4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.1](#openssl-31)
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 3.0 and 3.1 [xx XXX xxxx]
27 * s_client and s_server apps now explicitly say when the TLS version
28 does not include the renegotiation mechanism. This avoids confusion
29 between that scenario versus when the TLS version includes secure
30 renegotiation but the peer lacks support for it.
34 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
35 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
36 of 160 bits and above and less than 224 bits were previously accepted by
37 default but are now no longer allowed. By default TLS compression was
38 already disabled in previous OpenSSL versions. At security level 2 it cannot
43 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
48 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
49 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
50 will need to load the legacy crypto provider.
54 * The various OBJ_* functions have been made thread safe.
58 * CCM8 cipher suites in TLS have been downgraded to security level zero
59 because they use a short authentication tag which lowers their strength.
63 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
71 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
72 listed here are only a brief description.
73 The migration guide contains more detailed information related to new features,
74 breaking changes, and mappings for the large list of deprecated functions.
76 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
78 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
80 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
85 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
86 S390X capability vector to zero. This simplifies testing of different code
87 paths on S390X architecture.
91 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
92 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
93 SP 800-38D". The communication will fail at this point.
97 * The EC_GROUP_clear_free() function is deprecated as there is nothing
98 confidential in EC_GROUP data.
102 * The byte order mark (BOM) character is ignored if encountered at the
103 beginning of a PEM-formatted file.
107 * Added CMS support for the Russian GOST algorithms.
111 * Due to move of the implementation of cryptographic operations
112 to the providers, validation of various operation parameters can
113 be postponed until the actual operation is executed where previously
114 it happened immediately when an operation parameter was set.
116 For example when setting an unsupported curve with
117 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
118 fail but later keygen operations with the EVP_PKEY_CTX will fail.
120 *OpenSSL team members and many third party contributors*
122 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
123 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
124 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
125 instead to retrieve these algorithms from a provider.
129 * On build targets where the multilib postfix is set in the build
130 configuration the libdir directory was changing based on whether
131 the lib directory with the multilib postfix exists on the system
132 or not. This unpredictable behavior was removed and eventual
133 multilib postfix is now always added to the default libdir. Use
134 `--libdir=lib` to override the libdir if adding the postfix is
139 * The triple DES key wrap functionality now conforms to RFC 3217 but is
140 no longer interoperable with OpenSSL 1.1.1.
144 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
145 function codes, this function can only cause problems for calling
150 * Add a configurable flag to output date formats as ISO 8601. Does not
151 change the default date format.
155 * Version of MSVC earlier than 1300 could get link warnings, which could
156 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
157 Support for this flag has been removed.
161 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
162 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
163 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
164 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
165 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
169 * The signatures of the functions to get and set options on SSL and
170 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
171 Some source code changes may be required.
175 * The public definitions of conf_method_st and conf_st have been
176 deprecated. They will be made opaque in a future release.
178 *Rich Salz and Tomáš Mráz*
180 * Client-initiated renegotiation is disabled by default. To allow it, use
181 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
182 flag, or the "ClientRenegotiation" config parameter as appropriate.
186 * Add "abspath" and "includedir" pragma's to config files, to prevent,
187 or modify relative pathname inclusion.
191 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
192 validated. Please consult the README-FIPS and
193 README-PROVIDERS files, as well as the migration guide.
195 *OpenSSL team members and many third party contributors*
197 * For the key types DH and DHX the allowed settable parameters are now different.
201 * The openssl commands that read keys, certificates, and CRLs now
202 automatically detect the PEM or DER format of the input files.
204 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
206 * Added enhanced PKCS#12 APIs which accept a library context.
210 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
214 * Added support for Kernel TLS (KTLS).
216 *Boris Pismenny, John Baldwin and Andrew Gallatin*
218 * Support for RFC 5746 secure renegotiation is now required by default for
219 SSL or TLS connections to succeed.
223 * The signature of the `copy` functional parameter of the
224 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
225 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
226 the signature of the `pub_decode` functional parameter of the
227 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
228 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
232 * The error return values from some control calls (ctrl) have changed.
236 * A public key check is now performed during EVP_PKEY_derive_set_peer().
240 * Many functions in the EVP_ namespace that are getters of values from
241 implementations or contexts were renamed to include get or get0 in their
242 names. Old names are provided as macro aliases for compatibility and
247 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
248 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
249 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
254 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
257 * The output from the command line applications may have minor
262 * The output from numerous "printing" may have minor changes.
266 * Windows thread synchronization uses read/write primitives (SRWLock) when
267 supported by the OS, otherwise CriticalSection continues to be used.
271 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
272 work on read only BIO source/sinks that do not support these functions.
273 This allows piping or redirection of a file BIO using stdin to be buffered
274 into memory. This is used internally in OSSL_DECODER_from_bio().
278 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
279 this function would return one of the values OSSL_STORE_INFO_NAME,
280 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
281 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
282 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
283 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
284 using this function should be amended to handle the changed return value.
288 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
289 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
290 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
291 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
292 contains more than one certificate identifier: This means that all
293 certificates referenced there MUST be part of the validation chain.
297 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
298 RC5, DESX and DES have been moved to the legacy provider.
302 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
303 RIPEMD-160 have been moved to the legacy provider.
307 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
312 * The deprecated functions EVP_PKEY_get0_RSA(),
313 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
314 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
315 well as the similarly named "get1" functions behave differently in
320 * A number of functions handling low-level keys or engines were deprecated
321 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
322 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
323 EVP_PKEY_get0_siphash().
327 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
328 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
329 will need to load the legacy crypto provider. This includes these PBE
330 algorithms which use this KDF:
331 - NID_pbeWithMD2AndDES_CBC
332 - NID_pbeWithMD5AndDES_CBC
333 - NID_pbeWithSHA1AndRC2_CBC
334 - NID_pbeWithMD2AndRC2_CBC
335 - NID_pbeWithMD5AndRC2_CBC
336 - NID_pbeWithSHA1AndDES_CBC
340 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
341 BIO_debug_callback() functions.
345 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
346 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
350 * The RAND_METHOD APIs have been deprecated.
354 * The SRP APIs have been deprecated.
358 * Add a compile time option to prevent the caching of provider fetched
359 algorithms. This is enabled by including the no-cached-fetch option
360 at configuration time.
364 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
365 count of PKCS12_DEFAULT_ITER.
367 *Tomáš Mráz and Sahana Prasad*
369 * The openssl speed command does not use low-level API calls anymore.
373 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
376 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
378 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
382 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
383 providers may supply their own group implementations (using either the "key
384 exchange" or the "key encapsulation" methods) which will automatically be
385 detected and used by libssl.
387 *Matt Caswell, Nicola Tuveri*
389 * The undocumented function X509_certificate_type() has been deprecated;
393 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
397 * Removed RSA padding mode for SSLv23 (which was only used for
398 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
399 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
404 * Deprecated the obsolete X9.31 RSA key generation related functions.
406 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
407 is not allowed to return a value > 1, this is no more taken as failure.
409 *Viktor Dukhovni and David von Oheimb*
411 * Deprecated the obsolete X9.31 RSA key generation related functions
412 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
413 BN_X931_generate_prime_ex().
417 * The default key generation method for the regular 2-prime RSA keys was
418 changed to the FIPS 186-4 B.3.6 method.
422 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
426 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
430 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
431 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
433 *Rich Salz, Richard Levitte, and David von Oheimb*
435 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
439 * Deprecated `OCSP_parse_url()`.
443 * Validation of SM2 keys has been separated from the validation of regular EC
448 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
449 switches: a validation failure triggers an early exit, returning a failure
450 exit status to the parent process.
454 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
455 to ignore unknown ciphers.
459 * The `-cipher-commands` and `-digest-commands` options
460 of the command line utility `list` have been deprecated.
461 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
465 * Added convenience functions for generating asymmetric key pairs:
466 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
467 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
471 * All of the low level EC_KEY functions have been deprecated.
473 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
475 * Deprecated all the libcrypto and libssl error string loading
480 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
481 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
486 * The `-crypt` option to the `passwd` command line tool has been removed.
490 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
495 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
499 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
500 EVP_PKEY_get1_tls_encodedpoint().
504 * The security callback, which can be customised by application code, supports
505 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
506 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
510 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
511 interface. Their functionality remains unchanged.
515 * Added new option for 'openssl list', '-providers', which will display the
516 list of loaded providers, their names, version and status. It optionally
517 displays their gettable parameters.
521 * Removed EVP_PKEY_set_alias_type().
525 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
526 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
530 * Changed all "STACK" functions to be macros instead of inline functions. Macro
531 parameters are still checked for type safety at compile time via helper
536 * Remove the RAND_DRBG API
538 *Paul Dale and Matthias St. Pierre*
540 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
541 as well as actual hostnames.
545 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
546 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
547 conversely, silently ignore DTLS protocol version bounds when configuring
548 TLS-based contexts. The commands can be repeated to set bounds of both
549 types. The same applies with the corresponding "min_protocol" and
550 "max_protocol" command-line switches, in case some application uses both TLS
553 SSL_CTX instances that are created for a fixed protocol version (e.g.
554 `TLSv1_server_method()`) also silently ignore version bounds. Previously
555 attempts to apply bounds to these protocol versions would result in an
556 error. Now only the "version-flexible" SSL_CTX instances are subject to
557 limits in configuration files in command-line options.
561 * Deprecated the `ENGINE` API. Engines should be replaced with providers
566 * Reworked the recorded ERR codes to make better space for system errors.
567 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
568 given code is a system error (true) or an OpenSSL error (false).
572 * Reworked the test perl framework to better allow parallel testing.
574 *Nicola Tuveri and David von Oheimb*
576 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
577 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
581 * 'Configure' has been changed to figure out the configuration target if
582 none is given on the command line. Consequently, the 'config' script is
583 now only a mere wrapper. All documentation is changed to only mention
586 *Rich Salz and Richard Levitte*
588 * Added a library context `OSSL_LIB_CTX` that applications as well as
589 other libraries can use to form a separate context within which
590 libcrypto operations are performed.
594 * Added various `_ex` functions to the OpenSSL API that support using
595 a non-default `OSSL_LIB_CTX`.
599 * Handshake now fails if Extended Master Secret extension is dropped
604 * Dropped interactive mode from the `openssl` program.
608 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
610 *David von Oheimb and Shane Lontis*
612 * Deprecated `EC_METHOD_get_field_type()`.
616 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
617 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
618 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
622 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
626 * Add CAdES-BES signature verification support, mostly derived
627 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
629 *Filipe Raimundo da Silva*
631 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
635 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
636 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
640 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
644 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
645 EC_KEY_precompute_mult().
649 * Deprecated EC_POINTs_mul().
653 * Removed FIPS_mode() and FIPS_mode_set().
657 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
661 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
662 EC_POINT_get_Jprojective_coordinates_GFp().
666 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
667 arrays to be more easily constructed via a series of utility functions.
668 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
669 the various push functions and finally convert to a passable OSSL_PARAM
670 array using OSSL_PARAM_BLD_to_param().
674 * The security strength of SHA1 and MD5 based signatures in TLS has been
679 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
680 contain a provider side internal key.
684 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
688 * Project text documents not yet having a proper file name extension
689 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
690 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
694 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
695 have been converted to Markdown with the goal to produce documents
696 which not only look pretty when viewed online in the browser, but
697 remain well readable inside a plain text editor.
699 To achieve this goal, a 'minimalistic' Markdown style has been applied
700 which avoids formatting elements that interfere too much with the
701 reading flow in the text file. For example, it
703 * avoids [ATX headings][] and uses [setext headings][] instead
704 (which works for `<h1>` and `<h2>` headings only).
705 * avoids [inline links][] and uses [reference links][] instead.
706 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
708 [ATX headings]: https://github.github.com/gfm/#atx-headings
709 [setext headings]: https://github.github.com/gfm/#setext-headings
710 [inline links]: https://github.github.com/gfm/#inline-link
711 [reference links]: https://github.github.com/gfm/#reference-link
712 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
713 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
715 *Matthias St. Pierre*
717 * The test suite is changed to preserve results of each test recipe.
718 A new directory test-runs/ with subdirectories named like the
719 test recipes are created in the build tree for this purpose.
723 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
724 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
725 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
727 *David von Oheimb, Martin Peylo*
729 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
730 It supports arbitrary request and response content types, GET redirection,
731 TLS, connections via HTTP(S) proxies, connections and exchange via
732 user-defined BIOs (allowing implicit connections), persistent connections,
733 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
734 The legacy OCSP-focused (and only partly documented) API
735 is retained for backward compatibility, while most of it is deprecated.
739 * Added `util/check-format.pl`, a tool for checking adherence to the
740 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
741 The checks performed are incomplete and yield some false positives.
742 Still the tool should be useful for detecting most typical glitches.
746 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
747 If domain name resolution yields multiple IP addresses all of them are tried
748 after `connect()` failures.
752 * All of the low level RSA functions have been deprecated.
756 * X509 certificates signed using SHA1 are no longer allowed at security
761 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
762 modified to use PKEY APIs. These commands are now in maintenance mode
763 and no new features will be added to them.
767 * The command line utility rsautl has been deprecated.
771 * The command line utilities genrsa and rsa have been modified to use PKEY
772 APIs. They now write PKCS#8 keys by default. These commands are now in
773 maintenance mode and no new features will be added to them.
777 * All of the low level DH functions have been deprecated.
779 *Paul Dale and Matt Caswell*
781 * All of the low level DSA functions have been deprecated.
785 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
786 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
790 * Deprecated low level ECDH and ECDSA functions.
794 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
798 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
799 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
800 a new formulation to include all the things it can be used for,
801 as well as words of caution.
805 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
809 * All of the low level HMAC functions have been deprecated.
811 *Paul Dale and David von Oheimb*
813 * Over two thousand fixes were made to the documentation, including:
814 - Common options (such as -rand/-writerand, TLS version control, etc)
815 were refactored and point to newly-enhanced descriptions in openssl.pod.
816 - Added style conformance for all options (with help from Richard Levitte),
817 documented all reported missing options, added a CI build to check
818 that all options are documented and that no unimplemented options
820 - Documented some internals, such as all use of environment variables.
821 - Addressed all internal broken L<> references.
825 * All of the low level CMAC functions have been deprecated.
829 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
830 functions have been deprecated.
832 *Paul Dale and David von Oheimb*
834 * Corrected the documentation of the return values from the `EVP_DigestSign*`
835 set of functions. The documentation mentioned negative values for some
836 errors, but this was never the case, so the mention of negative values
839 Code that followed the documentation and thereby check with something
840 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
844 * All of the low level cipher functions have been deprecated.
846 *Matt Caswell and Paul Dale*
848 * Removed include/openssl/opensslconf.h.in and replaced it with
849 include/openssl/configuration.h.in, which differs in not including
850 <openssl/macros.h>. A short header include/openssl/opensslconf.h
851 was added to include both.
853 This allows internal hacks where one might need to modify the set
854 of configured macros, for example this if deprecated symbols are
855 still supposed to be available internally:
857 #include <openssl/configuration.h>
859 #undef OPENSSL_NO_DEPRECATED
860 #define OPENSSL_SUPPRESS_DEPRECATED
862 #include <openssl/macros.h>
864 This should not be used by applications that use the exported
865 symbols, as that will lead to linking errors.
869 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
870 used in exponentiation with 512-bit moduli. No EC algorithms are
871 affected. Analysis suggests that attacks against 2-prime RSA1024,
872 3-prime RSA1536, and DSA1024 as a result of this defect would be very
873 difficult to perform and are not believed likely. Attacks against DH512
874 are considered just feasible. However, for an attack the target would
875 have to re-use the DH512 private key, which is not recommended anyway.
876 Also applications directly using the low-level API BN_mod_exp may be
877 affected if they use BN_FLG_CONSTTIME.
882 * Most memory-debug features have been deprecated, and the functionality
883 replaced with no-ops.
887 * Added documentation for the STACK API.
891 * Introduced a new method type and API, OSSL_ENCODER, to represent
892 generic encoders. These do the same sort of job that PEM writers
893 and d2i functions do, but with support for methods supplied by
894 providers, and the possibility for providers to support other
899 * Introduced a new method type and API, OSSL_DECODER, to represent
900 generic decoders. These do the same sort of job that PEM readers
901 and i2d functions do, but with support for methods supplied by
902 providers, and the possibility for providers to support other
907 * Added a .pragma directive to the syntax of configuration files, to
908 allow varying behavior in a supported and predictable manner.
909 Currently added pragma:
913 This allows dollar signs to be a keyword character unless it's
914 followed by a opening brace or parenthesis. This is useful for
915 platforms where dollar signs are commonly used in names, such as
916 volume names and system directory names on VMS.
920 * Added functionality to create an EVP_PKEY from user data.
924 * Change the interpretation of the '--api' configuration option to
925 mean that this is a desired API compatibility level with no
926 further meaning. The previous interpretation, that this would
927 also mean to remove all deprecated symbols up to and including
928 the given version, no requires that 'no-deprecated' is also used
929 in the configuration.
931 When building applications, the desired API compatibility level
932 can be set with the OPENSSL_API_COMPAT macro like before. For
933 API compatibility version below 3.0, the old style numerical
934 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
935 For version 3.0 and on, the value is expected to be the decimal
936 value calculated from the major and minor version like this:
938 MAJOR * 10000 + MINOR * 100
942 -DOPENSSL_API_COMPAT=30000 For 3.0
943 -DOPENSSL_API_COMPAT=30200 For 3.2
945 To hide declarations that are deprecated up to and including the
946 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
947 given when building the application as well.
951 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
952 access to certificate and CRL stores via URIs and OSSL_STORE
955 This adds the following functions:
957 - X509_LOOKUP_store()
958 - X509_STORE_load_file()
959 - X509_STORE_load_path()
960 - X509_STORE_load_store()
961 - SSL_add_store_cert_subjects_to_stack()
962 - SSL_CTX_set_default_verify_store()
963 - SSL_CTX_load_verify_file()
964 - SSL_CTX_load_verify_dir()
965 - SSL_CTX_load_verify_store()
969 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
970 The presence of this system service is determined at run-time.
974 * Added functionality to create an EVP_PKEY context based on data
975 for methods from providers. This takes an algorithm name and a
976 property query string and simply stores them, with the intent
977 that any operation that uses this context will use those strings
978 to fetch the needed methods implicitly, thereby making the port
979 of application written for pre-3.0 OpenSSL easier.
983 * The undocumented function NCONF_WIN32() has been deprecated; for
984 conversion details see the HISTORY section of doc/man5/config.pod
988 * Introduced the new functions EVP_DigestSignInit_ex() and
989 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
990 EVP_DigestVerifyUpdate() have been converted to functions. See the man
991 pages for further details.
995 * Over two thousand fixes were made to the documentation, including:
996 adding missing command flags, better style conformance, documentation
999 *Rich Salz, Richard Levitte*
1001 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1002 X25519, X448, Ed25519 and Ed448.
1006 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1011 * Deprecated the public definition of `ERR_STATE` as well as the function
1012 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1017 * Added ERR functionality to give callers access to the stored function
1018 names that have replaced the older function code based functions.
1020 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1021 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1022 ERR_peek_error_all() and ERR_peek_last_error_all().
1024 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1025 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1026 ERR_func_error_string().
1030 * Extended testing to be verbose for failing tests only. The make variables
1031 VERBOSE_FAILURE or VF can be used to enable this:
1033 $ make VF=1 test # Unix
1034 $ mms /macro=(VF=1) test ! OpenVMS
1035 $ nmake VF=1 test # Windows
1039 * Added the `-copy_extensions` option to the `x509` command for use with
1040 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1041 all extensions in the request are copied to the certificate or vice versa.
1043 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1045 * Added the `-copy_extensions` option to the `req` command for use with
1046 `-x509`. When given with the `copy` or `copyall` argument,
1047 all extensions in the certification request are copied to the certificate.
1051 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1052 they generate are by default RFC 5280 compliant in the following sense:
1053 There is a subjectKeyIdentifier extension with a hash value of the public key
1054 and for not self-signed certs there is an authorityKeyIdentifier extension
1055 with a keyIdentifier field or issuer information identifying the signing key.
1056 This is done unless some configuration overrides the new default behavior,
1057 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1061 * Added several checks to `X509_verify_cert()` according to requirements in
1062 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1063 (which may be done by using the CLI option `-x509_strict`):
1064 * The basicConstraints of CA certificates must be marked critical.
1065 * CA certificates must explicitly include the keyUsage extension.
1066 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1067 * The issuer name of any certificate must not be empty.
1068 * The subject name of CA certs, certs with keyUsage crlSign,
1069 and certs without subjectAlternativeName must not be empty.
1070 * If a subjectAlternativeName extension is given it must not be empty.
1071 * The signatureAlgorithm field and the cert signature must be consistent.
1072 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1073 must not be marked critical.
1074 * The authorityKeyIdentifier must be given for X.509v3 certs
1075 unless they are self-signed.
1076 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1080 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1081 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1085 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1086 used even when parsing explicit parameters, when loading a encoded key
1087 or calling `EC_GROUP_new_from_ecpkparameters()`/
1088 `EC_GROUP_new_from_ecparameters()`.
1089 This prevents bypass of security hardening and performance gains,
1090 especially for curves with specialized EC_METHODs.
1091 By default, if a key encoded with explicit parameters is loaded and later
1092 encoded, the output is still encoded with explicit parameters, even if
1093 internally a "named" EC_GROUP is used for computation.
1097 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1098 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1099 NULL. After this change, only the cofactor parameter can be NULL. It also
1100 does some minimal sanity checks on the passed order.
1105 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1106 An attack is simple, if the first CMS_recipientInfo is valid but the
1107 second CMS_recipientInfo is chosen ciphertext. If the second
1108 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1109 encryption key will be replaced by garbage, and the message cannot be
1110 decoded, but if the RSA decryption fails, the correct encryption key is
1111 used and the recipient will not notice the attack.
1112 As a work around for this potential attack the length of the decrypted
1113 key must be equal to the cipher default key length, in case the
1114 certifiate is not given and all recipientInfo are tried out.
1115 The old behaviour can be re-enabled in the CMS code by setting the
1116 CMS_DEBUG_DECRYPT flag.
1120 * Early start up entropy quality from the DEVRANDOM seed source has been
1121 improved for older Linux systems. The RAND subsystem will wait for
1122 /dev/random to be producing output before seeding from /dev/urandom.
1123 The seeded state is stored for future library initialisations using
1124 a system global shared memory segment. The shared memory identifier
1125 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1126 the desired value. The default identifier is 114.
1130 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1131 when primes for RSA keys are computed.
1132 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1133 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1134 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1135 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1136 This avoids possible fingerprinting of newly generated RSA modules.
1140 * Correct the extended master secret constant on EBCDIC systems. Without this
1141 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1142 negotiate EMS will fail. Unfortunately this also means that TLS connections
1143 between EBCDIC systems with this fix, and EBCDIC systems without this
1144 fix will fail if they negotiate EMS.
1148 * Changed the library initialisation so that the config file is now loaded
1149 by default. This was already the case for libssl. It now occurs for both
1150 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1151 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1155 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1156 where the former acts as a replacement for `ERR_put_error()`, and the
1157 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1158 `ERR_raise_data()` adds more flexibility by taking a format string and
1159 an arbitrary number of arguments following it, to be processed with
1164 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1165 to check if a named provider is loaded and available. When called, it
1166 will also activate all fallback providers if such are still present.
1170 * Enforce a minimum DH modulus size of 512 bits.
1174 * Changed DH parameters to generate the order q subgroup instead of 2q.
1175 Previously generated DH parameters are still accepted by DH_check
1176 but DH_generate_key works around that by clearing bit 0 of the
1177 private key for those. This avoids leaking bit 0 of the private key.
1181 * Significantly reduce secure memory usage by the randomness pools.
1185 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1190 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1191 algorithms. An implementation of a key exchange algorithm can be obtained
1192 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1193 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1194 the older EVP_PKEY_derive_init() function. See the man pages for the new
1195 functions for further details.
1199 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1203 * Removed the function names from error messages and deprecated the
1208 * Removed NextStep support and the macro OPENSSL_UNISTD
1212 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1213 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1214 Also removed "export var as function" capability; we do not export
1215 variables, only functions.
1219 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1220 an error and 1 indicating success. In previous versions of OpenSSL this
1221 was a void type. If a key was set longer than the maximum possible this
1226 * Support SM2 signing and verification schemes with X509 certificate.
1230 * Use SHA256 as the default digest for TS query in the `ts` app.
1234 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1238 * Default cipher lists/suites are now available via a function, the
1239 #defines are deprecated.
1243 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1244 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1245 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1249 * Join the directories crypto/x509 and crypto/x509v3
1253 * Added command 'openssl kdf' that uses the EVP_KDF API.
1257 * Added command 'openssl mac' that uses the EVP_MAC API.
1261 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1262 as default directories. Also added the command 'openssl info'
1263 for scripting purposes.
1267 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1272 * Add prediction resistance to the DRBG reseeding process.
1276 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1277 mandated by IEEE Std 1619-2018.
1281 * Added newline escaping functionality to a filename when using openssl dgst.
1282 This output format is to replicate the output format found in the `*sum`
1283 checksum programs. This aims to preserve backward compatibility.
1285 *Matt Eaton, Richard Levitte, and Paul Dale*
1287 * Removed the heartbeat message in DTLS feature, as it has very
1288 little usage and doesn't seem to fulfill a valuable purpose.
1289 The configuration option is now deprecated.
1293 * Changed the output of 'openssl {digestname} < file' to display the
1294 digest name in its output.
1298 * Added a new generic trace API which provides support for enabling
1299 instrumentation through trace output.
1301 *Richard Levitte & Matthias St. Pierre*
1303 * Added build tests for C++. These are generated files that only do one
1304 thing, to include one public OpenSSL head file each. This tests that
1305 the public header files can be usefully included in a C++ application.
1307 This test isn't enabled by default. It can be enabled with the option
1308 'enable-buildtest-c++'.
1312 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1316 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1320 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1324 * Added KMAC to EVP_MAC.
1328 * Added property based algorithm implementation selection framework to
1333 * Added SCA hardening for modular field inversion in EC_GROUP through
1334 a new dedicated field_inv() pointer in EC_METHOD.
1335 This also addresses a leakage affecting conversions from projective
1336 to affine coordinates.
1338 *Billy Bob Brumley, Nicola Tuveri*
1340 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1341 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1342 those algorithms that were already supported through the EVP_PKEY API
1343 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1344 and scrypt are now wrappers that call EVP_KDF.
1348 * Build devcrypto engine as a dynamic engine.
1350 *Eneas U de Queiroz*
1352 * Add keyed BLAKE2 to EVP_MAC.
1356 * Fix a bug in the computation of the endpoint-pair shared secret used
1357 by DTLS over SCTP. This breaks interoperability with older versions
1358 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1359 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1360 interoperability with such broken implementations. However, enabling
1361 this switch breaks interoperability with correct implementations.
1363 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1364 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1368 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1372 * Changed the license to the Apache License v2.0.
1376 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1378 - Major releases (indicated by incrementing the MAJOR release number)
1379 may introduce incompatible API/ABI changes.
1380 - Minor releases (indicated by incrementing the MINOR release number)
1381 may introduce new features but retain API/ABI compatibility.
1382 - Patch releases (indicated by incrementing the PATCH number)
1383 are intended for bug fixes and other improvements of existing
1384 features only (like improving performance or adding documentation)
1385 and retain API/ABI compatibility.
1389 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1393 * Remove the 'dist' target and add a tarball building script. The
1394 'dist' target has fallen out of use, and it shouldn't be
1395 necessary to configure just to create a source distribution.
1399 * Recreate the OS390-Unix config target. It no longer relies on a
1400 special script like it did for OpenSSL pre-1.1.0.
1404 * Instead of having the source directories listed in Configure, add
1405 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1410 * Add GMAC to EVP_MAC.
1414 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1418 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1419 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1420 to facilitate the continued use of MACs through raw private keys in
1421 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1425 * Deprecate ECDH_KDF_X9_62().
1429 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1430 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1431 are retained for backwards compatibility.
1435 * AES-XTS mode now enforces that its two keys are different to mitigate
1436 the attacked described in "Efficient Instantiations of Tweakable
1437 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1438 Details of this attack can be obtained from:
1439 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1443 * Rename the object files, i.e. give them other names than in previous
1444 versions. Their names now include the name of the final product, as
1445 well as its type mnemonic (bin, lib, shlib).
1449 * Added new option for 'openssl list', '-objects', which will display the
1450 list of built in objects, i.e. OIDs with names.
1454 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1455 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1460 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1461 improves application performance by removing data copies and providing
1462 applications with zero-copy system calls such as sendfile and splice.
1466 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1470 * `PKCS12_parse` now maintains the order of the parsed certificates
1471 when outputting them via `*ca` (rather than reversing it).
1475 * Deprecated pthread fork support methods.
1479 * Added support for FFDHE key exchange in TLS 1.3.
1483 * Added a new concept for OpenSSL plugability: providers. This
1484 functionality is designed to replace the ENGINE API and ENGINE
1485 implementations, and to be much more dynamic, allowing provider
1486 authors to introduce new algorithms among other things, as long as
1487 there's an API that supports the algorithm type.
1489 With this concept comes a new core API for interaction between
1490 libcrypto and provider implementations. Public libcrypto functions
1491 that want to use providers do so through this core API.
1493 The main documentation for this core API is found in
1494 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1495 refer to other manuals describing the API specific for supported
1496 algorithm types (also called operations).
1503 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1505 * Fixed an SM2 Decryption Buffer Overflow.
1507 In order to decrypt SM2 encrypted data an application is expected to
1508 call the API function EVP_PKEY_decrypt(). Typically an application will
1509 call this function twice. The first time, on entry, the "out" parameter
1510 can be NULL and, on exit, the "outlen" parameter is populated with the
1511 buffer size required to hold the decrypted plaintext. The application
1512 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1513 again, but this time passing a non-NULL value for the "out" parameter.
1515 A bug in the implementation of the SM2 decryption code means that the
1516 calculation of the buffer size required to hold the plaintext returned
1517 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1518 size required by the second call. This can lead to a buffer overflow
1519 when EVP_PKEY_decrypt() is called by the application a second time with
1520 a buffer that is too small.
1522 A malicious attacker who is able present SM2 content for decryption to
1523 an application could cause attacker chosen data to overflow the buffer
1524 by up to a maximum of 62 bytes altering the contents of other data held
1525 after the buffer, possibly changing application behaviour or causing
1526 the application to crash. The location of the buffer is application
1527 dependent but is typically heap allocated.
1532 * Fixed various read buffer overruns processing ASN.1 strings
1534 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1535 structure which contains a buffer holding the string data and a field
1536 holding the buffer length. This contrasts with normal C strings which
1537 are repesented as a buffer for the string data which is terminated
1538 with a NUL (0) byte.
1540 Although not a strict requirement, ASN.1 strings that are parsed using
1541 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1542 well as any string whose value has been set with the ASN1_STRING_set()
1543 function will additionally NUL terminate the byte array in the
1544 ASN1_STRING structure.
1546 However, it is possible for applications to directly construct valid
1547 ASN1_STRING structures which do not NUL terminate the byte array by
1548 directly setting the "data" and "length" fields in the ASN1_STRING
1549 array. This can also happen by using the ASN1_STRING_set0() function.
1551 Numerous OpenSSL functions that print ASN.1 data have been found to
1552 assume that the ASN1_STRING byte array will be NUL terminated, even
1553 though this is not guaranteed for strings that have been directly
1554 constructed. Where an application requests an ASN.1 structure to be
1555 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1556 been directly constructed by the application without NUL terminating
1557 the "data" field, then a read buffer overrun can occur.
1559 The same thing can also occur during name constraints processing
1560 of certificates (for example if a certificate has been directly
1561 constructed by the application instead of loading it via the OpenSSL
1562 parsing functions, and the certificate contains non NUL terminated
1563 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1564 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1566 If a malicious actor can cause an application to directly construct an
1567 ASN1_STRING and then process it through one of the affected OpenSSL
1568 functions then this issue could be hit. This might result in a crash
1569 (causing a Denial of Service attack). It could also result in the
1570 disclosure of private memory contents (such as private keys, or
1571 sensitive plaintext).
1576 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1578 * Fixed a problem with verifying a certificate chain when using the
1579 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1580 the certificates present in a certificate chain. It is not set by default.
1582 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1583 the chain that have explicitly encoded elliptic curve parameters was added
1584 as an additional strict check.
1586 An error in the implementation of this check meant that the result of a
1587 previous check to confirm that certificates in the chain are valid CA
1588 certificates was overwritten. This effectively bypasses the check
1589 that non-CA certificates must not be able to issue other certificates.
1591 If a "purpose" has been configured then there is a subsequent opportunity
1592 for checks that the certificate is a valid CA. All of the named "purpose"
1593 values implemented in libcrypto perform this check. Therefore, where
1594 a purpose is set the certificate chain will still be rejected even when the
1595 strict flag has been used. A purpose is set by default in libssl client and
1596 server certificate verification routines, but it can be overridden or
1597 removed by an application.
1599 In order to be affected, an application must explicitly set the
1600 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1601 for the certificate verification or, in the case of TLS client or server
1602 applications, override the default purpose.
