it is given the same
15101 value as LN and vice versa), these are now added on the
15102 grounds that if an object has a name we should be able to
15103 look it up. Finally added warning output when duplicate
15104 short or long names are found.
15108 * Changes needed for Tandem NSK.
15110 *Scott Uroff <scott@xypro.com>*
15112 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15113 RSA_padding_check_SSLv23(), special padding was never detected
15114 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15115 version rollback attacks was not effective.
15117 In s23_clnt.c, don't use special rollback-attack detection padding
15118 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15119 client; similarly, in s23_srvr.c, don't do the rollback check if
15120 SSL 2.0 is the only protocol enabled in the server.
15124 * Make it possible to get hexdumps of unprintable data with 'openssl
15125 asn1parse'. By implication, the functions ASN1_parse_dump() and
15126 BIO_dump_indent() are added.
15130 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15131 these print out strings and name structures based on various
15132 flags including RFC2253 support and proper handling of
15133 multibyte characters. Added options to the 'x509' utility
15134 to allow the various flags to be set.
15138 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15139 Also change the functions X509_cmp_current_time() and
15140 X509_gmtime_adj() work with an ASN1_TIME structure,
15141 this will enable certificates using GeneralizedTime in validity
15142 dates to be checked.
15146 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15147 negative public key encodings) on by default,
15148 NO_NEG_PUBKEY_BUG can be set to disable it.
15152 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15153 content octets. An i2c_ASN1_OBJECT is unnecessary because
15154 the encoding can be trivially obtained from the structure.
15158 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15159 not read locks (`CRYPTO_r_[un]lock`).
15163 * A first attempt at creating official support for shared
15164 libraries through configuration. I've kept it so the
15165 default is static libraries only, and the OpenSSL programs
15166 are always statically linked for now, but there are
15167 preparations for dynamic linking in place.
15168 This has been tested on Linux and Tru64.
15172 * Randomness polling function for Win9x, as described in:
15173 Peter Gutmann, Software Generation of Practically Strong
15178 * Fix so PRNG is seeded in req if using an already existing
15183 * New options to smime application. -inform and -outform
15184 allow alternative formats for the S/MIME message including
15185 PEM and DER. The -content option allows the content to be
15186 specified separately. This should allow things like Netscape
15187 form signing output easier to verify.
15191 * Fix the ASN1 encoding of tags using the 'long form'.
15195 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15196 STRING types. These convert content octets to and from the
15197 underlying type. The actual tag and length octets are
15198 already assumed to have been read in and checked. These
15199 are needed because all other string types have virtually
15200 identical handling apart from the tag. By having versions
15201 of the ASN1 functions that just operate on content octets
15202 IMPLICIT tagging can be handled properly. It also allows
15203 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15204 and ASN1_INTEGER are identical apart from the tag.
15208 * Change the handling of OID objects as follows:
15210 - New object identifiers are inserted in objects.txt, following
15211 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15212 - objects.pl is used to process obj_mac.num and create a new
15214 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15217 This is currently kind of a hack, and the perl code in objects.pl
15218 isn't very elegant, but it works as I intended. The simplest way
15219 to check that it worked correctly is to look in obj_dat.h and
15220 check the array nid_objs and make sure the objects haven't moved
15221 around (this is important!). Additions are OK, as well as
15222 consistent name changes.
15226 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15230 * Addition of the command line parameter '-rand file' to 'openssl req'.
15231 The given file adds to whatever has already been seeded into the
15232 random pool through the RANDFILE configuration file option or
15233 environment variable, or the default random state file.
15237 * mkstack.pl now sorts each macro group into lexical order.
15238 Previously the output order depended on the order the files
15239 appeared in the directory, resulting in needless rewriting
15244 * Patches to make OpenSSL compile under Win32 again. Mostly
15245 work arounds for the VC++ problem that it treats func() as
15246 func(void). Also stripped out the parts of mkdef.pl that
15247 added extra typesafe functions: these no longer exist.
15251 * Reorganisation of the stack code. The macros are now all
15252 collected in safestack.h . Each macro is defined in terms of
15253 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15254 DEBUG_SAFESTACK is now handled in terms of function casts,
15255 this has the advantage of retaining type safety without the
15256 use of additional functions. If DEBUG_SAFESTACK is not defined
15257 then the non typesafe macros are used instead. Also modified the
15258 mkstack.pl script to handle the new form. Needs testing to see
15259 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15260 the default if no major problems. Similar behaviour for ASN1_SET_OF
15261 and PKCS12_STACK_OF.
15265 * When some versions of IIS use the 'NET' form of private key the
15266 key derivation algorithm is different. Normally MD5(password) is
15267 used as a 128 bit RC4 key. In the modified case
15268 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15269 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15270 as the old Netscape_RSA functions except they have an additional
15271 'sgckey' parameter which uses the modified algorithm. Also added
15272 an -sgckey command line option to the rsa utility. Thanks to
15273 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15274 algorithm to openssl-dev.
15278 * The evp_local.h macros were using 'c.##kname' which resulted in
15279 invalid expansion on some systems (SCO 5.0.5 for example).
15280 Corrected to 'c.kname'.
15282 *Phillip Porch <root@theporch.com>*
15284 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15285 a STACK of email addresses from a certificate or request, these look
15286 in the subject name and the subject alternative name extensions and
15287 omit any duplicate addresses.
15291 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15292 This makes DSA verification about 2 % faster.
15296 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15297 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15298 plus overhead for 1024 bit moduli).
15299 This makes exponentiations about 0.5 % faster for 1024 bit
15300 exponents (as measured by "openssl speed rsa2048").
15304 * Rename memory handling macros to avoid conflicts with other
15306 Malloc => OPENSSL_malloc
15307 Malloc_locked => OPENSSL_malloc_locked
15308 Realloc => OPENSSL_realloc
15309 Free => OPENSSL_free
15313 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15314 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15318 * CygWin32 support.
15320 *John Jarvie <jjarvie@newsguy.com>*
15322 * The type-safe stack code has been rejigged. It is now only compiled
15323 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15324 by default all type-specific stack functions are "#define"d back to
15325 standard stack functions. This results in more streamlined output
15326 but retains the type-safety checking possibilities of the original
15331 * The STACK code has been cleaned up, and certain type declarations
15332 that didn't make a lot of sense have been brought in line. This has
15333 also involved a cleanup of sorts in safestack.h to more correctly
15334 map type-safe stack functions onto their plain stack counterparts.
15335 This work has also resulted in a variety of "const"ifications of
15336 lots of the code, especially `_cmp` operations which should normally
15337 be prototyped with "const" parameters anyway.
15341 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15342 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15343 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15344 where all of 'md' is used each time the PRNG is used, but 'state'
15345 is used only indexed by a cyclic counter. As entropy may not be
15346 well distributed from the beginning, 'md' is important as a
15347 chaining variable. However, the output function chains only half
15348 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15349 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15350 in all of 'state' being rewritten, with the new values depending
15351 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15355 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15356 the handshake is continued after ssl_verify_cert_chain();
15357 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15358 can lead to 'unexplainable' connection aborts later.
15360 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15362 * Major EVP API cipher revision.
15363 Add hooks for extra EVP features. This allows various cipher
15364 parameters to be set in the EVP interface. Support added for variable
15365 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15366 setting of RC2 and RC5 parameters.
15368 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15371 Remove lots of duplicated code from the EVP library. For example *every*
15372 cipher init() function handles the 'iv' in the same way according to the
15373 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15374 for CFB and OFB modes they zero ctx->num.
15376 New functionality allows removal of S/MIME code RC2 hack.
15378 Most of the routines have the same form and so can be declared in terms
15381 By shifting this to the top level EVP_CipherInit() it can be removed from
15382 all individual ciphers. If the cipher wants to handle IVs or keys
15383 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15386 Change lots of functions like EVP_EncryptUpdate() to now return a
15387 value: although software versions of the algorithms cannot fail
15388 any installed hardware versions can.
15392 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15393 this option is set, tolerate broken clients that send the negotiated
15394 protocol version number instead of the requested protocol version
15399 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15400 i.e. non-zero for export ciphersuites, zero otherwise.
15401 Previous versions had this flag inverted, inconsistent with
15402 rsa_tmp_cb (..._TMP_RSA_CB).
15404 *Bodo Moeller; problem reported by Amit Chopra*
15406 * Add missing DSA library text string. Work around for some IIS
15407 key files with invalid SEQUENCE encoding.
15411 * Add a document (doc/standards.txt) that list all kinds of standards
15412 and so on that are implemented in OpenSSL.
15416 * Enhance c_rehash script. Old version would mishandle certificates
15417 with the same subject name hash and wouldn't handle CRLs at all.
15418 Added -fingerprint option to crl utility, to support new c_rehash
15423 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15427 * Fix for SSL server purpose checking. Server checking was
15428 rejecting certificates which had extended key usage present
15429 but no ssl client purpose.
15431 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15433 * Make PKCS#12 code work with no password. The PKCS#12 spec
15434 is a little unclear about how a blank password is handled.
15435 Since the password in encoded as a BMPString with terminating
15436 double NULL a zero length password would end up as just the
15437 double NULL. However no password at all is different and is
15438 handled differently in the PKCS#12 key generation code. NS
15439 treats a blank password as zero length. MSIE treats it as no
15440 password on export: but it will try both on import. We now do
15441 the same: PKCS12_parse() tries zero length and no password if
15442 the password is set to "" or NULL (NULL is now a valid password:
15443 it wasn't before) as does the pkcs12 application.
15447 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15448 perror when PEM_read_bio_X509_REQ fails, the error message must
15449 be obtained from the error queue.
15453 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15454 it in ERR_remove_state if appropriate, and change ERR_get_state
15455 accordingly to avoid race conditions (this is necessary because
15456 thread_hash is no longer constant once set).
15460 * Bugfix for linux-elf makefile.one.
15464 * RSA_get_default_method() will now cause a default
15465 RSA_METHOD to be chosen if one doesn't exist already.
15466 Previously this was only set during a call to RSA_new()
15467 or RSA_new_method(NULL) meaning it was possible for
15468 RSA_get_default_method() to return NULL.
15472 * Added native name translation to the existing DSO code
15473 that will convert (if the flag to do so is set) filenames
15474 that are sufficiently small and have no path information
15475 into a canonical native form. Eg. "blah" converted to
15476 "libblah.so" or "blah.dll" etc.
15480 * New function ERR_error_string_n(e, buf, len) which is like
15481 ERR_error_string(e, buf), but writes at most 'len' bytes
15482 including the 0 terminator. For ERR_error_string_n, 'buf'
15485 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15487 * CONF library reworked to become more general. A new CONF
15488 configuration file reader "class" is implemented as well as a
15489 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15490 old `CONF_*` functions are still there, but are reimplemented to
15491 work in terms of the new functions. Also, a set of functions
15492 to handle the internal storage of the configuration data is
15493 provided to make it easier to write new configuration file
15494 reader "classes" (I can definitely see something reading a
15495 configuration file in XML format, for example), called `_CONF_*`,
15496 or "the configuration storage API"...
15498 The new configuration file reading functions are:
15500 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15501 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15503 NCONF_default, NCONF_WIN32
15505 NCONF_dump_fp, NCONF_dump_bio
15507 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15508 NCONF_new creates a new CONF object. This works in the same way
15509 as other interfaces in OpenSSL, like the BIO interface.
15510 `NCONF_dump_*` dump the internal storage of the configuration file,
15511 which is useful for debugging. All other functions take the same
15512 arguments as the old `CONF_*` functions with the exception of the
15513 first that must be a `CONF *` instead of a `LHASH *`.
15515 To make it easier to use the new classes with the old `CONF_*` functions,
15516 the function CONF_set_default_method is provided.