1607 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1608 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1609 renegotiation ClientHello omits the signature_algorithms extension (where it
1610 was present in the initial ClientHello), but includes a
1611 signature_algorithms_cert extension then a NULL pointer dereference will
1612 result, leading to a crash and a denial of service attack.
1614 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1615 (which is the default configuration). OpenSSL TLS clients are not impacted by
1619 *Peter Kästle and Samuel Sapalski*
1621 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1623 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1624 create a unique hash value based on the issuer and serial number data
1625 contained within an X509 certificate. However it was failing to correctly
1626 handle any errors that may occur while parsing the issuer field (which might
1627 occur if the issuer field is maliciously constructed). This may subsequently
1628 result in a NULL pointer deref and a crash leading to a potential denial of
1634 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1635 padding mode to correctly check for rollback attacks. This is considered a
1636 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1641 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1642 functions. Previously they could overflow the output length argument in some
1643 cases where the input length is close to the maximum permissable length for
1644 an integer on the platform. In such cases the return value from the function
1645 call would be 1 (indicating success), but the output length value would be
1646 negative. This could cause applications to behave incorrectly or crash.
1651 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1652 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1653 could be exploited in a side channel attack to recover the password. Since
1654 the attack is local host only this is outside of the current OpenSSL
1655 threat model and therefore no CVE is assigned.
1657 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1662 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1664 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1665 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1666 If an attacker can control both items being compared then this could lead
1667 to a possible denial of service attack. OpenSSL itself uses the
1668 GENERAL_NAME_cmp function for two purposes:
1669 1) Comparing CRL distribution point names between an available CRL and a
1670 CRL distribution point embedded in an X509 certificate
1671 2) When verifying that a timestamp response token signer matches the
1672 timestamp authority name (exposed via the API functions
1673 TS_RESP_verify_response and TS_RESP_verify_token)
1678 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1680 * Certificates with explicit curve parameters are now disallowed in
1681 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1685 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1686 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1687 conversely, silently ignore DTLS protocol version bounds when configuring
1688 TLS-based contexts. The commands can be repeated to set bounds of both
1689 types. The same applies with the corresponding "min_protocol" and
1690 "max_protocol" command-line switches, in case some application uses both TLS
1693 SSL_CTX instances that are created for a fixed protocol version (e.g.
1694 TLSv1_server_method()) also silently ignore version bounds. Previously
1695 attempts to apply bounds to these protocol versions would result in an
1696 error. Now only the "version-flexible" SSL_CTX instances are subject to
1697 limits in configuration files in command-line options.
1701 * Handshake now fails if Extended Master Secret extension is dropped
1706 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1708 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1710 * Fixed segmentation fault in SSL_check_chain()
1711 Server or client applications that call the SSL_check_chain() function
1712 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1713 dereference as a result of incorrect handling of the
1714 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1715 or unrecognised signature algorithm is received from the peer. This could
1716 be exploited by a malicious peer in a Denial of Service attack.
1721 * Added AES consttime code for no-asm configurations
1722 an optional constant time support for AES was added
1723 when building openssl for no-asm.
1724 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1725 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1726 At this time this feature is by default disabled.
1727 It will be enabled by default in 3.0.
1731 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1733 * Revert the change of EOF detection while reading in libssl to avoid
1734 regressions in applications depending on the current way of reporting
1735 the EOF. As the existing method is not fully accurate the change to
1736 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1737 branch and will be present in the 3.0 release.
1741 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1742 when primes for RSA keys are computed.
1743 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1744 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1745 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1746 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1747 This avoids possible fingerprinting of newly generated RSA modules.
1751 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1753 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1754 while reading in libssl then we would report an error back to the
1755 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1756 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1757 therefore give a hint as to what went wrong.
1761 * Check that ed25519 and ed448 are allowed by the security level. Previously
1762 signature algorithms not using an MD were not being checked that they were
1763 allowed by the security level.
1767 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1768 was not quite right. The behaviour was not consistent between resumption
1769 and normal handshakes, and also not quite consistent with historical
1770 behaviour. The behaviour in various scenarios has been clarified and
1771 it has been updated to make it match historical behaviour as closely as
1776 * *[VMS only]* The header files that the VMS compilers include automatically,
1777 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1778 that the C++ compiler doesn't understand. This is a shortcoming in the
1779 compiler, but can be worked around with `__cplusplus` guards.
1781 C++ applications that use OpenSSL libraries must be compiled using the
1782 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1783 functions. Otherwise, only functions with symbols of less than 31
1784 characters can be used, as the linker will not be able to successfully
1785 resolve symbols with longer names.
1789 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1790 The presence of this system service is determined at run-time.
1794 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1799 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1801 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1802 number generator (RNG). This was intended to include protection in the
1803 event of a fork() system call in order to ensure that the parent and child
1804 processes did not share the same RNG state. However this protection was not
1805 being used in the default case.
1807 A partial mitigation for this issue is that the output from a high
1808 precision timer is mixed into the RNG state so the likelihood of a parent
1809 and child process sharing state is significantly reduced.
1811 If an application already calls OPENSSL_init_crypto() explicitly using
1812 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1815 *Matthias St. Pierre*
1817 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1818 used even when parsing explicit parameters, when loading a encoded key
1819 or calling `EC_GROUP_new_from_ecpkparameters()`/
1820 `EC_GROUP_new_from_ecparameters()`.
1821 This prevents bypass of security hardening and performance gains,
1822 especially for curves with specialized EC_METHODs.
1823 By default, if a key encoded with explicit parameters is loaded and later
1824 encoded, the output is still encoded with explicit parameters, even if
1825 internally a "named" EC_GROUP is used for computation.
1829 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1830 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1831 NULL. After this change, only the cofactor parameter can be NULL. It also
1832 does some minimal sanity checks on the passed order.
1837 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1838 An attack is simple, if the first CMS_recipientInfo is valid but the
1839 second CMS_recipientInfo is chosen ciphertext. If the second
1840 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1841 encryption key will be replaced by garbage, and the message cannot be
1842 decoded, but if the RSA decryption fails, the correct encryption key is
1843 used and the recipient will not notice the attack.
1844 As a work around for this potential attack the length of the decrypted
1845 key must be equal to the cipher default key length, in case the
1846 certifiate is not given and all recipientInfo are tried out.
1847 The old behaviour can be re-enabled in the CMS code by setting the
1848 CMS_DEBUG_DECRYPT flag.
1853 * Early start up entropy quality from the DEVRANDOM seed source has been
1854 improved for older Linux systems. The RAND subsystem will wait for
1855 /dev/random to be producing output before seeding from /dev/urandom.
1856 The seeded state is stored for future library initialisations using
1857 a system global shared memory segment. The shared memory identifier
1858 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1859 the desired value. The default identifier is 114.
1863 * Correct the extended master secret constant on EBCDIC systems. Without this
1864 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1865 negotiate EMS will fail. Unfortunately this also means that TLS connections
1866 between EBCDIC systems with this fix, and EBCDIC systems without this
1867 fix will fail if they negotiate EMS.
1871 * Use Windows installation paths in the mingw builds
1873 Mingw isn't a POSIX environment per se, which means that Windows
1874 paths should be used for installation.
1879 * Changed DH_check to accept parameters with order q and 2q subgroups.
1880 With order 2q subgroups the bit 0 of the private key is not secret
1881 but DH_generate_key works around that by clearing bit 0 of the
1882 private key for those. This avoids leaking bit 0 of the private key.
1886 * Significantly reduce secure memory usage by the randomness pools.
1890 * Revert the DEVRANDOM_WAIT feature for Linux systems
1892 The DEVRANDOM_WAIT feature added a select() call to wait for the
1893 /dev/random device to become readable before reading from the
1894 /dev/urandom device.
1896 It turned out that this change had negative side effects on
1897 performance which were not acceptable. After some discussion it
1898 was decided to revert this feature and leave it up to the OS
1899 resp. the platform maintainer to ensure a proper initialization
1900 during early boot time.
1902 *Matthias St. Pierre*
1904 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1906 * Add build tests for C++. These are generated files that only do one
1907 thing, to include one public OpenSSL head file each. This tests that
1908 the public header files can be usefully included in a C++ application.
1910 This test isn't enabled by default. It can be enabled with the option
1911 'enable-buildtest-c++'.
1915 * Enable SHA3 pre-hashing for ECDSA and DSA.
1919 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1920 This changes the size when using the `genpkey` command when no size is given.
1921 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1922 generation commands to use 2048 bits by default.
1926 * Reorganize the manual pages to consistently have RETURN VALUES,
1927 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1928 util/fix-doc-nits accordingly.
1930 *Paul Yang, Joshua Lock*
1932 * Add the missing accessor EVP_PKEY_get0_engine()
1936 * Have commands like `s_client` and `s_server` output the signature scheme
1937 along with other cipher suite parameters when debugging.
1941 * Make OPENSSL_config() error agnostic again.
1945 * Do the error handling in RSA decryption constant time.
1949 * Prevent over long nonces in ChaCha20-Poly1305.
1951 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1952 for every encryption operation. RFC 7539 specifies that the nonce value
1953 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1954 and front pads the nonce with 0 bytes if it is less than 12
1955 bytes. However it also incorrectly allows a nonce to be set of up to 16
1956 bytes. In this case only the last 12 bytes are significant and any
1957 additional leading bytes are ignored.
1959 It is a requirement of using this cipher that nonce values are
1960 unique. Messages encrypted using a reused nonce value are susceptible to
1961 serious confidentiality and integrity attacks. If an application changes
1962 the default nonce length to be longer than 12 bytes and then makes a
1963 change to the leading bytes of the nonce expecting the new value to be a
1964 new unique nonce then such an application could inadvertently encrypt
1965 messages with a reused nonce.
1967 Additionally the ignored bytes in a long nonce are not covered by the
1968 integrity guarantee of this cipher. Any application that relies on the
1969 integrity of these ignored leading bytes of a long nonce may be further
1970 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1971 is safe because no such use sets such a long nonce value. However user
1972 applications that use this cipher directly and set a non-default nonce
1973 length to be longer than 12 bytes may be vulnerable.
1975 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1981 * Add DEVRANDOM_WAIT feature for Linux systems
1983 On older Linux systems where the getrandom() system call is not available,
1984 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1985 Contrary to getrandom(), the /dev/urandom device will not block during
1986 early boot when the kernel CSPRNG has not been seeded yet.
1988 To mitigate this known weakness, use select() to wait for /dev/random to
1989 become readable before reading from /dev/urandom.
1991 * Ensure that SM2 only uses SM3 as digest algorithm
1995 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1997 * Change the info callback signals for the start and end of a post-handshake
1998 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1999 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2000 confused by this and assume that a TLSv1.2 renegotiation has started. This
2001 can break KeyUpdate handling. Instead we no longer signal the start and end
2002 of a post handshake message exchange (although the messages themselves are
2003 still signalled). This could break some applications that were expecting
2004 the old signals. However without this KeyUpdate is not usable for many
2009 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2011 * Timing vulnerability in DSA signature generation
2013 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2014 timing side channel attack. An attacker could use variations in the signing
2015 algorithm to recover the private key.
2017 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2022 * Timing vulnerability in ECDSA signature generation
2024 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2025 timing side channel attack. An attacker could use variations in the signing
2026 algorithm to recover the private key.
2028 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2033 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2034 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2035 of two gigabytes and the error handling improved.
2037 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2038 categorized as a normal bug, not a security issue, because the DRBG reseeds
2039 automatically and is fully functional even without additional randomness
2040 provided by the application.
2042 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2044 * Add a new ClientHello callback. Provides a callback interface that gives
2045 the application the ability to adjust the nascent SSL object at the
2046 earliest stage of ClientHello processing, immediately after extensions have
2047 been collected but before they have been processed. In particular, this
2048 callback can adjust the supported TLS versions in response to the contents
2053 * Add SM2 base algorithm support.
2057 * s390x assembly pack: add (improved) hardware-support for the following
2058 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2059 aes-cfb/cfb8, aes-ecb.
2063 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2064 parameter is no longer accepted, as it leads to a corrupt table. NULL
2065 pem_str is reserved for alias entries only.
2069 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2070 step for prime curves. The new implementation is based on formulae from
2071 differential addition-and-doubling in homogeneous projective coordinates
2072 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2073 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2074 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2075 to work in projective coordinates.
2077 *Billy Bob Brumley, Nicola Tuveri*
2079 * Change generating and checking of primes so that the error rate of not
2080 being prime depends on the intended use based on the size of the input.
2081 For larger primes this will result in more rounds of Miller-Rabin.
2082 The maximal error rate for primes with more than 1080 bits is lowered
2085 *Kurt Roeckx, Annie Yousar*
2087 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2091 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2092 moving between systems, and to avoid confusion when a Windows build is
2093 done with mingw vs with MSVC. For POSIX installs, there's still a
2094 symlink or copy named 'tsget' to avoid that confusion as well.
2098 * Revert blinding in ECDSA sign and instead make problematic addition
2099 length-invariant. Switch even to fixed-length Montgomery multiplication.
2103 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2104 step for binary curves. The new implementation is based on formulae from
2105 differential addition-and-doubling in mixed Lopez-Dahab projective
2106 coordinates, modified to independently blind the operands.
2108 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2110 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2111 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2112 EC_METHODs to implement their own specialized "ladder step", to take
2113 advantage of more favorable coordinate systems or more efficient
2114 differential addition-and-doubling algorithms.
2116 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2118 * Modified the random device based seed sources to keep the relevant
2119 file descriptors open rather than reopening them on each access.
2120 This allows such sources to operate in a chroot() jail without
2121 the associated device nodes being available. This behaviour can be
2122 controlled using RAND_keep_random_devices_open().
2126 * Numerous side-channel attack mitigations have been applied. This may have
2127 performance impacts for some algorithms for the benefit of improved
2128 security. Specific changes are noted in this change log by their respective
2133 * AIX shared library support overhaul. Switch to AIX "natural" way of
2134 handling shared libraries, which means collecting shared objects of
2135 different versions and bitnesses in one common archive. This allows to
2136 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2137 doesn't affect the way 3rd party applications are linked, only how
2138 multi-version installation is managed.
2142 * Make ec_group_do_inverse_ord() more robust and available to other
2143 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2144 mitigations are applied to the fallback BN_mod_inverse().
2145 When using this function rather than BN_mod_inverse() directly, new
2146 EC cryptosystem implementations are then safer-by-default.
2150 * Add coordinate blinding for EC_POINT and implement projective
2151 coordinate blinding for generic prime curves as a countermeasure to
2152 chosen point SCA attacks.
2154 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2156 * Add blinding to ECDSA and DSA signatures to protect against side channel
2157 attacks discovered by Keegan Ryan (NCC Group).
2161 * Enforce checking in the `pkeyutl` command to ensure that the input
2162 length does not exceed the maximum supported digest length when performing
2163 a sign, verify or verifyrecover operation.
2167 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2168 I/O in combination with something like select() or poll() will hang. This
2169 can be turned off again using SSL_CTX_clear_mode().
2170 Many applications do not properly handle non-application data records, and
2171 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2172 around the problems in those applications, but can also break some.
2173 It's recommended to read the manpages about SSL_read(), SSL_write(),
2174 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2175 SSL_CTX_set_read_ahead() again.
2179 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2180 now allow empty (zero character) pass phrases.
2184 * Apply blinding to binary field modular inversion and remove patent
2185 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2189 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2190 binary and prime elliptic curves.
2194 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2195 constant time fixed point multiplication.
2199 * Revise elliptic curve scalar multiplication with timing attack
2200 defenses: ec_wNAF_mul redirects to a constant time implementation
2201 when computing fixed point and variable point multiplication (which
2202 in OpenSSL are mostly used with secret scalars in keygen, sign,
2203 ECDH derive operations).
2204 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2207 * Updated CONTRIBUTING
2211 * Updated DRBG / RAND to request nonce and additional low entropy
2212 randomness from the system.
2214 *Matthias St. Pierre*
2216 * Updated 'openssl rehash' to use OpenSSL consistent default.
2220 * Moved the load of the ssl_conf module to libcrypto, which helps
2221 loading engines that libssl uses before libssl is initialised.
2225 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2229 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2231 *Ingo Schwarze, Rich Salz*
2233 * Added output of accepting IP address and port for 'openssl s_server'
2237 * Added a new API for TLSv1.3 ciphersuites:
2238 SSL_CTX_set_ciphersuites()
2239 SSL_set_ciphersuites()
2243 * Memory allocation failures consistently add an error to the error
2248 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2249 in libcrypto when run as setuid/setgid.
2253 * Load any config file by default when libssl is used.
2257 * Added new public header file <openssl/rand_drbg.h> and documentation
2258 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2260 *Matthias St. Pierre*
2262 * QNX support removed (cannot find contributors to get their approval
2263 for the license change).
2267 * TLSv1.3 replay protection for early data has been implemented. See the
2268 SSL_read_early_data() man page for further details.
2272 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2273 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2274 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2275 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2276 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2277 configuration has been separated out. See the ciphers man page or the
2278 SSL_CTX_set_ciphersuites() man page for more information.
2282 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2283 in responder mode now supports the new "-multi" option, which
2284 spawns the specified number of child processes to handle OCSP
2285 requests. The "-timeout" option now also limits the OCSP
2286 responder's patience to wait to receive the full client request
2287 on a newly accepted connection. Child processes are respawned
2288 as needed, and the CA index file is automatically reloaded
2289 when changed. This makes it possible to run the "ocsp" responder
2290 as a long-running service, making the OpenSSL CA somewhat more
2291 feature-complete. In this mode, most diagnostic messages logged
2292 after entering the event loop are logged via syslog(3) rather than
2297 * Added support for X448 and Ed448. Heavily based on original work by
2302 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2303 objects loaded. This adds the functions OSSL_STORE_expect() and
2304 OSSL_STORE_find() as well as needed tools to construct searches and
2305 get the search data out of them.
2309 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2310 version of OpenSSL should review their configuration settings to ensure
2311 that they are still appropriate for TLSv1.3. For further information see:
2312 <https://wiki.openssl.org/index.php/TLS1.3>
2316 * Grand redesign of the OpenSSL random generator
2318 The default RAND method now utilizes an AES-CTR DRBG according to
2319 NIST standard SP 800-90Ar1. The new random generator is essentially
2320 a port of the default random generator from the OpenSSL FIPS 2.0
2321 object module. It is a hybrid deterministic random bit generator
2322 using an AES-CTR bit stream and which seeds and reseeds itself
2323 automatically using trusted system entropy sources.
2325 Some of its new features are:
2326 - Support for multiple DRBG instances with seed chaining.
2327 - The default RAND method makes use of a DRBG.
2328 - There is a public and private DRBG instance.
2329 - The DRBG instances are fork-safe.
2330 - Keep all global DRBG instances on the secure heap if it is enabled.
2331 - The public and private DRBG instance are per thread for lock free
2334 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2336 * Changed Configure so it only says what it does and doesn't dump
2337 so much data. Instead, ./configdata.pm should be used as a script
2338 to display all sorts of configuration data.
2342 * Added processing of "make variables" to Configure.
2346 * Added SHA512/224 and SHA512/256 algorithm support.
2350 * The last traces of Netware support, first removed in 1.1.0, have
2355 * Get rid of Makefile.shared, and in the process, make the processing
2356 of certain files (rc.obj, or the .def/.map/.opt files produced from
2357 the ordinal files) more visible and hopefully easier to trace and
2358 debug (or make silent).
2362 * Make it possible to have environment variable assignments as
2363 arguments to config / Configure.
2367 * Add multi-prime RSA (RFC 8017) support.
2371 * Add SM3 implemented according to GB/T 32905-2016
2372 *Jack Lloyd <jack.lloyd@ribose.com>,*
2373 *Ronald Tse <ronald.tse@ribose.com>,*
2374 *Erick Borsboom <erick.borsboom@ribose.com>*
2376 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2377 as documented in RFC6066.
2378 Based on a patch from Tomasz Moń
2380 *Filipe Raimundo da Silva*
2382 * Add SM4 implemented according to GB/T 32907-2016.
2383 *Jack Lloyd <jack.lloyd@ribose.com>,*
2384 *Ronald Tse <ronald.tse@ribose.com>,*
2385 *Erick Borsboom <erick.borsboom@ribose.com>*
2387 * Reimplement -newreq-nodes and ERR_error_string_n; the
2388 original author does not agree with the license change.
2392 * Add ARIA AEAD TLS support.
2396 * Some macro definitions to support VS6 have been removed. Visual
2397 Studio 6 has not worked since 1.1.0
2401 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2402 without clearing the errors.
2406 * Add "atfork" functions. If building on a system that without
2407 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2408 requirements. The RAND facility now uses/requires this.
2416 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2417 not possible to disable entirely. However, it's still possible to
2418 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2421 To disable, configure with 'no-ui-console'. 'no-ui' is still
2422 possible to use as an alias. Check at compile time with the
2423 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2424 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2428 * Add a STORE module, which implements a uniform and URI based reader of
2429 stores that can contain keys, certificates, CRLs and numerous other
2430 objects. The main API is loosely based on a few stdio functions,
2431 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2432 OSSL_STORE_error and OSSL_STORE_close.
2433 The implementation uses backends called "loaders" to implement arbitrary
2434 URI schemes. There is one built in "loader" for the 'file' scheme.
2438 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2439 then adjusted to work on FreeBSD 8.4 as well.
2440 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2441 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2445 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2446 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2447 error code calls like this:
2449 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2451 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2452 that can be encoded in C. For the foreseeable future, this will only
2455 *Richard Levitte and Tim Hudson*
2457 * Removed BSD cryptodev engine.
2461 * Add a build target 'build_all_generated', to build all generated files
2462 and only that. This can be used to prepare everything that requires
2463 things like perl for a system that lacks perl and then move everything
2464 to that system and do the rest of the build there.
2468 * In the UI interface, make it possible to duplicate the user data. This
2469 can be used by engines that need to retain the data for a longer time
2470 than just the call where this user data is passed.
2474 * Ignore the '-named_curve auto' value for compatibility of applications
2477 *Tomáš Mráz <tmraz@fedoraproject.org>*
2479 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2480 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2481 alerts across multiple records (some of which could be empty). In practice
2482 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2483 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2484 support this at all. Supporting it adds significant complexity to the
2485 record layer, and its removal is unlikely to cause interoperability
2490 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2491 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2492 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2497 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2498 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2500 *Richard Levitte, Andy Polyakov*
2502 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2507 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2508 platform rather than 'mingw'.
2512 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2513 success if they are asked to add an object which already exists
2514 in the store. This change cascades to other functions which load
2515 certificates and CRLs.
2519 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2520 facilitate stack unwinding even from assembly subroutines.
2524 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2525 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2529 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2530 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2531 which is the minimum version we support.
2535 * Certificate time validation (X509_cmp_time) enforces stricter
2536 compliance with RFC 5280. Fractional seconds and timezone offsets
2537 are no longer allowed.
2541 * Add support for ARIA
2545 * s_client will now send the Server Name Indication (SNI) extension by
2546 default unless the new "-noservername" option is used. The server name is
2547 based on the host provided to the "-connect" option unless overridden by
2548 using "-servername".
2552 * Add support for SipHash
2556 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2557 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2558 prevent issues where no progress is being made and the peer continually
2559 sends unrecognised record types, using up resources processing them.
2563 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2564 using the algorithm defined in
2565 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2569 * Heartbeat support has been removed; the ABI is changed for now.
2571 *Richard Levitte, Rich Salz*
2573 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2577 * The RSA "null" method, which was partially supported to avoid patent
2578 issues, has been replaced to always returns NULL.
2585 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2587 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2588 used even when parsing explicit parameters, when loading a encoded key
2589 or calling `EC_GROUP_new_from_ecpkparameters()`/
2590 `EC_GROUP_new_from_ecparameters()`.
2591 This prevents bypass of security hardening and performance gains,
2592 especially for curves with specialized EC_METHODs.
2593 By default, if a key encoded with explicit parameters is loaded and later
2594 encoded, the output is still encoded with explicit parameters, even if
2595 internally a "named" EC_GROUP is used for computation.
2599 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2600 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2601 NULL. After this change, only the cofactor parameter can be NULL. It also
2602 does some minimal sanity checks on the passed order.
2607 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2608 An attack is simple, if the first CMS_recipientInfo is valid but the
2609 second CMS_recipientInfo is chosen ciphertext. If the second
2610 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2611 encryption key will be replaced by garbage, and the message cannot be
2612 decoded, but if the RSA decryption fails, the correct encryption key is
2613 used and the recipient will not notice the attack.
2614 As a work around for this potential attack the length of the decrypted
2615 key must be equal to the cipher default key length, in case the
2616 certifiate is not given and all recipientInfo are tried out.
2617 The old behaviour can be re-enabled in the CMS code by setting the
2618 CMS_DEBUG_DECRYPT flag.
2623 * Use Windows installation paths in the mingw builds
2625 Mingw isn't a POSIX environment per se, which means that Windows
2626 paths should be used for installation.
2631 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2633 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2634 This changes the size when using the `genpkey` command when no size is given.
2635 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2636 generation commands to use 2048 bits by default.
2640 * Prevent over long nonces in ChaCha20-Poly1305.
2642 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2643 for every encryption operation. RFC 7539 specifies that the nonce value
2644 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2645 and front pads the nonce with 0 bytes if it is less than 12
2646 bytes. However it also incorrectly allows a nonce to be set of up to 16
2647 bytes. In this case only the last 12 bytes are significant and any
2648 additional leading bytes are ignored.
2650 It is a requirement of using this cipher that nonce values are
2651 unique. Messages encrypted using a reused nonce value are susceptible to
2652 serious confidentiality and integrity attacks. If an application changes
2653 the default nonce length to be longer than 12 bytes and then makes a
2654 change to the leading bytes of the nonce expecting the new value to be a
2655 new unique nonce then such an application could inadvertently encrypt
2656 messages with a reused nonce.
2658 Additionally the ignored bytes in a long nonce are not covered by the
2659 integrity guarantee of this cipher. Any application that relies on the
2660 integrity of these ignored leading bytes of a long nonce may be further
2661 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2662 is safe because no such use sets such a long nonce value. However user
2663 applications that use this cipher directly and set a non-default nonce
2664 length to be longer than 12 bytes may be vulnerable.
2666 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2672 * Added SCA hardening for modular field inversion in EC_GROUP through
2673 a new dedicated field_inv() pointer in EC_METHOD.
2674 This also addresses a leakage affecting conversions from projective
2675 to affine coordinates.
2677 *Billy Bob Brumley, Nicola Tuveri*
2679 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2680 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2684 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2688 * Remove the 'dist' target and add a tarball building script. The
2689 'dist' target has fallen out of use, and it shouldn't be
2690 necessary to configure just to create a source distribution.
2694 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2696 * Timing vulnerability in DSA signature generation
2698 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2699 timing side channel attack. An attacker could use variations in the signing
2700 algorithm to recover the private key.
2702 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2707 * Timing vulnerability in ECDSA signature generation
2709 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2710 timing side channel attack. An attacker could use variations in the signing
2711 algorithm to recover the private key.
2713 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2718 * Add coordinate blinding for EC_POINT and implement projective
2719 coordinate blinding for generic prime curves as a countermeasure to
2720 chosen point SCA attacks.
2722 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2724 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2726 * Client DoS due to large DH parameter
2728 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2729 malicious server can send a very large prime value to the client. This will
2730 cause the client to spend an unreasonably long period of time generating a
2731 key for this prime resulting in a hang until the client has finished. This
2732 could be exploited in a Denial Of Service attack.
2734 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2739 * Cache timing vulnerability in RSA Key Generation
2741 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2742 a cache timing side channel attack. An attacker with sufficient access to
2743 mount cache timing attacks during the RSA key generation process could
2744 recover the private key.
2746 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2747 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2752 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2753 parameter is no longer accepted, as it leads to a corrupt table. NULL
2754 pem_str is reserved for alias entries only.
2758 * Revert blinding in ECDSA sign and instead make problematic addition
2759 length-invariant. Switch even to fixed-length Montgomery multiplication.
2763 * Change generating and checking of primes so that the error rate of not
2764 being prime depends on the intended use based on the size of the input.
2765 For larger primes this will result in more rounds of Miller-Rabin.
2766 The maximal error rate for primes with more than 1080 bits is lowered
2769 *Kurt Roeckx, Annie Yousar*
2771 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2775 * Add blinding to ECDSA and DSA signatures to protect against side channel
2776 attacks discovered by Keegan Ryan (NCC Group).
2780 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2781 now allow empty (zero character) pass phrases.
2785 * Certificate time validation (X509_cmp_time) enforces stricter
2786 compliance with RFC 5280. Fractional seconds and timezone offsets
2787 are no longer allowed.
2791 * Fixed a text canonicalisation bug in CMS
2793 Where a CMS detached signature is used with text content the text goes
2794 through a canonicalisation process first prior to signing or verifying a
2795 signature. This process strips trailing space at the end of lines, converts
2796 line terminators to CRLF and removes additional trailing line terminators
2797 at the end of a file. A bug in the canonicalisation process meant that
2798 some characters, such as form-feed, were incorrectly treated as whitespace
2799 and removed. This is contrary to the specification (RFC5485). This fix
2800 could mean that detached text data signed with an earlier version of
2801 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2802 signed with a fixed OpenSSL may fail to verify with an earlier version of
2803 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2804 and use the "-binary" flag (for the "cms" command line application) or set
2805 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2809 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2811 * Constructed ASN.1 types with a recursive definition could exceed the stack
2813 Constructed ASN.1 types with a recursive definition (such as can be found
2814 in PKCS7) could eventually exceed the stack given malicious input with
2815 excessive recursion. This could result in a Denial Of Service attack. There
2816 are no such structures used within SSL/TLS that come from untrusted sources
2817 so this is considered safe.
2819 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2825 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2827 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2828 effectively reduced to only comparing the least significant bit of each
2829 byte. This allows an attacker to forge messages that would be considered as
2830 authenticated in an amount of tries lower than that guaranteed by the
2831 security claims of the scheme. The module can only be compiled by the
2832 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2834 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2840 * Add a build target 'build_all_generated', to build all generated files
2841 and only that. This can be used to prepare everything that requires
2842 things like perl for a system that lacks perl and then move everything
2843 to that system and do the rest of the build there.
2847 * Backport SSL_OP_NO_RENGOTIATION
2849 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2850 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2851 changes this is no longer possible in 1.1.0. Therefore the new
2852 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2853 1.1.0 to provide equivalent functionality.
2855 Note that if an application built against 1.1.0h headers (or above) is run
2856 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2857 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2861 * Removed the OS390-Unix config target. It relied on a script that doesn't
2866 * rsaz_1024_mul_avx2 overflow bug on x86_64
2868 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2869 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2870 Analysis suggests that attacks against RSA and DSA as a result of this
2871 defect would be very difficult to perform and are not believed likely.
2872 Attacks against DH1024 are considered just feasible, because most of the
2873 work necessary to deduce information about a private key may be performed
2874 offline. The amount of resources required for such an attack would be
2875 significant. However, for an attack on TLS to be meaningful, the server
2876 would have to share the DH1024 private key among multiple clients, which is
2877 no longer an option since CVE-2016-0701.
2879 This only affects processors that support the AVX2 but not ADX extensions
2880 like Intel Haswell (4th generation).
2882 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2883 was originally found via the OSS-Fuzz project.
2888 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2890 * bn_sqrx8x_internal carry bug on x86_64
2892 There is a carry propagating bug in the x86_64 Montgomery squaring
2893 procedure. No EC algorithms are affected. Analysis suggests that attacks
2894 against RSA and DSA as a result of this defect would be very difficult to
2895 perform and are not believed likely. Attacks against DH are considered just
2896 feasible (although very difficult) because most of the work necessary to
2897 deduce information about a private key may be performed offline. The amount
2898 of resources required for such an attack would be very significant and
2899 likely only accessible to a limited number of attackers. An attacker would
2900 additionally need online access to an unpatched system using the target
2901 private key in a scenario with persistent DH parameters and a private
2902 key that is shared between multiple clients.
2904 This only affects processors that support the BMI1, BMI2 and ADX extensions
2905 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2907 This issue was reported to OpenSSL by the OSS-Fuzz project.
2912 * Malformed X.509 IPAddressFamily could cause OOB read
2914 If an X.509 certificate has a malformed IPAddressFamily extension,
2915 OpenSSL could do a one-byte buffer overread. The most likely result
2916 would be an erroneous display of the certificate in text format.
2918 This issue was reported to OpenSSL by the OSS-Fuzz project.
2923 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2925 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2926 platform rather than 'mingw'.
2930 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2931 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2932 which is the minimum version we support.
2936 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2938 * Encrypt-Then-Mac renegotiation crash
2940 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2941 negotiated where it was not in the original handshake (or vice-versa) then
2942 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2943 and servers are affected.
2945 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2950 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2952 * Truncated packet could crash via OOB read
2954 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2955 cipher is being used, then a truncated packet can cause that host to
2956 perform an out-of-bounds read, usually resulting in a crash.