15520 * Add '-tls1' option to 'openssl ciphers', which was already
15521 mentioned in the documentation but had not been implemented.
15522 (This option is not yet really useful because even the additional
15523 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15527 * Initial DSO code added into libcrypto for letting OpenSSL (and
15528 OpenSSL-based applications) load shared libraries and bind to
15529 them in a portable way.
15531 *Geoff Thorpe, with contributions from Richard Levitte*
15533 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15535 * Make sure _lrotl and _lrotr are only used with MSVC.
15537 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15538 (the default implementation of RAND_status).
15540 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15541 to '-clrext' (= clear extensions), as intended and documented.
15542 *Bodo Moeller; inconsistency pointed out by Michael Attili
15543 <attili@amaxo.com>*
15545 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15546 was larger than the MD block size.
15548 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15550 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15551 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15552 using the passed key: if the passed key was a private key the result
15553 of X509_print(), for example, would be to print out all the private key
15558 * des_quad_cksum() byte order bug fix.
15559 *Ulf Möller, using the problem description in krb4-0.9.7, where
15560 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15562 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15565 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15567 * For easily testing in shell scripts whether some command
15568 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15569 returns with exit code 0 iff no command of the given name is available.
15570 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15571 the output goes to stdout and nothing is printed to stderr.
15572 Additional arguments are always ignored.
15574 Since for each cipher there is a command of the same name,
15575 the 'no-cipher' compilation switches can be tested this way.
15577 ('openssl no-XXX' is not able to detect pseudo-commands such
15578 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15582 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15586 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15587 is set; it will be thrown away anyway because each handshake creates
15589 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15590 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15591 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15592 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15596 * New s_client option -ign_eof: EOF at stdin is ignored, and
15597 'Q' and 'R' lose their special meanings (quit/renegotiate).
15598 This is part of what -quiet does; unlike -quiet, -ign_eof
15599 does not suppress any output.
15603 * Add compatibility options to the purpose and trust code. The
15604 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15605 accepts a certificate or CA, this was the previous behaviour,
15606 with all the associated security issues.
15608 X509_TRUST_COMPAT is the old trust behaviour: only and
15609 automatically trust self signed roots in certificate store. A
15610 new trust setting X509_TRUST_DEFAULT is used to specify that
15611 a purpose has no associated trust setting and it should instead
15612 use the value in the default purpose.
15616 * Fix the PKCS#8 DSA private key code so it decodes keys again
15617 and fix a memory leak.
15621 * In util/mkerr.pl (which implements 'make errors'), preserve
15622 reason strings from the previous version of the .c file, as
15623 the default to have only downcase letters (and digits) in
15624 automatically generated reasons codes is not always appropriate.
15628 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15629 using strerror. Previously, ERR_reason_error_string() returned
15630 library names as reason strings for SYSerr; but SYSerr is a special
15631 case where small numbers are errno values, not library numbers.
15635 * Add '-dsaparam' option to 'openssl dhparam' application. This
15636 converts DSA parameters into DH parameters. (When creating parameters,
15637 DSA_generate_parameters is used.)
15641 * Include 'length' (recommended exponent length) in C code generated
15642 by 'openssl dhparam -C'.
15646 * The second argument to set_label in perlasm was already being used
15647 so couldn't be used as a "file scope" flag. Moved to third argument
15652 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15653 instead of RAND_bytes for encryption IVs and salts.
15657 * Include RAND_status() into RAND_METHOD instead of implementing
15658 it only for md_rand.c Otherwise replacing the PRNG by calling
15659 RAND_set_rand_method would be impossible.
15663 * Don't let DSA_generate_key() enter an infinite loop if the random
15664 number generation fails.
15668 * New 'rand' application for creating pseudo-random output.
15672 * Added configuration support for Linux/IA64
15674 *Rolf Haberrecker <rolf@suse.de>*
15676 * Assembler module support for Mingw32.
15680 * Shared library support for HPUX (in shlib/).
15682 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15684 * Shared library support for Solaris gcc.
15686 *Lutz Behnke <behnke@trustcenter.de>*
15688 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15690 * PKCS7_encrypt() was adding text MIME headers twice because they
15691 were added manually and by SMIME_crlf_copy().
15695 * In bntest.c don't call BN_rand with zero bits argument.
15697 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15699 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15700 case was implemented. This caused BN_div_recp() to fail occasionally.
15704 * Add an optional second argument to the set_label() in the perl
15705 assembly language builder. If this argument exists and is set
15706 to 1 it signals that the assembler should use a symbol whose
15707 scope is the entire file, not just the current function. This
15708 is needed with MASM which uses the format label:: for this scope.
15710 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15712 * Change the ASN1 types so they are typedefs by default. Before
15713 almost all types were #define'd to ASN1_STRING which was causing
15714 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15719 * Change names of new functions to the new get1/get0 naming
15720 convention: After 'get1', the caller owns a reference count
15721 and has to call `..._free`; 'get0' returns a pointer to some
15722 data structure without incrementing reference counters.
15723 (Some of the existing 'get' functions increment a reference
15724 counter, some don't.)
15725 Similarly, 'set1' and 'add1' functions increase reference
15726 counters or duplicate objects.
15730 * Allow for the possibility of temp RSA key generation failure:
15731 the code used to assume it always worked and crashed on failure.
15735 * Fix potential buffer overrun problem in BIO_printf().
15736 *Ulf Möller, using public domain code by Patrick Powell; problem
15737 pointed out by David Sacerdote <das33@cornell.edu>*
15739 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15740 RAND_egd() and RAND_status(). In the command line application,
15741 the EGD socket can be specified like a seed file using RANDFILE
15746 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15747 Some CAs (e.g. Verisign) distribute certificates in this form.
15751 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15752 list to exclude them. This means that no special compilation option
15753 is needed to use anonymous DH: it just needs to be included in the
15758 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15759 EVP_MD_type. The old functionality is available in a new macro called
15760 EVP_MD_md(). Change code that uses it and update docs.
15764 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15765 where the `void *` argument is replaced by a function pointer argument.
15766 Previously `void *` was abused to point to functions, which works on
15767 many platforms, but is not correct. As these functions are usually
15768 called by macros defined in OpenSSL header files, most source code
15769 should work without changes.
15773 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15774 sections with information on -D... compiler switches used for
15775 compiling the library so that applications can see them. To enable
15776 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15777 must be defined. E.g.,
15778 #define OPENSSL_ALGORITHM_DEFINES
15779 #include <openssl/opensslconf.h>
15780 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15782 *Richard Levitte, Ulf and Bodo Möller*
15784 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15789 * Change the 'other' type in certificate aux info to a STACK_OF
15790 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15791 the required ASN1 format: arbitrary types determined by an OID.
15795 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15796 argument to 'req'. This is not because the function is newer or
15797 better than others it just uses the work 'NEW' in the certificate
15798 request header lines. Some software needs this.
15802 * Reorganise password command line arguments: now passwords can be
15803 obtained from various sources. Delete the PEM_cb function and make
15804 it the default behaviour: i.e. if the callback is NULL and the
15805 usrdata argument is not NULL interpret it as a null terminated pass
15806 phrase. If usrdata and the callback are NULL then the pass phrase
15807 is prompted for as usual.
15811 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15812 the support is automatically enabled. The resulting binaries will
15813 autodetect the card and use it if present.
15815 *Ben Laurie and Compaq Inc.*
15817 * Work around for Netscape hang bug. This sends certificate request
15818 and server done in one record. Since this is perfectly legal in the
15819 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15820 the bugs/SSLv3 entry for more info.
15824 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15828 * Add -rand argument to smime and pkcs12 applications and read/write
15833 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15837 * Add command line password options to the remaining applications.
15841 * Bug fix for BN_div_recp() for numerators with an even number of
15846 * More tests in bntest.c, and changed test_bn output.
15850 * ./config recognizes MacOS X now.
15854 * Bug fix for BN_div() when the first words of num and divisor are
15855 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15859 * Add support for various broken PKCS#8 formats, and command line
15860 options to produce them.
15864 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15865 get temporary BIGNUMs from a BN_CTX.
15869 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15874 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15875 include a #define from the old name to the new. The original intent
15876 was that statically linked binaries could for example just call
15877 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15878 link with digests. This never worked because SSLeay_add_all_digests()
15879 and SSLeay_add_all_ciphers() were in the same source file so calling
15880 one would link with the other. They are now in separate source files.
15884 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15888 * Use a less unusual form of the Miller-Rabin primality test (it used
15889 a binary algorithm for exponentiation integrated into the Miller-Rabin
15890 loop, our standard modexp algorithms are faster).
15894 * Support for the EBCDIC character set completed.
15896 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15898 * Source code cleanups: use const where appropriate, eliminate casts,
15899 use `void *` instead of `char *` in lhash.
15903 * Bugfix: ssl3_send_server_key_exchange was not restartable
15904 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15905 this the server could overwrite ephemeral keys that the client
15910 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15911 using 50 iterations of the Rabin-Miller test.
15913 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15914 iterations of the Rabin-Miller test as required by the appendix
15915 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15916 As BN_is_prime_fasttest includes trial division, DSA parameter
15917 generation becomes much faster.
15919 This implies a change for the callback functions in DSA_is_prime
15920 and DSA_generate_parameters: The callback function is called once
15921 for each positive witness in the Rabin-Miller test, not just
15922 occasionally in the inner loop; and the parameters to the
15923 callback function now provide an iteration count for the outer
15924 loop rather than for the current invocation of the inner loop.
15925 DSA_generate_parameters additionally can call the callback
15926 function with an 'iteration count' of -1, meaning that a
15927 candidate has passed the trial division test (when q is generated
15928 from an application-provided seed, trial division is skipped).
15932 * New function BN_is_prime_fasttest that optionally does trial
15933 division before starting the Rabin-Miller test and has
15934 an additional BN_CTX * argument (whereas BN_is_prime always
15935 has to allocate at least one BN_CTX).
15936 'callback(1, -1, cb_arg)' is called when a number has passed the
15937 trial division stage.
15941 * Fix for bug in CRL encoding. The validity dates weren't being handled
15946 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15950 * New function BN_pseudo_rand().
15954 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15955 bignum version of BN_from_montgomery() with the working code from
15956 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15961 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15962 made it impossible to use the same SSL_SESSION data structure in
15963 SSL2 clients in multiple threads.
15967 * The return value of RAND_load_file() no longer counts bytes obtained
15968 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15969 to seed the PRNG (previously an explicit byte count was required).
15971 *Ulf Möller, Bodo Möller*
15973 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15974 used `char *` instead of `void *` and had casts all over the place.
15978 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15982 * Retain source code compatibility for BN_prime_checks macro:
15983 BN_is_prime(..., BN_prime_checks, ...) now uses
15984 BN_prime_checks_for_size to determine the appropriate number of
15985 Rabin-Miller iterations.
15989 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15990 DH_CHECK_P_NOT_SAFE_PRIME.
15991 (Check if this is true? OpenPGP calls them "strong".)
15995 * Merge the functionality of "dh" and "gendh" programs into a new program
15996 "dhparam". The old programs are retained for now but will handle DH keys
15997 (instead of parameters) in future.
16001 * Make the ciphers, s_server and s_client programs check the return values
16002 when a new cipher list is set.
16006 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16007 ciphers. Before when the 56bit ciphers were enabled the sorting was
16010 The syntax for the cipher sorting has been extended to support sorting by
16011 cipher-strength (using the strength_bits hard coded in the tables).
16012 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16014 Fix a bug in the cipher-command parser: when supplying a cipher command
16015 string with an "undefined" symbol (neither command nor alphanumeric
16016 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16017 an error is flagged.
16019 Due to the strength-sorting extension, the code of the
16020 ssl_create_cipher_list() function was completely rearranged. I hope that
16021 the readability was also increased :-)
16023 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16025 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16026 for the first serial number and places 2 in the serial number file. This
16027 avoids problems when the root CA is created with serial number zero and
16028 the first user certificate has the same issuer name and serial number
16033 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16034 the new code. Add documentation for this stuff.
16038 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16039 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16040 structures and behave in an analogous way to the X509v3 functions:
16041 they shouldn't be called directly but wrapper functions should be used
16044 So we also now have some wrapper functions that call the X509at functions
16045 when passed certificate requests. (TO DO: similar things can be done with
16046 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16047 things. Some of these need some d2i or i2d and print functionality
16048 because they handle more complex structures.)
16052 * Add missing #ifndefs that caused missing symbols when building libssl
16053 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16054 NO_RSA in `ssl/s2*.c`.
16056 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16058 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16059 has a return value which indicates the quality of the random data
16060 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16061 error queue. New function RAND_pseudo_bytes() generates output that is
16062 guaranteed to be unique but not unpredictable. RAND_add is like
16063 RAND_seed, but takes an extra argument for an entropy estimate
16064 (RAND_seed always assumes full entropy).
16068 * Do more iterations of Rabin-Miller probable prime test (specifically,
16069 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16070 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16071 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16072 false-positive rate of at most 2^-80 for random input.
16076 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16080 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16081 in the 0.9.5 release), this returns the chain
16082 from an X509_CTX structure with a dup of the stack and all
16083 the X509 reference counts upped: so the stack will exist
16084 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16087 Also make SSL_SESSION_print() print out the verify return
16092 * Add manpage for the pkcs12 command. Also change the default
16093 behaviour so MAC iteration counts are used unless the new
16094 -nomaciter option is used. This improves file security and
16095 only older versions of MSIE (4.0 for example) need it.
16099 * Honor the no-xxx Configure options when creating .DEF files.
16103 * Add PKCS#10 attributes to field table: challengePassword,
16104 unstructuredName and unstructuredAddress. These are taken from
16105 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16106 international characters are used.
16108 More changes to X509_ATTRIBUTE code: allow the setting of types
16109 based on strings. Remove the 'loc' parameter when adding
16110 attributes because these will be a SET OF encoding which is sorted
16115 * Initial changes to the 'req' utility to allow request generation
16116 automation. This will allow an application to just generate a template
16117 file containing all the field values and have req construct the
16120 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16121 used all over the place including certificate requests and PKCS#7
16122 structures. They are currently handled manually where necessary with
16123 some primitive wrappers for PKCS#7. The new functions behave in a
16124 manner analogous to the X509 extension functions: they allow
16125 attributes to be looked up by NID and added.
16127 Later something similar to the X509V3 code would be desirable to
16128 automatically handle the encoding, decoding and printing of the
16129 more complex types. The string types like challengePassword can
16130 be handled by the string table functions.
16132 Also modified the multi byte string table handling. Now there is
16133 a 'global mask' which masks out certain types. The table itself
16134 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16135 is useful when for example there is only one permissible type
16136 (as in countryName) and using the mask might result in no valid
16141 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16142 SSL_get_peer_finished to allow applications to obtain the latest
16143 Finished messages sent to the peer or expected from the peer,
16144 respectively. (SSL_get_peer_finished is usually the Finished message
16145 actually received from the peer, otherwise the protocol will be aborted.)
16147 As the Finished message are message digests of the complete handshake
16148 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16149 be used for external authentication procedures when the authentication
16150 provided by SSL/TLS is not desired or is not enough.
16154 * Enhanced support for Alpha Linux is added. Now ./config checks if
16155 the host supports BWX extension and if Compaq C is present on the
16156 $PATH. Just exploiting of the BWX extension results in 20-30%
16157 performance kick for some algorithms, e.g. DES and RC4 to mention
16158 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16163 * Add support for MS "fast SGC". This is arguably a violation of the
16164 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16165 weak crypto and after checking the certificate is SGC a second one
16166 with strong crypto. MS SGC stops the first handshake after receiving
16167 the server certificate message and sends a second client hello. Since
16168 a server will typically do all the time consuming operations before
16169 expecting any further messages from the client (server key exchange
16170 is the most expensive) there is little difference between the two.
16172 To get OpenSSL to support MS SGC we have to permit a second client
16173 hello message after we have sent server done. In addition we have to
16174 reset the MAC if we do get this second client hello.
16178 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16179 if a DER encoded private key is RSA or DSA traditional format. Changed
16180 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16181 format DER encoded private key. Newer code should use PKCS#8 format which
16182 has the key type encoded in the ASN1 structure. Added DER private key
16183 support to pkcs8 application.
16187 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16188 ciphersuites has been selected (as required by the SSL 3/TLS 1
16189 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16190 is set, we interpret this as a request to violate the specification
16191 (the worst that can happen is a handshake failure, and 'correct'
16192 behaviour would result in a handshake failure anyway).
16196 * In SSL_CTX_add_session, take into account that there might be multiple
16197 SSL_SESSION structures with the same session ID (e.g. when two threads
16198 concurrently obtain them from an external cache).
16199 The internal cache can handle only one SSL_SESSION with a given ID,
16200 so if there's a conflict, we now throw out the old one to achieve
16205 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16206 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16207 some routines that use cipher OIDs: some ciphers do not have OIDs
16208 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16213 * Simplify the trust setting structure and code. Now we just have
16214 two sequences of OIDs for trusted and rejected settings. These will
16215 typically have values the same as the extended key usage extension
16216 and any application specific purposes.
16218 The trust checking code now has a default behaviour: it will just
16219 check for an object with the same NID as the passed id. Functions can
16220 be provided to override either the default behaviour or the behaviour
16221 for a given id. SSL client, server and email already have functions
16222 in place for compatibility: they check the NID and also return "trusted"
16223 if the certificate is self signed.
16227 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16228 traditional format into an EVP_PKEY structure.
16232 * Add a password callback function PEM_cb() which either prompts for
16233 a password if usr_data is NULL or otherwise assumes it is a null
16234 terminated password. Allow passwords to be passed on command line
16235 environment or config files in a few more utilities.
16239 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16240 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16241 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16242 Update documentation.
16246 * Support for ASN1 "NULL" type. This could be handled before by using
16247 ASN1_TYPE but there wasn't any function that would try to read a NULL
16248 and produce an error if it couldn't. For compatibility we also have
16249 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16250 don't allocate anything because they don't need to.
16254 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16257 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16259 * Rebuild of the memory allocation routines used by OpenSSL code and
16260 possibly others as well. The purpose is to make an interface that
16261 provide hooks so anyone can build a separate set of allocation and
16262 deallocation routines to be used by OpenSSL, for example memory
16263 pool implementations, or something else, which was previously hard
16264 since Malloc(), Realloc() and Free() were defined as macros having
16265 the values malloc, realloc and free, respectively (except for Win32
16266 compilations). The same is provided for memory debugging code.
16267 OpenSSL already comes with functionality to find memory leaks, but
16268 this gives people a chance to debug other memory problems.
16270 With these changes, a new set of functions and macros have appeared:
16272 CRYPTO_set_mem_debug_functions() [F]
16273 CRYPTO_get_mem_debug_functions() [F]
16274 CRYPTO_dbg_set_options() [F]
16275 CRYPTO_dbg_get_options() [F]
16276 CRYPTO_malloc_debug_init() [M]
16278 The memory debug functions are NULL by default, unless the library
16279 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16280 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16281 gives the standard debugging functions that come with OpenSSL) or
16282 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16283 provided by the library user) must be used. When the standard
16284 debugging functions are used, CRYPTO_dbg_set_options can be used to
16285 request additional information:
16286 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16287 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16289 Also, things like CRYPTO_set_mem_functions will always give the
16290 expected result (the new set of functions is used for allocation
16291 and deallocation) at all times, regardless of platform and compiler
16294 To finish it up, some functions that were never use in any other
16295 way than through macros have a new API and new semantic:
16297 CRYPTO_dbg_malloc()
16298 CRYPTO_dbg_realloc()
16301 All macros of value have retained their old syntax.
16303 *Richard Levitte and Bodo Moeller*
16305 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16306 ordering of SMIMECapabilities wasn't in "strength order" and there
16307 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16312 * Some ASN1 types with illegal zero length encoding (INTEGER,
16313 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16315 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16317 * Merge in my S/MIME library for OpenSSL. This provides a simple
16318 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16319 functionality to handle multipart/signed properly) and a utility
16320 called 'smime' to call all this stuff. This is based on code I
16321 originally wrote for Celo who have kindly allowed it to be
16322 included in OpenSSL.
16326 * Add variants des_set_key_checked and des_set_key_unchecked of
16327 des_set_key (aka des_key_sched). Global variable des_check_key
16328 decides which of these is called by des_set_key; this way
16329 des_check_key behaves as it always did, but applications and
16330 the library itself, which was buggy for des_check_key == 1,
16331 have a cleaner way to pick the version they need.
16335 * New function PKCS12_newpass() which changes the password of a
16340 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16341 dynamic mix. In both cases the ids can be used as an index into the
16342 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16343 functions so they accept a list of the field values and the
16344 application doesn't need to directly manipulate the X509_TRUST
16349 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16354 * Modify the way the V3 extension code looks up extensions. This now
16355 works in a similar way to the object code: we have some "standard"
16356 extensions in a static table which is searched with OBJ_bsearch()
16357 and the application can add dynamic ones if needed. The file
16358 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16359 updated whenever a new extension is added to the core code and kept
16360 in ext_nid order. There is a simple program 'tabtest.c' which checks
16361 this. New extensions are not added too often so this file can readily
16362 be maintained manually.
16364 There are two big advantages in doing things this way. The extensions
16365 can be looked up immediately and no longer need to be "added" using
16366 X509V3_add_standard_extensions(): this function now does nothing.
16367 Side note: I get *lots* of email saying the extension code doesn't
16368 work because people forget to call this function.
16369 Also no dynamic allocation is done unless new extensions are added:
16370 so if we don't add custom extensions there is no need to call
16371 X509V3_EXT_cleanup().
16375 * Modify enc utility's salting as follows: make salting the default. Add a
16376 magic header, so unsalted files fail gracefully instead of just decrypting
16377 to garbage. This is because not salting is a big security hole, so people
16378 should be discouraged from doing it.
16382 * Fixes and enhancements to the 'x509' utility. It allowed a message
16383 digest to be passed on the command line but it only used this
16384 parameter when signing a certificate. Modified so all relevant
16385 operations are affected by the digest parameter including the
16386 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16387 DSA key was used because it didn't fix the digest.
16391 * Initial certificate chain verify code. Currently tests the untrusted
16392 certificates for consistency with the verify purpose (which is set
16393 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16395 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16396 this is because it will reject chains with invalid extensions whereas
16397 every previous version of OpenSSL and SSLeay made no checks at all.
16399 Trust code: checks the root CA for the relevant trust settings. Trust
16400 settings have an initial value consistent with the verify purpose: e.g.
16401 if the verify purpose is for SSL client use it expects the CA to be
16402 trusted for SSL client use. However the default value can be changed to
16403 permit custom trust settings: one example of this would be to only trust
16404 certificates from a specific "secure" set of CAs.
16406 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16407 which should be used for version portability: especially since the
16408 verify structure is likely to change more often now.
16410 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16411 to set them. If not set then assume SSL clients will verify SSL servers
16414 Two new options to the verify program: -untrusted allows a set of
16415 untrusted certificates to be passed in and -purpose which sets the
16416 intended purpose of the certificate. If a purpose is set then the
16417 new chain verify code is used to check extension consistency.
16421 * Support for the authority information access extension.