2958 This issue was reported to OpenSSL by Robert Święcki of Google.
2963 * Bad (EC)DHE parameters cause a client crash
2965 If a malicious server supplies bad parameters for a DHE or ECDHE key
2966 exchange then this can result in the client attempting to dereference a
2967 NULL pointer leading to a client crash. This could be exploited in a Denial
2970 This issue was reported to OpenSSL by Guido Vranken.
2975 * BN_mod_exp may produce incorrect results on x86_64
2977 There is a carry propagating bug in the x86_64 Montgomery squaring
2978 procedure. No EC algorithms are affected. Analysis suggests that attacks
2979 against RSA and DSA as a result of this defect would be very difficult to
2980 perform and are not believed likely. Attacks against DH are considered just
2981 feasible (although very difficult) because most of the work necessary to
2982 deduce information about a private key may be performed offline. The amount
2983 of resources required for such an attack would be very significant and
2984 likely only accessible to a limited number of attackers. An attacker would
2985 additionally need online access to an unpatched system using the target
2986 private key in a scenario with persistent DH parameters and a private
2987 key that is shared between multiple clients. For example this can occur by
2988 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2989 similar to CVE-2015-3193 but must be treated as a separate problem.
2991 This issue was reported to OpenSSL by the OSS-Fuzz project.
2996 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2998 * ChaCha20/Poly1305 heap-buffer-overflow
3000 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3001 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3002 crash. This issue is not considered to be exploitable beyond a DoS.
3004 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3009 * CMS Null dereference
3011 Applications parsing invalid CMS structures can crash with a NULL pointer
3012 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3013 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3014 structure callback if an attempt is made to free certain invalid encodings.
3015 Only CHOICE structures using a callback which do not handle NULL value are
3018 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3023 * Montgomery multiplication may produce incorrect results
3025 There is a carry propagating bug in the Broadwell-specific Montgomery
3026 multiplication procedure that handles input lengths divisible by, but
3027 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3028 and DH private keys are impossible. This is because the subroutine in
3029 question is not used in operations with the private key itself and an input
3030 of the attacker's direct choice. Otherwise the bug can manifest itself as
3031 transient authentication and key negotiation failures or reproducible
3032 erroneous outcome of public-key operations with specially crafted input.
3033 Among EC algorithms only Brainpool P-512 curves are affected and one
3034 presumably can attack ECDH key negotiation. Impact was not analyzed in
3035 detail, because pre-requisites for attack are considered unlikely. Namely
3036 multiple clients have to choose the curve in question and the server has to
3037 share the private key among them, neither of which is default behaviour.
3038 Even then only clients that chose the curve will be affected.
3040 This issue was publicly reported as transient failures and was not
3041 initially recognized as a security issue. Thanks to Richard Morgan for
3042 providing reproducible case.
3047 * Removed automatic addition of RPATH in shared libraries and executables,
3048 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3052 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3054 * Fix Use After Free for large message sizes
3056 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3057 message larger than approx 16k is received then the underlying buffer to
3058 store the incoming message is reallocated and moved. Unfortunately a
3059 dangling pointer to the old location is left which results in an attempt to
3060 write to the previously freed location. This is likely to result in a
3061 crash, however it could potentially lead to execution of arbitrary code.
3063 This issue only affects OpenSSL 1.1.0a.
3065 This issue was reported to OpenSSL by Robert Święcki.
3070 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3072 * OCSP Status Request extension unbounded memory growth
3074 A malicious client can send an excessively large OCSP Status Request
3075 extension. If that client continually requests renegotiation, sending a
3076 large OCSP Status Request extension each time, then there will be unbounded
3077 memory growth on the server. This will eventually lead to a Denial Of
3078 Service attack through memory exhaustion. Servers with a default
3079 configuration are vulnerable even if they do not support OCSP. Builds using
3080 the "no-ocsp" build time option are not affected.
3082 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3087 * SSL_peek() hang on empty record
3089 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3090 sends an empty record. This could be exploited by a malicious peer in a
3091 Denial Of Service attack.
3093 This issue was reported to OpenSSL by Alex Gaynor.
3098 * Excessive allocation of memory in tls_get_message_header() and
3099 dtls1_preprocess_fragment()
3101 A (D)TLS message includes 3 bytes for its length in the header for the
3102 message. This would allow for messages up to 16Mb in length. Messages of
3103 this length are excessive and OpenSSL includes a check to ensure that a
3104 peer is sending reasonably sized messages in order to avoid too much memory
3105 being consumed to service a connection. A flaw in the logic of version
3106 1.1.0 means that memory for the message is allocated too early, prior to
3107 the excessive message length check. Due to way memory is allocated in
3108 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3109 to service a connection. This could lead to a Denial of Service through
3110 memory exhaustion. However, the excessive message length check still takes
3111 place, and this would cause the connection to immediately fail. Assuming
3112 that the application calls SSL_free() on the failed connection in a timely
3113 manner then the 21Mb of allocated memory will then be immediately freed
3114 again. Therefore the excessive memory allocation will be transitory in
3115 nature. This then means that there is only a security impact if:
3117 1) The application does not call SSL_free() in a timely manner in the event
3118 that the connection fails
3120 2) The application is working in a constrained environment where there is
3121 very little free memory
3123 3) The attacker initiates multiple connection attempts such that there are
3124 multiple connections in a state where memory has been allocated for the
3125 connection; SSL_free() has not yet been called; and there is insufficient
3126 memory to service the multiple requests.
3128 Except in the instance of (1) above any Denial Of Service is likely to be
3129 transitory because as soon as the connection fails the memory is
3130 subsequently freed again in the SSL_free() call. However there is an
3131 increased risk during this period of application crashes due to the lack of
3132 memory - which would then mean a more serious Denial of Service.
3134 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3135 (CVE-2016-6307 and CVE-2016-6308)
3139 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3140 had to be removed. Primary reason is that vendor assembler can't
3141 assemble our modules with -KPIC flag. As result it, assembly
3142 support, was not even available as option. But its lack means
3143 lack of side-channel resistant code, which is incompatible with
3144 security by todays standards. Fortunately gcc is readily available
3145 prepackaged option, which we firmly point at...
3149 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3151 * Windows command-line tool supports UTF-8 opt-in option for arguments
3152 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3153 (to any value) allows Windows user to access PKCS#12 file generated
3154 with Windows CryptoAPI and protected with non-ASCII password, as well
3155 as files generated under UTF-8 locale on Linux also protected with
3160 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3161 have been disabled by default and removed from DEFAULT, just like RC4.
3162 See the RC4 item below to re-enable both.
3166 * The method for finding the storage location for the Windows RAND seed file
3167 has changed. First we check %RANDFILE%. If that is not set then we check
3168 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3169 all else fails we fall back to C:\.
3173 * The EVP_EncryptUpdate() function has had its return type changed from void
3174 to int. A return of 0 indicates and error while a return of 1 indicates
3179 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3180 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3181 off the constant time implementation for RSA, DSA and DH have been made
3182 no-ops and deprecated.
3186 * Windows RAND implementation was simplified to only get entropy by
3187 calling CryptGenRandom(). Various other RAND-related tickets
3190 *Joseph Wylie Yandle, Rich Salz*
3192 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3193 and `OPENSSL_LH_`, respectively. The old names are available
3194 with API compatibility. They new names are now completely documented.
3198 * Unify TYPE_up_ref(obj) methods signature.
3199 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3200 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3201 int (instead of void) like all others TYPE_up_ref() methods.
3202 So now these methods also check the return value of CRYPTO_atomic_add(),
3203 and the validity of object reference counter.
3205 *fdasilvayy@gmail.com*
3207 * With Windows Visual Studio builds, the .pdb files are installed
3208 alongside the installed libraries and executables. For a static
3209 library installation, ossl_static.pdb is the associate compiler
3210 generated .pdb file to be used when linking programs.
3214 * Remove openssl.spec. Packaging files belong with the packagers.
3218 * Automatic Darwin/OSX configuration has had a refresh, it will now
3219 recognise x86_64 architectures automatically. You can still decide
3220 to build for a different bitness with the environment variable
3221 KERNEL_BITS (can be 32 or 64), for example:
3223 KERNEL_BITS=32 ./config
3227 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3228 256 bit AES and HMAC with SHA256.
3232 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3236 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3240 * To enable users to have their own config files and build file templates,
3241 Configure looks in the directory indicated by the environment variable
3242 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3243 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3244 name and is used as is.
3248 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3249 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3250 X509_CERT_FILE_CTX was removed.
3254 * "shared" builds are now the default. To create only static libraries use
3255 the "no-shared" Configure option.
3259 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3260 All of these option have not worked for some while and are fundamental
3265 * Make various cleanup routines no-ops and mark them as deprecated. Most
3266 global cleanup functions are no longer required because they are handled
3267 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3268 Explicitly de-initing can cause problems (e.g. where a library that uses
3269 OpenSSL de-inits, but an application is still using it). The affected
3270 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3271 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3272 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3273 COMP_zlib_cleanup().
3277 * --strict-warnings no longer enables runtime debugging options
3278 such as REF_DEBUG. Instead, debug options are automatically
3279 enabled with '--debug' builds.
3281 *Andy Polyakov, Emilia Käsper*
3283 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3284 have been moved out of the public header files. New functions for managing
3285 these have been added.
3289 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3290 objects have been moved out of the public header files. New
3291 functions for managing these have been added.
3295 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3296 have been moved out of the public header files. New functions for managing
3297 these have been added.
3301 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3302 moved out of the public header files. New functions for managing these
3307 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3311 * Removed the mk1mf build scripts.
3315 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3316 it is always safe to #include a header now.
3320 * Removed the aged BC-32 config and all its supporting scripts
3324 * Removed support for Ultrix, Netware, and OS/2.
3328 * Add support for HKDF.
3330 *Alessandro Ghedini*
3332 * Add support for blake2b and blake2s
3336 * Added support for "pipelining". Ciphers that have the
3337 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3338 encryptions/decryptions simultaneously. There are currently no built-in
3339 ciphers with this property but the expectation is that engines will be able
3340 to offer it to significantly improve throughput. Support has been extended
3341 into libssl so that multiple records for a single connection can be
3342 processed in one go (for >=TLS 1.1).
3346 * Added the AFALG engine. This is an async capable engine which is able to
3347 offload work to the Linux kernel. In this initial version it only supports
3348 AES128-CBC. The kernel must be version 4.1.0 or greater.
3352 * OpenSSL now uses a new threading API. It is no longer necessary to
3353 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3354 are two supported threading models: pthreads and windows threads. It is
3355 also possible to configure OpenSSL at compile time for "no-threads". The
3356 old threading API should no longer be used. The functions have been
3357 replaced with "no-op" compatibility macros.
3359 *Alessandro Ghedini, Matt Caswell*
3361 * Modify behavior of ALPN to invoke callback after SNI/servername
3362 callback, such that updates to the SSL_CTX affect ALPN.
3366 * Add SSL_CIPHER queries for authentication and key-exchange.
3370 * Changes to the DEFAULT cipherlist:
3371 - Prefer (EC)DHE handshakes over plain RSA.
3372 - Prefer AEAD ciphers over legacy ciphers.
3373 - Prefer ECDSA over RSA when both certificates are available.
3374 - Prefer TLSv1.2 ciphers/PRF.
3375 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3380 * Change the ECC default curve list to be this, in order: x25519,
3381 secp256r1, secp521r1, secp384r1.
3385 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3386 disabled by default. They can be re-enabled using the
3387 enable-weak-ssl-ciphers option to Configure.
3391 * If the server has ALPN configured, but supports no protocols that the
3392 client advertises, send a fatal "no_application_protocol" alert.
3393 This behaviour is SHALL in RFC 7301, though it isn't universally
3394 implemented by other servers.
3398 * Add X25519 support.
3399 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3400 for public and private key encoding using the format documented in
3401 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3402 key generation and key derivation.
3404 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3409 * Deprecate SRP_VBASE_get_by_user.
3410 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3411 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3412 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3413 seed, even if the seed is configured.
3415 Users should use SRP_VBASE_get1_by_user instead. Note that in
3416 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3417 also that even though configuring the SRP seed attempts to hide
3418 invalid usernames by continuing the handshake with fake
3419 credentials, this behaviour is not constant time and no strong
3420 guarantees are made that the handshake is indistinguishable from
3421 that of a valid user.
3425 * Configuration change; it's now possible to build dynamic engines
3426 without having to build shared libraries and vice versa. This
3427 only applies to the engines in `engines/`, those in `crypto/engine/`
3428 will always be built into libcrypto (i.e. "static").
3430 Building dynamic engines is enabled by default; to disable, use
3431 the configuration option "disable-dynamic-engine".
3433 The only requirements for building dynamic engines are the
3434 presence of the DSO module and building with position independent
3435 code, so they will also automatically be disabled if configuring
3436 with "disable-dso" or "disable-pic".
3438 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3439 are also taken away from openssl/opensslconf.h, as they are
3444 * Configuration change; if there is a known flag to compile
3445 position independent code, it will always be applied on the
3446 libcrypto and libssl object files, and never on the application
3447 object files. This means other libraries that use routines from
3448 libcrypto / libssl can be made into shared libraries regardless
3449 of how OpenSSL was configured.
3451 If this isn't desirable, the configuration options "disable-pic"
3452 or "no-pic" can be used to disable the use of PIC. This will
3453 also disable building shared libraries and dynamic engines.
3457 * Removed JPAKE code. It was experimental and has no wide use.
3461 * The INSTALL_PREFIX Makefile variable has been renamed to
3462 DESTDIR. That makes for less confusion on what this variable
3463 is for. Also, the configuration option --install_prefix is
3468 * Heartbeat for TLS has been removed and is disabled by default
3469 for DTLS; configure with enable-heartbeats. Code that uses the
3470 old #define's might need to be updated.
3472 *Emilia Käsper, Rich Salz*
3474 * Rename REF_CHECK to REF_DEBUG.
3478 * New "unified" build system
3480 The "unified" build system is aimed to be a common system for all
3481 platforms we support. With it comes new support for VMS.
3483 This system builds supports building in a different directory tree
3484 than the source tree. It produces one Makefile (for unix family
3485 or lookalikes), or one descrip.mms (for VMS).
3487 The source of information to make the Makefile / descrip.mms is
3488 small files called 'build.info', holding the necessary
3489 information for each directory with source to compile, and a
3490 template in Configurations, like unix-Makefile.tmpl or
3493 With this change, the library names were also renamed on Windows
3494 and on VMS. They now have names that are closer to the standard
3495 on Unix, and include the major version number, and in certain
3496 cases, the architecture they are built for. See "Notes on shared
3497 libraries" in INSTALL.
3499 We rely heavily on the perl module Text::Template.
3503 * Added support for auto-initialisation and de-initialisation of the library.
3504 OpenSSL no longer requires explicit init or deinit routines to be called,
3505 except in certain circumstances. See the OPENSSL_init_crypto() and
3506 OPENSSL_init_ssl() man pages for further information.
3510 * The arguments to the DTLSv1_listen function have changed. Specifically the
3511 "peer" argument is now expected to be a BIO_ADDR object.
3513 * Rewrite of BIO networking library. The BIO library lacked consistent
3514 support of IPv6, and adding it required some more extensive
3515 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3516 which hold all types of addresses and chains of address information.
3517 It also introduces a new API, with functions like BIO_socket,
3518 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3519 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3520 have been adapted accordingly.
3524 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3529 * CRIME protection: disable compression by default, even if OpenSSL is
3530 compiled with zlib enabled. Applications can still enable compression
3531 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3532 using the SSL_CONF library to configure compression.
3536 * The signature of the session callback configured with
3537 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3538 was explicitly marked as `const unsigned char*` instead of
3543 * Always DPURIFY. Remove the use of uninitialized memory in the
3544 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3548 * Removed many obsolete configuration items, including
3549 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3550 MD2_CHAR, MD2_INT, MD2_LONG
3552 IDEA_SHORT, IDEA_LONG
3553 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3555 *Rich Salz, with advice from Andy Polyakov*
3557 * Many BN internals have been moved to an internal header file.
3559 *Rich Salz with help from Andy Polyakov*
3561 * Configuration and writing out the results from it has changed.
3562 Files such as Makefile include/openssl/opensslconf.h and are now
3563 produced through general templates, such as Makefile.in and
3564 crypto/opensslconf.h.in and some help from the perl module
3567 Also, the center of configuration information is no longer
3568 Makefile. Instead, Configure produces a perl module in
3569 configdata.pm which holds most of the config data (in the hash
3570 table %config), the target data that comes from the target
3571 configuration in one of the `Configurations/*.conf` files (in
3576 * To clarify their intended purposes, the Configure options
3577 --prefix and --openssldir change their semantics, and become more
3578 straightforward and less interdependent.
3580 --prefix shall be used exclusively to give the location INSTALLTOP
3581 where programs, scripts, libraries, include files and manuals are
3582 going to be installed. The default is now /usr/local.
3584 --openssldir shall be used exclusively to give the default
3585 location OPENSSLDIR where certificates, private keys, CRLs are
3586 managed. This is also where the default openssl.cnf gets
3588 If the directory given with this option is a relative path, the
3589 values of both the --prefix value and the --openssldir value will
3590 be combined to become OPENSSLDIR.
3591 The default for --openssldir is INSTALLTOP/ssl.
3593 Anyone who uses --openssldir to specify where OpenSSL is to be
3594 installed MUST change to use --prefix instead.
3598 * The GOST engine was out of date and therefore it has been removed. An up
3599 to date GOST engine is now being maintained in an external repository.
3600 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3601 support for GOST ciphersuites (these are only activated if a GOST engine
3606 * EGD is no longer supported by default; use enable-egd when
3609 *Ben Kaduk and Rich Salz*
3611 * The distribution now has Makefile.in files, which are used to
3612 create Makefile's when Configure is run. *Configure must be run
3613 before trying to build now.*
3617 * The return value for SSL_CIPHER_description() for error conditions
3622 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3624 Obtaining and performing DNSSEC validation of TLSA records is
3625 the application's responsibility. The application provides
3626 the TLSA records of its choice to OpenSSL, and these are then
3627 used to authenticate the peer.
3629 The TLSA records need not even come from DNS. They can, for
3630 example, be used to implement local end-entity certificate or
3631 trust-anchor "pinning", where the "pin" data takes the form
3632 of TLSA records, which can augment or replace verification
3633 based on the usual WebPKI public certification authorities.
3637 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3638 continues to support deprecated interfaces in default builds.
3639 However, applications are strongly advised to compile their
3640 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3641 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3642 or the 1.1.0 releases.
3644 In environments in which all applications have been ported to
3645 not use any deprecated interfaces OpenSSL's Configure script
3646 should be used with the --api=1.1.0 option to entirely remove
3647 support for the deprecated features from the library and
3648 unconditionally disable them in the installed headers.
3649 Essentially the same effect can be achieved with the "no-deprecated"
3650 argument to Configure, except that this will always restrict
3651 the build to just the latest API, rather than a fixed API
3654 As applications are ported to future revisions of the API,
3655 they should update their compile-time OPENSSL_API_COMPAT define
3656 accordingly, but in most cases should be able to continue to
3657 compile with later releases.
3659 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3660 0x10000000L and 0x00908000L, respectively. However those
3661 versions did not support the OPENSSL_API_COMPAT feature, and
3662 so applications are not typically tested for explicit support
3663 of just the undeprecated features of either release.
3667 * Add support for setting the minimum and maximum supported protocol.
3668 It can bet set via the SSL_set_min_proto_version() and
3669 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3670 MaxProtocol. It's recommended to use the new APIs to disable
3671 protocols instead of disabling individual protocols using
3672 SSL_set_options() or SSL_CONF's Protocol. This change also
3673 removes support for disabling TLS 1.2 in the OpenSSL TLS
3674 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3678 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3682 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3683 and integrates ECDSA and ECDH functionality into EC. Implementations can
3684 now redirect key generation and no longer need to convert to or from
3687 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3688 include the ec.h header file instead.
3692 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3693 ciphers who are no longer supported and drops support the ephemeral RSA key
3694 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3698 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3699 opaque. For HMAC_CTX, the following constructors and destructors
3702 HMAC_CTX *HMAC_CTX_new(void);
3703 void HMAC_CTX_free(HMAC_CTX *ctx);
3705 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3706 destroy such methods has been added. See EVP_MD_meth_new(3) and
3707 EVP_CIPHER_meth_new(3) for documentation.
3710 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3711 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3712 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3713 an already created structure.
3714 2) For consistency with the majority of our object creators and
3715 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3716 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3717 for deprecated builds.
3721 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3722 cryptographic operations to be performed asynchronously as long as an
3723 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3724 further details. Libssl has also had this capability integrated with the
3725 introduction of the new mode SSL_MODE_ASYNC and associated error
3726 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3727 pages. This work was developed in partnership with Intel Corp.
3731 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3732 always enabled now. If you want to disable the support you should
3733 exclude it using the list of supported ciphers. This also means that the
3734 "-no_ecdhe" option has been removed from s_server.
3738 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3739 SSL_{CTX_}set1_curves() which can set a list.
3743 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3744 curve you want to support using SSL_{CTX_}set1_curves().
3748 * State machine rewrite. The state machine code has been significantly
3749 refactored in order to remove much duplication of code and solve issues
3750 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3751 further details). This change does have some associated API changes.
3752 Notably the SSL_state() function has been removed and replaced by
3753 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3754 SSL_set_state() has been removed altogether. The previous handshake states
3755 defined in ssl.h and ssl3.h have also been removed.
3759 * All instances of the string "ssleay" in the public API were replaced
3760 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3761 Some error codes related to internal RSA_eay API's were renamed.
3765 * The demo files in crypto/threads were moved to demo/threads.
3769 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3772 *Matt Caswell, Rich Salz*
3774 * New ASN.1 embed macro.
3776 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3777 structure is not allocated: it is part of the parent. That is instead of
3785 This reduces memory fragmentation and make it impossible to accidentally
3786 set a mandatory field to NULL.
3788 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3789 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3790 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3795 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3799 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3800 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3801 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3802 DES and RC4 ciphersuites.
3806 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3807 This changes the decoding behaviour for some invalid messages,
3808 though the change is mostly in the more lenient direction, and
3809 legacy behaviour is preserved as much as possible.
3813 * Fix no-stdio build.
3814 *David Woodhouse <David.Woodhouse@intel.com> and also*
3815 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3817 * New testing framework
3818 The testing framework has been largely rewritten and is now using
3819 perl and the perl modules Test::Harness and an extended variant of
3820 Test::More called OpenSSL::Test to do its work. All test scripts in
3821 test/ have been rewritten into test recipes, and all direct calls to
3822 executables in test/Makefile have become individual recipes using the
3823 simplified testing OpenSSL::Test::Simple.
3825 For documentation on our testing modules, do:
3827 perldoc test/testlib/OpenSSL/Test/Simple.pm
3828 perldoc test/testlib/OpenSSL/Test.pm
3832 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3833 are used; the latter aborts on memory leaks (usually checked on exit).
3834 Some undocumented "set malloc, etc., hooks" functions were removed
3835 and others were changed. All are now documented.
3839 * In DSA_generate_parameters_ex, if the provided seed is too short,
3842 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3844 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3845 from RFC4279, RFC4785, RFC5487, RFC5489.
3847 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3848 original RSA_PSK patch.
3852 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3853 era flag was never set throughout the codebase (only read). Also removed
3854 SSL3_FLAGS_POP_BUFFER which was only used if
3855 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3859 * Changed the default name options in the "ca", "crl", "req" and "x509"
3860 to be "oneline" instead of "compat".
3864 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3865 not aware of clients that still exhibit this bug, and the workaround
3866 hasn't been working properly for a while.
3870 * The return type of BIO_number_read() and BIO_number_written() as well as
3871 the corresponding num_read and num_write members in the BIO structure has
3872 changed from unsigned long to uint64_t. On platforms where an unsigned
3873 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3878 * Given the pervasive nature of TLS extensions it is inadvisable to run
3879 OpenSSL without support for them. It also means that maintaining
3880 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3881 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3885 * Removed support for the two export grade static DH ciphersuites
3886 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3887 were newly added (along with a number of other static DH ciphersuites) to
3888 1.0.2. However the two export ones have *never* worked since they were
3889 introduced. It seems strange in any case to be adding new export
3890 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3894 * Version negotiation has been rewritten. In particular SSLv23_method(),
3895 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3896 and turned into macros which simply call the new preferred function names
3897 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3898 should use the new names instead. Also as part of this change the ssl23.h
3899 header file has been removed.
3903 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3904 code and the associated standard is no longer considered fit-for-purpose.
3908 * RT2547 was closed. When generating a private key, try to make the
3909 output file readable only by the owner. This behavior change might
3910 be noticeable when interacting with other software.
3912 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3917 * Added HTTP GET support to the ocsp command.
3921 * Changed default digest for the dgst and enc commands from MD5 to
3926 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3930 * Added support for TLS extended master secret from
3931 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3932 initial patch which was a great help during development.
3936 * All libssl internal structures have been removed from the public header
3937 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3938 now redundant). Users should not attempt to access internal structures
3939 directly. Instead they should use the provided API functions.
3943 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3944 Access to deprecated functions can be re-enabled by running config with
3945 "enable-deprecated". In addition applications wishing to use deprecated
3946 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3947 will, by default, disable some transitive includes that previously existed
3948 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3952 * Added support for OCB mode. OpenSSL has been granted a patent license
3953 compatible with the OpenSSL license for use of OCB. Details are available
3954 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3955 for OCB can be removed by calling config with no-ocb.
3959 * SSLv2 support has been removed. It still supports receiving a SSLv2
3960 compatible client hello.
3964 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3965 done while fixing the error code for the key-too-small case.
3967 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3969 * CA.sh has been removed; use CA.pl instead.
3973 * Removed old DES API.
3977 * Remove various unsupported platforms:
3983 Sinix/ReliantUNIX RM400
3988 16-bit platforms such as WIN16
3992 * Clean up OPENSSL_NO_xxx #define's
3993 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3994 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3995 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3996 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3997 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3998 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3999 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4000 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4001 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4002 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4006 * Cleaned up dead code
4007 Remove all but one '#ifdef undef' which is to be looked at.
4011 * Clean up calling of xxx_free routines.
4012 Just like free(), fix most of the xxx_free routines to accept
4013 NULL. Remove the non-null checks from callers. Save much code.
4017 * Add secure heap for storage of private keys (when possible).
4018 Add BIO_s_secmem(), CBIGNUM, etc.
4019 Contributed by Akamai Technologies under our Corporate CLA.
4023 * Experimental support for a new, fast, unbiased prime candidate generator,
4024 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4026 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4028 * New output format NSS in the sess_id command line tool. This allows
4029 exporting the session id and the master key in NSS keylog format.
4031 *Martin Kaiser <martin@kaiser.cx>*
4033 * Harmonize version and its documentation. -f flag is used to display
4036 *mancha <mancha1@zoho.com>*
4038 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4039 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4041 *mancha <mancha1@zoho.com>*
4043 * Fix some double frees. These are not thought to be exploitable.
4045 *mancha <mancha1@zoho.com>*
4047 * A missing bounds check in the handling of the TLS heartbeat extension
4048 can be used to reveal up to 64k of memory to a connected client or
4051 Thanks for Neel Mehta of Google Security for discovering this bug and to
4052 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4053 preparing the fix ([CVE-2014-0160])
4055 *Adam Langley, Bodo Moeller*
4057 * Fix for the attack described in the paper "Recovering OpenSSL
4058 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4059 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4060 <http://eprint.iacr.org/2014/140>
4062 Thanks to Yuval Yarom and Naomi Benger for discovering this
4063 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4065 *Yuval Yarom and Naomi Benger*
4067 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4068 this fixes a limitation in previous versions of OpenSSL.
4072 * Experimental encrypt-then-mac support.
4074 Experimental support for encrypt then mac from
4075 draft-gutmann-tls-encrypt-then-mac-02.txt
4077 To enable it set the appropriate extension number (0x42 for the test
4078 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4080 For non-compliant peers (i.e. just about everything) this should have no
4083 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4087 * Add EVP support for key wrapping algorithms, to avoid problems with
4088 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4089 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4090 algorithms and include tests cases.
4094 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4099 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4100 MGF1 digest and OAEP label.
4104 * Make openssl verify return errors.
4106 *Chris Palmer <palmer@google.com> and Ben Laurie*
4108 * New function ASN1_TIME_diff to calculate the difference between two
4109 ASN1_TIME structures or one structure and the current time.
4113 * Update fips_test_suite to support multiple command line options. New
4114 test to induce all self test errors in sequence and check expected
4119 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4120 sign or verify all in one operation.
4124 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4125 test programs and fips_test_suite. Includes functionality to parse
4126 the minimal script output of fipsalgest.pl directly.
4130 * Add authorisation parameter to FIPS_module_mode_set().
4134 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4138 * Use separate DRBG fields for internal and external flags. New function
4139 FIPS_drbg_health_check() to perform on demand health checking. Add
4140 generation tests to fips_test_suite with reduced health check interval to
4141 demonstrate periodic health checking. Add "nodh" option to
4142 fips_test_suite to skip very slow DH test.
4146 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4151 * More extensive health check for DRBG checking many more failure modes.
4152 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4153 combination: call this in fips_test_suite.
4157 * Add support for canonical generation of DSA parameter 'g'. See
4160 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4161 POST to handle HMAC cases.
4165 * Add functions FIPS_module_version() and FIPS_module_version_text()
4166 to return numerical and string versions of the FIPS module number.
4170 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4171 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4172 outside the validated module in the FIPS capable OpenSSL.
4176 * Minor change to DRBG entropy callback semantics. In some cases
4177 there is no multiple of the block length between min_len and
4178 max_len. Allow the callback to return more than max_len bytes
4179 of entropy but discard any extra: it is the callback's responsibility
4180 to ensure that the extra data discarded does not impact the
4181 requested amount of entropy.
4185 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4186 information in FIPS186-3, SP800-57 and SP800-131A.
4190 * CCM support via EVP. Interface is very similar to GCM case except we
4191 must supply all data in one chunk (i.e. no update, final) and the
4192 message length must be supplied if AAD is used. Add algorithm test
4197 * Initial version of POST overhaul. Add POST callback to allow the status
4198 of POST to be monitored and/or failures induced. Modify fips_test_suite
4199 to use callback. Always run all selftests even if one fails.
4203 * XTS support including algorithm test driver in the fips_gcmtest program.
4204 Note: this does increase the maximum key length from 32 to 64 bytes but
4205 there should be no binary compatibility issues as existing applications
4206 will never use XTS mode.
4210 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4211 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4212 performs algorithm blocking for unapproved PRNG types. Also do not
4213 set PRNG type in FIPS_mode_set(): leave this to the application.
4214 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4215 the standard OpenSSL PRNG: set additional data to a date time vector.
4219 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4220 This shouldn't present any incompatibility problems because applications
4221 shouldn't be using these directly and any that are will need to rethink
4222 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4226 * Extensive self tests and health checking required by SP800-90 DRBG.
4227 Remove strength parameter from FIPS_drbg_instantiate and always
4228 instantiate at maximum supported strength.
4232 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4236 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4240 * New function DH_compute_key_padded() to compute a DH key and pad with
4241 leading zeroes if needed: this complies with SP800-56A et al.
4245 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4246 anything, incomplete, subject to change and largely untested at present.
4250 * Modify fipscanisteronly build option to only build the necessary object
4251 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4255 * Add experimental option FIPSSYMS to give all symbols in
4256 fipscanister.o and FIPS or fips prefix. This will avoid
4257 conflicts with future versions of OpenSSL. Add perl script
4258 util/fipsas.pl to preprocess assembly language source files
4259 and rename any affected symbols.
4263 * Add selftest checks and algorithm block of non-fips algorithms in
4264 FIPS mode. Remove DES2 from selftests.
4268 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4269 return internal method without any ENGINE dependencies. Add new
4270 tiny fips sign and verify functions.
4274 * New build option no-ec2m to disable characteristic 2 code.
4278 * New build option "fipscanisteronly". This only builds fipscanister.o
4279 and (currently) associated fips utilities. Uses the file Makefile.fips
4280 instead of Makefile.org as the prototype.
4284 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4285 Update fips_gcmtest to use IV generator.
4289 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4290 setting output buffer to NULL. The `*Final` function must be
4291 called although it will not retrieve any additional data. The tag
4292 can be set or retrieved with a ctrl. The IV length is by default 12
4293 bytes (96 bits) but can be set to an alternative value. If the IV
4294 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4299 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4300 underlying do_cipher function handles all cipher semantics itself
4301 including padding and finalisation. This is useful if (for example)
4302 an ENGINE cipher handles block padding itself. The behaviour of
4303 do_cipher is subtly changed if this flag is set: the return value
4304 is the number of characters written to the output buffer (zero is
4305 no longer an error code) or a negative error code. Also if the
4306 input buffer is NULL and length 0 finalisation should be performed.
4310 * If a candidate issuer certificate is already part of the constructed
4311 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4315 * Improve forward-security support: add functions
4317 void SSL_CTX_set_not_resumable_session_callback(
4318 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4319 void SSL_set_not_resumable_session_callback(
4320 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4322 for use by SSL/TLS servers; the callback function will be called whenever a
4323 new session is created, and gets to decide whether the session may be
4324 cached to make it resumable (return 0) or not (return 1). (As by the
4325 SSL/TLS protocol specifications, the session_id sent by the server will be
4326 empty to indicate that the session is not resumable; also, the server will
4327 not generate RFC 4507 (RFC 5077) session tickets.)