16425 * Modify RSA and DSA PEM read routines to transparently handle
16426 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16427 public keys in a format compatible with certificate
16428 SubjectPublicKeyInfo structures. Unfortunately there were already
16429 functions called *_PublicKey_* which used various odd formats so
16430 these are retained for compatibility: however the DSA variants were
16431 never in a public release so they have been deleted. Changed dsa/rsa
16432 utilities to handle the new format: note no releases ever handled public
16433 keys so we should be OK.
16435 The primary motivation for this change is to avoid the same fiasco
16436 that dogs private keys: there are several incompatible private key
16437 formats some of which are standard and some OpenSSL specific and
16438 require various evil hacks to allow partial transparent handling and
16439 even then it doesn't work with DER formats. Given the option anything
16440 other than PKCS#8 should be dumped: but the other formats have to
16441 stay in the name of compatibility.
16443 With public keys and the benefit of hindsight one standard format
16444 is used which works with EVP_PKEY, RSA or DSA structures: though
16445 it clearly returns an error if you try to read the wrong kind of key.
16447 Added a -pubkey option to the 'x509' utility to output the public key.
16448 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16449 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16450 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16451 that do the same as the `EVP_PKEY_assign_*()` except they up the
16452 reference count of the added key (they don't "swallow" the
16457 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16458 CRLs would fail if the file contained no certificates or no CRLs:
16459 added a new function to read in both types and return the number
16460 read: this means that if none are read it will be an error. The
16461 DER versions of the certificate and CRL reader would always fail
16462 because it isn't possible to mix certificates and CRLs in DER format
16463 without choking one or the other routine. Changed this to just read
16464 a certificate: this is the best we can do. Also modified the code
16465 in `apps/verify.c` to take notice of return codes: it was previously
16466 attempting to read in certificates from NULL pointers and ignoring
16467 any errors: this is one reason why the cert and CRL reader seemed
16468 to work. It doesn't check return codes from the default certificate
16469 routines: these may well fail if the certificates aren't installed.
16473 * Code to support otherName option in GeneralName.
16477 * First update to verify code. Change the verify utility
16478 so it warns if it is passed a self signed certificate:
16479 for consistency with the normal behaviour. X509_verify
16480 has been modified to it will now verify a self signed
16481 certificate if *exactly* the same certificate appears
16482 in the store: it was previously impossible to trust a
16483 single self signed certificate. This means that:
16484 openssl verify ss.pem
16485 now gives a warning about a self signed certificate but
16486 openssl verify -CAfile ss.pem ss.pem
16491 * For servers, store verify_result in SSL_SESSION data structure
16492 (and add it to external session representation).
16493 This is needed when client certificate verifications fails,
16494 but an application-provided verification callback (set by
16495 SSL_CTX_set_cert_verify_callback) allows accepting the session
16496 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16497 but returns 1): When the session is reused, we have to set
16498 ssl->verify_result to the appropriate error code to avoid
16501 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16503 * Fix a bug in the new PKCS#7 code: it didn't consider the
16504 case in PKCS7_dataInit() where the signed PKCS7 structure
16505 didn't contain any existing data because it was being created.
16507 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16509 * Add a salt to the key derivation routines in enc.c. This
16510 forms the first 8 bytes of the encrypted file. Also add a
16511 -S option to allow a salt to be input on the command line.
16515 * New function X509_cmp(). Oddly enough there wasn't a function
16516 to compare two certificates. We do this by working out the SHA1
16517 hash and comparing that. X509_cmp() will be needed by the trust
16522 * SSL_get1_session() is like SSL_get_session(), but increments
16523 the reference count in the SSL_SESSION returned.
16525 *Geoff Thorpe <geoff@eu.c2.net>*
16527 * Fix for 'req': it was adding a null to request attributes.
16528 Also change the X509_LOOKUP and X509_INFO code to handle
16529 certificate auxiliary information.
16533 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16538 * Add the possibility to add extra information to the memory leak
16539 detecting output, to form tracebacks, showing from where each
16540 allocation was originated: CRYPTO_push_info("constant string") adds
16541 the string plus current file name and line number to a per-thread
16542 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16543 is like calling CYRPTO_pop_info() until the stack is empty.
16544 Also updated memory leak detection code to be multi-thread-safe.
16548 * Add options -text and -noout to pkcs7 utility and delete the
16549 encryption options which never did anything. Update docs.
16553 * Add options to some of the utilities to allow the pass phrase
16554 to be included on either the command line (not recommended on
16555 OSes like Unix) or read from the environment. Update the
16556 manpages and fix a few bugs.
16560 * Add a few manpages for some of the openssl commands.
16564 * Fix the -revoke option in ca. It was freeing up memory twice,
16565 leaking and not finding already revoked certificates.
16569 * Extensive changes to support certificate auxiliary information.
16570 This involves the use of X509_CERT_AUX structure and X509_AUX
16571 functions. An X509_AUX function such as PEM_read_X509_AUX()
16572 can still read in a certificate file in the usual way but it
16573 will also read in any additional "auxiliary information". By
16574 doing things this way a fair degree of compatibility can be
16575 retained: existing certificates can have this information added
16576 using the new 'x509' options.
16578 Current auxiliary information includes an "alias" and some trust
16579 settings. The trust settings will ultimately be used in enhanced
16580 certificate chain verification routines: currently a certificate
16581 can only be trusted if it is self signed and then it is trusted
16586 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16587 The problem was that one of the replacement routines had not been working
16588 since SSLeay releases. For now the offending routine has been replaced
16589 with non-optimised assembler. Even so, this now gives around 95%
16590 performance improvement for 1024 bit RSA signs.
16594 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16595 handling. Most clients have the effective key size in bits equal to
16596 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16597 A few however don't do this and instead use the size of the decrypted key
16598 to determine the RC2 key length and the AlgorithmIdentifier to determine
16599 the effective key length. In this case the effective key length can still
16600 be 40 bits but the key length can be 168 bits for example. This is fixed
16601 by manually forcing an RC2 key into the EVP_PKEY structure because the
16602 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16603 the key length and effective key length are equal.
16607 * Add a bunch of functions that should simplify the creation of
16608 X509_NAME structures. Now you should be able to do:
16609 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16610 and have it automatically work out the correct field type and fill in
16611 the structures. The more adventurous can try:
16612 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16613 and it will (hopefully) work out the correct multibyte encoding.
16617 * Change the 'req' utility to use the new field handling and multibyte
16618 copy routines. Before the DN field creation was handled in an ad hoc
16619 way in req, ca, and x509 which was rather broken and didn't support
16620 BMPStrings or UTF8Strings. Since some software doesn't implement
16621 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16622 using the dirstring_type option. See the new comment in the default
16623 openssl.cnf for more info.
16627 * Make crypto/rand/md_rand.c more robust:
16628 - Assure unique random numbers after fork().
16629 - Make sure that concurrent threads access the global counter and
16630 md serializably so that we never lose entropy in them
16631 or use exactly the same state in multiple threads.
16632 Access to the large state is not always serializable because
16633 the additional locking could be a performance killer, and
16634 md should be large enough anyway.
16638 * New file `apps/app_rand.c` with commonly needed functionality
16639 for handling the random seed file.
16641 Use the random seed file in some applications that previously did not:
16643 dsaparam -genkey (which also ignored its '-rand' option),
16646 x509 (when signing).
16647 Except on systems with /dev/urandom, it is crucial to have a random
16648 seed file at least for key creation, DSA signing, and for DH exchanges;
16649 for RSA signatures we could do without one.
16651 gendh and gendsa (unlike genrsa) used to read only the first byte
16652 of each file listed in the '-rand' option. The function as previously
16653 found in genrsa is now in app_rand.c and is used by all programs
16654 that support '-rand'.
16658 * In RAND_write_file, use mode 0600 for creating files;
16659 don't just chmod when it may be too late.
16663 * Report an error from X509_STORE_load_locations
16664 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16668 * New function ASN1_mbstring_copy() this copies a string in either
16669 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16670 into an ASN1_STRING type. A mask of permissible types is passed
16671 and it chooses the "minimal" type to use or an error if not type
16676 * Add function equivalents to the various macros in asn1.h. The old
16677 macros are retained with an `M_` prefix. Code inside the library can
16678 use the `M_` macros. External code (including the openssl utility)
16679 should *NOT* in order to be "shared library friendly".
16683 * Add various functions that can check a certificate's extensions
16684 to see if it usable for various purposes such as SSL client,
16685 server or S/MIME and CAs of these types. This is currently
16686 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16687 verification. Also added a -purpose flag to x509 utility to
16688 print out all the purposes.
16692 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16697 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16698 for, obtain and decode and extension and obtain its critical flag.
16699 This allows all the necessary extension code to be handled in a
16700 single function call.
16704 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16705 platforms. See crypto/rc4/rc4_enc.c for further details.
16709 * New -noout option to asn1parse. This causes no output to be produced
16710 its main use is when combined with -strparse and -out to extract data
16711 from a file (which may not be in ASN.1 format).
16715 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16716 when producing the local key id.
16718 *Richard Levitte <levitte@stacken.kth.se>*
16720 * New option -dhparam in s_server. This allows a DH parameter file to be
16721 stated explicitly. If it is not stated then it tries the first server
16722 certificate file. The previous behaviour hard coded the filename
16727 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16728 a public key to be input or output. For example:
16729 openssl rsa -in key.pem -pubout -out pubkey.pem
16730 Also added necessary DSA public key functions to handle this.
16734 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16735 in the message. This was handled by allowing
16736 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16738 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16740 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16741 to the end of the strings whereas this didn't. This would cause problems
16742 if strings read with d2i_ASN1_bytes() were later modified.
16744 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16746 * Fix for base64 decode bug. When a base64 bio reads only one line of
16747 data and it contains EOF it will end up returning an error. This is
16748 caused by input 46 bytes long. The cause is due to the way base64
16749 BIOs find the start of base64 encoded data. They do this by trying a
16750 trial decode on each line until they find one that works. When they
16751 do a flag is set and it starts again knowing it can pass all the
16752 data directly through the decoder. Unfortunately it doesn't reset
16753 the context it uses. This means that if EOF is reached an attempt
16754 is made to pass two EOFs through the context and this causes the
16755 resulting error. This can also cause other problems as well. As is
16756 usual with these problems it takes *ages* to find and the fix is
16757 trivial: move one line.
16759 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16761 * Ugly workaround to get s_client and s_server working under Windows. The
16762 old code wouldn't work because it needed to select() on sockets and the
16763 tty (for keypresses and to see if data could be written). Win32 only
16764 supports select() on sockets so we select() with a 1s timeout on the
16765 sockets and then see if any characters are waiting to be read, if none
16766 are present then we retry, we also assume we can always write data to
16767 the tty. This isn't nice because the code then blocks until we've
16768 received a complete line of data and it is effectively polling the
16769 keyboard at 1s intervals: however it's quite a bit better than not
16770 working at all :-) A dedicated Windows application might handle this
16771 with an event loop for example.
16775 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16776 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16777 will be called when RSA_sign() and RSA_verify() are used. This is useful
16778 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16779 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16780 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16781 This necessitated the support of an extra signature type NID_md5_sha1
16782 for SSL signatures and modifications to the SSL library to use it instead
16783 of calling RSA_public_decrypt() and RSA_private_encrypt().
16787 * Add new -verify -CAfile and -CApath options to the crl program, these
16788 will lookup a CRL issuers certificate and verify the signature in a
16789 similar way to the verify program. Tidy up the crl program so it
16790 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16791 less strict. It will now permit CRL extensions even if it is not
16792 a V2 CRL: this will allow it to tolerate some broken CRLs.