4329 A simple reasonable callback implementation is to return is_forward_secure.
4330 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4331 by the SSL/TLS server library, indicating whether it can provide forward
4334 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4336 * New -verify_name option in command line utilities to set verification
4341 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4342 Add CMAC pkey methods.
4346 * Experimental renegotiation in s_server -www mode. If the client
4347 browses /reneg connection is renegotiated. If /renegcert it is
4348 renegotiated requesting a certificate.
4352 * Add an "external" session cache for debugging purposes to s_server. This
4353 should help trace issues which normally are only apparent in deployed
4354 multi-process servers.
4358 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4359 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4360 BIO_set_cipher() and some obscure PEM functions were changed so they
4361 can now return an error. The RAND changes required a change to the
4362 RAND_METHOD structure.
4366 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4367 a gcc attribute to warn if the result of a function is ignored. This
4368 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4369 whose return value is often ignored.
4373 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4374 These allow SCTs (signed certificate timestamps) to be requested and
4375 validated when establishing a connection.
4377 *Rob Percival <robpercival@google.com>*
4382 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4384 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4385 used even when parsing explicit parameters, when loading a encoded key
4386 or calling `EC_GROUP_new_from_ecpkparameters()`/
4387 `EC_GROUP_new_from_ecparameters()`.
4388 This prevents bypass of security hardening and performance gains,
4389 especially for curves with specialized EC_METHODs.
4390 By default, if a key encoded with explicit parameters is loaded and later
4391 encoded, the output is still encoded with explicit parameters, even if
4392 internally a "named" EC_GROUP is used for computation.
4396 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4397 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4398 NULL. After this change, only the cofactor parameter can be NULL. It also
4399 does some minimal sanity checks on the passed order.
4404 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4405 An attack is simple, if the first CMS_recipientInfo is valid but the
4406 second CMS_recipientInfo is chosen ciphertext. If the second
4407 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4408 encryption key will be replaced by garbage, and the message cannot be
4409 decoded, but if the RSA decryption fails, the correct encryption key is
4410 used and the recipient will not notice the attack.
4411 As a work around for this potential attack the length of the decrypted
4412 key must be equal to the cipher default key length, in case the
4413 certifiate is not given and all recipientInfo are tried out.
4414 The old behaviour can be re-enabled in the CMS code by setting the
4415 CMS_DEBUG_DECRYPT flag.
4420 * Document issue with installation paths in diverse Windows builds
4422 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4423 binaries and run-time config file.
4428 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4430 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4431 This changes the size when using the `genpkey` command when no size is given.
4432 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4433 generation commands to use 2048 bits by default.
4437 * Add FIPS support for Android Arm 64-bit
4439 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4440 Module in Version 2.0.10. For some reason, the corresponding target
4441 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4442 built with FIPS support on Android Arm 64-bit. This omission has been
4445 *Matthias St. Pierre*
4447 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4449 * 0-byte record padding oracle
4451 If an application encounters a fatal protocol error and then calls
4452 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4453 then OpenSSL can respond differently to the calling application if a 0 byte
4454 record is received with invalid padding compared to if a 0 byte record is
4455 received with an invalid MAC. If the application then behaves differently
4456 based on that in a way that is detectable to the remote peer, then this
4457 amounts to a padding oracle that could be used to decrypt data.
4459 In order for this to be exploitable "non-stitched" ciphersuites must be in
4460 use. Stitched ciphersuites are optimised implementations of certain
4461 commonly used ciphersuites. Also the application must call SSL_shutdown()
4462 twice even if a protocol error has occurred (applications should not do
4463 this but some do anyway).
4465 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4466 Aviram, with additional investigation by Steven Collison and Andrew
4467 Hourselt. It was reported to OpenSSL on 10th December 2018.
4472 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4476 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4478 * Microarchitecture timing vulnerability in ECC scalar multiplication
4480 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4481 shown to be vulnerable to a microarchitecture timing side channel attack.
4482 An attacker with sufficient access to mount local timing attacks during
4483 ECDSA signature generation could recover the private key.
4485 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4486 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4492 * Timing vulnerability in DSA signature generation
4494 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4495 timing side channel attack. An attacker could use variations in the signing
4496 algorithm to recover the private key.
4498 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4503 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4504 Module, accidentally introduced while backporting security fixes from the
4505 development branch and hindering the use of ECC in FIPS mode.
4509 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4511 * Client DoS due to large DH parameter
4513 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4514 malicious server can send a very large prime value to the client. This will
4515 cause the client to spend an unreasonably long period of time generating a
4516 key for this prime resulting in a hang until the client has finished. This
4517 could be exploited in a Denial Of Service attack.
4519 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4524 * Cache timing vulnerability in RSA Key Generation
4526 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4527 a cache timing side channel attack. An attacker with sufficient access to
4528 mount cache timing attacks during the RSA key generation process could
4529 recover the private key.
4531 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4532 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4537 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4538 parameter is no longer accepted, as it leads to a corrupt table. NULL
4539 pem_str is reserved for alias entries only.
4543 * Revert blinding in ECDSA sign and instead make problematic addition
4544 length-invariant. Switch even to fixed-length Montgomery multiplication.
4548 * Change generating and checking of primes so that the error rate of not
4549 being prime depends on the intended use based on the size of the input.
4550 For larger primes this will result in more rounds of Miller-Rabin.
4551 The maximal error rate for primes with more than 1080 bits is lowered
4554 *Kurt Roeckx, Annie Yousar*
4556 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4560 * Add blinding to ECDSA and DSA signatures to protect against side channel
4561 attacks discovered by Keegan Ryan (NCC Group).
4565 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4566 now allow empty (zero character) pass phrases.
4570 * Certificate time validation (X509_cmp_time) enforces stricter
4571 compliance with RFC 5280. Fractional seconds and timezone offsets
4572 are no longer allowed.
4576 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4578 * Constructed ASN.1 types with a recursive definition could exceed the stack
4580 Constructed ASN.1 types with a recursive definition (such as can be found
4581 in PKCS7) could eventually exceed the stack given malicious input with
4582 excessive recursion. This could result in a Denial Of Service attack. There
4583 are no such structures used within SSL/TLS that come from untrusted sources
4584 so this is considered safe.
4586 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4592 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4594 * Read/write after SSL object in error state
4596 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4597 mechanism. The intent was that if a fatal error occurred during a handshake
4598 then OpenSSL would move into the error state and would immediately fail if
4599 you attempted to continue the handshake. This works as designed for the
4600 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4601 SSL_connect()), however due to a bug it does not work correctly if
4602 SSL_read() or SSL_write() is called directly. In that scenario, if the
4603 handshake fails then a fatal error will be returned in the initial function
4604 call. If SSL_read()/SSL_write() is subsequently called by the application
4605 for the same SSL object then it will succeed and the data is passed without
4606 being decrypted/encrypted directly from the SSL/TLS record layer.
4608 In order to exploit this issue an application bug would have to be present
4609 that resulted in a call to SSL_read()/SSL_write() being issued after having
4610 already received a fatal error.
4612 This issue was reported to OpenSSL by David Benjamin (Google).
4617 * rsaz_1024_mul_avx2 overflow bug on x86_64
4619 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4620 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4621 Analysis suggests that attacks against RSA and DSA as a result of this
4622 defect would be very difficult to perform and are not believed likely.
4623 Attacks against DH1024 are considered just feasible, because most of the
4624 work necessary to deduce information about a private key may be performed
4625 offline. The amount of resources required for such an attack would be
4626 significant. However, for an attack on TLS to be meaningful, the server
4627 would have to share the DH1024 private key among multiple clients, which is
4628 no longer an option since CVE-2016-0701.
4630 This only affects processors that support the AVX2 but not ADX extensions
4631 like Intel Haswell (4th generation).
4633 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4634 was originally found via the OSS-Fuzz project.
4639 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4641 * bn_sqrx8x_internal carry bug on x86_64
4643 There is a carry propagating bug in the x86_64 Montgomery squaring
4644 procedure. No EC algorithms are affected. Analysis suggests that attacks
4645 against RSA and DSA as a result of this defect would be very difficult to
4646 perform and are not believed likely. Attacks against DH are considered just
4647 feasible (although very difficult) because most of the work necessary to
4648 deduce information about a private key may be performed offline. The amount
4649 of resources required for such an attack would be very significant and
4650 likely only accessible to a limited number of attackers. An attacker would
4651 additionally need online access to an unpatched system using the target
4652 private key in a scenario with persistent DH parameters and a private
4653 key that is shared between multiple clients.
4655 This only affects processors that support the BMI1, BMI2 and ADX extensions
4656 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4658 This issue was reported to OpenSSL by the OSS-Fuzz project.
4663 * Malformed X.509 IPAddressFamily could cause OOB read
4665 If an X.509 certificate has a malformed IPAddressFamily extension,
4666 OpenSSL could do a one-byte buffer overread. The most likely result
4667 would be an erroneous display of the certificate in text format.
4669 This issue was reported to OpenSSL by the OSS-Fuzz project.
4673 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4675 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4676 platform rather than 'mingw'.
4680 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4682 * Truncated packet could crash via OOB read
4684 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4685 cipher is being used, then a truncated packet can cause that host to
4686 perform an out-of-bounds read, usually resulting in a crash.
4688 This issue was reported to OpenSSL by Robert Święcki of Google.
4693 * BN_mod_exp may produce incorrect results on x86_64
4695 There is a carry propagating bug in the x86_64 Montgomery squaring
4696 procedure. No EC algorithms are affected. Analysis suggests that attacks
4697 against RSA and DSA as a result of this defect would be very difficult to
4698 perform and are not believed likely. Attacks against DH are considered just
4699 feasible (although very difficult) because most of the work necessary to
4700 deduce information about a private key may be performed offline. The amount
4701 of resources required for such an attack would be very significant and
4702 likely only accessible to a limited number of attackers. An attacker would
4703 additionally need online access to an unpatched system using the target
4704 private key in a scenario with persistent DH parameters and a private
4705 key that is shared between multiple clients. For example this can occur by
4706 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4707 similar to CVE-2015-3193 but must be treated as a separate problem.
4709 This issue was reported to OpenSSL by the OSS-Fuzz project.
4714 * Montgomery multiplication may produce incorrect results
4716 There is a carry propagating bug in the Broadwell-specific Montgomery
4717 multiplication procedure that handles input lengths divisible by, but
4718 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4719 and DH private keys are impossible. This is because the subroutine in
4720 question is not used in operations with the private key itself and an input
4721 of the attacker's direct choice. Otherwise the bug can manifest itself as
4722 transient authentication and key negotiation failures or reproducible
4723 erroneous outcome of public-key operations with specially crafted input.
4724 Among EC algorithms only Brainpool P-512 curves are affected and one
4725 presumably can attack ECDH key negotiation. Impact was not analyzed in
4726 detail, because pre-requisites for attack are considered unlikely. Namely
4727 multiple clients have to choose the curve in question and the server has to
4728 share the private key among them, neither of which is default behaviour.
4729 Even then only clients that chose the curve will be affected.
4731 This issue was publicly reported as transient failures and was not
4732 initially recognized as a security issue. Thanks to Richard Morgan for
4733 providing reproducible case.
4738 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4739 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4740 prevent issues where no progress is being made and the peer continually
4741 sends unrecognised record types, using up resources processing them.
4745 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4747 * Missing CRL sanity check
4749 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4750 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4751 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4753 This issue only affects the OpenSSL 1.0.2i
4758 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4760 * OCSP Status Request extension unbounded memory growth
4762 A malicious client can send an excessively large OCSP Status Request
4763 extension. If that client continually requests renegotiation, sending a
4764 large OCSP Status Request extension each time, then there will be unbounded
4765 memory growth on the server. This will eventually lead to a Denial Of
4766 Service attack through memory exhaustion. Servers with a default
4767 configuration are vulnerable even if they do not support OCSP. Builds using
4768 the "no-ocsp" build time option are not affected.
4770 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4775 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4778 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4784 * OOB write in MDC2_Update()
4786 An overflow can occur in MDC2_Update() either if called directly or
4787 through the EVP_DigestUpdate() function using MDC2. If an attacker
4788 is able to supply very large amounts of input data after a previous
4789 call to EVP_EncryptUpdate() with a partial block then a length check
4790 can overflow resulting in a heap corruption.
4792 The amount of data needed is comparable to SIZE_MAX which is impractical
4795 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4800 * Malformed SHA512 ticket DoS
4802 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4803 DoS attack where a malformed ticket will result in an OOB read which will
4806 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4807 a custom server callback and ticket lookup mechanism.
4809 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4814 * OOB write in BN_bn2dec()
4816 The function BN_bn2dec() does not check the return value of BN_div_word().
4817 This can cause an OOB write if an application uses this function with an
4818 overly large BIGNUM. This could be a problem if an overly large certificate
4819 or CRL is printed out from an untrusted source. TLS is not affected because
4820 record limits will reject an oversized certificate before it is parsed.
4822 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4827 * OOB read in TS_OBJ_print_bio()
4829 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4830 the total length the OID text representation would use and not the amount
4831 of data written. This will result in OOB reads when large OIDs are
4834 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4839 * Pointer arithmetic undefined behaviour
4841 Avoid some undefined pointer arithmetic
4843 A common idiom in the codebase is to check limits in the following manner:
4846 Where "p" points to some malloc'd data of SIZE bytes and
4849 "len" here could be from some externally supplied data (e.g. from a TLS
4852 The rules of C pointer arithmetic are such that "p + len" is only well
4853 defined where len <= SIZE. Therefore the above idiom is actually
4854 undefined behaviour.
4856 For example this could cause problems if some malloc implementation
4857 provides an address for "p" such that "p + len" actually overflows for
4858 values of len that are too big and therefore p + len < limit.
4860 This issue was reported to OpenSSL by Guido Vranken
4865 * Constant time flag not preserved in DSA signing
4867 Operations in the DSA signing algorithm should run in constant time in
4868 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4869 implementation means that a non-constant time codepath is followed for
4870 certain operations. This has been demonstrated through a cache-timing
4871 attack to be sufficient for an attacker to recover the private DSA key.
4873 This issue was reported by César Pereida (Aalto University), Billy Brumley
4874 (Tampere University of Technology), and Yuval Yarom (The University of
4875 Adelaide and NICTA).
4880 * DTLS buffered message DoS
4882 In a DTLS connection where handshake messages are delivered out-of-order
4883 those messages that OpenSSL is not yet ready to process will be buffered
4884 for later use. Under certain circumstances, a flaw in the logic means that
4885 those messages do not get removed from the buffer even though the handshake
4886 has been completed. An attacker could force up to approx. 15 messages to
4887 remain in the buffer when they are no longer required. These messages will
4888 be cleared when the DTLS connection is closed. The default maximum size for
4889 a message is 100k. Therefore the attacker could force an additional 1500k
4890 to be consumed per connection. By opening many simulataneous connections an
4891 attacker could cause a DoS attack through memory exhaustion.
4893 This issue was reported to OpenSSL by Quan Luo.
4898 * DTLS replay protection DoS
4900 A flaw in the DTLS replay attack protection mechanism means that records
4901 that arrive for future epochs update the replay protection "window" before
4902 the MAC for the record has been validated. This could be exploited by an
4903 attacker by sending a record for the next epoch (which does not have to
4904 decrypt or have a valid MAC), with a very large sequence number. This means
4905 that all subsequent legitimate packets are dropped causing a denial of
4906 service for a specific DTLS connection.
4908 This issue was reported to OpenSSL by the OCAP audit team.
4913 * Certificate message OOB reads
4915 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4916 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4917 theoretical DoS risk but this has not been observed in practice on common
4920 The messages affected are client certificate, client certificate request
4921 and server certificate. As a result the attack can only be performed
4922 against a client or a server which enables client authentication.
4924 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4929 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4931 * Prevent padding oracle in AES-NI CBC MAC check
4933 A MITM attacker can use a padding oracle attack to decrypt traffic
4934 when the connection uses an AES CBC cipher and the server support
4937 This issue was introduced as part of the fix for Lucky 13 padding
4938 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4939 constant time by making sure that always the same bytes are read and
4940 compared against either the MAC or padding bytes. But it no longer
4941 checked that there was enough data to have both the MAC and padding
4944 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4948 * Fix EVP_EncodeUpdate overflow
4950 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4951 Base64 encoding of binary data. If an attacker is able to supply very large
4952 amounts of input data then a length check can overflow resulting in a heap
4955 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4956 the `PEM_write_bio*` family of functions. These are mainly used within the
4957 OpenSSL command line applications, so any application which processes data
4958 from an untrusted source and outputs it as a PEM file should be considered
4959 vulnerable to this issue. User applications that call these APIs directly
4960 with large amounts of untrusted data may also be vulnerable.
4962 This issue was reported by Guido Vranken.
4967 * Fix EVP_EncryptUpdate overflow
4969 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4970 is able to supply very large amounts of input data after a previous call to
4971 EVP_EncryptUpdate() with a partial block then a length check can overflow
4972 resulting in a heap corruption. Following an analysis of all OpenSSL
4973 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4974 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4975 the first called function after an EVP_EncryptInit(), and therefore that
4976 specific call must be safe. The second form is where the length passed to
4977 EVP_EncryptUpdate() can be seen from the code to be some small value and
4978 therefore there is no possibility of an overflow. Since all instances are
4979 one of these two forms, it is believed that there can be no overflows in
4980 internal code due to this problem. It should be noted that
4981 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4982 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4983 of these calls have also been analysed too and it is believed there are no
4984 instances in internal usage where an overflow could occur.
4986 This issue was reported by Guido Vranken.
4991 * Prevent ASN.1 BIO excessive memory allocation
4993 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4994 a short invalid encoding can cause allocation of large amounts of memory
4995 potentially consuming excessive resources or exhausting memory.
4997 Any application parsing untrusted data through d2i BIO functions is
4998 affected. The memory based functions such as d2i_X509() are *not* affected.
4999 Since the memory based functions are used by the TLS library, TLS
5000 applications are not affected.
5002 This issue was reported by Brian Carpenter.
5009 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5010 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5011 in arbitrary stack data being returned in the buffer.
5013 This issue was reported by Guido Vranken.
5018 * Modify behavior of ALPN to invoke callback after SNI/servername
5019 callback, such that updates to the SSL_CTX affect ALPN.
5023 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5028 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5029 methods are enabled and ssl2 is disabled the methods return NULL.
5033 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5035 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5036 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5037 provide any "EXPORT" or "LOW" strength ciphers.
5041 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5042 is by default disabled at build-time. Builds that are not configured with
5043 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5044 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5045 will need to explicitly call either of:
5047 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5049 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5051 as appropriate. Even if either of those is used, or the application
5052 explicitly uses the version-specific SSLv2_method() or its client and
5053 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5054 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5055 ciphers, and SSLv2 56-bit DES are no longer available.
5060 * Fix a double-free in DSA code
5062 A double free bug was discovered when OpenSSL parses malformed DSA private
5063 keys and could lead to a DoS attack or memory corruption for applications
5064 that receive DSA private keys from untrusted sources. This scenario is
5067 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5073 * Disable SRP fake user seed to address a server memory leak.
5075 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5077 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5078 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5079 was changed to ignore the "fake user" SRP seed, even if the seed
5082 Users should use SRP_VBASE_get1_by_user instead. Note that in
5083 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5084 also that even though configuring the SRP seed attempts to hide
5085 invalid usernames by continuing the handshake with fake
5086 credentials, this behaviour is not constant time and no strong
5087 guarantees are made that the handshake is indistinguishable from
5088 that of a valid user.
5093 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5095 In the BN_hex2bn function the number of hex digits is calculated using an
5096 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5097 large values of `i` this can result in `bn_expand` not allocating any
5098 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5099 field as NULL leading to a subsequent NULL ptr deref. For very large values
5100 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5101 In this case memory is allocated to the internal BIGNUM data field, but it
5102 is insufficiently sized leading to heap corruption. A similar issue exists
5103 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5104 is ever called by user applications with very large untrusted hex/dec data.
5105 This is anticipated to be a rare occurrence.
5107 All OpenSSL internal usage of these functions use data that is not expected
5108 to be untrusted, e.g. config file data or application command line
5109 arguments. If user developed applications generate config file data based
5110 on untrusted data then it is possible that this could also lead to security
5111 consequences. This is also anticipated to be rare.
5113 This issue was reported to OpenSSL by Guido Vranken.
5118 * Fix memory issues in `BIO_*printf` functions
5120 The internal `fmtstr` function used in processing a "%s" format string in
5121 the `BIO_*printf` functions could overflow while calculating the length of a
5122 string and cause an OOB read when printing very long strings.
5124 Additionally the internal `doapr_outch` function can attempt to write to an
5125 OOB memory location (at an offset from the NULL pointer) in the event of a
5126 memory allocation failure. In 1.0.2 and below this could be caused where
5127 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5128 could be in processing a very long "%s" format string. Memory leaks can
5131 The first issue may mask the second issue dependent on compiler behaviour.
5132 These problems could enable attacks where large amounts of untrusted data
5133 is passed to the `BIO_*printf` functions. If applications use these functions
5134 in this way then they could be vulnerable. OpenSSL itself uses these
5135 functions when printing out human-readable dumps of ASN.1 data. Therefore
5136 applications that print this data could be vulnerable if the data is from
5137 untrusted sources. OpenSSL command line applications could also be
5138 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5139 as command line arguments.
5141 Libssl is not considered directly vulnerable. Additionally certificates etc
5142 received via remote connections via libssl are also unlikely to be able to
5143 trigger these issues because of message size limits enforced within libssl.
5145 This issue was reported to OpenSSL Guido Vranken.
5150 * Side channel attack on modular exponentiation
5152 A side-channel attack was found which makes use of cache-bank conflicts on
5153 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5154 of RSA keys. The ability to exploit this issue is limited as it relies on
5155 an attacker who has control of code in a thread running on the same
5156 hyper-threaded core as the victim thread which is performing decryptions.
5158 This issue was reported to OpenSSL by Yuval Yarom, The University of
5159 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5160 Nadia Heninger, University of Pennsylvania with more information at
5161 <http://cachebleed.info>.
5166 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5167 if no keysize is specified with default_bits. This fixes an
5168 omission in an earlier change that changed all RSA/DSA key generation
5169 commands to use 2048 bits by default.
5173 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5175 * DH small subgroups
5177 Historically OpenSSL only ever generated DH parameters based on "safe"
5178 primes. More recently (in version 1.0.2) support was provided for
5179 generating X9.42 style parameter files such as those required for RFC 5114
5180 support. The primes used in such files may not be "safe". Where an
5181 application is using DH configured with parameters based on primes that are
5182 not "safe" then an attacker could use this fact to find a peer's private
5183 DH exponent. This attack requires that the attacker complete multiple
5184 handshakes in which the peer uses the same private DH exponent. For example
5185 this could be used to discover a TLS server's private DH exponent if it's
5186 reusing the private DH exponent or it's using a static DH ciphersuite.
5188 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5189 TLS. It is not on by default. If the option is not set then the server
5190 reuses the same private DH exponent for the life of the server process and
5191 would be vulnerable to this attack. It is believed that many popular
5192 applications do set this option and would therefore not be at risk.
5194 The fix for this issue adds an additional check where a "q" parameter is
5195 available (as is the case in X9.42 based parameters). This detects the
5196 only known attack, and is the only possible defense for static DH
5197 ciphersuites. This could have some performance impact.
5199 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5200 default and cannot be disabled. This could have some performance impact.
5202 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5207 * SSLv2 doesn't block disabled ciphers
5209 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5210 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5211 been disabled, provided that the SSLv2 protocol was not also disabled via
5214 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5215 and Sebastian Schinzel.
5220 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5222 * BN_mod_exp may produce incorrect results on x86_64
5224 There is a carry propagating bug in the x86_64 Montgomery squaring
5225 procedure. No EC algorithms are affected. Analysis suggests that attacks
5226 against RSA and DSA as a result of this defect would be very difficult to
5227 perform and are not believed likely. Attacks against DH are considered just
5228 feasible (although very difficult) because most of the work necessary to
5229 deduce information about a private key may be performed offline. The amount
5230 of resources required for such an attack would be very significant and
5231 likely only accessible to a limited number of attackers. An attacker would
5232 additionally need online access to an unpatched system using the target
5233 private key in a scenario with persistent DH parameters and a private
5234 key that is shared between multiple clients. For example this can occur by
5235 default in OpenSSL DHE based SSL/TLS ciphersuites.
5237 This issue was reported to OpenSSL by Hanno Böck.
5242 * Certificate verify crash with missing PSS parameter
5244 The signature verification routines will crash with a NULL pointer
5245 dereference if presented with an ASN.1 signature using the RSA PSS
5246 algorithm and absent mask generation function parameter. Since these
5247 routines are used to verify certificate signature algorithms this can be
5248 used to crash any certificate verification operation and exploited in a
5249 DoS attack. Any application which performs certificate verification is
5250 vulnerable including OpenSSL clients and servers which enable client
5253 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5258 * X509_ATTRIBUTE memory leak
5260 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5261 memory. This structure is used by the PKCS#7 and CMS routines so any
5262 application which reads PKCS#7 or CMS data from untrusted sources is
5263 affected. SSL/TLS is not affected.
5265 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5271 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5272 This changes the decoding behaviour for some invalid messages,
5273 though the change is mostly in the more lenient direction, and
5274 legacy behaviour is preserved as much as possible.
5278 * In DSA_generate_parameters_ex, if the provided seed is too short,
5281 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5283 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5285 * Alternate chains certificate forgery
5287 During certificate verification, OpenSSL will attempt to find an
5288 alternative certificate chain if the first attempt to build such a chain
5289 fails. An error in the implementation of this logic can mean that an
5290 attacker could cause certain checks on untrusted certificates to be
5291 bypassed, such as the CA flag, enabling them to use a valid leaf
5292 certificate to act as a CA and "issue" an invalid certificate.
5294 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5299 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5301 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5302 incompatibility in the handling of HMAC. The previous ABI has now been
5307 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5309 * Malformed ECParameters causes infinite loop
5311 When processing an ECParameters structure OpenSSL enters an infinite loop
5312 if the curve specified is over a specially malformed binary polynomial
5315 This can be used to perform denial of service against any
5316 system which processes public keys, certificate requests or
5317 certificates. This includes TLS clients and TLS servers with
5318 client authentication enabled.
5320 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5325 * Exploitable out-of-bounds read in X509_cmp_time
5327 X509_cmp_time does not properly check the length of the ASN1_TIME
5328 string and can read a few bytes out of bounds. In addition,
5329 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5332 An attacker can use this to craft malformed certificates and CRLs of
5333 various sizes and potentially cause a segmentation fault, resulting in
5334 a DoS on applications that verify certificates or CRLs. TLS clients
5335 that verify CRLs are affected. TLS clients and servers with client
5336 authentication enabled may be affected if they use custom verification
5339 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5340 independently by Hanno Böck.
5345 * PKCS7 crash with missing EnvelopedContent
5347 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5348 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5349 with missing content and trigger a NULL pointer dereference on parsing.
5351 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5352 structures from untrusted sources are affected. OpenSSL clients and
5353 servers are not affected.
5355 This issue was reported to OpenSSL by Michal Zalewski (Google).
5360 * CMS verify infinite loop with unknown hash function
5362 When verifying a signedData message the CMS code can enter an infinite loop
5363 if presented with an unknown hash function OID. This can be used to perform
5364 denial of service against any system which verifies signedData messages using
5366 This issue was reported to OpenSSL by Johannes Bauer.
5371 * Race condition handling NewSessionTicket
5373 If a NewSessionTicket is received by a multi-threaded client when attempting to
5374 reuse a previous ticket then a race condition can occur potentially leading to
5375 a double free of the ticket data.
5380 * Only support 256-bit or stronger elliptic curves with the
5381 'ecdh_auto' setting (server) or by default (client). Of supported
5382 curves, prefer P-256 (both).
5386 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5388 * ClientHello sigalgs DoS fix
5390 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5391 invalid signature algorithms extension a NULL pointer dereference will
5392 occur. This can be exploited in a DoS attack against the server.
5394 This issue was was reported to OpenSSL by David Ramos of Stanford
5398 *Stephen Henson and Matt Caswell*
5400 * Multiblock corrupted pointer fix
5402 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5403 feature only applies on 64 bit x86 architecture platforms that support AES
5404 NI instructions. A defect in the implementation of "multiblock" can cause
5405 OpenSSL's internal write buffer to become incorrectly set to NULL when
5406 using non-blocking IO. Typically, when the user application is using a
5407 socket BIO for writing, this will only result in a failed connection.
5408 However if some other BIO is used then it is likely that a segmentation
5409 fault will be triggered, thus enabling a potential DoS attack.
5411 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5416 * Segmentation fault in DTLSv1_listen fix
5418 The DTLSv1_listen function is intended to be stateless and processes the
5419 initial ClientHello from many peers. It is common for user code to loop
5420 over the call to DTLSv1_listen until a valid ClientHello is received with
5421 an associated cookie. A defect in the implementation of DTLSv1_listen means
5422 that state is preserved in the SSL object from one invocation to the next
5423 that can lead to a segmentation fault. Errors processing the initial
5424 ClientHello can trigger this scenario. An example of such an error could be
5425 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5428 This issue was reported to OpenSSL by Per Allansson.
5433 * Segmentation fault in ASN1_TYPE_cmp fix
5435 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5436 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5437 certificate signature algorithm consistency this can be used to crash any
5438 certificate verification operation and exploited in a DoS attack. Any
5439 application which performs certificate verification is vulnerable including
5440 OpenSSL clients and servers which enable client authentication.
5445 * Segmentation fault for invalid PSS parameters fix
5447 The signature verification routines will crash with a NULL pointer
5448 dereference if presented with an ASN.1 signature using the RSA PSS
5449 algorithm and invalid parameters. Since these routines are used to verify
5450 certificate signature algorithms this can be used to crash any
5451 certificate verification operation and exploited in a DoS attack. Any
5452 application which performs certificate verification is vulnerable including
5453 OpenSSL clients and servers which enable client authentication.
5455 This issue was was reported to OpenSSL by Brian Carpenter.
5460 * ASN.1 structure reuse memory corruption fix
5462 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5463 memory corruption via an invalid write. Such reuse is and has been
5464 strongly discouraged and is believed to be rare.
5466 Applications that parse structures containing CHOICE or ANY DEFINED BY
5467 components may be affected. Certificate parsing (d2i_X509 and related
5468 functions) are however not affected. OpenSSL clients and servers are
5474 * PKCS7 NULL pointer dereferences fix
5476 The PKCS#7 parsing code does not handle missing outer ContentInfo
5477 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5478 missing content and trigger a NULL pointer dereference on parsing.
5480 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5481 otherwise parse PKCS#7 structures from untrusted sources are
5482 affected. OpenSSL clients and servers are not affected.
5484 This issue was reported to OpenSSL by Michal Zalewski (Google).
5489 * DoS via reachable assert in SSLv2 servers fix
5491 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5492 servers that both support SSLv2 and enable export cipher suites by sending
5493 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5495 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5496 (OpenSSL development team).
5501 * Empty CKE with client auth and DHE fix
5503 If client auth is used then a server can seg fault in the event of a DHE
5504 ciphersuite being selected and a zero length ClientKeyExchange message
5505 being sent by the client. This could be exploited in a DoS attack.
5510 * Handshake with unseeded PRNG fix
5512 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5513 with an unseeded PRNG. The conditions are:
5514 - The client is on a platform where the PRNG has not been seeded
5515 automatically, and the user has not seeded manually
5516 - A protocol specific client method version has been used (i.e. not
5517 SSL_client_methodv23)
5518 - A ciphersuite is used that does not require additional random data from
5519 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5521 If the handshake succeeds then the client random that has been used will
5522 have been generated from a PRNG with insufficient entropy and therefore the
5523 output may be predictable.
5525 For example using the following command with an unseeded openssl will
5526 succeed on an unpatched platform:
5528 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5533 * Use After Free following d2i_ECPrivatekey error fix
5535 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5536 could cause a use after free condition. This, in turn, could cause a double
5537 free in several private key parsing functions (such as d2i_PrivateKey
5538 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5539 for applications that receive EC private keys from untrusted
5540 sources. This scenario is considered rare.
5542 This issue was discovered by the BoringSSL project and fixed in their
5548 * X509_to_X509_REQ NULL pointer deref fix
5550 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5551 the certificate key is invalid. This function is rarely used in practice.
5553 This issue was discovered by Brian Carpenter.
5558 * Removed the export ciphers from the DEFAULT ciphers
5562 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5564 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5565 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5566 So far those who have to target multiple platforms would compromise
5567 and argue that binary targeting say ARMv5 would still execute on
5568 ARMv8. "Universal" build resolves this compromise by providing
5569 near-optimal performance even on newer platforms.
5573 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5574 (other platforms pending).
5576 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5578 * Add support for the SignedCertificateTimestampList certificate and
5579 OCSP response extensions from RFC6962.
5583 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5584 for corner cases. (Certain input points at infinity could lead to
5585 bogus results, with non-infinity inputs mapped to infinity too.)
5589 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5590 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5591 common cases are optimized and there still is room for further
5592 improvements. Vector Permutation AES for Altivec is also added.