16796 * Initialize all non-automatic variables each time one of the openssl
16797 sub-programs is started (this is necessary as they may be started
16798 multiple times from the "OpenSSL>" prompt).
16800 *Lennart Bang, Bodo Moeller*
16802 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16803 removing all other RSA functionality (this is what NO_RSA does). This
16804 is so (for example) those in the US can disable those operations covered
16805 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16810 * Non-copying interface to BIO pairs.
16811 (still largely untested)
16815 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16816 ASCII string. This was handled independently in various places before.
16820 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16821 UTF8 strings a character at a time.
16825 * Use client_version from client hello to select the protocol
16826 (s23_srvr.c) and for RSA client key exchange verification
16827 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16831 * Add various utility functions to handle SPKACs, these were previously
16832 handled by poking round in the structure internals. Added new function
16833 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16834 print, verify and generate SPKACs. Based on an original idea from
16835 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16839 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16843 * Allow the config file extension section to be overwritten on the
16844 command line. Based on an original idea from Massimiliano Pala
16845 <madwolf@comune.modena.it>. The new option is called -extensions
16846 and can be applied to ca, req and x509. Also -reqexts to override
16847 the request extensions in req and -crlexts to override the crl extensions
16852 * Add new feature to the SPKAC handling in ca. Now you can include
16853 the same field multiple times by preceding it by "XXXX." for example:
16856 this is the same syntax as used in the req config file.
16860 * Allow certificate extensions to be added to certificate requests. These
16861 are specified in a 'req_extensions' option of the req section of the
16862 config file. They can be printed out with the -text option to req but
16863 are otherwise ignored at present.
16867 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16868 data read consists of only the final block it would not decrypted because
16869 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16870 A misplaced 'break' also meant the decrypted final block might not be
16871 copied until the next read.
16875 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16876 a few extra parameters to the DH structure: these will be useful if
16877 for example we want the value of 'q' or implement X9.42 DH.
16881 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16882 provides hooks that allow the default DSA functions or functions on a
16883 "per key" basis to be replaced. This allows hardware acceleration and
16884 hardware key storage to be handled without major modification to the
16885 library. Also added low-level modexp hooks and CRYPTO_EX structure and
16886 associated functions.
16890 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16891 as "read only": it can't be written to and the buffer it points to will
16892 not be freed. Reading from a read only BIO is much more efficient than
16893 a normal memory BIO. This was added because there are several times when
16894 an area of memory needs to be read from a BIO. The previous method was
16895 to create a memory BIO and write the data to it, this results in two
16896 copies of the data and an O(n^2) reading algorithm. There is a new
16897 function BIO_new_mem_buf() which creates a read only memory BIO from
16898 an area of memory. Also modified the PKCS#7 routines to use read only
16903 * Bugfix: ssl23_get_client_hello did not work properly when called in
16904 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16905 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16906 but a retry condition occurred while trying to read the rest.
16910 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16911 NID_pkcs7_encrypted by default: this was wrong since this should almost
16912 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16913 the encrypted data type: this is a more sensible place to put it and it
16914 allows the PKCS#12 code to be tidied up that duplicated this
16919 * Changed obj_dat.pl script so it takes its input and output files on
16920 the command line. This should avoid shell escape redirection problems
16925 * Initial support for certificate extension requests, these are included
16926 in things like Xenroll certificate requests. Included functions to allow
16927 extensions to be obtained and added.
16931 * -crlf option to s_client and s_server for sending newlines as
16932 CRLF (as required by many protocols).
16936 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16938 * Install libRSAglue.a when OpenSSL is built with RSAref.
16940 *Ralf S. Engelschall*
16942 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16944 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16946 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16951 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16952 DH parameters/keys (q is lost during that conversion, but the resulting
16953 DH parameters contain its length).
16955 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16956 much faster than DH_generate_parameters (which creates parameters
16957 where `p = 2*q + 1`), and also the smaller q makes DH computations
16958 much more efficient (160-bit exponentiation instead of 1024-bit
16959 exponentiation); so this provides a convenient way to support DHE
16960 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16961 utter importance to use
16962 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16964 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16965 when such DH parameters are used, because otherwise small subgroup
16966 attacks may become possible!
16970 * Avoid memory leak in i2d_DHparams.
16974 * Allow the -k option to be used more than once in the enc program:
16975 this allows the same encrypted message to be read by multiple recipients.
16979 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16980 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16981 it will always use the numerical form of the OID, even if it has a short
16986 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16987 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16988 otherwise bn_mod_exp was called. In the case of hardware keys for example
16989 no private key components need be present and it might store extra data
16990 in the RSA structure, which cannot be accessed from bn_mod_exp.
16991 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16992 private key operations.
16996 * Added support for SPARC Linux.
17000 * pem_password_cb function type incompatibly changed from
17001 typedef int pem_password_cb(char *buf, int size, int rwflag);
17003 ....(char *buf, int size, int rwflag, void *userdata);
17004 so that applications can pass data to their callbacks:
17005 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17006 additional void * argument, which is just handed through whenever
17007 the password callback is called.
17009 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17011 New function SSL_CTX_set_default_passwd_cb_userdata.
17013 Compatibility note: As many C implementations push function arguments
17014 onto the stack in reverse order, the new library version is likely to
17015 interoperate with programs that have been compiled with the old
17016 pem_password_cb definition (PEM_whatever takes some data that
17017 happens to be on the stack as its last argument, and the callback
17018 just ignores this garbage); but there is no guarantee whatsoever that
17021 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17022 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17023 problems not only on Windows, but also on some Unix platforms.
17024 To avoid problematic command lines, these definitions are now in an
17025 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17026 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17030 * MIPS III/IV assembler module is reimplemented.
17034 * More DES library cleanups: remove references to srand/rand and
17035 delete an unused file.
17039 * Add support for the free Netwide assembler (NASM) under Win32,
17040 since not many people have MASM (ml) and it can be hard to obtain.
17041 This is currently experimental but it seems to work OK and pass all
17042 the tests. Check out INSTALL.W32 for info.
17046 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17047 without temporary keys kept an extra copy of the server key,
17048 and connections with temporary keys did not free everything in case
17053 * New function RSA_check_key and new openssl rsa option -check
17054 for verifying the consistency of RSA keys.
17056 *Ulf Moeller, Bodo Moeller*
17058 * Various changes to make Win32 compile work:
17059 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17060 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17061 comparison" warnings.
17062 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17066 * Add a debugging option to PKCS#5 v2 key generation function: when
17067 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17068 derived keys are printed to stderr.
17072 * Copy the flags in ASN1_STRING_dup().
17074 *Roman E. Pavlov <pre@mo.msk.ru>*
17076 * The x509 application mishandled signing requests containing DSA
17077 keys when the signing key was also DSA and the parameters didn't match.
17079 It was supposed to omit the parameters when they matched the signing key:
17080 the verifying software was then supposed to automatically use the CA's
17081 parameters if they were absent from the end user certificate.
17083 Omitting parameters is no longer recommended. The test was also
17084 the wrong way round! This was probably due to unusual behaviour in
17085 EVP_cmp_parameters() which returns 1 if the parameters match.
17086 This meant that parameters were omitted when they *didn't* match and
17087 the certificate was useless. Certificates signed with 'ca' didn't have
17090 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17092 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17093 The interface is as follows:
17094 Applications can use
17095 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17096 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17097 "off" is now the default.
17098 The library internally uses
17099 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17100 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17101 to disable memory-checking temporarily.
17103 Some inconsistent states that previously were possible (and were
17104 even the default) are now avoided.
17106 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17107 with each memory chunk allocated; this is occasionally more helpful
17108 than just having a counter.
17110 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17112 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17117 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17118 which largely parallels "options", but is for changing API behaviour,
17119 whereas "options" are about protocol behaviour.
17120 Initial "mode" flags are:
17122 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17123 a single record has been written.
17124 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17125 retries use the same buffer location.
17126 (But all of the contents must be
17131 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17134 * Fix problems with no-hmac etc.
17136 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17138 * New functions RSA_get_default_method(), RSA_set_method() and
17139 RSA_get_method(). These allows replacement of RSA_METHODs without having
17140 to mess around with the internals of an RSA structure.
17144 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17145 Also really enable memory leak checks in openssl.c and in some
17148 *Chad C. Mulligan, Bodo Moeller*
17150 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17151 up the length of negative integers. This has now been simplified to just
17152 store the length when it is first determined and use it later, rather
17153 than trying to keep track of where data is copied and updating it to
17155 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17157 * Add a new function PKCS7_signatureVerify. This allows the verification
17158 of a PKCS#7 signature but with the signing certificate passed to the
17159 function itself. This contrasts with PKCS7_dataVerify which assumes the
17160 certificate is present in the PKCS#7 structure. This isn't always the
17161 case: certificates can be omitted from a PKCS#7 structure and be
17162 distributed by "out of band" means (such as a certificate database).
17166 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17167 function prototypes in pem.h, also change util/mkdef.pl to add the
17168 necessary function names.
17172 * mk1mf.pl (used by Windows builds) did not properly read the
17173 options set by Configure in the top level Makefile, and Configure
17174 was not even able to write more than one option correctly.
17175 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17179 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17180 file to be loaded from a BIO or FILE pointer. The BIO version will
17181 for example allow memory BIOs to contain config info.
17185 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17186 Whoever hopes to achieve shared-library compatibility across versions
17187 must use this, not the compile-time macro.
17188 (Exercise 0.9.4: Which is the minimum library version required by
17190 Note: All this applies only to multi-threaded programs, others don't
17195 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17196 through a BIO pair triggered the default case, i.e.
17197 SSLerr(...,SSL_R_UNKNOWN_STATE).
17201 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17202 can use the SSL library even if none of the specific BIOs is
17207 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17208 for the encoded length.
17210 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17212 * Add initial documentation of the X509V3 functions.
17216 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17217 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17218 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17219 secure PKCS#8 private key format with a high iteration count.
17223 * Fix determination of Perl interpreter: A perl or perl5
17224 *directory* in $PATH was also accepted as the interpreter.
17226 *Ralf S. Engelschall*
17228 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17229 wrong with it but it was very old and did things like calling
17230 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17231 unusual formatting.
17235 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17236 to use the new extension code.
17240 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17241 with macros. This should make it easier to change their form, add extra
17242 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17247 * Add to configuration table a new entry that can specify an alternative
17248 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17249 according to Mark Crispin <MRC@Panda.COM>.
17253 * DES CBC did not update the IV. Weird.
17257 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17258 Changing the behaviour of the former might break existing programs --
17259 where IV updating is needed, des_ncbc_encrypt can be used.
17262 * When bntest is run from "make test" it drives bc to check its
17263 calculations, as well as internally checking them. If an internal check
17264 fails, it needs to cause bc to give a non-zero result or make test carries
17265 on without noticing the failure. Fixed.
17269 * DES library cleanups.
17273 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17274 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17275 ciphers. NOTE: although the key derivation function has been verified
17276 against some published test vectors it has not been extensively tested
17277 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17282 * Instead of "mkdir -p", which is not fully portable, use new
17283 Perl script "util/mkdir-p.pl".
17287 * Rewrite the way password based encryption (PBE) is handled. It used to
17288 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17289 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17290 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17291 the 'parameter' field of the AlgorithmIdentifier is passed to the
17292 underlying key generation function so it must do its own ASN1 parsing.
17293 This has also changed the EVP_PBE_CipherInit() function which now has a
17294 'parameter' argument instead of literal salt and iteration count values
17295 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17299 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17300 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17301 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17302 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17303 value was just used as a "magic string" and not used directly its
17304 value doesn't matter.
17308 * Introduce some semblance of const correctness to BN. Shame C doesn't
17313 * "linux-sparc64" configuration (ultrapenguin).
17315 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17316 "linux-sparc" configuration.
17318 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17320 * config now generates no-xxx options for missing ciphers.
17324 * Support the EBCDIC character set (work in progress).
17325 File ebcdic.c not yet included because it has a different license.
17327 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17329 * Support BS2000/OSD-POSIX.
17331 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17333 * Make callbacks for key generation use `void *` instead of `char *`.
17337 * Make S/MIME samples compile (not yet tested).
17341 * Additional typesafe stacks.
17345 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17349 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17351 * New configuration variant "sco5-gcc".
17353 * Updated some demos.
17355 *Sean O Riordain, Wade Scholine*
17357 * Add missing BIO_free at exit of pkcs12 application.
17361 * Fix memory leak in conf.c.
17365 * Updates for Win32 to assembler version of MD5.
17369 * Set #! path to perl in `apps/der_chop` to where we found it
17370 instead of using a fixed path.
17374 * SHA library changes for irix64-mips4-cc.
17378 * Improvements for VMS support.
17382 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17384 * Bignum library bug fix. IRIX 6 passes "make test" now!
17385 This also avoids the problems with SC4.2 and unpatched SC5.
17387 *Andy Polyakov <appro@fy.chalmers.se>*
17389 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17390 These are required because of the typesafe stack would otherwise break
17391 existing code. If old code used a structure member which used to be STACK
17392 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17393 sk_num or sk_value it would produce an error because the num, data members
17394 are not present in STACK_OF. Now it just produces a warning. sk_set
17395 replaces the old method of assigning a value to sk_value
17396 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17397 that does this will no longer work (and should use sk_set instead) but
17398 this could be regarded as a "questionable" behaviour anyway.
17402 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17403 correctly handle encrypted S/MIME data.
17407 * Change type of various DES function arguments from des_cblock
17408 (which means, in function argument declarations, pointer to char)
17409 to des_cblock * (meaning pointer to array with 8 char elements),
17410 which allows the compiler to do more typechecking; it was like
17411 that back in SSLeay, but with lots of ugly casts.
17413 Introduce new type const_des_cblock.
17417 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17418 problems: find RecipientInfo structure that matches recipient certificate
17419 and initialise the ASN1 structures properly based on passed cipher.
17423 * Belatedly make the BN tests actually check the results.
17427 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17428 to and from BNs: it was completely broken. New compilation option
17429 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17430 key elements as negative integers.
17434 * Reorganize and speed up MD5.
17436 *Andy Polyakov <appro@fy.chalmers.se>*
17440 *Richard Levitte <richard@levitte.org>*
17442 * New option -out to asn1parse to allow the parsed structure to be
17443 output to a file. This is most useful when combined with the -strparse
17444 option to examine the output of things like OCTET STRINGS.
17448 * Make SSL library a little more fool-proof by not requiring any longer
17449 that `SSL_set_{accept,connect}_state` be called before
17450 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17451 in many applications because usually everything *appeared* to work as
17452 intended anyway -- now it really works as intended).
17456 * Move openssl.cnf out of lib/.
17460 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17461 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17462 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17464 *Ralf S. Engelschall*
17466 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17467 handle PKCS#7 enveloped data properly.
17469 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17471 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17472 copying pointers. The cert_st handling is changed by this in
17473 various ways (and thus what used to be known as ctx->default_cert
17474 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17475 any longer when s->cert does not give us what we need).
17476 ssl_cert_instantiate becomes obsolete by this change.
17477 As soon as we've got the new code right (possibly it already is?),
17478 we have solved a couple of bugs of the earlier code where s->cert
17479 was used as if it could not have been shared with other SSL structures.
17481 Note that using the SSL API in certain dirty ways now will result
17482 in different behaviour than observed with earlier library versions:
17483 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17484 does not influence s as it used to.
17486 In order to clean up things more thoroughly, inside SSL_SESSION
17487 we don't use CERT any longer, but a new structure SESS_CERT
17488 that holds per-session data (if available); currently, this is
17489 the peer's certificate chain and, for clients, the server's certificate
17490 and temporary key. CERT holds only those values that can have
17491 meaningful defaults in an SSL_CTX.
17495 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17496 from the internal representation. Various PKCS#7 fixes: remove some
17497 evil casts and set the enc_dig_alg field properly based on the signing
17502 * Allow PKCS#12 password to be set from the command line or the
17503 environment. Let 'ca' get its config file name from the environment
17504 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17509 * Allow certificate policies extension to use an IA5STRING for the
17510 organization field. This is contrary to the PKIX definition but
17511 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17516 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17517 without disallowing inline assembler and the like for non-pedantic builds.
17521 * Support Borland C++ builder.
17523 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17529 * SHA-1 cleanups and performance enhancements.
17531 *Andy Polyakov <appro@fy.chalmers.se>*
17533 * Sparc v8plus assembler for the bignum library.
17535 *Andy Polyakov <appro@fy.chalmers.se>*
17537 * Accept any -xxx and +xxx compiler options in Configure.
17541 * Update HPUX configuration.
17545 * Add missing `sk_<type>_unshift()` function to safestack.h
17547 *Ralf S. Engelschall*
17549 * New function SSL_CTX_use_certificate_chain_file that sets the
17550 "extra_cert"s in addition to the certificate. (This makes sense
17551 only for "PEM" format files, as chains as a whole are not
17556 * Support verify_depth from the SSL API.
17557 x509_vfy.c had what can be considered an off-by-one-error:
17558 Its depth (which was not part of the external interface)
17559 was actually counting the number of certificates in a chain;
17560 now it really counts the depth.
17564 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17565 instead of X509err, which often resulted in confusing error
17566 messages since the error codes are not globally unique
17567 (e.g. an alleged error in ssl3_accept when a certificate
17568 didn't match the private key).
17570 * New function SSL_CTX_set_session_id_context that allows to set a default
17571 value (so that you don't need SSL_set_session_id_context for each
17572 connection using the SSL_CTX).
17576 * OAEP decoding bug fix.
17580 * Support INSTALL_PREFIX for package builders, as proposed by
17585 * New Configure options "threads" and "no-threads". For systems
17586 where the proper compiler options are known (currently Solaris
17587 and Linux), "threads" is the default.
17591 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17595 * Install various scripts to $(OPENSSLDIR)/misc, not to
17596 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17597 such as /usr/local/bin.
17601 * "make linux-shared" to build shared libraries.
17603 *Niels Poppe <niels@netbox.org>*
17605 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17609 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17610 extension adding in x509 utility.
17614 * Remove NOPROTO sections and error code comments.
17618 * Partial rewrite of the DEF file generator to now parse the ANSI
17623 * New Configure options --prefix=DIR and --openssldir=DIR.
17627 * Complete rewrite of the error code script(s). It is all now handled
17628 by one script at the top level which handles error code gathering,
17629 header rewriting and C source file generation. It should be much better
17630 than the old method: it now uses a modified version of Ulf's parser to
17631 read the ANSI prototypes in all header files (thus the old K&R definitions
17632 aren't needed for error creation any more) and do a better job of
17633 translating function codes into names. The old 'ASN1 error code embedded
17634 in a comment' is no longer necessary and it doesn't use .err files which
17635 have now been deleted. Also the error code call doesn't have to appear all
17636 on one line (which resulted in some large lines...).
17640 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17644 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17645 0 (which usually indicates a closed connection), but continue reading.
17649 * Fix some race conditions.
17653 * Add support for CRL distribution points extension. Add Certificate
17654 Policies and CRL distribution points documentation.
17658 * Move the autogenerated header file parts to crypto/opensslconf.h.
17662 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17663 8 of keying material. Merlin has also confirmed interop with this fix
17664 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17666 *Merlin Hughes <merlin@baltimore.ie>*
17668 * Fix lots of warnings.
17670 *Richard Levitte <levitte@stacken.kth.se>*
17672 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17673 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17675 *Richard Levitte <levitte@stacken.kth.se>*
17677 * Fix problems with sizeof(long) == 8.
17679 *Andy Polyakov <appro@fy.chalmers.se>*
17681 * Change functions to ANSI C.
17685 * Fix typos in error codes.
17687 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17689 * Remove defunct assembler files from Configure.
17693 * SPARC v8 assembler BIGNUM implementation.
17695 *Andy Polyakov <appro@fy.chalmers.se>*
17697 * Support for Certificate Policies extension: both print and set.
17698 Various additions to support the r2i method this uses.
17702 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17703 return a const string when you are expecting an allocated buffer.
17707 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17708 types DirectoryString and DisplayText.
17712 * Add code to allow r2i extensions to access the configuration database,
17713 add an LHASH database driver and add several ctx helper functions.
17717 * Fix an evil bug in bn_expand2() which caused various BN functions to
17718 fail when they extended the size of a BIGNUM.
17722 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17723 support typesafe stack.
17727 * Fix typo in SSL_[gs]et_options().
17729 *Nils Frostberg <nils@medcom.se>*
17731 * Delete various functions and files that belonged to the (now obsolete)
17732 old X509V3 handling code.
17736 * New Configure option "rsaref".
17740 * Don't auto-generate pem.h.
17744 * Introduce type-safe ASN.1 SETs.
17748 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17750 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17752 * Introduce type-safe STACKs. This will almost certainly break lots of code
17753 that links with OpenSSL (well at least cause lots of warnings), but fear
17754 not: the conversion is trivial, and it eliminates loads of evil casts. A
17755 few STACKed things have been converted already. Feel free to convert more.
17756 In the fullness of time, I'll do away with the STACK type altogether.
17760 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17761 specified in `<certfile>` by updating the entry in the index.txt file.
17762 This way one no longer has to edit the index.txt file manually for
17763 revoking a certificate. The -revoke option does the gory details now.
17765 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17767 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17768 `-text` option at all and this way the `-noout -text` combination was
17769 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17771 *Ralf S. Engelschall*
17773 * Make sure a corresponding plain text error message exists for the
17774 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17775 verify callback function determined that a certificate was revoked.
17777 *Ralf S. Engelschall*
17779 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17780 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17781 all available ciphers including rc5, which was forgotten until now.
17782 In order to let the testing shell script know which algorithms
17783 are available, a new (up to now undocumented) command
17784 `openssl list-cipher-commands` is used.
17788 * Bugfix: s_client occasionally would sleep in select() when
17789 it should have checked SSL_pending() first.
17793 * New functions DSA_do_sign and DSA_do_verify to provide access to
17794 the raw DSA values prior to ASN.1 encoding.
17798 * Tweaks to Configure
17800 *Niels Poppe <niels@netbox.org>*
17802 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17807 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17811 * New config option to avoid instructions that are illegal on the 80386.
17812 The default code is faster, but requires at least a 486.
17816 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17817 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17818 same as SSL2_VERSION anyway.
17822 * New "-showcerts" option for s_client.
17826 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17827 application. Various cleanups and fixes.
17831 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17832 modify error routines to work internally. Add error codes and PBE init
17833 to library startup routines.
17837 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17838 packing functions to asn1 and evp. Changed function names and error
17839 codes along the way.
17843 * PKCS12 integration: and so it begins... First of several patches to
17844 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17845 objects to objects.h
17849 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17850 and display support for Thawte strong extranet extension.
17854 * Add LinuxPPC support.
17856 *Jeff Dubrule <igor@pobox.org>*
17858 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17859 bn_div_words in alpha.s.
17861 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17863 * Make sure the RSA OAEP test is skipped under -DRSAref because
17864 OAEP isn't supported when OpenSSL is built with RSAref.
17866 *Ulf Moeller <ulf@fitug.de>*
17868 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17869 so they no longer are missing under -DNOPROTO.
17871 *Soren S. Jorvang <soren@t.dk>*
17873 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17875 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17876 doesn't work when the session is reused. Coming soon!