5596 * Add support for little-endian ppc64 Linux target.
5598 *Marcelo Cerri (IBM)*
5600 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5601 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5602 are optimized and there still is room for further improvements.
5603 Both 32- and 64-bit modes are supported.
5605 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5607 * Improved ARMv7 NEON support.
5611 * Support for SPARC Architecture 2011 crypto extensions, first
5612 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5613 SHA256/512, MD5, GHASH and modular exponentiation.
5615 *Andy Polyakov, David Miller*
5617 * Accelerated modular exponentiation for Intel processors, a.k.a.
5620 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5622 * Support for new and upcoming Intel processors, including AVX2,
5623 BMI and SHA ISA extensions. This includes additional "stitched"
5624 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5627 This work was sponsored by Intel Corp.
5631 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5632 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5633 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5637 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5638 this fixes a limitation in previous versions of OpenSSL.
5642 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5643 MGF1 digest and OAEP label.
5647 * Add EVP support for key wrapping algorithms, to avoid problems with
5648 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5649 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5650 algorithms and include tests cases.
5654 * Add functions to allocate and set the fields of an ECDSA_METHOD
5657 *Douglas E. Engert, Steve Henson*
5659 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5660 difference in days and seconds between two tm or ASN1_TIME structures.
5664 * Add -rev test option to s_server to just reverse order of characters
5665 received by client and send back to server. Also prints an abbreviated
5666 summary of the connection parameters.
5670 * New option -brief for s_client and s_server to print out a brief summary
5671 of connection parameters.
5675 * Add callbacks for arbitrary TLS extensions.
5677 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5679 * New option -crl_download in several openssl utilities to download CRLs
5680 from CRLDP extension in certificates.
5684 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5688 * New function X509_CRL_diff to generate a delta CRL from the difference
5689 of two full CRLs. Add support to "crl" utility.
5693 * New functions to set lookup_crls function and to retrieve
5694 X509_STORE from X509_STORE_CTX.
5698 * Print out deprecated issuer and subject unique ID fields in
5703 * Extend OCSP I/O functions so they can be used for simple general purpose
5704 HTTP as well as OCSP. New wrapper function which can be used to download
5705 CRLs using the OCSP API.
5709 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5713 * `SSL_CONF*` functions. These provide a common framework for application
5714 configuration using configuration files or command lines.
5718 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5719 message callback and prints the results. Needs compile time option
5720 "enable-ssl-trace". New options to s_client and s_server to enable
5725 * New ctrl and macro to retrieve supported points extensions.
5726 Print out extension in s_server and s_client.
5730 * New functions to retrieve certificate signature and signature
5735 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5740 * New Suite B modes for TLS code. These use and enforce the requirements
5741 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5742 only use Suite B curves. The Suite B modes can be set by using the
5743 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5747 * New chain verification flags for Suite B levels of security. Check
5748 algorithms are acceptable when flags are set in X509_verify_cert.
5752 * Make tls1_check_chain return a set of flags indicating checks passed
5753 by a certificate chain. Add additional tests to handle client
5754 certificates: checks for matching certificate type and issuer name
5759 * If an attempt is made to use a signature algorithm not in the peer
5760 preference list abort the handshake. If client has no suitable
5761 signature algorithms in response to a certificate request do not
5762 use the certificate.
5766 * If server EC tmp key is not in client preference list abort handshake.
5770 * Add support for certificate stores in CERT structure. This makes it
5771 possible to have different stores per SSL structure or one store in
5772 the parent SSL_CTX. Include distinct stores for certificate chain
5773 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5774 to build and store a certificate chain in CERT structure: returning
5775 an error if the chain cannot be built: this will allow applications
5776 to test if a chain is correctly configured.
5778 Note: if the CERT based stores are not set then the parent SSL_CTX
5779 store is used to retain compatibility with existing behaviour.
5783 * New function ssl_set_client_disabled to set a ciphersuite disabled
5784 mask based on the current session, check mask when sending client
5785 hello and checking the requested ciphersuite.
5789 * New ctrls to retrieve and set certificate types in a certificate
5790 request message. Print out received values in s_client. If certificate
5791 types is not set with custom values set sensible values based on
5792 supported signature algorithms.
5796 * Support for distinct client and server supported signature algorithms.
5800 * Add certificate callback. If set this is called whenever a certificate
5801 is required by client or server. An application can decide which
5802 certificate chain to present based on arbitrary criteria: for example
5803 supported signature algorithms. Add very simple example to s_server.
5804 This fixes many of the problems and restrictions of the existing client
5805 certificate callback: for example you can now clear an existing
5806 certificate and specify the whole chain.
5810 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5811 the certificate can be used for (if anything). Set valid_flags field
5812 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5813 to have similar checks in it.
5815 Add new "cert_flags" field to CERT structure and include a "strict mode".
5816 This enforces some TLS certificate requirements (such as only permitting
5817 certificate signature algorithms contained in the supported algorithms
5818 extension) which some implementations ignore: this option should be used
5819 with caution as it could cause interoperability issues.
5823 * Update and tidy signature algorithm extension processing. Work out
5824 shared signature algorithms based on preferences and peer algorithms
5825 and print them out in s_client and s_server. Abort handshake if no
5826 shared signature algorithms.
5830 * Add new functions to allow customised supported signature algorithms
5831 for SSL and SSL_CTX structures. Add options to s_client and s_server
5836 * New function SSL_certs_clear() to delete all references to certificates
5837 from an SSL structure. Before this once a certificate had been added
5838 it couldn't be removed.
5842 * Integrate hostname, email address and IP address checking with certificate
5843 verification. New verify options supporting checking in openssl utility.
5847 * Fixes and wildcard matching support to hostname and email checking
5848 functions. Add manual page.
5850 *Florian Weimer (Red Hat Product Security Team)*
5852 * New functions to check a hostname email or IP address against a
5853 certificate. Add options x509 utility to print results of checks against
5858 * Fix OCSP checking.
5860 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5862 * Initial experimental support for explicitly trusted non-root CAs.
5863 OpenSSL still tries to build a complete chain to a root but if an
5864 intermediate CA has a trust setting included that is used. The first
5865 setting is used: whether to trust (e.g., -addtrust option to the x509
5870 * Add -trusted_first option which attempts to find certificates in the
5871 trusted store even if an untrusted chain is also supplied.
5875 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5876 platform support for Linux and Android.
5880 * Support for linux-x32, ILP32 environment in x86_64 framework.
5884 * Experimental multi-implementation support for FIPS capable OpenSSL.
5885 When in FIPS mode the approved implementations are used as normal,
5886 when not in FIPS mode the internal unapproved versions are used instead.
5887 This means that the FIPS capable OpenSSL isn't forced to use the
5888 (often lower performance) FIPS implementations outside FIPS mode.
5892 * Transparently support X9.42 DH parameters when calling
5893 PEM_read_bio_DHparameters. This means existing applications can handle
5894 the new parameter format automatically.
5898 * Initial experimental support for X9.42 DH parameter format: mainly
5899 to support use of 'q' parameter for RFC5114 parameters.
5903 * Add DH parameters from RFC5114 including test data to dhtest.
5907 * Support for automatic EC temporary key parameter selection. If enabled
5908 the most preferred EC parameters are automatically used instead of
5909 hardcoded fixed parameters. Now a server just has to call:
5910 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5911 support ECDH and use the most appropriate parameters.
5915 * Enhance and tidy EC curve and point format TLS extension code. Use
5916 static structures instead of allocation if default values are used.
5917 New ctrls to set curves we wish to support and to retrieve shared curves.
5918 Print out shared curves in s_server. New options to s_server and s_client
5919 to set list of supported curves.
5923 * New ctrls to retrieve supported signature algorithms and
5924 supported curve values as an array of NIDs. Extend openssl utility
5925 to print out received values.
5929 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5930 between NIDs and the more common NIST names such as "P-256". Enhance
5931 ecparam utility and ECC method to recognise the NIST names for curves.
5935 * Enhance SSL/TLS certificate chain handling to support different
5936 chains for each certificate instead of one chain in the parent SSL_CTX.
5940 * Support for fixed DH ciphersuite client authentication: where both
5941 server and client use DH certificates with common parameters.
5945 * Support for fixed DH ciphersuites: those requiring DH server
5950 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5952 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5953 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5954 X509_CINF_get_signature were reverted post internal team review.
5959 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5961 * OCSP Status Request extension unbounded memory growth
5963 A malicious client can send an excessively large OCSP Status Request
5964 extension. If that client continually requests renegotiation, sending a
5965 large OCSP Status Request extension each time, then there will be unbounded
5966 memory growth on the server. This will eventually lead to a Denial Of
5967 Service attack through memory exhaustion. Servers with a default
5968 configuration are vulnerable even if they do not support OCSP. Builds using
5969 the "no-ocsp" build time option are not affected.
5971 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5976 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5979 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5985 * OOB write in MDC2_Update()
5987 An overflow can occur in MDC2_Update() either if called directly or
5988 through the EVP_DigestUpdate() function using MDC2. If an attacker
5989 is able to supply very large amounts of input data after a previous
5990 call to EVP_EncryptUpdate() with a partial block then a length check
5991 can overflow resulting in a heap corruption.
5993 The amount of data needed is comparable to SIZE_MAX which is impractical
5996 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6001 * Malformed SHA512 ticket DoS
6003 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6004 DoS attack where a malformed ticket will result in an OOB read which will
6007 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6008 a custom server callback and ticket lookup mechanism.
6010 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6015 * OOB write in BN_bn2dec()
6017 The function BN_bn2dec() does not check the return value of BN_div_word().
6018 This can cause an OOB write if an application uses this function with an
6019 overly large BIGNUM. This could be a problem if an overly large certificate
6020 or CRL is printed out from an untrusted source. TLS is not affected because
6021 record limits will reject an oversized certificate before it is parsed.
6023 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6028 * OOB read in TS_OBJ_print_bio()
6030 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6031 the total length the OID text representation would use and not the amount
6032 of data written. This will result in OOB reads when large OIDs are
6035 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6040 * Pointer arithmetic undefined behaviour
6042 Avoid some undefined pointer arithmetic
6044 A common idiom in the codebase is to check limits in the following manner:
6047 Where "p" points to some malloc'd data of SIZE bytes and
6050 "len" here could be from some externally supplied data (e.g. from a TLS
6053 The rules of C pointer arithmetic are such that "p + len" is only well
6054 defined where len <= SIZE. Therefore the above idiom is actually
6055 undefined behaviour.
6057 For example this could cause problems if some malloc implementation
6058 provides an address for "p" such that "p + len" actually overflows for
6059 values of len that are too big and therefore p + len < limit.
6061 This issue was reported to OpenSSL by Guido Vranken
6066 * Constant time flag not preserved in DSA signing
6068 Operations in the DSA signing algorithm should run in constant time in
6069 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6070 implementation means that a non-constant time codepath is followed for
6071 certain operations. This has been demonstrated through a cache-timing
6072 attack to be sufficient for an attacker to recover the private DSA key.
6074 This issue was reported by César Pereida (Aalto University), Billy Brumley
6075 (Tampere University of Technology), and Yuval Yarom (The University of
6076 Adelaide and NICTA).
6081 * DTLS buffered message DoS
6083 In a DTLS connection where handshake messages are delivered out-of-order
6084 those messages that OpenSSL is not yet ready to process will be buffered
6085 for later use. Under certain circumstances, a flaw in the logic means that
6086 those messages do not get removed from the buffer even though the handshake
6087 has been completed. An attacker could force up to approx. 15 messages to
6088 remain in the buffer when they are no longer required. These messages will
6089 be cleared when the DTLS connection is closed. The default maximum size for
6090 a message is 100k. Therefore the attacker could force an additional 1500k
6091 to be consumed per connection. By opening many simulataneous connections an
6092 attacker could cause a DoS attack through memory exhaustion.
6094 This issue was reported to OpenSSL by Quan Luo.
6099 * DTLS replay protection DoS
6101 A flaw in the DTLS replay attack protection mechanism means that records
6102 that arrive for future epochs update the replay protection "window" before
6103 the MAC for the record has been validated. This could be exploited by an
6104 attacker by sending a record for the next epoch (which does not have to
6105 decrypt or have a valid MAC), with a very large sequence number. This means
6106 that all subsequent legitimate packets are dropped causing a denial of
6107 service for a specific DTLS connection.
6109 This issue was reported to OpenSSL by the OCAP audit team.
6114 * Certificate message OOB reads
6116 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6117 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6118 theoretical DoS risk but this has not been observed in practice on common
6121 The messages affected are client certificate, client certificate request
6122 and server certificate. As a result the attack can only be performed
6123 against a client or a server which enables client authentication.
6125 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6130 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6132 * Prevent padding oracle in AES-NI CBC MAC check
6134 A MITM attacker can use a padding oracle attack to decrypt traffic
6135 when the connection uses an AES CBC cipher and the server support
6138 This issue was introduced as part of the fix for Lucky 13 padding
6139 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6140 constant time by making sure that always the same bytes are read and
6141 compared against either the MAC or padding bytes. But it no longer
6142 checked that there was enough data to have both the MAC and padding
6145 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6150 * Fix EVP_EncodeUpdate overflow
6152 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6153 Base64 encoding of binary data. If an attacker is able to supply very large
6154 amounts of input data then a length check can overflow resulting in a heap
6157 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6158 the `PEM_write_bio*` family of functions. These are mainly used within the
6159 OpenSSL command line applications, so any application which processes data
6160 from an untrusted source and outputs it as a PEM file should be considered
6161 vulnerable to this issue. User applications that call these APIs directly
6162 with large amounts of untrusted data may also be vulnerable.
6164 This issue was reported by Guido Vranken.
6169 * Fix EVP_EncryptUpdate overflow
6171 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6172 is able to supply very large amounts of input data after a previous call to
6173 EVP_EncryptUpdate() with a partial block then a length check can overflow
6174 resulting in a heap corruption. Following an analysis of all OpenSSL
6175 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6176 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6177 the first called function after an EVP_EncryptInit(), and therefore that
6178 specific call must be safe. The second form is where the length passed to
6179 EVP_EncryptUpdate() can be seen from the code to be some small value and
6180 therefore there is no possibility of an overflow. Since all instances are
6181 one of these two forms, it is believed that there can be no overflows in
6182 internal code due to this problem. It should be noted that
6183 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6184 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6185 of these calls have also been analysed too and it is believed there are no
6186 instances in internal usage where an overflow could occur.
6188 This issue was reported by Guido Vranken.
6193 * Prevent ASN.1 BIO excessive memory allocation
6195 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6196 a short invalid encoding can casuse allocation of large amounts of memory
6197 potentially consuming excessive resources or exhausting memory.
6199 Any application parsing untrusted data through d2i BIO functions is
6200 affected. The memory based functions such as d2i_X509() are *not* affected.
6201 Since the memory based functions are used by the TLS library, TLS
6202 applications are not affected.
6204 This issue was reported by Brian Carpenter.
6211 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6212 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6213 in arbitrary stack data being returned in the buffer.
6215 This issue was reported by Guido Vranken.
6220 * Modify behavior of ALPN to invoke callback after SNI/servername
6221 callback, such that updates to the SSL_CTX affect ALPN.
6225 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6230 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6231 methods are enabled and ssl2 is disabled the methods return NULL.
6235 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6237 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6238 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6239 provide any "EXPORT" or "LOW" strength ciphers.
6243 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6244 is by default disabled at build-time. Builds that are not configured with
6245 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6246 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6247 will need to explicitly call either of:
6249 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6251 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6253 as appropriate. Even if either of those is used, or the application
6254 explicitly uses the version-specific SSLv2_method() or its client and
6255 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6256 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6257 ciphers, and SSLv2 56-bit DES are no longer available.
6262 * Fix a double-free in DSA code
6264 A double free bug was discovered when OpenSSL parses malformed DSA private
6265 keys and could lead to a DoS attack or memory corruption for applications
6266 that receive DSA private keys from untrusted sources. This scenario is
6269 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6275 * Disable SRP fake user seed to address a server memory leak.
6277 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6279 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6280 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6281 was changed to ignore the "fake user" SRP seed, even if the seed
6284 Users should use SRP_VBASE_get1_by_user instead. Note that in
6285 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6286 also that even though configuring the SRP seed attempts to hide
6287 invalid usernames by continuing the handshake with fake
6288 credentials, this behaviour is not constant time and no strong
6289 guarantees are made that the handshake is indistinguishable from
6290 that of a valid user.
6295 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6297 In the BN_hex2bn function the number of hex digits is calculated using an
6298 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6299 large values of `i` this can result in `bn_expand` not allocating any
6300 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6301 field as NULL leading to a subsequent NULL ptr deref. For very large values
6302 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6303 In this case memory is allocated to the internal BIGNUM data field, but it
6304 is insufficiently sized leading to heap corruption. A similar issue exists
6305 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6306 is ever called by user applications with very large untrusted hex/dec data.
6307 This is anticipated to be a rare occurrence.
6309 All OpenSSL internal usage of these functions use data that is not expected
6310 to be untrusted, e.g. config file data or application command line
6311 arguments. If user developed applications generate config file data based
6312 on untrusted data then it is possible that this could also lead to security
6313 consequences. This is also anticipated to be rare.
6315 This issue was reported to OpenSSL by Guido Vranken.
6320 * Fix memory issues in `BIO_*printf` functions
6322 The internal `fmtstr` function used in processing a "%s" format string in
6323 the `BIO_*printf` functions could overflow while calculating the length of a
6324 string and cause an OOB read when printing very long strings.
6326 Additionally the internal `doapr_outch` function can attempt to write to an
6327 OOB memory location (at an offset from the NULL pointer) in the event of a
6328 memory allocation failure. In 1.0.2 and below this could be caused where
6329 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6330 could be in processing a very long "%s" format string. Memory leaks can
6333 The first issue may mask the second issue dependent on compiler behaviour.
6334 These problems could enable attacks where large amounts of untrusted data
6335 is passed to the `BIO_*printf` functions. If applications use these functions
6336 in this way then they could be vulnerable. OpenSSL itself uses these
6337 functions when printing out human-readable dumps of ASN.1 data. Therefore
6338 applications that print this data could be vulnerable if the data is from
6339 untrusted sources. OpenSSL command line applications could also be
6340 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6341 as command line arguments.
6343 Libssl is not considered directly vulnerable. Additionally certificates etc
6344 received via remote connections via libssl are also unlikely to be able to
6345 trigger these issues because of message size limits enforced within libssl.
6347 This issue was reported to OpenSSL Guido Vranken.
6352 * Side channel attack on modular exponentiation
6354 A side-channel attack was found which makes use of cache-bank conflicts on
6355 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6356 of RSA keys. The ability to exploit this issue is limited as it relies on
6357 an attacker who has control of code in a thread running on the same
6358 hyper-threaded core as the victim thread which is performing decryptions.
6360 This issue was reported to OpenSSL by Yuval Yarom, The University of
6361 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6362 Nadia Heninger, University of Pennsylvania with more information at
6363 <http://cachebleed.info>.
6368 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6369 if no keysize is specified with default_bits. This fixes an
6370 omission in an earlier change that changed all RSA/DSA key generation
6371 commands to use 2048 bits by default.
6375 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6377 * Protection for DH small subgroup attacks
6379 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6380 switched on by default and cannot be disabled. This could have some
6385 * SSLv2 doesn't block disabled ciphers
6387 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6388 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6389 been disabled, provided that the SSLv2 protocol was not also disabled via
6392 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6393 and Sebastian Schinzel.
6398 * Reject DH handshakes with parameters shorter than 1024 bits.
6402 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6404 * Certificate verify crash with missing PSS parameter
6406 The signature verification routines will crash with a NULL pointer
6407 dereference if presented with an ASN.1 signature using the RSA PSS
6408 algorithm and absent mask generation function parameter. Since these
6409 routines are used to verify certificate signature algorithms this can be
6410 used to crash any certificate verification operation and exploited in a
6411 DoS attack. Any application which performs certificate verification is
6412 vulnerable including OpenSSL clients and servers which enable client
6415 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6420 * X509_ATTRIBUTE memory leak
6422 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6423 memory. This structure is used by the PKCS#7 and CMS routines so any
6424 application which reads PKCS#7 or CMS data from untrusted sources is
6425 affected. SSL/TLS is not affected.
6427 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6433 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6434 This changes the decoding behaviour for some invalid messages,
6435 though the change is mostly in the more lenient direction, and
6436 legacy behaviour is preserved as much as possible.
6440 * In DSA_generate_parameters_ex, if the provided seed is too short,
6441 use a random seed, as already documented.
6443 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6445 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6447 * Alternate chains certificate forgery
6449 During certificate verfification, OpenSSL will attempt to find an
6450 alternative certificate chain if the first attempt to build such a chain
6451 fails. An error in the implementation of this logic can mean that an
6452 attacker could cause certain checks on untrusted certificates to be
6453 bypassed, such as the CA flag, enabling them to use a valid leaf
6454 certificate to act as a CA and "issue" an invalid certificate.
6456 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6462 * Race condition handling PSK identify hint
6464 If PSK identity hints are received by a multi-threaded client then
6465 the values are wrongly updated in the parent SSL_CTX structure. This can
6466 result in a race condition potentially leading to a double free of the
6472 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6474 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6475 incompatibility in the handling of HMAC. The previous ABI has now been
6478 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6480 * Malformed ECParameters causes infinite loop
6482 When processing an ECParameters structure OpenSSL enters an infinite loop
6483 if the curve specified is over a specially malformed binary polynomial
6486 This can be used to perform denial of service against any
6487 system which processes public keys, certificate requests or
6488 certificates. This includes TLS clients and TLS servers with
6489 client authentication enabled.
6491 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6496 * Exploitable out-of-bounds read in X509_cmp_time
6498 X509_cmp_time does not properly check the length of the ASN1_TIME
6499 string and can read a few bytes out of bounds. In addition,
6500 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6503 An attacker can use this to craft malformed certificates and CRLs of
6504 various sizes and potentially cause a segmentation fault, resulting in
6505 a DoS on applications that verify certificates or CRLs. TLS clients
6506 that verify CRLs are affected. TLS clients and servers with client
6507 authentication enabled may be affected if they use custom verification
6510 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6511 independently by Hanno Böck.
6516 * PKCS7 crash with missing EnvelopedContent
6518 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6519 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6520 with missing content and trigger a NULL pointer dereference on parsing.
6522 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6523 structures from untrusted sources are affected. OpenSSL clients and
6524 servers are not affected.
6526 This issue was reported to OpenSSL by Michal Zalewski (Google).
6531 * CMS verify infinite loop with unknown hash function
6533 When verifying a signedData message the CMS code can enter an infinite loop
6534 if presented with an unknown hash function OID. This can be used to perform
6535 denial of service against any system which verifies signedData messages using
6537 This issue was reported to OpenSSL by Johannes Bauer.
6542 * Race condition handling NewSessionTicket
6544 If a NewSessionTicket is received by a multi-threaded client when attempting to
6545 reuse a previous ticket then a race condition can occur potentially leading to
6546 a double free of the ticket data.
6551 * Reject DH handshakes with parameters shorter than 768 bits.
6553 *Kurt Roeckx and Emilia Kasper*
6555 * dhparam: generate 2048-bit parameters by default.
6557 *Kurt Roeckx and Emilia Kasper*
6559 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6561 * Segmentation fault in ASN1_TYPE_cmp fix
6563 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6564 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6565 certificate signature algorithm consistency this can be used to crash any
6566 certificate verification operation and exploited in a DoS attack. Any
6567 application which performs certificate verification is vulnerable including
6568 OpenSSL clients and servers which enable client authentication.
6573 * ASN.1 structure reuse memory corruption fix
6575 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6576 memory corruption via an invalid write. Such reuse is and has been
6577 strongly discouraged and is believed to be rare.
6579 Applications that parse structures containing CHOICE or ANY DEFINED BY
6580 components may be affected. Certificate parsing (d2i_X509 and related
6581 functions) are however not affected. OpenSSL clients and servers are
6587 * PKCS7 NULL pointer dereferences fix
6589 The PKCS#7 parsing code does not handle missing outer ContentInfo
6590 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6591 missing content and trigger a NULL pointer dereference on parsing.
6593 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6594 otherwise parse PKCS#7 structures from untrusted sources are
6595 affected. OpenSSL clients and servers are not affected.
6597 This issue was reported to OpenSSL by Michal Zalewski (Google).
6602 * DoS via reachable assert in SSLv2 servers fix
6604 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6605 servers that both support SSLv2 and enable export cipher suites by sending
6606 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6608 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6609 (OpenSSL development team).
6614 * Use After Free following d2i_ECPrivatekey error fix
6616 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6617 could cause a use after free condition. This, in turn, could cause a double
6618 free in several private key parsing functions (such as d2i_PrivateKey
6619 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6620 for applications that receive EC private keys from untrusted
6621 sources. This scenario is considered rare.
6623 This issue was discovered by the BoringSSL project and fixed in their
6629 * X509_to_X509_REQ NULL pointer deref fix
6631 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6632 the certificate key is invalid. This function is rarely used in practice.
6634 This issue was discovered by Brian Carpenter.
6639 * Removed the export ciphers from the DEFAULT ciphers
6643 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6645 * Build fixes for the Windows and OpenVMS platforms
6647 *Matt Caswell and Richard Levitte*
6649 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6651 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6652 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6653 dereference. This could lead to a Denial Of Service attack. Thanks to
6654 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6659 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6660 dtls1_buffer_record function under certain conditions. In particular this
6661 could occur if an attacker sent repeated DTLS records with the same
6662 sequence number but for the next epoch. The memory leak could be exploited
6663 by an attacker in a Denial of Service attack through memory exhaustion.
6664 Thanks to Chris Mueller for reporting this issue.
6669 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6670 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6671 method would be set to NULL which could later result in a NULL pointer
6672 dereference. Thanks to Frank Schmirler for reporting this issue.
6677 * Abort handshake if server key exchange message is omitted for ephemeral
6680 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6681 reporting this issue.
6686 * Remove non-export ephemeral RSA code on client and server. This code
6687 violated the TLS standard by allowing the use of temporary RSA keys in
6688 non-export ciphersuites and could be used by a server to effectively
6689 downgrade the RSA key length used to a value smaller than the server
6690 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6691 INRIA or reporting this issue.
6696 * Fixed issue where DH client certificates are accepted without verification.
6697 An OpenSSL server will accept a DH certificate for client authentication
6698 without the certificate verify message. This effectively allows a client to
6699 authenticate without the use of a private key. This only affects servers
6700 which trust a client certificate authority which issues certificates
6701 containing DH keys: these are extremely rare and hardly ever encountered.
6702 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6708 * Ensure that the session ID context of an SSL is updated when its
6709 SSL_CTX is updated via SSL_set_SSL_CTX.
6711 The session ID context is typically set from the parent SSL_CTX,
6712 and can vary with the CTX.
6716 * Fix various certificate fingerprint issues.
6718 By using non-DER or invalid encodings outside the signed portion of a
6719 certificate the fingerprint can be changed without breaking the signature.
6720 Although no details of the signed portion of the certificate can be changed
6721 this can cause problems with some applications: e.g. those using the
6722 certificate fingerprint for blacklists.
6724 1. Reject signatures with non zero unused bits.
6726 If the BIT STRING containing the signature has non zero unused bits reject
6727 the signature. All current signature algorithms require zero unused bits.
6729 2. Check certificate algorithm consistency.
6731 Check the AlgorithmIdentifier inside TBS matches the one in the
6732 certificate signature. NB: this will result in signature failure
6733 errors for some broken certificates.
6735 Thanks to Konrad Kraszewski from Google for reporting this issue.
6737 3. Check DSA/ECDSA signatures use DER.
6739 Re-encode DSA/ECDSA signatures and compare with the original received
6740 signature. Return an error if there is a mismatch.
6742 This will reject various cases including garbage after signature
6743 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6744 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6745 (negative or with leading zeroes).
6747 Further analysis was conducted and fixes were developed by Stephen Henson
6748 of the OpenSSL core team.
6754 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6755 results on some platforms, including x86_64. This bug occurs at random
6756 with a very low probability, and is not known to be exploitable in any
6757 way, though its exact impact is difficult to determine. Thanks to Pieter
6758 Wuille (Blockstream) who reported this issue and also suggested an initial
6759 fix. Further analysis was conducted by the OpenSSL development team and
6760 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6761 the OpenSSL core team.
6766 * Do not resume sessions on the server if the negotiated protocol
6767 version does not match the session's version. Resuming with a different
6768 version, while not strictly forbidden by the RFC, is of questionable
6769 sanity and breaks all known clients.
6771 *David Benjamin, Emilia Käsper*
6773 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6774 early CCS messages during renegotiation. (Note that because
6775 renegotiation is encrypted, this early CCS was not exploitable.)
6779 * Tighten client-side session ticket handling during renegotiation:
6780 ensure that the client only accepts a session ticket if the server sends
6781 the extension anew in the ServerHello. Previously, a TLS client would
6782 reuse the old extension state and thus accept a session ticket if one was
6783 announced in the initial ServerHello.
6785 Similarly, ensure that the client requires a session ticket if one
6786 was advertised in the ServerHello. Previously, a TLS client would
6787 ignore a missing NewSessionTicket message.
6791 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6795 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6796 sends a carefully crafted handshake message, to cause OpenSSL to fail
6797 to free up to 64k of memory causing a memory leak. This could be
6798 exploited in a Denial Of Service attack. This issue affects OpenSSL
6799 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6800 whether SRTP is used or configured. Implementations of OpenSSL that
6801 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6803 The fix was developed by the OpenSSL team.
6808 * Session Ticket Memory Leak.
6810 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6811 integrity of that ticket is first verified. In the event of a session
6812 ticket integrity check failing, OpenSSL will fail to free memory
6813 causing a memory leak. By sending a large number of invalid session
6814 tickets an attacker could exploit this issue in a Denial Of Service
6820 * Build option no-ssl3 is incomplete.
6822 When OpenSSL is configured with "no-ssl3" as a build option, servers
6823 could accept and complete a SSL 3.0 handshake, and clients could be
6824 configured to send them.
6827 *Akamai and the OpenSSL team*
6829 * Add support for TLS_FALLBACK_SCSV.
6830 Client applications doing fallback retries should call
6831 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6834 *Adam Langley, Bodo Moeller*
6836 * Add additional DigestInfo checks.
6838 Re-encode DigestInto in DER and check against the original when
6839 verifying RSA signature: this will reject any improperly encoded
6840 DigestInfo structures.
6842 Note: this is a precautionary measure and no attacks are currently known.
6846 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6848 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6849 SRP code can be overrun an internal buffer. Add sanity check that
6850 g, A, B < N to SRP code.
6852 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6853 Group for discovering this issue.
6858 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6859 TLS 1.0 instead of higher protocol versions when the ClientHello message
6860 is badly fragmented. This allows a man-in-the-middle attacker to force a
6861 downgrade to TLS 1.0 even if both the server and the client support a
6862 higher protocol version, by modifying the client's TLS records.
6864 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6865 researching this issue.
6870 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6871 to a denial of service attack. A malicious server can crash the client
6872 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6873 ciphersuite and sending carefully crafted handshake messages.
6875 Thanks to Felix Gröbert (Google) for discovering and researching this
6881 * By sending carefully crafted DTLS packets an attacker could cause openssl
6882 to leak memory. This can be exploited through a Denial of Service attack.
6883 Thanks to Adam Langley for discovering and researching this issue.
6888 * An attacker can force openssl to consume large amounts of memory whilst
6889 processing DTLS handshake messages. This can be exploited through a
6890 Denial of Service attack.
6891 Thanks to Adam Langley for discovering and researching this issue.
6896 * An attacker can force an error condition which causes openssl to crash
6897 whilst processing DTLS packets due to memory being freed twice. This
6898 can be exploited through a Denial of Service attack.
6899 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6905 * If a multithreaded client connects to a malicious server using a resumed
6906 session and the server sends an ec point format extension it could write
6907 up to 255 bytes to freed memory.
6909 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6915 * A malicious server can crash an OpenSSL client with a null pointer
6916 dereference (read) by specifying an SRP ciphersuite even though it was not
6917 properly negotiated with the client. This can be exploited through a
6918 Denial of Service attack.
6920 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6921 discovering and researching this issue.
6926 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6927 X509_name_oneline, X509_name_print_ex et al. to leak some information
6928 from the stack. Applications may be affected if they echo pretty printing
6929 output to the attacker.
6931 Thanks to Ivan Fratric (Google) for discovering this issue.
6934 *Emilia Käsper, and Steve Henson*
6936 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6937 for corner cases. (Certain input points at infinity could lead to
6938 bogus results, with non-infinity inputs mapped to infinity too.)
6942 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6944 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6945 handshake can force the use of weak keying material in OpenSSL
6946 SSL/TLS clients and servers.
6948 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6949 researching this issue. ([CVE-2014-0224])
6951 *KIKUCHI Masashi, Steve Henson*
6953 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6954 OpenSSL DTLS client the code can be made to recurse eventually crashing
6957 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6960 *Imre Rad, Steve Henson*
6962 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6963 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6964 client or server. This is potentially exploitable to run arbitrary
6965 code on a vulnerable client or server.