17880 * Fix a security hole, that allows sessions to be reused in the wrong
17881 context thus bypassing client cert protection! All software that uses
17882 client certs and session caches in multiple contexts NEEDS PATCHING to
17883 allow session reuse! A fuller solution is in the works.
17885 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17887 * Some more source tree cleanups (removed obsolete files
17888 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17889 permission on "config" script to be executable) and a fix for the INSTALL
17892 *Ulf Moeller <ulf@fitug.de>*
17894 * Remove some legacy and erroneous uses of malloc, free instead of
17897 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17899 * Make rsa_oaep_test return non-zero on error.
17901 *Ulf Moeller <ulf@fitug.de>*
17903 * Add support for native Solaris shared libraries. Configure
17904 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17905 if someone would make that last step automatic.
17907 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17909 * ctx_size was not built with the right compiler during "make links". Fixed.
17913 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17914 except NULL ciphers". This means the default cipher list will no longer
17915 enable NULL ciphers. They need to be specifically enabled e.g. with
17916 the string "DEFAULT:eNULL".
17920 * Fix to RSA private encryption routines: if p < q then it would
17921 occasionally produce an invalid result. This will only happen with
17922 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17926 * Be less restrictive and allow also `perl util/perlpath.pl
17927 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17928 because this way one can also use an interpreter named `perl5` (which is
17929 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17930 installed as `perl`).
17932 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17934 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17936 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17938 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17939 advapi32.lib to Win32 build and change the pem test comparison
17940 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17941 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17942 and crypto/des/ede_cbcm_enc.c.
17946 * DES quad checksum was broken on big-endian architectures. Fixed.
17950 * Comment out two functions in bio.h that aren't implemented. Fix up the
17951 Win32 test batch file so it (might) work again. The Win32 test batch file
17952 is horrible: I feel ill....
17956 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17957 in e_os.h. Audit of header files to check ANSI and non ANSI
17958 sections: 10 functions were absent from non ANSI section and not exported
17959 from Windows DLLs. Fixed up libeay.num for new functions.
17963 * Make `openssl version` output lines consistent.
17965 *Ralf S. Engelschall*
17967 * Fix Win32 symbol export lists for BIO functions: Added
17968 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17969 to ms/libeay{16,32}.def.
17971 *Ralf S. Engelschall*
17973 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17974 fine under Unix and passes some trivial tests I've now added. But the
17975 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17976 added to make sure no one expects that this stuff really works in the
17977 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17978 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17981 *Ralf S. Engelschall*
17983 * Fix the generation of two part addresses in perl.
17985 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17987 * Add config entry for Linux on MIPS.
17989 *John Tobey <jtobey@channel1.com>*
17991 * Make links whenever Configure is run, unless we are on Windoze.
17995 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17996 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18001 * Add a useful kludge to allow package maintainers to specify compiler and
18002 other platforms details on the command line without having to patch the
18003 Configure script every time: One now can use
18004 `perl Configure <id>:<details>`,
18005 i.e. platform ids are allowed to have details appended
18006 to them (separated by colons). This is treated as there would be a static
18007 pre-configured entry in Configure's %table under key `<id>` with value
18008 `<details>` and `perl Configure <id>` is called. So, when you want to
18009 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18010 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18011 now, which overrides the FreeBSD-elf entry on-the-fly.
18013 *Ralf S. Engelschall*
18015 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18019 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18020 on the `perl Configure ...` command line. This way one can compile
18021 OpenSSL libraries with Position Independent Code (PIC) which is needed
18022 for linking it into DSOs.
18024 *Ralf S. Engelschall*
18026 * Remarkably, export ciphers were totally broken and no-one had noticed!
18031 * Cleaned up the LICENSE document: The official contact for any license
18032 questions now is the OpenSSL core team under openssl-core@openssl.org.
18033 And add a paragraph about the dual-license situation to make sure people
18034 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18035 to the OpenSSL toolkit.
18037 *Ralf S. Engelschall*
18039 * General source tree makefile cleanups: Made `making xxx in yyy...`
18040 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18041 Additionally cleaned up the `make links` target: Remove unnecessary
18042 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18043 to speed processing and no longer clutter the display with confusing
18044 stuff. Instead only the actually done links are displayed.
18046 *Ralf S. Engelschall*
18048 * Permit null encryption ciphersuites, used for authentication only. It used
18049 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18050 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18055 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18056 signed attributes when verifying signatures (this would break them),
18057 the detached data encoding was wrong and public keys obtained using
18058 X509_get_pubkey() weren't freed.
18062 * Add text documentation for the BUFFER functions. Also added a work around
18063 to a Win95 console bug. This was triggered by the password read stuff: the
18064 last character typed gets carried over to the next fread(). If you were
18065 generating a new cert request using 'req' for example then the last
18066 character of the passphrase would be CR which would then enter the first
18071 * Added the new 'Includes OpenSSL Cryptography Software' button as
18072 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18073 button and can be used by applications based on OpenSSL to show the
18074 relationship to the OpenSSL project.
18076 *Ralf S. Engelschall*
18078 * Remove confusing variables in function signatures in files
18079 ssl/ssl_lib.c and ssl/ssl.h.
18081 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18083 * Don't install bss_file.c under PREFIX/include/
18085 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18087 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18088 functions that return function pointers and has support for NT specific
18089 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18090 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18091 unsigned to signed types: this was killing the Win32 compile.
18095 * Add new certificate file to stack functions,
18096 SSL_add_dir_cert_subjects_to_stack() and
18097 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18098 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18099 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18100 This means that Apache-SSL and similar packages don't have to mess around
18101 to add as many CAs as they want to the preferred list.
18105 * Experiment with doxygen documentation. Currently only partially applied to
18107 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18108 openssl.doxy as the configuration file.
18112 * Get rid of remaining C++-style comments which strict C compilers hate.
18114 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18116 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18117 compiled in by default: it has problems with large keys.
18121 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18122 DH private keys and/or callback functions which directly correspond to
18123 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18124 is needed for applications which have to configure certificates on a
18125 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18127 For the RSA certificate situation is makes no difference, but
18128 for the DSA certificate situation this fixes the "no shared cipher"
18129 problem where the OpenSSL cipher selection procedure failed because the
18130 temporary keys were not overtaken from the context and the API provided
18131 no way to reconfigure them.
18132 The new functions now let applications reconfigure the stuff and they
18133 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18134 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18135 non-public-API function ssl_cert_instantiate() is used as a helper
18136 function and also to reduce code redundancy inside ssl_rsa.c.
18138 *Ralf S. Engelschall*
18140 * Move s_server -dcert and -dkey options out of the undocumented feature
18141 area because they are useful for the DSA situation and should be
18142 recognized by the users.
18144 *Ralf S. Engelschall*
18146 * Fix the cipher decision scheme for export ciphers: the export bits are
18147 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18148 SSL_EXP_MASK. So, the original variable has to be used instead of the
18149 already masked variable.
18151 *Richard Levitte <levitte@stacken.kth.se>*
18153 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18155 *Richard Levitte <levitte@stacken.kth.se>*
18157 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18158 from `int` to `unsigned int` because it is a length and initialized by
18159 EVP_DigestFinal() which expects an `unsigned int *`.
18161 *Richard Levitte <levitte@stacken.kth.se>*
18163 * Don't hard-code path to Perl interpreter on shebang line of Configure
18164 script. Instead use the usual Shell->Perl transition trick.
18166 *Ralf S. Engelschall*
18168 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18169 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18170 -noout -modulus` as it's already the case for `openssl rsa -noout
18171 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18172 currently the public key is printed (a decision which was already done by
18173 `openssl dsa -modulus` in the past) which serves a similar purpose.
18174 Additionally the NO_RSA no longer completely removes the whole -modulus
18175 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18178 *Ralf S. Engelschall*
18180 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18181 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18183 *Arne Ansper <arne@ats.cyber.ee>*
18185 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18186 to be added. Now both 'req' and 'ca' can use new objects defined in the
18191 * Add cool BIO that does syslog (or event log on NT).
18193 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18195 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18196 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18197 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18198 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18202 * Add preliminary config info for new extension code.
18206 * Make RSA_NO_PADDING really use no padding.
18208 *Ulf Moeller <ulf@fitug.de>*
18210 * Generate errors when private/public key check is done.
18214 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18215 for some CRL extensions and new objects added.
18219 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18220 key usage extension and fuller support for authority key id.
18224 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18225 padding method for RSA, which is recommended for new applications in PKCS
18226 #1 v2.0 (RFC 2437, October 1998).
18227 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18228 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18229 against Bleichbacher's attack on RSA.
18230 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18233 * Updates to the new SSL compression code
18235 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18237 * Fix so that the version number in the master secret, when passed
18238 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18239 (because the server will not accept higher), that the version number
18240 is 0x03,0x01, not 0x03,0x00
18242 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18244 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18245 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18246 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18250 * Support for RAW extensions where an arbitrary extension can be
18251 created by including its DER encoding. See `apps/openssl.cnf` for
18256 * Make sure latest Perl versions don't interpret some generated C array
18257 code as Perl array code in the crypto/err/err_genc.pl script.
18259 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18261 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18262 not many people have the assembler. Various Win32 compilation fixes and
18263 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18264 build instructions.
18268 * Modify configure script 'Configure' to automatically create crypto/date.h
18269 file under Win32 and also build pem.h from pem.org. New script
18270 util/mkfiles.pl to create the MINFO file on environments that can't do a
18271 'make files': perl util/mkfiles.pl >MINFO should work.
18275 * Major rework of DES function declarations, in the pursuit of correctness
18276 and purity. As a result, many evil casts evaporated, and some weirdness,
18277 too. You may find this causes warnings in your code. Zapping your evil
18278 casts will probably fix them. Mostly.
18282 * Fix for a typo in asn1.h. Bug fix to object creation script
18283 obj_dat.pl. It considered a zero in an object definition to mean
18284 "end of object": none of the objects in objects.h have any zeros
18285 so it wasn't spotted.
18287 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18289 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18290 Masking (CBCM). In the absence of test vectors, the best I have been able
18291 to do is check that the decrypt undoes the encrypt, so far. Send me test
18292 vectors if you have them.
18296 * Correct calculation of key length for export ciphers (too much space was
18297 allocated for null ciphers). This has not been tested!
18301 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18302 message is now correct (it understands "crypto" and "ssl" on its
18303 command line). There is also now an "update" option. This will update
18304 the util/ssleay.num and util/libeay.num files with any new functions.
18306 perl util/mkdef.pl crypto ssl update
18307 it will update them.
18311 * Overhauled the Perl interface:
18312 - ported BN stuff to OpenSSL's different BN library
18313 - made the perl/ source tree CVS-aware
18314 - renamed the package from SSLeay to OpenSSL (the files still contain
18315 their history because I've copied them in the repository)
18316 - removed obsolete files (the test scripts will be replaced
18317 by better Test::Harness variants in the future)
18319 *Ralf S. Engelschall*
18321 * First cut for a very conservative source tree cleanup:
18322 1. merge various obsolete readme texts into doc/ssleay.txt
18323 where we collect the old documents and readme texts.
18324 2. remove the first part of files where I'm already sure that we no
18325 longer need them because of three reasons: either they are just temporary
18326 files which were left by Eric or they are preserved original files where
18327 I've verified that the diff is also available in the CVS via "cvs diff
18328 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18329 the crypto/md/ stuff).
18331 *Ralf S. Engelschall*
18333 * More extension code. Incomplete support for subject and issuer alt
18334 name, issuer and authority key id. Change the i2v function parameters
18335 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18336 what that's for :-) Fix to ASN1 macro which messed up
18337 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18341 * Preliminary support for ENUMERATED type. This is largely copied from the
18346 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18348 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18350 * Make sure `make rehash` target really finds the `openssl` program.
18352 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18354 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18355 like to hear about it if this slows down other processors.
18359 * Add CygWin32 platform information to Configure script.
18361 *Alan Batie <batie@aahz.jf.intel.com>*
18363 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18365 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18367 * New program nseq to manipulate netscape certificate sequences
18371 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18376 * Fixes to BN code. Previously the default was to define BN_RECURSION
18377 but the BN code had some problems that would cause failures when
18378 doing certificate verification and some other functions.
18380 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18382 * Add ASN1 and PEM code to support netscape certificate sequences.
18386 * Add ASN1 and PEM code to support netscape certificate sequences.
18390 * Add several PKIX and private extended key usage OIDs.
18394 * Modify the 'ca' program to handle the new extension code. Modify
18395 openssl.cnf for new extension format, add comments.
18399 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18400 and add a sample to openssl.cnf so req -x509 now adds appropriate
18405 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18406 error code, add initial support to X509_print() and x509 application.
18410 * Takes a deep breath and start adding X509 V3 extension support code. Add
18411 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18412 stuff is currently isolated and isn't even compiled yet.
18416 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18417 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18418 Removed the versions check from X509 routines when loading extensions:
18419 this allows certain broken certificates that don't set the version
18420 properly to be processed.
18424 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18425 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18426 can still be regenerated with "make depend".
18430 * Spelling mistake in C version of CAST-128.
18432 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18434 * Changes to the error generation code. The perl script err-code.pl
18435 now reads in the old error codes and retains the old numbers, only
18436 adding new ones if necessary. It also only changes the .err files if new
18437 codes are added. The makefiles have been modified to only insert errors
18438 when needed (to avoid needlessly modifying header files). This is done
18439 by only inserting errors if the .err file is newer than the auto generated
18440 C file. To rebuild all the error codes from scratch (the old behaviour)
18441 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18442 or delete all the .err files.
18446 * CAST-128 was incorrectly implemented for short keys. The C version has
18447 been fixed, but is untested. The assembler versions are also fixed, but
18448 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18449 to regenerate it if needed.
18450 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18451 Hagino <itojun@kame.net>*
18453 * File was opened incorrectly in randfile.c.
18455 *Ulf Möller <ulf@fitug.de>*
18457 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18458 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18459 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18460 al: it's just almost always a UTCTime. Note this patch adds new error
18461 codes so do a "make errors" if there are problems.
18465 * Correct Linux 1 recognition in config.
18467 *Ulf Möller <ulf@fitug.de>*
18469 * Remove pointless MD5 hash when using DSA keys in ca.
18471 *Anonymous <nobody@replay.com>*
18473 * Generate an error if given an empty string as a cert directory. Also
18474 generate an error if handed NULL (previously returned 0 to indicate an
18475 error, but didn't set one).
18477 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18479 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18483 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18484 parameters. This was causing a warning which killed off the Win32 compile.
18488 * Remove C++ style comments from crypto/bn/bn_local.h.
18490 *Neil Costigan <neil.costigan@celocom.com>*
18492 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18493 based on a text string, looking up short and long names and finally
18494 "dot" format. The "dot" format stuff didn't work. Added new function
18495 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18496 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18497 OID is not part of the table.
18501 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18502 X509_LOOKUP_by_alias().
18506 * Sort openssl functions by name.
18510 * Get the `gendsa` command working and add it to the `list` command. Remove
18511 encryption from sample DSA keys (in case anyone is interested the password
18516 * Make *all* `*_free` functions accept a NULL pointer.
18518 *Frans Heymans <fheymans@isaserver.be>*
18520 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18523 *Anonymous <nobody@replay.com>*
18525 * s_server should send the CAfile as acceptable CAs, not its own cert.
18527 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18529 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18531 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18533 * Temp key "for export" tests were wrong in s3_srvr.c.
18535 *Anonymous <nobody@replay.com>*
18537 * Add prototype for temp key callback functions
18538 SSL_CTX_set_tmp_{rsa,dh}_callback().
18542 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18543 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18547 * X509_name_add_entry() freed the wrong thing after an error.
18549 *Arne Ansper <arne@ats.cyber.ee>*
18551 * rsa_eay.c would attempt to free a NULL context.
18553 *Arne Ansper <arne@ats.cyber.ee>*
18555 * BIO_s_socket() had a broken should_retry() on Windoze.
18557 *Arne Ansper <arne@ats.cyber.ee>*
18559 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18561 *Arne Ansper <arne@ats.cyber.ee>*
18563 * Make sure the already existing X509_STORE->depth variable is initialized
18564 in X509_STORE_new(), but document the fact that this variable is still
18565 unused in the certificate verification process.
18567 *Ralf S. Engelschall*
18569 * Fix the various library and `apps/` files to free up pkeys obtained from
18570 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18574 * Fix reference counting in X509_PUBKEY_get(). This makes
18575 demos/maurice/example2.c work, amongst others, probably.
18577 *Steve Henson and Ben Laurie*
18579 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18580 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18581 are no longer created. This way we have a single and consistent command
18582 line interface `openssl <command>`, similar to `cvs <command>`.
18584 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18586 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18587 BIT STRING wrapper always have zero unused bits.
18591 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18595 * Make the top-level INSTALL documentation easier to understand.
18599 * Makefiles updated to exit if an error occurs in a sub-directory
18600 make (including if user presses ^C) [Paul Sutton]
18602 * Make Montgomery context stuff explicit in RSA data structure.
18606 * Fix build order of pem and err to allow for generated pem.h.
18610 * Fix renumbering bug in X509_NAME_delete_entry().
18614 * Enhanced the err-ins.pl script so it makes the error library number
18615 global and can add a library name. This is needed for external ASN1 and
18616 other error libraries.
18620 * Fixed sk_insert which never worked properly.
18624 * Fix ASN1 macros so they can handle indefinite length constructed
18625 EXPLICIT tags. Some non standard certificates use these: they can now
18630 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18631 into a single doc/ssleay.txt bundle. This way the information is still
18632 preserved but no longer messes up this directory. Now it's new room for
18633 the new set of documentation files.
18635 *Ralf S. Engelschall*
18637 * SETs were incorrectly DER encoded. This was a major pain, because they
18638 shared code with SEQUENCEs, which aren't coded the same. This means that
18639 almost everything to do with SETs or SEQUENCEs has either changed name or
18640 number of arguments.
18642 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18644 * Fix test data to work with the above.
18648 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18649 was already fixed by Eric for 0.9.1 it seems.
18651 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18653 * Autodetect FreeBSD3.
18657 * Fix various bugs in Configure. This affects the following platforms:
18661 unixware-2.0-pentium
18666 * Eliminate generated files from CVS. Reorder tests to regenerate files
18667 before they are needed.
18671 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18675 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18677 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18678 changed SSLeay to OpenSSL in version strings.
18680 *Ralf S. Engelschall*
18682 * Some fixups to the top-level documents.
18686 * Fixed the nasty bug where rsaref.h was not found under compile-time
18687 because the symlink to include/ was missing.
18689 *Ralf S. Engelschall*
18691 * Incorporated the popular no-RSA/DSA-only patches
18692 which allow to compile a RSA-free SSLeay.
18694 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18696 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18697 when "ssleay" is still not found.
18699 *Ralf S. Engelschall*
18701 * Added more platforms to Configure: Cray T3E, HPUX 11,
18703 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18705 * Updated the README file.
18707 *Ralf S. Engelschall*
18709 * Added various .cvsignore files in the CVS repository subdirs
18710 to make a "cvs update" really silent.
18712 *Ralf S. Engelschall*
18714 * Recompiled the error-definition header files and added
18715 missing symbols to the Win32 linker tables.
18717 *Ralf S. Engelschall*
18719 * Cleaned up the top-level documents;
18720 o new files: CHANGES and LICENSE
18721 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18722 o merged COPYRIGHT into LICENSE
18723 o removed obsolete TODO file
18724 o renamed MICROSOFT to INSTALL.W32
18726 *Ralf S. Engelschall*
18728 * Removed dummy files from the 0.9.1b source tree:
18729 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18730 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18731 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18732 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18733 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18735 *Ralf S. Engelschall*
18737 * Added various platform portability fixes.
18741 * The Genesis of the OpenSSL rpject:
18742 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18743 Young and Tim J. Hudson created while they were working for C2Net until
18746 *The OpenSSL Project*
18748 ### Changes between 0.9.0b and 0.9.1b [not released]
18750 * Updated a few CA certificates under certs/
18754 * Changed some BIGNUM api stuff.
18758 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18759 DGUX x86, Linux Alpha, etc.
18763 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18764 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18769 * Add -strparse option to asn1pars program which parses nested
18772 *Dr Stephen Henson <shenson@bigfoot.com>*
18774 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18778 * DSA fix for "ca" program.
18782 * Added "-genkey" option to "dsaparam" program.
18786 * Added RIPE MD160 (rmd160) message digest.
18790 * Added -a (all) option to "ssleay version" command.
18794 * Added PLATFORM define which is the id given to Configure.
18798 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18802 * Extended the ASN.1 parser routines.
18806 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18810 * Added a BN_CTX to the BN library.
18814 * Fixed the weak key values in DES library
18818 * Changed API in EVP library for cipher aliases.
18822 * Added support for RC2/64bit cipher.
18826 * Converted the lhash library to the crypto/mem.c functions.
18830 * Added more recognized ASN.1 object ids.
18834 * Added more RSA padding checks for SSL/TLS.
18838 * Added BIO proxy/filter functionality.
18842 * Added extra_certs to SSL_CTX which can be used
18843 send extra CA certificates to the client in the CA cert chain sending
18844 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18848 * Now Fortezza is denied in the authentication phase because
18849 this is key exchange mechanism is not supported by SSLeay at all.
18853 * Additional PKCS1 checks.
18857 * Support the string "TLSv1" for all TLS v1 ciphers.
18861 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18862 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18866 * Fixed a few memory leaks.
18870 * Fixed various code and comment typos.
18874 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18875 bytes sent in the client random.
18877 *Edward Bishop <ebishop@spyglass.com>*
18881 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18882 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18883 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18884 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18885 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18886 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18887 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18888 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18889 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18890 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18891 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18892 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18893 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18894 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18895 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18896 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18897 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18898 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18899 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18900 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18901 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18902 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18903 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18904 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18905 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18906 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18907 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18908 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18909 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18910 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18911 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18912 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18913 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18914 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18915 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18916 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18917 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18918 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18919 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18920 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18921 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18922 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18923 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18924 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18925 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18926 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18927 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18928 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18929 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18930 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18931 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18932 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18933 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18934 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18935 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18936 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18937 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18938 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18939 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18940 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18941 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18942 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18943 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18944 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18945 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18946 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18947 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18948 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18949 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18950 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18951 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18952 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18953 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18954 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18955 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18956 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18957 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18958 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18959 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18960 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18961 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18962 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18963 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18964 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18965 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18966 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18967 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18968 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18969 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18970 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18971 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18972 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18973 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18974 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18975 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18976 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18977 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18978 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18979 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18980 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18981 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18982 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18983 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18984 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18985 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18986 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18987 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18988 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18989 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18990 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18991 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18992 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18993 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18994 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18995 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18996 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18997 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18998 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18999 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19000 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19001 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19002 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19003 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19004 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19005 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19006 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19007 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19008 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19009 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19010 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19011 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19012 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19013 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19014 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19015 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19016 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19017 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19018 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19019 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19020 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19021 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19022 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19023 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19024 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19025 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19026 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19027 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19028 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19029 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19030 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19031 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19032 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19033 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19034 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19035 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19036 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19037 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19038 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19039 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19040 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19041 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19042 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655