6967 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6969 *Jüri Aedla, Steve Henson*
6971 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6972 are subject to a denial of service attack.
6974 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6975 this issue. ([CVE-2014-3470])
6977 *Felix Gröbert, Ivan Fratric, Steve Henson*
6979 * Harmonize version and its documentation. -f flag is used to display
6982 *mancha <mancha1@zoho.com>*
6984 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6985 in i2d_ECPrivateKey.
6987 *mancha <mancha1@zoho.com>*
6989 * Fix some double frees. These are not thought to be exploitable.
6991 *mancha <mancha1@zoho.com>*
6993 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6995 * A missing bounds check in the handling of the TLS heartbeat extension
6996 can be used to reveal up to 64k of memory to a connected client or
6999 Thanks for Neel Mehta of Google Security for discovering this bug and to
7000 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7001 preparing the fix ([CVE-2014-0160])
7003 *Adam Langley, Bodo Moeller*
7005 * Fix for the attack described in the paper "Recovering OpenSSL
7006 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7007 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7008 <http://eprint.iacr.org/2014/140>
7010 Thanks to Yuval Yarom and Naomi Benger for discovering this
7011 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7013 *Yuval Yarom and Naomi Benger*
7015 * TLS pad extension: draft-agl-tls-padding-03
7017 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7018 TLS client Hello record length value would otherwise be > 255 and
7019 less that 512 pad with a dummy extension containing zeroes so it
7020 is at least 512 bytes long.
7022 *Adam Langley, Steve Henson*
7024 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7026 * Fix for TLS record tampering bug. A carefully crafted invalid
7027 handshake could crash OpenSSL with a NULL pointer exception.
7028 Thanks to Anton Johansson for reporting this issues.
7031 * Keep original DTLS digest and encryption contexts in retransmission
7032 structures so we can use the previous session parameters if they need
7033 to be resent. ([CVE-2013-6450])
7037 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7038 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7039 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7040 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7041 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7042 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7044 *Rob Stradling, Adam Langley*
7046 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7048 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7049 supporting platforms or when small records were transferred.
7051 *Andy Polyakov, Steve Henson*
7053 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7055 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7057 This addresses the flaw in CBC record processing discovered by
7058 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7059 at: <http://www.isg.rhul.ac.uk/tls/>
7061 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7062 Security Group at Royal Holloway, University of London
7063 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7064 Emilia Käsper for the initial patch.
7067 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7069 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7070 ciphersuites which can be exploited in a denial of service attack.
7071 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7072 and detecting this bug and to Wolfgang Ettlinger
7073 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7078 * Return an error when checking OCSP signatures when key is NULL.
7079 This fixes a DoS attack. ([CVE-2013-0166])
7083 * Make openssl verify return errors.
7085 *Chris Palmer <palmer@google.com> and Ben Laurie*
7087 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7088 the right response is stapled. Also change SSL_get_certificate()
7089 so it returns the certificate actually sent.
7090 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7092 *Rob Stradling <rob.stradling@comodo.com>*
7094 * Fix possible deadlock when decoding public keys.
7098 * Don't use TLS 1.0 record version number in initial client hello
7103 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7105 * Sanity check record length before skipping explicit IV in TLS
7106 1.2, 1.1 and DTLS to fix DoS attack.
7108 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7109 fuzzing as a service testing platform.
7114 * Initialise tkeylen properly when encrypting CMS messages.
7115 Thanks to Solar Designer of Openwall for reporting this issue.
7119 * In FIPS mode don't try to use composite ciphers as they are not
7124 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7126 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7127 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7128 mean any application compiled against OpenSSL 1.0.0 headers setting
7129 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7130 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7131 0x10000000L Any application which was previously compiled against
7132 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7133 will need to be recompiled as a result. Letting be results in
7134 inability to disable specifically TLS 1.1 and in client context,
7135 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7139 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7140 disable just protocol X, but all protocols above X *if* there are
7141 protocols *below* X still enabled. In more practical terms it means
7142 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7143 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7144 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7149 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7151 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7152 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7153 in CRYPTO_realloc_clean.
7155 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7156 issue and to Adam Langley <agl@chromium.org> for fixing it.
7159 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7161 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7165 * Workarounds for some broken servers that "hang" if a client hello
7166 record length exceeds 255 bytes.
7168 1. Do not use record version number > TLS 1.0 in initial client
7169 hello: some (but not all) hanging servers will now work.
7170 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7171 the number of ciphers sent in the client hello. This should be
7172 set to an even number, such as 50, for example by passing:
7173 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7174 Most broken servers should now work.
7175 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7176 TLS 1.2 client support entirely.
7180 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7184 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7186 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7187 STRING form instead of a DigestInfo.
7191 * The format used for MDC2 RSA signatures is inconsistent between EVP
7192 and the RSA_sign/RSA_verify functions. This was made more apparent when
7193 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7194 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7195 the correct format in RSA_verify so both forms transparently work.
7199 * Some servers which support TLS 1.0 can choke if we initially indicate
7200 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7201 encrypted premaster secret. As a workaround use the maximum permitted
7202 client version in client hello, this should keep such servers happy
7203 and still work with previous versions of OpenSSL.
7207 * Add support for TLS/DTLS heartbeats.
7209 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7211 * Add support for SCTP.
7213 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7215 * Improved PRNG seeding for VOS.
7217 *Paul Green <Paul.Green@stratus.com>*
7219 * Extensive assembler packs updates, most notably:
7221 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7222 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7223 - x86_64: bit-sliced AES implementation;
7224 - ARM: NEON support, contemporary platforms optimizations;
7225 - s390x: z196 support;
7226 - `*`: GHASH and GF(2^m) multiplication implementations;
7230 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7231 (removal of unnecessary code)
7233 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7235 * Add TLS key material exporter from RFC 5705.
7239 * Add DTLS-SRTP negotiation from RFC 5764.
7243 * Add Next Protocol Negotiation,
7244 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7245 disabled with a no-npn flag to config or Configure. Code donated
7248 *Adam Langley <agl@google.com> and Ben Laurie*
7250 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7251 NIST-P256, NIST-P521, with constant-time single point multiplication on
7252 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7253 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7254 Code made available under Apache License version 2.0.
7256 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7257 line to include this in your build of OpenSSL, and run "make depend" (or
7258 "make update"). This enables the following EC_METHODs:
7260 EC_GFp_nistp224_method()
7261 EC_GFp_nistp256_method()
7262 EC_GFp_nistp521_method()
7264 EC_GROUP_new_by_curve_name() will automatically use these (while
7265 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7268 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7270 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7271 all platforms. Move ssize_t definition from e_os.h to the public
7272 header file e_os2.h as it now appears in public header file cms.h
7276 * New -sigopt option to the ca, req and x509 utilities. Additional
7277 signature parameters can be passed using this option and in
7282 * Add RSA PSS signing function. This will generate and set the
7283 appropriate AlgorithmIdentifiers for PSS based on those in the
7284 corresponding EVP_MD_CTX structure. No application support yet.
7288 * Support for companion algorithm specific ASN1 signing routines.
7289 New function ASN1_item_sign_ctx() signs a pre-initialised
7290 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7291 the appropriate parameters.
7295 * Add new algorithm specific ASN1 verification initialisation function
7296 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7297 handling will be the same no matter what EVP_PKEY_METHOD is used.
7298 Add a PSS handler to support verification of PSS signatures: checked
7299 against a number of sample certificates.
7303 * Add signature printing for PSS. Add PSS OIDs.
7305 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7307 * Add algorithm specific signature printing. An individual ASN1 method
7308 can now print out signatures instead of the standard hex dump.
7310 More complex signatures (e.g. PSS) can print out more meaningful
7311 information. Include DSA version that prints out the signature
7316 * Password based recipient info support for CMS library: implementing
7321 * Split password based encryption into PBES2 and PBKDF2 functions. This
7322 neatly separates the code into cipher and PBE sections and is required
7323 for some algorithms that split PBES2 into separate pieces (such as
7324 password based CMS).
7328 * Session-handling fixes:
7329 - Fix handling of connections that are resuming with a session ID,
7330 but also support Session Tickets.
7331 - Fix a bug that suppressed issuing of a new ticket if the client
7332 presented a ticket with an expired session.
7333 - Try to set the ticket lifetime hint to something reasonable.
7334 - Make tickets shorter by excluding irrelevant information.
7335 - On the client side, don't ignore renewed tickets.
7337 *Adam Langley, Bodo Moeller (Google)*
7339 * Fix PSK session representation.
7343 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7345 This work was sponsored by Intel.
7349 * Add GCM support to TLS library. Some custom code is needed to split
7350 the IV between the fixed (from PRF) and explicit (from TLS record)
7351 portions. This adds all GCM ciphersuites supported by RFC5288 and
7352 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7353 add a special AESGCM string for GCM only.
7357 * Expand range of ctrls for AES GCM. Permit setting invocation
7358 field on decrypt and retrieval of invocation field only on encrypt.
7362 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7363 As required by RFC5289 these ciphersuites cannot be used if for
7364 versions of TLS earlier than 1.2.
7368 * For FIPS capable OpenSSL interpret a NULL default public key method
7369 as unset and return the appropriate default but do *not* set the default.
7370 This means we can return the appropriate method in applications that
7371 switch between FIPS and non-FIPS modes.
7375 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7376 ENGINE is used then we cannot handle that in the FIPS module so we
7377 keep original code iff non-FIPS operations are allowed.
7381 * Add -attime option to openssl utilities.
7383 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7385 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7389 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7390 FIPS EC methods unconditionally for now.
7394 * New build option no-ec2m to disable characteristic 2 code.
7398 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7399 all cases can be covered as some introduce binary incompatibilities.
7403 * Redirect RSA operations to FIPS module including keygen,
7404 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7408 * Add similar low-level API blocking to ciphers.
7412 * low-level digest APIs are not approved in FIPS mode: any attempt
7413 to use these will cause a fatal error. Applications that *really* want
7414 to use them can use the `private_*` version instead.
7418 * Redirect cipher operations to FIPS module for FIPS builds.
7422 * Redirect digest operations to FIPS module for FIPS builds.
7426 * Update build system to add "fips" flag which will link in fipscanister.o
7427 for static and shared library builds embedding a signature if needed.
7431 * Output TLS supported curves in preference order instead of numerical
7432 order. This is currently hardcoded for the highest order curves first.
7433 This should be configurable so applications can judge speed vs strength.
7437 * Add TLS v1.2 server support for client authentication.
7441 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7446 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7447 FIPS modules versions.
7451 * Add TLS v1.2 client side support for client authentication. Keep cache
7452 of handshake records longer as we don't know the hash algorithm to use
7453 until after the certificate request message is received.
7457 * Initial TLS v1.2 client support. Add a default signature algorithms
7458 extension including all the algorithms we support. Parse new signature
7459 format in client key exchange. Relax some ECC signing restrictions for
7460 TLS v1.2 as indicated in RFC5246.
7464 * Add server support for TLS v1.2 signature algorithms extension. Switch
7465 to new signature format when needed using client digest preference.
7466 All server ciphersuites should now work correctly in TLS v1.2. No client
7467 support yet and no support for client certificates.
7471 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7472 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7473 ciphersuites. At present only RSA key exchange ciphersuites work with
7474 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7475 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7476 and version checking.
7480 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7481 with this defined it will not be affected by any changes to ssl internal
7482 structures. Add several utility functions to allow openssl application
7483 to work with OPENSSL_NO_SSL_INTERN defined.
7487 * A long standing patch to add support for SRP from EdelWeb (Peter
7488 Sylvester and Christophe Renou) was integrated.
7489 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7490 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7493 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7497 * Permit abbreviated handshakes when renegotiating using the function
7498 SSL_renegotiate_abbreviated().
7500 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7502 * Add call to ENGINE_register_all_complete() to
7503 ENGINE_load_builtin_engines(), so some implementations get used
7504 automatically instead of needing explicit application support.
7508 * Add support for TLS key exporter as described in RFC5705.
7510 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7512 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7513 a few changes are required:
7515 Add SSL_OP_NO_TLSv1_1 flag.
7516 Add TLSv1_1 methods.
7517 Update version checking logic to handle version 1.1.
7518 Add explicit IV handling (ported from DTLS code).
7519 Add command line options to s_client/s_server.
7526 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7528 * X509_ATTRIBUTE memory leak
7530 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7531 memory. This structure is used by the PKCS#7 and CMS routines so any
7532 application which reads PKCS#7 or CMS data from untrusted sources is
7533 affected. SSL/TLS is not affected.
7535 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7541 * Race condition handling PSK identify hint
7543 If PSK identity hints are received by a multi-threaded client then
7544 the values are wrongly updated in the parent SSL_CTX structure. This can
7545 result in a race condition potentially leading to a double free of the
7551 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7553 * Malformed ECParameters causes infinite loop
7555 When processing an ECParameters structure OpenSSL enters an infinite loop
7556 if the curve specified is over a specially malformed binary polynomial
7559 This can be used to perform denial of service against any
7560 system which processes public keys, certificate requests or
7561 certificates. This includes TLS clients and TLS servers with
7562 client authentication enabled.
7564 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7569 * Exploitable out-of-bounds read in X509_cmp_time
7571 X509_cmp_time does not properly check the length of the ASN1_TIME
7572 string and can read a few bytes out of bounds. In addition,
7573 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7576 An attacker can use this to craft malformed certificates and CRLs of
7577 various sizes and potentially cause a segmentation fault, resulting in
7578 a DoS on applications that verify certificates or CRLs. TLS clients
7579 that verify CRLs are affected. TLS clients and servers with client
7580 authentication enabled may be affected if they use custom verification
7583 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7584 independently by Hanno Böck.
7589 * PKCS7 crash with missing EnvelopedContent
7591 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7592 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7593 with missing content and trigger a NULL pointer dereference on parsing.
7595 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7596 structures from untrusted sources are affected. OpenSSL clients and
7597 servers are not affected.
7599 This issue was reported to OpenSSL by Michal Zalewski (Google).
7604 * CMS verify infinite loop with unknown hash function
7606 When verifying a signedData message the CMS code can enter an infinite loop
7607 if presented with an unknown hash function OID. This can be used to perform
7608 denial of service against any system which verifies signedData messages using
7610 This issue was reported to OpenSSL by Johannes Bauer.
7615 * Race condition handling NewSessionTicket
7617 If a NewSessionTicket is received by a multi-threaded client when attempting to
7618 reuse a previous ticket then a race condition can occur potentially leading to
7619 a double free of the ticket data.
7624 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7626 * Segmentation fault in ASN1_TYPE_cmp fix
7628 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7629 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7630 certificate signature algorithm consistency this can be used to crash any
7631 certificate verification operation and exploited in a DoS attack. Any
7632 application which performs certificate verification is vulnerable including
7633 OpenSSL clients and servers which enable client authentication.
7638 * ASN.1 structure reuse memory corruption fix
7640 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7641 memory corruption via an invalid write. Such reuse is and has been
7642 strongly discouraged and is believed to be rare.
7644 Applications that parse structures containing CHOICE or ANY DEFINED BY
7645 components may be affected. Certificate parsing (d2i_X509 and related
7646 functions) are however not affected. OpenSSL clients and servers are
7652 * PKCS7 NULL pointer dereferences fix
7654 The PKCS#7 parsing code does not handle missing outer ContentInfo
7655 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7656 missing content and trigger a NULL pointer dereference on parsing.
7658 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7659 otherwise parse PKCS#7 structures from untrusted sources are
7660 affected. OpenSSL clients and servers are not affected.
7662 This issue was reported to OpenSSL by Michal Zalewski (Google).
7667 * DoS via reachable assert in SSLv2 servers fix
7669 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7670 servers that both support SSLv2 and enable export cipher suites by sending
7671 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7673 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7674 (OpenSSL development team).
7679 * Use After Free following d2i_ECPrivatekey error fix
7681 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7682 could cause a use after free condition. This, in turn, could cause a double
7683 free in several private key parsing functions (such as d2i_PrivateKey
7684 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7685 for applications that receive EC private keys from untrusted
7686 sources. This scenario is considered rare.
7688 This issue was discovered by the BoringSSL project and fixed in their
7694 * X509_to_X509_REQ NULL pointer deref fix
7696 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7697 the certificate key is invalid. This function is rarely used in practice.
7699 This issue was discovered by Brian Carpenter.
7704 * Removed the export ciphers from the DEFAULT ciphers
7708 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7710 * Build fixes for the Windows and OpenVMS platforms
7712 *Matt Caswell and Richard Levitte*
7714 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7716 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7717 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7718 dereference. This could lead to a Denial Of Service attack. Thanks to
7719 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7724 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7725 dtls1_buffer_record function under certain conditions. In particular this
7726 could occur if an attacker sent repeated DTLS records with the same
7727 sequence number but for the next epoch. The memory leak could be exploited
7728 by an attacker in a Denial of Service attack through memory exhaustion.
7729 Thanks to Chris Mueller for reporting this issue.
7734 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7735 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7736 method would be set to NULL which could later result in a NULL pointer
7737 dereference. Thanks to Frank Schmirler for reporting this issue.
7742 * Abort handshake if server key exchange message is omitted for ephemeral
7745 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7746 reporting this issue.
7751 * Remove non-export ephemeral RSA code on client and server. This code
7752 violated the TLS standard by allowing the use of temporary RSA keys in
7753 non-export ciphersuites and could be used by a server to effectively
7754 downgrade the RSA key length used to a value smaller than the server
7755 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7756 INRIA or reporting this issue.
7761 * Fixed issue where DH client certificates are accepted without verification.
7762 An OpenSSL server will accept a DH certificate for client authentication
7763 without the certificate verify message. This effectively allows a client to
7764 authenticate without the use of a private key. This only affects servers
7765 which trust a client certificate authority which issues certificates
7766 containing DH keys: these are extremely rare and hardly ever encountered.
7767 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7773 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7774 results on some platforms, including x86_64. This bug occurs at random
7775 with a very low probability, and is not known to be exploitable in any
7776 way, though its exact impact is difficult to determine. Thanks to Pieter
7777 Wuille (Blockstream) who reported this issue and also suggested an initial
7778 fix. Further analysis was conducted by the OpenSSL development team and
7779 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7780 the OpenSSL core team.
7785 * Fix various certificate fingerprint issues.
7787 By using non-DER or invalid encodings outside the signed portion of a
7788 certificate the fingerprint can be changed without breaking the signature.
7789 Although no details of the signed portion of the certificate can be changed
7790 this can cause problems with some applications: e.g. those using the
7791 certificate fingerprint for blacklists.
7793 1. Reject signatures with non zero unused bits.
7795 If the BIT STRING containing the signature has non zero unused bits reject
7796 the signature. All current signature algorithms require zero unused bits.
7798 2. Check certificate algorithm consistency.
7800 Check the AlgorithmIdentifier inside TBS matches the one in the
7801 certificate signature. NB: this will result in signature failure
7802 errors for some broken certificates.
7804 Thanks to Konrad Kraszewski from Google for reporting this issue.
7806 3. Check DSA/ECDSA signatures use DER.
7808 Reencode DSA/ECDSA signatures and compare with the original received
7809 signature. Return an error if there is a mismatch.
7811 This will reject various cases including garbage after signature
7812 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7813 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7814 (negative or with leading zeroes).
7816 Further analysis was conducted and fixes were developed by Stephen Henson
7817 of the OpenSSL core team.
7823 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7825 * Session Ticket Memory Leak.
7827 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7828 integrity of that ticket is first verified. In the event of a session
7829 ticket integrity check failing, OpenSSL will fail to free memory
7830 causing a memory leak. By sending a large number of invalid session
7831 tickets an attacker could exploit this issue in a Denial Of Service
7837 * Build option no-ssl3 is incomplete.
7839 When OpenSSL is configured with "no-ssl3" as a build option, servers
7840 could accept and complete a SSL 3.0 handshake, and clients could be
7841 configured to send them.
7844 *Akamai and the OpenSSL team*
7846 * Add support for TLS_FALLBACK_SCSV.
7847 Client applications doing fallback retries should call
7848 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7851 *Adam Langley, Bodo Moeller*
7853 * Add additional DigestInfo checks.
7855 Reencode DigestInto in DER and check against the original when
7856 verifying RSA signature: this will reject any improperly encoded
7857 DigestInfo structures.
7859 Note: this is a precautionary measure and no attacks are currently known.
7863 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7865 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7866 to a denial of service attack. A malicious server can crash the client
7867 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7868 ciphersuite and sending carefully crafted handshake messages.
7870 Thanks to Felix Gröbert (Google) for discovering and researching this
7876 * By sending carefully crafted DTLS packets an attacker could cause openssl
7877 to leak memory. This can be exploited through a Denial of Service attack.
7878 Thanks to Adam Langley for discovering and researching this issue.
7883 * An attacker can force openssl to consume large amounts of memory whilst
7884 processing DTLS handshake messages. This can be exploited through a
7885 Denial of Service attack.
7886 Thanks to Adam Langley for discovering and researching this issue.
7891 * An attacker can force an error condition which causes openssl to crash
7892 whilst processing DTLS packets due to memory being freed twice. This
7893 can be exploited through a Denial of Service attack.
7894 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7900 * If a multithreaded client connects to a malicious server using a resumed
7901 session and the server sends an ec point format extension it could write
7902 up to 255 bytes to freed memory.
7904 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7910 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7911 X509_name_oneline, X509_name_print_ex et al. to leak some information
7912 from the stack. Applications may be affected if they echo pretty printing
7913 output to the attacker.
7915 Thanks to Ivan Fratric (Google) for discovering this issue.
7918 *Emilia Käsper, and Steve Henson*
7920 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7921 for corner cases. (Certain input points at infinity could lead to
7922 bogus results, with non-infinity inputs mapped to infinity too.)
7926 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7928 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7929 handshake can force the use of weak keying material in OpenSSL
7930 SSL/TLS clients and servers.
7932 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7933 researching this issue. ([CVE-2014-0224])
7935 *KIKUCHI Masashi, Steve Henson*
7937 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7938 OpenSSL DTLS client the code can be made to recurse eventually crashing
7941 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7944 *Imre Rad, Steve Henson*
7946 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7947 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7948 client or server. This is potentially exploitable to run arbitrary
7949 code on a vulnerable client or server.
7951 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7953 *Jüri Aedla, Steve Henson*
7955 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7956 are subject to a denial of service attack.
7958 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7959 this issue. ([CVE-2014-3470])
7961 *Felix Gröbert, Ivan Fratric, Steve Henson*
7963 * Harmonize version and its documentation. -f flag is used to display
7966 *mancha <mancha1@zoho.com>*
7968 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7969 in i2d_ECPrivateKey.
7971 *mancha <mancha1@zoho.com>*
7973 * Fix some double frees. These are not thought to be exploitable.
7975 *mancha <mancha1@zoho.com>*
7977 * Fix for the attack described in the paper "Recovering OpenSSL
7978 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7979 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7980 <http://eprint.iacr.org/2014/140>
7982 Thanks to Yuval Yarom and Naomi Benger for discovering this
7983 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7985 *Yuval Yarom and Naomi Benger*
7987 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7989 * Keep original DTLS digest and encryption contexts in retransmission
7990 structures so we can use the previous session parameters if they need
7991 to be resent. ([CVE-2013-6450])
7995 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7996 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7997 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7998 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7999 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8000 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8002 *Rob Stradling, Adam Langley*
8004 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8006 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8008 This addresses the flaw in CBC record processing discovered by
8009 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8010 at: <http://www.isg.rhul.ac.uk/tls/>
8012 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8013 Security Group at Royal Holloway, University of London
8014 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8015 Emilia Käsper for the initial patch.
8018 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8020 * Return an error when checking OCSP signatures when key is NULL.
8021 This fixes a DoS attack. ([CVE-2013-0166])
8025 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8026 the right response is stapled. Also change SSL_get_certificate()
8027 so it returns the certificate actually sent.
8028 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8029 (This is a backport)
8031 *Rob Stradling <rob.stradling@comodo.com>*
8033 * Fix possible deadlock when decoding public keys.
8037 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8039 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8042 * Sanity check record length before skipping explicit IV in DTLS
8045 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8046 fuzzing as a service testing platform.
8051 * Initialise tkeylen properly when encrypting CMS messages.
8052 Thanks to Solar Designer of Openwall for reporting this issue.
8056 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8058 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8059 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8060 in CRYPTO_realloc_clean.
8062 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8063 issue and to Adam Langley <agl@chromium.org> for fixing it.
8066 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8068 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8070 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8071 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8072 content decryption and always return the same error. Note: this attack
8073 needs on average 2^20 messages so it only affects automated senders. The
8074 old behaviour can be re-enabled in the CMS code by setting the
8075 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8076 an MMA defence is not necessary.
8077 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8078 this issue. ([CVE-2012-0884])
8082 * Fix CVE-2011-4619: make sure we really are receiving a
8083 client hello before rejecting multiple SGC restarts. Thanks to
8084 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8088 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8090 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8091 Thanks to Antonio Martin, Enterprise Secure Access Research and
8092 Development, Cisco Systems, Inc. for discovering this bug and
8093 preparing a fix. ([CVE-2012-0050])
8097 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8099 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8100 of the Vaudenay padding oracle attack on CBC mode encryption
8101 which enables an efficient plaintext recovery attack against
8102 the OpenSSL implementation of DTLS. Their attack exploits timing
8103 differences arising during decryption processing. A research
8104 paper describing this attack can be found at:
8105 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8106 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8107 Security Group at Royal Holloway, University of London
8108 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8109 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8110 for preparing the fix. ([CVE-2011-4108])
8112 *Robin Seggelmann, Michael Tuexen*
8114 * Clear bytes used for block padding of SSL 3.0 records.
8117 *Adam Langley (Google)*
8119 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8120 Kadianakis <desnacked@gmail.com> for discovering this issue and
8121 Adam Langley for preparing the fix. ([CVE-2011-4619])
8123 *Adam Langley (Google)*
8125 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8127 *Andrey Kulikov <amdeich@gmail.com>*
8129 * Prevent malformed RFC3779 data triggering an assertion failure.
8130 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8131 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8133 *Rob Austein <sra@hactrn.net>*
8135 * Improved PRNG seeding for VOS.
8137 *Paul Green <Paul.Green@stratus.com>*
8139 * Fix ssl_ciph.c set-up race.
8141 *Adam Langley (Google)*
8143 * Fix spurious failures in ecdsatest.c.
8145 *Emilia Käsper (Google)*
8147 * Fix the BIO_f_buffer() implementation (which was mixing different
8148 interpretations of the `..._len` fields).
8150 *Adam Langley (Google)*
8152 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8153 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8154 threads won't reuse the same blinding coefficients.
8156 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8157 lock to call BN_BLINDING_invert_ex, and avoids one use of
8158 BN_BLINDING_update for each BN_BLINDING structure (previously,
8159 the last update always remained unused).
8161 *Emilia Käsper (Google)*
8163 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8165 *Bob Buckholz (Google)*
8167 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8169 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8170 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8172 *Kaspar Brand <ossl@velox.ch>*
8174 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8175 for multi-threaded use of ECDH. ([CVE-2011-3210])
8177 *Adam Langley (Google)*
8179 * Fix x509_name_ex_d2i memory leak on bad inputs.
8183 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8184 signature public key algorithm by using OID xref utilities instead.
8185 Before this you could only use some ECC ciphersuites with SHA1 only.
8189 * Add protection against ECDSA timing attacks as mentioned in the paper
8190 by Billy Bob Brumley and Nicola Tuveri, see:
8191 <http://eprint.iacr.org/2011/232.pdf>
8193 *Billy Bob Brumley and Nicola Tuveri*
8195 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8197 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8199 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8201 * Fix bug in string printing code: if *any* escaping is enabled we must
8202 escape the escape character (backslash) or the resulting string is
8207 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8209 * Disable code workaround for ancient and obsolete Netscape browsers
8210 and servers: an attacker can use it in a ciphersuite downgrade attack.
8211 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8215 * Fixed J-PAKE implementation error, originally discovered by
8216 Sebastien Martini, further info and confirmation from Stefan
8217 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8221 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8223 * Fix extension code to avoid race conditions which can result in a buffer
8224 overrun vulnerability: resumed sessions must not be modified as they can
8225 be shared by multiple threads. CVE-2010-3864
8229 * Fix WIN32 build system to correctly link an ENGINE directory into
8234 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8236 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8239 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8241 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8243 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8244 context. The operation can be customised via the ctrl mechanism in
8245 case ENGINEs want to include additional functionality.
8249 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8253 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8254 output hashes compatible with older versions of OpenSSL.
8256 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8258 * Fix compression algorithm handling: if resuming a session use the
8259 compression algorithm of the resumed session instead of determining
8260 it from client hello again. Don't allow server to change algorithm.
8264 * Add load_crls() function to commands tidying load_certs() too. Add option
8265 to verify utility to allow additional CRLs to be included.
8269 * Update OCSP request code to permit adding custom headers to the request:
8270 some responders need this.
8274 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8277 *Julia Lawall <julia@diku.dk>*
8279 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8280 needlessly dereferenced structures, used obsolete functions and
8281 didn't handle all updated verify codes correctly.
8285 * Disable MD2 in the default configuration.
8289 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8290 indicate the initial BIO being pushed or popped. This makes it possible
8291 to determine whether the BIO is the one explicitly called or as a result
8292 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8293 it handles reference counts correctly and doesn't zero out the I/O bio
8294 when it is not being explicitly popped. WARNING: applications which
8295 included workarounds for the old buggy behaviour will need to be modified
8296 or they could free up already freed BIOs.
8300 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8301 renaming to all platforms (within the 0.9.8 branch, this was
8302 done conditionally on Netware platforms to avoid a name clash).
8304 *Guenter <lists@gknw.net>*
8306 * Add ECDHE and PSK support to DTLS.
8308 *Michael Tuexen <tuexen@fh-muenster.de>*
8310 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8315 * Add "missing" function EVP_MD_flags() (without this the only way to
8316 retrieve a digest flags is by accessing the structure directly. Update
8317 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8318 or cipher is registered as in the "from" argument. Print out all
8319 registered digests in the dgst usage message instead of manually
8320 attempting to work them out.
8324 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8325 this allows the use of compression and extensions. Change default cipher
8326 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8327 by default unless an application cipher string requests it.
8331 * Alter match criteria in PKCS12_parse(). It used to try to use local
8332 key ids to find matching certificates and keys but some PKCS#12 files
8333 don't follow the (somewhat unwritten) rules and this strategy fails.
8334 Now just gather all certificates together and the first private key
8335 then look for the first certificate that matches the key.
8339 * Support use of registered digest and cipher names for dgst and cipher
8340 commands instead of having to add each one as a special case. So now
8347 openssl dgst -sha256 foo
8349 and this works for ENGINE based algorithms too.
8353 * Update Gost ENGINE to support parameter files.
8355 *Victor B. Wagner <vitus@cryptocom.ru>*
8357 * Support GeneralizedTime in ca utility.
8359 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8361 * Enhance the hash format used for certificate directory links. The new
8362 form uses the canonical encoding (meaning equivalent names will work
8363 even if they aren't identical) and uses SHA1 instead of MD5. This form
8364 is incompatible with the older format and as a result c_rehash should
8365 be used to rebuild symbolic links.
8369 * Make PKCS#8 the default write format for private keys, replacing the
8370 traditional format. This form is standardised, more secure and doesn't
8371 include an implicit MD5 dependency.
8375 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8376 committed to OpenSSL should pass this lot as a minimum.
8380 * Add session ticket override functionality for use by EAP-FAST.
8382 *Jouni Malinen <j@w1.fi>*
8384 * Modify HMAC functions to return a value. Since these can be implemented
8385 in an ENGINE errors can occur.
8389 * Type-checked OBJ_bsearch_ex.
8393 * Type-checked OBJ_bsearch. Also some constification necessitated
8394 by type-checking. Still to come: TXT_DB, bsearch(?),
8395 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8400 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8401 seconds to a tm structure directly, instead of going through OS
8402 specific date routines. This avoids any issues with OS routines such
8403 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8404 and X509_time_adj_ex() to cover the extended range. The existing
8405 X509_time_adj() is still usable and will no longer have any date issues.
8409 * Delta CRL support. New use deltas option which will attempt to locate
8410 and search any appropriate delta CRLs available.
8412 This work was sponsored by Google.
8416 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8417 code and add additional score elements. Validate alternate CRL paths
8418 as part of the CRL checking and indicate a new error "CRL path validation
8419 error" in this case. Applications wanting additional details can use
8420 the verify callback and check the new "parent" field. If this is not
8421 NULL CRL path validation is taking place. Existing applications won't
8422 see this because it requires extended CRL support which is off by
8425 This work was sponsored by Google.
8429 * Support for freshest CRL extension.
8431 This work was sponsored by Google.
8435 * Initial indirect CRL support. Currently only supported in the CRLs
8436 passed directly and not via lookup. Process certificate issuer
8437 CRL entry extension and lookup CRL entries by bother issuer name
8438 and serial number. Check and process CRL issuer entry in IDP extension.
8440 This work was sponsored by Google.
8444 * Add support for distinct certificate and CRL paths. The CRL issuer
8445 certificate is validated separately in this case. Only enabled if
8446 an extended CRL support flag is set: this flag will enable additional
8447 CRL functionality in future.
8449 This work was sponsored by Google.
8453 * Add support for policy mappings extension.
8455 This work was sponsored by Google.
8459 * Fixes to pathlength constraint, self issued certificate handling,
8460 policy processing to align with RFC3280 and PKITS tests.
8462 This work was sponsored by Google.
8466 * Support for name constraints certificate extension. DN, email, DNS
8467 and URI types are currently supported.
8469 This work was sponsored by Google.
8473 * To cater for systems that provide a pointer-based thread ID rather
8474 than numeric, deprecate the current numeric thread ID mechanism and
8475 replace it with a structure and associated callback type. This
8476 mechanism allows a numeric "hash" to be extracted from a thread ID in
8477 either case, and on platforms where pointers are larger than 'long',
8478 mixing is done to help ensure the numeric 'hash' is usable even if it
8479 can't be guaranteed unique. The default mechanism is to use "&errno"
8480 as a pointer-based thread ID to distinguish between threads.
8482 Applications that want to provide their own thread IDs should now use
8483 CRYPTO_THREADID_set_callback() to register a callback that will call
8484 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8486 Note that ERR_remove_state() is now deprecated, because it is tied
8487 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8488 to free the current thread's error state should be replaced by
8489 ERR_remove_thread_state(NULL).
8491 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8492 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8493 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8494 application was previously providing a numeric thread callback that
8495 was inappropriate for distinguishing threads, then uniqueness might
8496 have been obtained with &errno that happened immediately in the
8497 intermediate development versions of OpenSSL; this is no longer the
8498 case, the numeric thread callback will now override the automatic use
8501 *Geoff Thorpe, with help from Bodo Moeller*
8503 * Initial support for different CRL issuing certificates. This covers a
8504 simple case where the self issued certificates in the chain exist and
8505 the real CRL issuer is higher in the existing chain.
8507 This work was sponsored by Google.
8511 * Removed effectively defunct crypto/store from the build.
8515 * Revamp of STACK to provide stronger type-checking. Still to come:
8516 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8517 ASN1_STRING, CONF_VALUE.
8521 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8522 RAM on SSL connections. This option can save about 34k per idle SSL.
8526 * Revamp of LHASH to provide stronger type-checking. Still to come:
8527 STACK, TXT_DB, bsearch, qsort.
8531 * Initial support for Cryptographic Message Syntax (aka CMS) based
8532 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8533 support for data, signedData, compressedData, digestedData and
8534 encryptedData, envelopedData types included. Scripts to check against
8535 RFC4134 examples draft and interop and consistency checks of many
8536 content types and variants.
8540 * Add options to enc utility to support use of zlib compression BIO.
8544 * Extend mk1mf to support importing of options and assembly language
8545 files from Configure script, currently only included in VC-WIN32.
8546 The assembly language rules can now optionally generate the source
8547 files from the associated perl scripts.
8551 * Implement remaining functionality needed to support GOST ciphersuites.
8552 Interop testing has been performed using CryptoPro implementations.
8554 *Victor B. Wagner <vitus@cryptocom.ru>*
8556 * s390x assembler pack.
8560 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8565 * Implement Opaque PRF Input TLS extension as specified in
8566 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8567 official specification yet and no extension type assignment by
8568 IANA exists, this extension (for now) will have to be explicitly
8569 enabled when building OpenSSL by providing the extension number
8570 to use. For example, specify an option
8572 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8574 to the "config" or "Configure" script to enable the extension,
8575 assuming extension number 0x9527 (which is a completely arbitrary
8576 and unofficial assignment based on the MD5 hash of the Internet
8577 Draft). Note that by doing so, you potentially lose
8578 interoperability with other TLS implementations since these might
8579 be using the same extension number for other purposes.
8581 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8582 opaque PRF input value to use in the handshake. This will create
8583 an internal copy of the length-'len' string at 'src', and will
8584 return non-zero for success.
8586 To get more control and flexibility, provide a callback function
8589 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8590 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8594 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8597 Callback function 'cb' will be called in handshakes, and is
8598 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8599 Argument 'arg' is for application purposes (the value as given to
8600 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8601 be provided to the callback function). The callback function
8602 has to return non-zero to report success: usually 1 to use opaque
8603 PRF input just if possible, or 2 to enforce use of the opaque PRF
8604 input. In the latter case, the library will abort the handshake
8605 if opaque PRF input is not successfully negotiated.
8607 Arguments 'peerinput' and 'len' given to the callback function
8608 will always be NULL and 0 in the case of a client. A server will
8609 see the client's opaque PRF input through these variables if
8610 available (NULL and 0 otherwise). Note that if the server
8611 provides an opaque PRF input, the length must be the same as the
8612 length of the client's opaque PRF input.
8614 Note that the callback function will only be called when creating
8615 a new session (session resumption can resume whatever was
8616 previously negotiated), and will not be called in SSL 2.0
8617 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8618 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8619 for applications that need to enforce opaque PRF input.
8623 * Update ssl code to support digests other than SHA1+MD5 for handshake
8626 *Victor B. Wagner <vitus@cryptocom.ru>*
8628 * Add RFC4507 support to OpenSSL. This includes the corrections in
8629 RFC4507bis. The encrypted ticket format is an encrypted encoded
8630 SSL_SESSION structure, that way new session features are automatically
8633 If a client application caches session in an SSL_SESSION structure
8634 support is transparent because tickets are now stored in the encoded
8637 The SSL_CTX structure automatically generates keys for ticket
8638 protection in servers so again support should be possible
8639 with no application modification.
8641 If a client or server wishes to disable RFC4507 support then the option
8642 SSL_OP_NO_TICKET can be set.
8644 Add a TLS extension debugging callback to allow the contents of any client
8645 or server extensions to be examined.
8647 This work was sponsored by Google.
8651 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8652 OpenSSL should now compile cleanly on gcc 4.2
8654 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8656 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8657 support including streaming MAC support: this is required for GOST
8658 ciphersuite support.
8660 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8662 * Add option -stream to use PKCS#7 streaming in smime utility. New
8663 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8664 to output in BER and PEM format.
8668 * Experimental support for use of HMAC via EVP_PKEY interface. This
8669 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8670 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8671 ENGINE support for HMAC keys which are unextractable. New -mac and
8672 -macopt options to dgst utility.
8676 * New option -sigopt to dgst utility. Update dgst to use
8677 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8678 alternative signing parameters such as X9.31 or PSS in the dgst
8683 * Change ssl_cipher_apply_rule(), the internal function that does
8684 the work each time a ciphersuite string requests enabling
8685 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8686 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8687 the order of disabled ciphersuites such that those ciphersuites
8688 that most recently went from enabled to disabled not only stay
8689 in order with respect to each other, but also have higher priority
8690 than other disabled ciphersuites the next time ciphersuites are
8693 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8694 the same ciphersuites as with "HIGH" alone, but in a specific
8695 order where the PSK ciphersuites come first (since they are the
8696 most recently disabled ciphersuites when "HIGH" is parsed).
8698 Also, change ssl_create_cipher_list() (using this new
8699 functionality) such that between otherwise identical
8700 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8705 * Change ssl_create_cipher_list() so that it automatically
8706 arranges the ciphersuites in reasonable order before starting
8707 to process the rule string. Thus, the definition for "DEFAULT"
8708 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8709 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8710 This makes it much easier to arrive at a reasonable default order
8711 in applications for which anonymous ciphers are OK (meaning
8712 that you can't actually use DEFAULT).
8714 *Bodo Moeller; suggested by Victor Duchovni*
8716 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8717 processing) into multiple integers instead of setting
8718 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8719 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8720 (These masks as well as the individual bit definitions are hidden
8721 away into the non-exported interface ssl/ssl_locl.h, so this
8722 change to the definition of the SSL_CIPHER structure shouldn't
8723 affect applications.) This give us more bits for each of these
8724 categories, so there is no longer a need to coagulate AES128 and
8725 AES256 into a single algorithm bit, and to coagulate Camellia128
8726 and Camellia256 into a single algorithm bit, which has led to all
8729 Thus, among other things, the kludge introduced in 0.9.7m and
8730 0.9.8e for masking out AES256 independently of AES128 or masking
8731 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8733 With the change, we also introduce new ciphersuite aliases that
8734 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8739 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8740 Use the leftmost N bytes of the signature input if the input is
8741 larger than the prime q (with N being the size in bytes of q).
8745 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8746 it yet and it is largely untested.
8750 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8754 * Initial incomplete changes to avoid need for function casts in OpenSSL
8755 some compilers (gcc 4.2 and later) reject their use. Safestack is
8756 reimplemented. Update ASN1 to avoid use of legacy functions.
8760 * Win32/64 targets are linked with Winsock2.
8764 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8765 to external functions. This can be used to increase CRL handling
8766 efficiency especially when CRLs are very large by (for example) storing
8767 the CRL revoked certificates in a database.
8771 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8772 new CRLs added to a directory can be used. New command line option
8773 -verify_return_error to s_client and s_server. This causes real errors
8774 to be returned by the verify callback instead of carrying on no matter
8775 what. This reflects the way a "real world" verify callback would behave.
8779 * GOST engine, supporting several GOST algorithms and public key formats.
8780 Kindly donated by Cryptocom.
8784 * Partial support for Issuing Distribution Point CRL extension. CRLs
8785 partitioned by DP are handled but no indirect CRL or reason partitioning
8786 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8787 selected via a scoring technique which handles IDP and AKID in CRLs.
8791 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8792 will ultimately be used for all verify operations: this will remove the
8793 X509_STORE dependency on certificate verification and allow alternative
8794 lookup methods. X509_STORE based implementations of these two callbacks.
8798 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8799 Modify get_crl() to find a valid (unexpired) CRL if possible.
8803 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8804 this would be called X509_CRL_cmp() but that name is already used by
8805 a function that just compares CRL issuer names. Cache several CRL
8806 extensions in X509_CRL structure and cache CRLDP in X509.
8810 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8811 this maps equivalent X509_NAME structures into a consistent structure.
8812 Name comparison can then be performed rapidly using memcmp().
8816 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8821 * Allow digests to supply their own micalg string for S/MIME type using
8822 the ctrl EVP_MD_CTRL_MICALG.
8826 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8827 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8828 ctrl. It can then customise the structure before and/or after signing
8833 * New function OBJ_add_sigid() to allow application defined signature OIDs
8834 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8835 to free up any added signature OIDs.
8839 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8840 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8841 digest and cipher tables. New options added to openssl utility:
8842 list-message-digest-algorithms and list-cipher-algorithms.
8846 * Change the array representation of binary polynomials: the list
8847 of degrees of non-zero coefficients is now terminated with -1.
8848 Previously it was terminated with 0, which was also part of the
8849 value; thus, the array representation was not applicable to
8850 polynomials where t^0 has coefficient zero. This change makes
8851 the array representation useful in a more general context.
8855 * Various modifications and fixes to SSL/TLS cipher string
8856 handling. For ECC, the code now distinguishes between fixed ECDH
8857 with RSA certificates on the one hand and with ECDSA certificates
8858 on the other hand, since these are separate ciphersuites. The
8859 unused code for Fortezza ciphersuites has been removed.
8861 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8862 (not "ECDHE"). For consistency with the code for DH
8863 certificates, use of ECDH certificates is now considered ECDH
8864 authentication, not RSA or ECDSA authentication (the latter is
8865 merely the CA's signing algorithm and not actively used in the
8868 The temporary ciphersuite alias "ECCdraft" is no longer
8869 available, and ECC ciphersuites are no longer excluded from "ALL"
8870 and "DEFAULT". The following aliases now exist for RFC 4492
8871 ciphersuites, most of these by analogy with the DH case:
8873 kECDHr - ECDH cert, signed with RSA
8874 kECDHe - ECDH cert, signed with ECDSA
8875 kECDH - ECDH cert (signed with either RSA or ECDSA)
8876 kEECDH - ephemeral ECDH
8877 ECDH - ECDH cert or ephemeral ECDH
8883 AECDH - anonymous ECDH
8884 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8888 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8889 Use correct micalg parameters depending on digest(s) in signed message.
8893 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8894 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8898 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8899 an engine to register a method. Add ENGINE lookups for methods and
8900 functional reference processing.
8904 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8905 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8910 * New -resign option to smime utility. This adds one or more signers
8911 to an existing PKCS#7 signedData structure. Also -md option to use an
8912 alternative message digest algorithm for signing.
8916 * Tidy up PKCS#7 routines and add new functions to make it easier to
8917 create PKCS7 structures containing multiple signers. Update smime
8918 application to support multiple signers.
8922 * New -macalg option to pkcs12 utility to allow setting of an alternative
8927 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8928 Reorganize PBE internals to lookup from a static table using NIDs,
8929 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8930 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8931 PRF which will be automatically used with PBES2.
8935 * Replace the algorithm specific calls to generate keys in "req" with the
8940 * Update PKCS#7 enveloped data routines to use new API. This is now
8941 supported by any public key method supporting the encrypt operation. A
8942 ctrl is added to allow the public key algorithm to examine or modify
8943 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8948 * Add a ctrl to asn1 method to allow a public key algorithm to express
8949 a default digest type to use. In most cases this will be SHA1 but some
8950 algorithms (such as GOST) need to specify an alternative digest. The
8951 return value indicates how strong the preference is 1 means optional and
8952 2 is mandatory (that is it is the only supported type). Modify
8953 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8954 use the default md. Update openssl utilities to use the default digest
8955 type for signing if it is not explicitly indicated.
8959 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8960 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8961 signing method from the key type. This effectively removes the link
8962 between digests and public key types.
8966 * Add an OID cross reference table and utility functions. Its purpose is to
8967 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8968 rsaEncryption. This will allow some of the algorithm specific hackery
8969 needed to use the correct OID to be removed.
8973 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8974 structures for PKCS7_sign(). They are now set up by the relevant public
8979 * Add provisional EC pkey method with support for ECDSA and ECDH.
8983 * Add support for key derivation (agreement) in the API, DH method and
8988 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8989 public and private key formats. As a side effect these add additional
8990 command line functionality not previously available: DSA signatures can be
8991 generated and verified using pkeyutl and DH key support and generation in
8998 *Oliver Tappe <zooey@hirschkaefer.de>*
9000 * New make target "install_html_docs" installs HTML renditions of the
9003 *Oliver Tappe <zooey@hirschkaefer.de>*
9005 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9006 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9007 support key and parameter generation and add initial key generation
9008 functionality for RSA.
9012 * Add functions for main EVP_PKEY_method operations. The undocumented
9013 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9014 `EVP_PKEY_{encrypt,decrypt}_old`.
9018 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9019 key API, doesn't do much yet.
9023 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9024 public key algorithms. New option to openssl utility:
9025 "list-public-key-algorithms" to print out info.
9029 * Implement the Supported Elliptic Curves Extension for
9030 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9034 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9035 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9039 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9040 utilities such as rsa, dsa, dsaparam etc except they process any key
9045 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9046 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9047 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9052 * Initial support for pluggable public key ASN1.
9053 De-spaghettify the public key ASN1 handling. Move public and private
9054 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9055 algorithm specific handling to a single module within the relevant
9056 algorithm directory. Add functions to allow (near) opaque processing
9057 of public and private key structures.
9061 * Implement the Supported Point Formats Extension for
9062 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9066 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9067 for the psk identity [hint] and the psk callback functions to the
9068 SSL_SESSION, SSL and SSL_CTX structure.
9071 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9075 SSL_CTX_use_psk_identity_hint
9076 SSL_get_psk_identity_hint
9077 SSL_get_psk_identity
9078 SSL_use_psk_identity_hint
9080 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9082 * Add RFC 3161 compliant time stamp request creation, response generation
9083 and response verification functionality.
9085 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9087 * Add initial support for TLS extensions, specifically for the server_name
9088 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9089 have new members for a host name. The SSL data structure has an
9090 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9091 stored in that context to allow for session resumption, even after the
9092 SSL has been switched to a new SSL_CTX in reaction to a client's
9093 server_name extension.
9095 New functions (subject to change):
9097 SSL_get_servername()
9098 SSL_get_servername_type()
9101 New CTRL codes and macros (subject to change):
9103 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9104 - SSL_CTX_set_tlsext_servername_callback()
9105 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9106 - SSL_CTX_set_tlsext_servername_arg()
9107 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9109 openssl s_client has a new '-servername ...' option.
9111 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9112 '-key2 ...', '-servername_fatal' (subject to change). This allows
9113 testing the HostName extension for a specific single host name ('-cert'
9114 and '-key' remain fallbacks for handshakes without HostName
9115 negotiation). If the unrecognized_name alert has to be sent, this by
9116 default is a warning; it becomes fatal with the '-servername_fatal'
9119 *Peter Sylvester, Remy Allais, Christophe Renou*
9121 * Whirlpool hash implementation is added.
9125 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9126 bn(64,32). Because of instruction set limitations it doesn't have
9127 any negative impact on performance. This was done mostly in order
9128 to make it possible to share assembler modules, such as bn_mul_mont
9129 implementations, between 32- and 64-bit builds without hassle.
9133 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9134 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9139 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9140 dedicated Montgomery multiplication procedure, is introduced.
9141 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9142 "64-bit" performance on certain 32-bit targets.
9146 * New option SSL_OP_NO_COMP to disable use of compression selectively
9147 in SSL structures. New SSL ctrl to set maximum send fragment size.
9148 Save memory by setting the I/O buffer sizes dynamically instead of
9149 using the maximum available value.
9153 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9154 in addition to the text details.
9158 * Very, very preliminary EXPERIMENTAL support for printing of general
9159 ASN1 structures. This currently produces rather ugly output and doesn't
9160 handle several customised structures at all.
9164 * Integrated support for PVK file format and some related formats such
9165 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9166 these in the 'rsa' and 'dsa' utilities.
9170 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9174 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9175 place for the (very old) "NETSCAPE" format certificates which are now
9176 handled using new ASN1 code equivalents.
9180 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9181 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9182 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9186 * Modify CRL distribution points extension code to print out previously
9187 unsupported fields. Enhance extension setting code to allow setting of
9192 * Add print and set support for Issuing Distribution Point CRL extension.
9196 * Change 'Configure' script to enable Camellia by default.
9203 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9205 * When rejecting SSL/TLS records due to an incorrect version number, never
9206 update s->server with a new major version number. As of
9207 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9208 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9209 the previous behavior could result in a read attempt at NULL when
9210 receiving specific incorrect SSL/TLS records once record payload
9211 protection is active. ([CVE-2010-0740])
9213 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9215 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9216 could be crashed if the relevant tables were not present (e.g. chrooted).
9218 *Tomas Hoger <thoger@redhat.com>*
9220 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9222 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9224 *Martin Olsson, Neel Mehta*
9226 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9227 accommodate for stack sorting, always a write lock!).
9231 * On some versions of WIN32 Heap32Next is very slow. This can cause
9232 excessive delays in the RAND_poll(): over a minute. As a workaround
9233 include a time check in the inner Heap32Next loop too.
9237 * The code that handled flushing of data in SSL/TLS originally used the
9238 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9239 the problem outlined in PR#1949. The fix suggested there however can
9240 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9241 of Apache). So instead simplify the code to flush unconditionally.
9242 This should be fine since flushing with no data to flush is a no op.
9246 * Handle TLS versions 2.0 and later properly and correctly use the
9247 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9248 off ancient servers have a habit of sticking around for a while...
9252 * Modify compression code so it frees up structures without using the
9253 ex_data callbacks. This works around a problem where some applications
9254 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9255 restarting) then use compression (e.g. SSL with compression) later.
9256 This results in significant per-connection memory leaks and
9257 has caused some security issues including CVE-2008-1678 and
9262 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9263 change when encrypting or decrypting.
9267 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9268 connect and renegotiate with servers which do not support RI.
9269 Until RI is more widely deployed this option is enabled by default.
9273 * Add "missing" ssl ctrls to clear options and mode.
9277 * If client attempts to renegotiate and doesn't support RI respond with
9278 a no_renegotiation alert as required by RFC5746. Some renegotiating
9279 TLS clients will continue a connection gracefully when they receive
9280 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9281 waiting for a server hello which it will never receive. Now we treat a
9282 received no_renegotiation alert as a fatal error. This is because
9283 applications requesting a renegotiation might well expect it to succeed
9284 and would have no code in place to handle the server denying it so the
9285 only safe thing to do is to terminate the connection.
9289 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9290 peer supports secure renegotiation and 0 otherwise. Print out peer
9291 renegotiation support in s_client/s_server.
9295 * Replace the highly broken and deprecated SPKAC certification method with
9296 the updated NID creation version. This should correctly handle UTF8.
9300 * Implement RFC5746. Re-enable renegotiation but require the extension
9301 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9302 turns out to be a bad idea. It has been replaced by
9303 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9304 SSL_CTX_set_options(). This is really not recommended unless you
9305 know what you are doing.
9307 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9309 * Fixes to stateless session resumption handling. Use initial_ctx when
9310 issuing and attempting to decrypt tickets in case it has changed during
9311 servername handling. Use a non-zero length session ID when attempting
9312 stateless session resumption: this makes it possible to determine if
9313 a resumption has occurred immediately after receiving server hello
9314 (several places in OpenSSL subtly assume this) instead of later in
9319 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9320 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9321 fixes for a few places where the return code is not checked
9324 *Julia Lawall <julia@diku.dk>*
9326 * Add --strict-warnings option to Configure script to include devteam
9327 warnings in other configurations.
9331 * Add support for --libdir option and LIBDIR variable in makefiles. This
9332 makes it possible to install openssl libraries in locations which
9333 have names other than "lib", for example "/usr/lib64" which some
9336 *Steve Henson, based on patch from Jeremy Utley*
9338 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9339 X690 8.9.12 and can produce some misleading textual output of OIDs.
9341 *Steve Henson, reported by Dan Kaminsky*
9343 * Delete MD2 from algorithm tables. This follows the recommendation in
9344 several standards that it is not used in new applications due to
9345 several cryptographic weaknesses. For binary compatibility reasons
9346 the MD2 API is still compiled in by default.
9350 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9355 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9356 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9359 *Guenter <lists@gknw.net>*
9361 * Fix the server certificate chain building code to use X509_verify_cert(),
9362 it used to have an ad-hoc builder which was unable to cope with anything
9363 other than a simple chain.
9365 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9367 * Don't check self signed certificate signatures in X509_verify_cert()
9368 by default (a flag can override this): it just wastes time without
9369 adding any security. As a useful side effect self signed root CAs
9370 with non-FIPS digests are now usable in FIPS mode.
9374 * In dtls1_process_out_of_seq_message() the check if the current message
9375 is already buffered was missing. For every new message was memory
9376 allocated, allowing an attacker to perform an denial of service attack
9377 with sending out of seq handshake messages until there is no memory
9378 left. Additionally every future message was buffered, even if the
9379 sequence number made no sense and would be part of another handshake.
9380 So only messages with sequence numbers less than 10 in advance will be
9381 buffered. ([CVE-2009-1378])
9383 *Robin Seggelmann, discovered by Daniel Mentz*
9385 * Records are buffered if they arrive with a future epoch to be
9386 processed after finishing the corresponding handshake. There is
9387 currently no limitation to this buffer allowing an attacker to perform
9388 a DOS attack with sending records with future epochs until there is no
9389 memory left. This patch adds the pqueue_size() function to determine
9390 the size of a buffer and limits the record buffer to 100 entries.
9393 *Robin Seggelmann, discovered by Daniel Mentz*
9395 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9396 parent structure is freed. ([CVE-2009-1379])
9400 * Handle non-blocking I/O properly in SSL_shutdown() call.
9402 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9404 * Add `2.5.4.*` OIDs
9406 *Ilya O. <vrghost@gmail.com>*
9408 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9410 * Disable renegotiation completely - this fixes a severe security
9411 problem ([CVE-2009-3555]) at the cost of breaking all
9412 renegotiation. Renegotiation can be re-enabled by setting
9413 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9414 run-time. This is really not recommended unless you know what
9419 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9421 * Don't set val to NULL when freeing up structures, it is freed up by
9422 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9423 zeroing past the valid field. ([CVE-2009-0789])
9425 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9427 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9428 checked correctly. This would allow some invalid signed attributes to
9429 appear to verify correctly. ([CVE-2009-0591])
9431 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9433 * Reject UniversalString and BMPString types with invalid lengths. This
9434 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9435 a legal length. ([CVE-2009-0590])
9439 * Set S/MIME signing as the default purpose rather than setting it
9440 unconditionally. This allows applications to override it at the store
9445 * Permit restricted recursion of ASN1 strings. This is needed in practice
9446 to handle some structures.
9450 * Improve efficiency of mem_gets: don't search whole buffer each time
9453 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9455 * New -hex option for openssl rand.
9459 * Print out UTF8String and NumericString when parsing ASN1.
9463 * Support NumericString type for name components.
9467 * Allow CC in the environment to override the automatically chosen
9468 compiler. Note that nothing is done to ensure flags work with the
9473 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9475 * Properly check EVP_VerifyFinal() and similar return values
9478 *Ben Laurie, Bodo Moeller, Google Security Team*
9480 * Enable TLS extensions by default.
9484 * Allow the CHIL engine to be loaded, whether the application is
9485 multithreaded or not. (This does not release the developer from the
9486 obligation to set up the dynamic locking callbacks.)
9488 *Sander Temme <sander@temme.net>*
9490 * Use correct exit code if there is an error in dgst command.
9492 *Steve Henson; problem pointed out by Roland Dirlewanger*
9494 * Tweak Configure so that you need to say "experimental-jpake" to enable
9495 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9499 * Add experimental JPAKE support, including demo authentication in
9500 s_client and s_server.
9504 * Set the comparison function in v3_addr_canonize().
9506 *Rob Austein <sra@hactrn.net>*
9508 * Add support for XMPP STARTTLS in s_client.
9510 *Philip Paeps <philip@freebsd.org>*
9512 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9513 to ensure that even with this option, only ciphersuites in the
9514 server's preference list will be accepted. (Note that the option
9515 applies only when resuming a session, so the earlier behavior was
9516 just about the algorithm choice for symmetric cryptography.)
9520 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9522 * Fix NULL pointer dereference if a DTLS server received
9523 ChangeCipherSpec as first record ([CVE-2009-1386]).
9527 * Fix a state transition in s3_srvr.c and d1_srvr.c
9528 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9532 * The fix in 0.9.8c that supposedly got rid of unsafe
9533 double-checked locking was incomplete for RSA blinding,
9534 addressing just one layer of what turns out to have been
9535 doubly unsafe triple-checked locking.
9537 So now fix this for real by retiring the MONT_HELPER macro
9538 in crypto/rsa/rsa_eay.c.
9540 *Bodo Moeller; problem pointed out by Marius Schilder*
9542 * Various precautionary measures:
9544 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9546 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9547 (NB: This would require knowledge of the secret session ticket key
9548 to exploit, in which case you'd be SOL either way.)
9550 - Change bn_nist.c so that it will properly handle input BIGNUMs
9551 outside the expected range.
9553 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9556 *Neel Mehta, Bodo Moeller*
9558 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9559 the load fails. Useful for distros.
9561 *Ben Laurie and the FreeBSD team*
9563 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9567 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9571 * Expand ENGINE to support engine supplied SSL client certificate functions.
9573 This work was sponsored by Logica.
9577 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9578 keystores. Support for SSL/TLS client authentication too.
9579 Not compiled unless enable-capieng specified to Configure.
9581 This work was sponsored by Logica.
9585 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9586 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9587 attribute creation routines such as certificate requests and PKCS#12
9592 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9594 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9595 handshake which could lead to a client crash as found using the
9596 Codenomicon TLS test suite ([CVE-2008-1672])
9598 *Steve Henson, Mark Cox*
9600 * Fix double free in TLS server name extensions which could lead to
9601 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9605 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9607 Clear the error queue to ensure that error entries left from
9608 older function calls do not interfere with the correct operation.
9610 *Lutz Jaenicke, Erik de Castro Lopo*
9612 * Remove root CA certificates of commercial CAs:
9614 The OpenSSL project does not recommend any specific CA and does not
9615 have any policy with respect to including or excluding any CA.
9616 Therefore it does not make any sense to ship an arbitrary selection
9617 of root CA certificates with the OpenSSL software.
9621 * RSA OAEP patches to fix two separate invalid memory reads.
9622 The first one involves inputs when 'lzero' is greater than
9623 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9624 before the beginning of from). The second one involves inputs where
9625 the 'db' section contains nothing but zeroes (there is a one-byte
9626 invalid read after the end of 'db').
9628 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9630 * Partial backport from 0.9.9-dev:
9632 Introduce bn_mul_mont (dedicated Montgomery multiplication
9633 procedure) as a candidate for BIGNUM assembler implementation.
9634 While 0.9.9-dev uses assembler for various architectures, only
9635 x86_64 is available by default here in the 0.9.8 branch, and
9636 32-bit x86 is available through a compile-time setting.
9638 To try the 32-bit x86 assembler implementation, use Configure
9639 option "enable-montasm" (which exists only for this backport).
9641 As "enable-montasm" for 32-bit x86 disclaims code stability
9642 anyway, in this constellation we activate additional code
9643 backported from 0.9.9-dev for further performance improvements,
9644 namely BN_from_montgomery_word. (To enable this otherwise,
9645 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9647 *Andy Polyakov (backport partially by Bodo Moeller)*
9649 * Add TLS session ticket callback. This allows an application to set
9650 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9651 values. This is useful for key rollover for example where several key
9652 sets may exist with different names.
9656 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9657 This was broken until now in 0.9.8 releases, such that the only way
9658 a registered ENGINE could be used (assuming it initialises
9659 successfully on the host) was to explicitly set it as the default
9660 for the relevant algorithms. This is in contradiction with 0.9.7
9661 behaviour and the documentation. With this fix, when an ENGINE is
9662 registered into a given algorithm's table of implementations, the
9663 'uptodate' flag is reset so that auto-discovery will be used next
9664 time a new context for that algorithm attempts to select an
9667 *Ian Lister (tweaked by Geoff Thorpe)*
9669 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9670 implementation in the following ways:
9672 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9675 Lack of BER streaming support means one pass streaming processing is
9676 only supported if data is detached: setting the streaming flag is
9677 ignored for embedded content.
9679 CMS support is disabled by default and must be explicitly enabled
9680 with the enable-cms configuration option.
9684 * Update the GMP engine glue to do direct copies between BIGNUM and
9685 mpz_t when openssl and GMP use the same limb size. Otherwise the
9686 existing "conversion via a text string export" trick is still used.
9688 *Paul Sheer <paulsheer@gmail.com>*
9690 * Zlib compression BIO. This is a filter BIO which compressed and
9691 uncompresses any data passed through it.
9695 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9696 RFC3394 compatible AES key wrapping.
9700 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9701 sets string data without copying. X509_ALGOR_set0() and
9702 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9703 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9704 from an X509_ATTRIBUTE structure optionally checking it occurs only
9705 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9710 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9711 to get the expected BN_FLG_CONSTTIME behavior.
9713 *Bodo Moeller (Google)*
9717 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9718 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9719 - added some more tests to do_tests.pl
9720 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9721 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9722 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9723 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9724 - various changes to netware.pl to enable gcc-cross builds on Win32
9726 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9727 - various changes to fix missing prototype warnings
9728 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9729 - added AES, WHIRLPOOL and CPUID assembler code to build files
9730 - added missing AES assembler make rules to mk1mf.pl
9731 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9733 *Guenter Knauf <eflash@gmx.net>*
9735 * Implement certificate status request TLS extension defined in RFC3546.
9736 A client can set the appropriate parameters and receive the encoded
9737 OCSP response via a callback. A server can query the supplied parameters
9738 and set the encoded OCSP response in the callback. Add simplified examples
9739 to s_client and s_server.
9743 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9746 + Binary incompatibility of ssl_ctx_st structure
9747 + DTLS interoperation with non-compliant servers
9748 + Don't call get_session_cb() without proposed session
9749 + Fix ia64 assembler code
9751 *Andy Polyakov, Steve Henson*
9753 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9755 * DTLS Handshake overhaul. There were longstanding issues with
9756 OpenSSL DTLS implementation, which were making it impossible for
9757 RFC 4347 compliant client to communicate with OpenSSL server.
9758 Unfortunately just fixing these incompatibilities would "cut off"
9759 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9760 server keeps tolerating non RFC compliant syntax. The opposite is
9761 not true, 0.9.8f client can not communicate with earlier server.
9762 This update even addresses CVE-2007-4995.
9766 * Changes to avoid need for function casts in OpenSSL: some compilers
9767 (gcc 4.2 and later) reject their use.
9768 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9771 * Add RFC4507 support to OpenSSL. This includes the corrections in
9772 RFC4507bis. The encrypted ticket format is an encrypted encoded
9773 SSL_SESSION structure, that way new session features are automatically
9776 If a client application caches session in an SSL_SESSION structure
9777 support is transparent because tickets are now stored in the encoded
9780 The SSL_CTX structure automatically generates keys for ticket
9781 protection in servers so again support should be possible
9782 with no application modification.
9784 If a client or server wishes to disable RFC4507 support then the option
9785 SSL_OP_NO_TICKET can be set.
9787 Add a TLS extension debugging callback to allow the contents of any client
9788 or server extensions to be examined.
9790 This work was sponsored by Google.
9794 * Add initial support for TLS extensions, specifically for the server_name
9795 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9796 have new members for a host name. The SSL data structure has an
9797 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9798 stored in that context to allow for session resumption, even after the
9799 SSL has been switched to a new SSL_CTX in reaction to a client's
9800 server_name extension.
9802 New functions (subject to change):
9804 SSL_get_servername()
9805 SSL_get_servername_type()
9808 New CTRL codes and macros (subject to change):
9810 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9811 - SSL_CTX_set_tlsext_servername_callback()
9812 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9813 - SSL_CTX_set_tlsext_servername_arg()
9814 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9816 openssl s_client has a new '-servername ...' option.
9818 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9819 '-key2 ...', '-servername_fatal' (subject to change). This allows
9820 testing the HostName extension for a specific single host name ('-cert'
9821 and '-key' remain fallbacks for handshakes without HostName
9822 negotiation). If the unrecognized_name alert has to be sent, this by
9823 default is a warning; it becomes fatal with the '-servername_fatal'
9826 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9828 * Add AES and SSE2 assembly language support to VC++ build.
9832 * Mitigate attack on final subtraction in Montgomery reduction.
9836 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9837 (which previously caused an internal error).
9841 * Squeeze another 10% out of IGE mode when in != out.
9845 * AES IGE mode speedup.
9847 *Dean Gaudet (Google)*
9849 * Add the Korean symmetric 128-bit cipher SEED (see
9850 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9851 add SEED ciphersuites from RFC 4162:
9853 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9854 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9855 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9856 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9858 To minimize changes between patchlevels in the OpenSSL 0.9.8
9859 series, SEED remains excluded from compilation unless OpenSSL
9860 is configured with 'enable-seed'.
9862 *KISA, Bodo Moeller*
9864 * Mitigate branch prediction attacks, which can be practical if a
9865 single processor is shared, allowing a spy process to extract
9866 information. For detailed background information, see
9867 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9868 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9869 and Necessary Software Countermeasures"). The core of the change
9870 are new versions BN_div_no_branch() and
9871 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9872 respectively, which are slower, but avoid the security-relevant
9873 conditional branches. These are automatically called by BN_div()
9874 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9875 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9876 remove a conditional branch.
9878 BN_FLG_CONSTTIME is the new name for the previous
9879 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9880 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9881 in the exponent causes BN_mod_exp_mont() to use the alternative
9882 implementation in BN_mod_exp_mont_consttime().) The old name
9883 remains as a deprecated alias.
9885 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9886 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9887 constant-time implementations for more than just exponentiation.
9888 Here too the old name is kept as a deprecated alias.
9890 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9891 the BN_BLINDING structure gets an independent copy of the
9892 modulus. This means that the previous `BIGNUM *m` argument to
9893 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9894 essentially becomes `const BIGNUM *m`, although we can't actually
9895 change this in the header file before 0.9.9. It allows
9896 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9897 enable BN_FLG_CONSTTIME.
9899 *Matthew D Wood (Intel Corp)*
9901 * In the SSL/TLS server implementation, be strict about session ID
9902 context matching (which matters if an application uses a single
9903 external cache for different purposes). Previously,
9904 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9905 set. This did ensure strict client verification, but meant that,
9906 with applications using a single external cache for quite
9907 different requirements, clients could circumvent ciphersuite
9908 restrictions for a given session ID context by starting a session
9909 in a different context.
9913 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9914 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9915 authentication-only ciphersuites.
9919 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9920 not complete and could lead to a possible single byte overflow
9921 ([CVE-2007-5135]) [Ben Laurie]
9923 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9925 * Since AES128 and AES256 (and similarly Camellia128 and
9926 Camellia256) share a single mask bit in the logic of
9927 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9928 kludge to work properly if AES128 is available and AES256 isn't
9929 (or if Camellia128 is available and Camellia256 isn't).
9933 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9934 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9935 When a point or a seed is encoded in a BIT STRING, we need to
9936 prevent the removal of trailing zero bits to get the proper DER
9937 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9938 of a NamedBitList, for which trailing 0 bits need to be removed.)
9942 * Have SSL/TLS server implementation tolerate "mismatched" record
9943 protocol version while receiving ClientHello even if the
9944 ClientHello is fragmented. (The server can't insist on the
9945 particular protocol version it has chosen before the ServerHello
9946 message has informed the client about his choice.)
9950 * Add RFC 3779 support.
9952 *Rob Austein for ARIN, Ben Laurie*
9954 * Load error codes if they are not already present instead of using a
9955 static variable. This allows them to be cleanly unloaded and reloaded.
9956 Improve header file function name parsing.
9960 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9961 or CAPABILITY handshake as required by RFCs.
9965 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9967 * Introduce limits to prevent malicious keys being able to
9968 cause a denial of service. ([CVE-2006-2940])
9970 *Steve Henson, Bodo Moeller*
9972 * Fix ASN.1 parsing of certain invalid structures that can result
9973 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9975 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9976 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9978 * Fix SSL client code which could crash if connecting to a
9979 malicious SSLv2 server. ([CVE-2006-4343])
9981 *Tavis Ormandy and Will Drewry, Google Security Team*
9983 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9984 match only those. Before that, "AES256-SHA" would be interpreted
9985 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9986 the same strength classification in 0.9.7h) as we currently only
9987 have a single AES bit in the ciphersuite description bitmap.
9988 That change, however, also applied to ciphersuite strings such as
9989 "RC4-MD5" that intentionally matched multiple ciphersuites --
9990 namely, SSL 2.0 ciphersuites in addition to the more common ones
9991 from SSL 3.0/TLS 1.0.
9993 So we change the selection algorithm again: Naming an explicit
9994 ciphersuite selects this one ciphersuite, and any other similar
9995 ciphersuite (same bitmap) from *other* protocol versions.
9996 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9997 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9999 Since SSL 2.0 does not have any ciphersuites for which the
10000 128/256 bit distinction would be relevant, this works for now.
10001 The proper fix will be to use different bits for AES128 and
10002 AES256, which would have avoided the problems from the beginning;
10003 however, bits are scarce, so we can only do this in a new release
10004 (not just a patchlevel) when we can change the SSL_CIPHER
10005 definition to split the single 'unsigned long mask' bitmap into
10006 multiple values to extend the available space.
10010 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10012 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10013 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10015 * Add AES IGE and biIGE modes.
10019 * Change the Unix randomness entropy gathering to use poll() when
10020 possible instead of select(), since the latter has some
10021 undesirable limitations.
10023 *Darryl Miles via Richard Levitte and Bodo Moeller*
10025 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10026 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10027 cannot be implicitly activated as part of, e.g., the "AES" alias.
10028 However, please upgrade to OpenSSL 0.9.9[-dev] for
10029 non-experimental use of the ECC ciphersuites to get TLS extension
10030 support, which is required for curve and point format negotiation
10031 to avoid potential handshake problems.
10035 * Disable rogue ciphersuites:
10037 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10038 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10039 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10041 The latter two were purportedly from
10042 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10045 Also deactivate the remaining ciphersuites from
10046 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10047 unofficial, and the ID has long expired.
10051 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10052 dual-core machines) and other potential thread-safety issues.
10056 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10057 versions), which is now available for royalty-free use
10058 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10059 Also, add Camellia TLS ciphersuites from RFC 4132.
10061 To minimize changes between patchlevels in the OpenSSL 0.9.8
10062 series, Camellia remains excluded from compilation unless OpenSSL
10063 is configured with 'enable-camellia'.
10067 * Disable the padding bug check when compression is in use. The padding
10068 bug check assumes the first packet is of even length, this is not
10069 necessarily true if compression is enabled and can result in false
10070 positives causing handshake failure. The actual bug test is ancient
10071 code so it is hoped that implementations will either have fixed it by
10072 now or any which still have the bug do not support compression.
10076 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10078 * When applying a cipher rule check to see if string match is an explicit
10079 cipher suite and only match that one cipher suite if it is.
10083 * Link in manifests for VC++ if needed.
10085 *Austin Ziegler <halostatue@gmail.com>*
10087 * Update support for ECC-based TLS ciphersuites according to
10088 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10089 TLS extensions, which are supported starting with the 0.9.9
10090 branch, not in the OpenSSL 0.9.8 branch).
10094 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10095 opaque EVP_CIPHER_CTX handling.
10099 * Fixes and enhancements to zlib compression code. We now only use
10100 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10101 to conform with the standards mentioned here:
10102 <http://www.zlib.net/DLL_FAQ.txt>
10103 Static zlib linking now works on Windows and the new --with-zlib-include
10104 --with-zlib-lib options to Configure can be used to supply the location
10105 of the headers and library. Gracefully handle case where zlib library
10110 * Several fixes and enhancements to the OID generation code. The old code
10111 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10112 handle numbers larger than ULONG_MAX, truncated printing and had a
10113 non standard OBJ_obj2txt() behaviour.
10117 * Add support for building of engines under engine/ as shared libraries
10118 under VC++ build system.
10122 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10123 Hopefully, we will not see any false combination of paths any more.
10127 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10129 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10130 (part of SSL_OP_ALL). This option used to disable the
10131 countermeasure against man-in-the-middle protocol-version
10132 rollback in the SSL 2.0 server implementation, which is a bad
10133 idea. ([CVE-2005-2969])
10135 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10136 for Information Security, National Institute of Advanced Industrial
10137 Science and Technology [AIST], Japan)*
10139 * Add two function to clear and return the verify parameter flags.
10143 * Keep cipherlists sorted in the source instead of sorting them at
10144 runtime, thus removing the need for a lock.
10148 * Avoid some small subgroup attacks in Diffie-Hellman.
10150 *Nick Mathewson and Ben Laurie*
10152 * Add functions for well-known primes.
10156 * Extended Windows CE support.
10158 *Satoshi Nakamura and Andy Polyakov*
10160 * Initialize SSL_METHOD structures at compile time instead of during
10161 runtime, thus removing the need for a lock.
10165 * Make PKCS7_decrypt() work even if no certificate is supplied by
10166 attempting to decrypt each encrypted key in turn. Add support to
10171 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10173 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10176 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10180 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10181 key into the same file any more.
10185 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10189 * Add -utf8 command line and config file option to 'ca'.
10191 *Stefan <stf@udoma.org*
10193 * Removed the macro des_crypt(), as it seems to conflict with some
10194 libraries. Use DES_crypt().
10198 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10199 involves renaming the source and generated shared-libs for
10200 both. The engines will accept the corrected or legacy ids
10201 ('ncipher' and '4758_cca' respectively) when binding. NB,
10202 this only applies when building 'shared'.
10204 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10206 * Add attribute functions to EVP_PKEY structure. Modify
10207 PKCS12_create() to recognize a CSP name attribute and
10208 use it. Make -CSP option work again in pkcs12 utility.
10212 * Add new functionality to the bn blinding code:
10213 - automatic re-creation of the BN_BLINDING parameters after
10214 a fixed number of uses (currently 32)
10215 - add new function for parameter creation
10216 - introduce flags to control the update behaviour of the
10217 BN_BLINDING parameters
10218 - hide BN_BLINDING structure
10219 Add a second BN_BLINDING slot to the RSA structure to improve
10220 performance when a single RSA object is shared among several
10225 * Add support for DTLS.
10227 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10229 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10230 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10234 * Remove buggy and incomplete DH cert support from
10235 ssl/ssl_rsa.c and ssl/s3_both.c
10239 * Use SHA-1 instead of MD5 as the default digest algorithm for
10240 the `apps/openssl` commands.
10244 * Compile clean with "-Wall -Wmissing-prototypes
10245 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10246 DEBUG_SAFESTACK must also be set.
10250 * Change ./Configure so that certain algorithms can be disabled by default.
10251 The new counterpiece to "no-xxx" is "enable-xxx".
10253 The patented RC5 and MDC2 algorithms will now be disabled unless
10254 "enable-rc5" and "enable-mdc2", respectively, are specified.
10256 (IDEA remains enabled despite being patented. This is because IDEA
10257 is frequently required for interoperability, and there is no license
10258 fee for non-commercial use. As before, "no-idea" can be used to
10259 avoid this algorithm.)
10263 * Add processing of proxy certificates (see RFC 3820). This work was
10264 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10265 EGEE (Enabling Grids for E-science in Europe).
10269 * RC4 performance overhaul on modern architectures/implementations, such
10270 as Intel P4, IA-64 and AMD64.
10274 * New utility extract-section.pl. This can be used specify an alternative
10275 section number in a pod file instead of having to treat each file as
10276 a separate case in Makefile. This can be done by adding two lines to the
10279 =for comment openssl_section:XXX
10281 The blank line is mandatory.
10285 * New arguments -certform, -keyform and -pass for s_client and s_server
10286 to allow alternative format key and certificate files and passphrase
10291 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10292 update associated structures and add various utility functions.
10294 Add new policy related verify parameters, include policy checking in
10295 standard verify code. Enhance 'smime' application with extra parameters
10296 to support policy checking and print out.
10300 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10301 Nehemiah processors. These extensions support AES encryption in hardware
10302 as well as RNG (though RNG support is currently disabled).
10304 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10306 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10310 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10312 *Andy Polyakov and a number of other people*
10314 * Improved PowerPC platform support. Most notably BIGNUM assembler
10315 implementation contributed by IBM.
10317 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10319 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10320 exponent rather than 'unsigned long'. There is a corresponding change to
10321 the new 'rsa_keygen' element of the RSA_METHOD structure.
10323 *Jelte Jansen, Geoff Thorpe*
10325 * Functionality for creating the initial serial number file is now
10326 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10328 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10329 number file to 1, which is bound to cause problems. To avoid
10330 the problems while respecting compatibility between different 0.9.7
10331 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10332 CA.pl for serial number initialization. With the new release 0.9.8,
10333 we can fix the problem directly in the 'ca' utility.)
10337 * Reduced header interdependencies by declaring more opaque objects in
10338 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10339 give fewer recursive includes, which could break lazy source code - so
10340 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10341 developers should define this symbol when building and using openssl to
10342 ensure they track the recommended behaviour, interfaces, [etc], but
10343 backwards-compatible behaviour prevails when this isn't defined.
10347 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10351 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10352 This will generate a random key of the appropriate length based on the
10353 cipher context. The EVP_CIPHER can provide its own random key generation
10354 routine to support keys of a specific form. This is used in the des and
10355 3des routines to generate a key of the correct parity. Update S/MIME
10356 code to use new functions and hence generate correct parity DES keys.
10357 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10358 valid (weak or incorrect parity).
10362 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10363 as looking them up. This is useful when the verified structure may contain
10364 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10365 present unless the new PKCS7_NO_CRL flag is asserted.
10369 * Extend ASN1 oid configuration module. It now additionally accepts the
10372 shortName = some long name, 1.2.3.4
10376 * Reimplemented the BN_CTX implementation. There is now no more static
10377 limitation on the number of variables it can handle nor the depth of the
10378 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10379 information can now expand as required, and rather than having a single
10380 static array of bignums, BN_CTX now uses a linked-list of such arrays
10381 allowing it to expand on demand whilst maintaining the usefulness of
10382 BN_CTX's "bundling".
10386 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10387 to allow all RSA operations to function using a single BN_CTX.
10391 * Preliminary support for certificate policy evaluation and checking. This
10392 is initially intended to pass the tests outlined in "Conformance Testing
10393 of Relying Party Client Certificate Path Processing Logic" v1.07.
10397 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10398 remained unused and not that useful. A variety of other little bignum
10399 tweaks and fixes have also been made continuing on from the audit (see
10404 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10405 associated ASN1, EVP and SSL functions and old ASN1 macros.
10409 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10410 and this should never fail. So the return value from the use of
10411 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10412 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10416 * BN_CTX_get() should return zero-valued bignums, providing the same
10417 initialised value as BN_new().
10419 *Geoff Thorpe, suggested by Ulf Möller*
10421 * Support for inhibitAnyPolicy certificate extension.
10425 * An audit of the BIGNUM code is underway, for which debugging code is
10426 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10427 is considered valid when processing BIGNUMs, and causes execution to
10428 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10429 further steps are taken to deliberately pollute unused data in BIGNUM
10430 structures to try and expose faulty code further on. For now, openssl will
10431 (in its default mode of operation) continue to tolerate the inconsistent
10432 forms that it has tolerated in the past, but authors and packagers should
10433 consider trying openssl and their own applications when compiled with
10434 these debugging symbols defined. It will help highlight potential bugs in
10435 their own code, and will improve the test coverage for OpenSSL itself. At
10436 some point, these tighter rules will become openssl's default to improve
10437 maintainability, though the assert()s and other overheads will remain only
10438 in debugging configurations. See bn.h for more details.
10440 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10442 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10443 that can only be obtained through BN_CTX_new() (which implicitly
10444 initialises it). The presence of this function only made it possible
10445 to overwrite an existing structure (and cause memory leaks).
10449 * Because of the callback-based approach for implementing LHASH as a
10450 template type, lh_insert() adds opaque objects to hash-tables and
10451 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10452 to clean up those corresponding objects before destroying the hash table
10453 (and losing the object pointers). So some over-zealous constifications in
10454 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10455 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10456 prototyped to have "const" restrictions on the object pointers they are
10457 given (and so aren't required to cast them away any more).
10461 * The tmdiff.h API was so ugly and minimal that our own timing utility
10462 (speed) prefers to use its own implementation. The two implementations
10463 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10464 its object type properly exposed (MS_TM) instead of casting to/from
10465 `char *`. This may still change yet if someone realises MS_TM and
10467 aren't necessarily the greatest nomenclatures - but this is what was used
10468 internally to the implementation so I've used that for now.
10472 * Ensure that deprecated functions do not get compiled when
10473 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10474 the self-tests were still using deprecated key-generation functions so
10475 these have been updated also.
10479 * Reorganise PKCS#7 code to separate the digest location functionality
10480 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10481 New function PKCS7_set_digest() to set the digest type for PKCS#7
10482 digestedData type. Add additional code to correctly generate the
10483 digestedData type and add support for this type in PKCS7 initialization
10488 * New function PKCS7_set0_type_other() this initializes a PKCS7
10489 structure of type "other".
10493 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10494 sure the loop does correctly stop and breaking ("division by zero")
10495 modulus operations are not performed. The (pre-generated) prime
10496 table crypto/bn/bn_prime.h was already correct, but it could not be
10497 re-generated on some platforms because of the "division by zero"
10498 situation in the script.
10500 *Ralf S. Engelschall*
10502 * Update support for ECC-based TLS ciphersuites according to
10503 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10504 SHA-1 now is only used for "small" curves (where the
10505 representation of a field element takes up to 24 bytes); for
10506 larger curves, the field element resulting from ECDH is directly
10507 used as premaster secret.
10509 *Douglas Stebila (Sun Microsystems Laboratories)*
10511 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10512 curve secp160r1 to the tests.
10514 *Douglas Stebila (Sun Microsystems Laboratories)*
10516 * Add the possibility to load symbols globally with DSO.
10518 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10520 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10521 control of the error stack.
10525 * Add support for STORE in ENGINE.
10529 * Add the STORE type. The intention is to provide a common interface
10530 to certificate and key stores, be they simple file-based stores, or
10531 HSM-type store, or LDAP stores, or...
10532 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10536 * Add a generic structure called OPENSSL_ITEM. This can be used to
10537 pass a list of arguments to any function as well as provide a way
10538 for a function to pass data back to the caller.
10542 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10543 works like BUF_strdup() but can be used to duplicate a portion of
10544 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10549 * Add the function sk_find_ex() which works like sk_find(), but will
10550 return an index to an element even if an exact match couldn't be
10551 found. The index is guaranteed to point at the element where the
10552 searched-for key would be inserted to preserve sorting order.
10556 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10557 takes an extra flags argument for optional functionality. Currently,
10558 the following flags are defined:
10560 OBJ_BSEARCH_VALUE_ON_NOMATCH
10561 This one gets OBJ_bsearch_ex() to return a pointer to the first
10562 element where the comparing function returns a negative or zero
10565 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10566 This one gets OBJ_bsearch_ex() to return a pointer to the first
10567 element where the comparing function returns zero. This is useful
10568 if there are more than one element where the comparing function
10573 * Make it possible to create self-signed certificates with 'openssl ca'
10574 in such a way that the self-signed certificate becomes part of the
10575 CA database and uses the same mechanisms for serial number generation
10576 as all other certificate signing. The new flag '-selfsign' enables
10577 this functionality. Adapt CA.sh and CA.pl.in.
10581 * Add functionality to check the public key of a certificate request
10582 against a given private. This is useful to check that a certificate
10583 request can be signed by that key (self-signing).
10587 * Make it possible to have multiple active certificates with the same
10588 subject in the CA index file. This is done only if the keyword
10589 'unique_subject' is set to 'no' in the main CA section (default
10590 if 'CA_default') of the configuration file. The value is saved
10591 with the database itself in a separate index attribute file,
10592 named like the index file with '.attr' appended to the name.
10596 * Generate multi-valued AVAs using '+' notation in config files for
10601 * Support for nameConstraints certificate extension.
10605 * Support for policyConstraints certificate extension.
10609 * Support for policyMappings certificate extension.
10613 * Make sure the default DSA_METHOD implementation only uses its
10614 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10615 and change its own handlers to be NULL so as to remove unnecessary
10616 indirection. This lets alternative implementations fallback to the
10617 default implementation more easily.
10621 * Support for directoryName in GeneralName related extensions
10626 * Make it possible to link applications using Makefile.shared.
10627 Make that possible even when linking against static libraries!
10631 * Support for single pass processing for S/MIME signing. This now
10632 means that S/MIME signing can be done from a pipe, in addition
10633 cleartext signing (multipart/signed type) is effectively streaming
10634 and the signed data does not need to be all held in memory.
10636 This is done with a new flag PKCS7_STREAM. When this flag is set
10637 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10638 is done after the data is output (and digests calculated) in
10639 SMIME_write_PKCS7().
10643 * Add full support for -rpath/-R, both in shared libraries and
10644 applications, at least on the platforms where it's known how
10649 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10650 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10651 will now compute a table of multiples of the generator that
10652 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10653 faster (notably in the case of a single point multiplication,
10654 scalar * generator).
10656 *Nils Larsch, Bodo Moeller*
10658 * IPv6 support for certificate extensions. The various extensions
10659 which use the IP:a.b.c.d can now take IPv6 addresses using the
10660 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10665 * Added an ENGINE that implements RSA by performing private key
10666 exponentiations with the GMP library. The conversions to and from
10667 GMP's mpz_t format aren't optimised nor are any montgomery forms
10668 cached, and on x86 it appears OpenSSL's own performance has caught up.
10669 However there are likely to be other architectures where GMP could
10670 provide a boost. This ENGINE is not built in by default, but it can be
10671 specified at Configure time and should be accompanied by the necessary
10672 linker additions, eg;
10673 ./config -DOPENSSL_USE_GMP -lgmp
10677 * "openssl engine" will not display ENGINE/DSO load failure errors when
10678 testing availability of engines with "-t" - the old behaviour is
10679 produced by increasing the feature's verbosity with "-tt".
10683 * ECDSA routines: under certain error conditions uninitialized BN objects
10684 could be freed. Solution: make sure initialization is performed early
10685 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10690 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10691 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10692 software implementations. For DSA and DH, parameter generation can
10693 also be overridden by providing the appropriate method callbacks.
10697 * Change the "progress" mechanism used in key-generation and
10698 primality testing to functions that take a new BN_GENCB pointer in
10699 place of callback/argument pairs. The new API functions have `_ex`
10700 postfixes and the older functions are reimplemented as wrappers for
10701 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10702 declarations of the old functions to help (graceful) attempts to
10703 migrate to the new functions. Also, the new key-generation API
10704 functions operate on a caller-supplied key-structure and return
10705 success/failure rather than returning a key or NULL - this is to
10706 help make "keygen" another member function of RSA_METHOD etc.
10708 Example for using the new callback interface:
10710 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10711 void *my_arg = ...;
10714 BN_GENCB_set(&my_cb, my_callback, my_arg);
10716 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10717 /* For the meaning of a, b in calls to my_callback(), see the
10718 * documentation of the function that calls the callback.
10719 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10720 * my_callback should return 1 if it wants BN_is_prime_ex()
10721 * to continue, or 0 to stop.
10726 * Change the ZLIB compression method to be stateful, and make it
10727 available to TLS with the number defined in
10728 draft-ietf-tls-compression-04.txt.
10732 * Add the ASN.1 structures and functions for CertificatePair, which
10733 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10735 CertificatePair ::= SEQUENCE {
10736 forward [0] Certificate OPTIONAL,
10737 reverse [1] Certificate OPTIONAL,
10738 -- at least one of the pair shall be present -- }
10740 Also implement the PEM functions to read and write certificate
10741 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10743 This needed to be defined, mostly for the sake of the LDAP
10744 attribute crossCertificatePair, but may prove useful elsewhere as
10749 * Make it possible to inhibit symlinking of shared libraries in
10750 Makefile.shared, for Cygwin's sake.
10754 * Extend the BIGNUM API by creating a function
10755 void BN_set_negative(BIGNUM *a, int neg);
10756 and a macro that behave like
10757 int BN_is_negative(const BIGNUM *a);
10759 to avoid the need to access 'a->neg' directly in applications.
10763 * Implement fast modular reduction for pseudo-Mersenne primes
10764 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10765 EC_GROUP_new_curve_GFp() will now automatically use this
10768 *Nils Larsch <nla@trustcenter.de>*
10770 * Add new lock type (CRYPTO_LOCK_BN).
10774 * Change the ENGINE framework to automatically load engines
10775 dynamically from specific directories unless they could be
10776 found to already be built in or loaded. Move all the
10777 current engines except for the cryptodev one to a new
10778 directory engines/.
10779 The engines in engines/ are built as shared libraries if
10780 the "shared" options was given to ./Configure or ./config.
10781 Otherwise, they are inserted in libcrypto.a.
10782 /usr/local/ssl/engines is the default directory for dynamic
10783 engines, but that can be overridden at configure time through
10784 the usual use of --prefix and/or --openssldir, and at run
10785 time with the environment variable OPENSSL_ENGINES.
10787 *Geoff Thorpe and Richard Levitte*
10789 * Add Makefile.shared, a helper makefile to build shared
10790 libraries. Adapt Makefile.org.
10794 * Add version info to Win32 DLLs.
10796 *Peter 'Luna' Runestig" <peter@runestig.com>*
10798 * Add new 'medium level' PKCS#12 API. Certificates and keys
10799 can be added using this API to created arbitrary PKCS#12
10800 files while avoiding the low-level API.
10802 New options to PKCS12_create(), key or cert can be NULL and
10803 will then be omitted from the output file. The encryption
10804 algorithm NIDs can be set to -1 for no encryption, the mac
10805 iteration count can be set to 0 to omit the mac.
10807 Enhance pkcs12 utility by making the -nokeys and -nocerts
10808 options work when creating a PKCS#12 file. New option -nomac
10809 to omit the mac, NONE can be set for an encryption algorithm.
10810 New code is modified to use the enhanced PKCS12_create()
10811 instead of the low-level API.
10815 * Extend ASN1 encoder to support indefinite length constructed
10816 encoding. This can output sequences tags and octet strings in
10817 this form. Modify pk7_asn1.c to support indefinite length
10818 encoding. This is experimental and needs additional code to
10819 be useful, such as an ASN1 bio and some enhanced streaming
10822 Extend template encode functionality so that tagging is passed
10823 down to the template encoder.
10827 * Let 'openssl req' fail if an argument to '-newkey' is not
10828 recognized instead of using RSA as a default.
10832 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10833 As these are not official, they are not included in "ALL";
10834 the "ECCdraft" ciphersuite group alias can be used to select them.
10836 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10838 * Add ECDH engine support.
10840 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10842 * Add ECDH in new directory crypto/ecdh/.
10844 *Douglas Stebila (Sun Microsystems Laboratories)*
10846 * Let BN_rand_range() abort with an error after 100 iterations
10847 without success (which indicates a broken PRNG).
10851 * Change BN_mod_sqrt() so that it verifies that the input value
10852 is really the square of the return value. (Previously,
10853 BN_mod_sqrt would show GIGO behaviour.)
10857 * Add named elliptic curves over binary fields from X9.62, SECG,
10858 and WAP/WTLS; add OIDs that were still missing.
10860 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10862 * Extend the EC library for elliptic curves over binary fields
10863 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10866 EC_GF2m_simple_method
10870 EC_GROUP_new_curve_GF2m
10871 EC_GROUP_set_curve_GF2m
10872 EC_GROUP_get_curve_GF2m
10873 EC_POINT_set_affine_coordinates_GF2m
10874 EC_POINT_get_affine_coordinates_GF2m
10875 EC_POINT_set_compressed_coordinates_GF2m
10877 Point compression for binary fields is disabled by default for
10878 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10881 As binary polynomials are represented as BIGNUMs, various members
10882 of the EC_GROUP and EC_POINT data structures can be shared
10883 between the implementations for prime fields and binary fields;
10884 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10885 are essentially identical to their `..._GFp` counterparts.
10886 (For simplicity, the `..._GFp` prefix has been dropped from
10887 various internal method names.)
10889 An internal 'field_div' method (similar to 'field_mul' and
10890 'field_sqr') has been added; this is used only for binary fields.
10892 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10894 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10895 through methods ('mul', 'precompute_mult').
10897 The generic implementations (now internally called 'ec_wNAF_mul'
10898 and 'ec_wNAF_precomputed_mult') remain the default if these
10899 methods are undefined.
10901 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10903 * New function EC_GROUP_get_degree, which is defined through
10904 EC_METHOD. For curves over prime fields, this returns the bit
10905 length of the modulus.
10907 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10909 * New functions EC_GROUP_dup, EC_POINT_dup.
10910 (These simply call ..._new and ..._copy).
10912 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10914 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10915 Polynomials are represented as BIGNUMs (where the sign bit is not
10916 used) in the following functions [macros]:
10919 BN_GF2m_sub [= BN_GF2m_add]
10920 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10921 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10922 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10924 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10925 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10926 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10927 BN_GF2m_cmp [= BN_ucmp]
10929 (Note that only the 'mod' functions are actually for fields GF(2^m).
10930 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10932 For some functions, an the irreducible polynomial defining a
10933 field can be given as an 'unsigned int[]' with strictly
10934 decreasing elements giving the indices of those bits that are set;
10935 i.e., p[] represents the polynomial
10936 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10938 p[0] > p[1] > ... > p[k] = 0.
10939 This applies to the following functions:
10942 BN_GF2m_mod_mul_arr
10943 BN_GF2m_mod_sqr_arr
10944 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10945 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10946 BN_GF2m_mod_exp_arr
10947 BN_GF2m_mod_sqrt_arr
10948 BN_GF2m_mod_solve_quad_arr
10952 Conversion can be performed by the following functions:
10957 bntest.c has additional tests for binary polynomial arithmetic.
10959 Two implementations for BN_GF2m_mod_div() are available.
10960 The default algorithm simply uses BN_GF2m_mod_inv() and
10961 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10962 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10963 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10965 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10967 * Add new error code 'ERR_R_DISABLED' that can be used when some
10968 functionality is disabled at compile-time.
10970 *Douglas Stebila <douglas.stebila@sun.com>*
10972 * Change default behaviour of 'openssl asn1parse' so that more
10973 information is visible when viewing, e.g., a certificate:
10975 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10976 mode the content of non-printable OCTET STRINGs is output in a
10977 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10978 avoid the appearance of a printable string.
10980 *Nils Larsch <nla@trustcenter.de>*
10982 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10984 EC_GROUP_set_asn1_flag()
10985 EC_GROUP_get_asn1_flag()
10986 EC_GROUP_set_point_conversion_form()
10987 EC_GROUP_get_point_conversion_form()
10988 These control ASN1 encoding details:
10989 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10990 has been set to OPENSSL_EC_NAMED_CURVE.
10991 - Points are encoded in uncompressed form by default; options for
10992 asn1_for are as for point2oct, namely
10993 POINT_CONVERSION_COMPRESSED
10994 POINT_CONVERSION_UNCOMPRESSED
10995 POINT_CONVERSION_HYBRID
10997 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10999 EC_GROUP_set_seed()
11000 EC_GROUP_get0_seed()
11001 EC_GROUP_get_seed_len()
11002 This is used only for ASN1 purposes (so far).
11004 *Nils Larsch <nla@trustcenter.de>*
11006 * Add 'field_type' member to EC_METHOD, which holds the NID
11007 of the appropriate field type OID. The new function
11008 EC_METHOD_get_field_type() returns this value.
11010 *Nils Larsch <nla@trustcenter.de>*
11013 EC_POINT_point2bn()
11014 EC_POINT_bn2point()
11015 EC_POINT_point2hex()
11016 EC_POINT_hex2point()
11017 providing useful interfaces to EC_POINT_point2oct() and
11018 EC_POINT_oct2point().
11020 *Nils Larsch <nla@trustcenter.de>*
11022 * Change internals of the EC library so that the functions
11023 EC_GROUP_set_generator()
11024 EC_GROUP_get_generator()
11025 EC_GROUP_get_order()
11026 EC_GROUP_get_cofactor()
11027 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11028 to methods, which would lead to unnecessary code duplication when
11029 adding different types of curves.
11031 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11033 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11034 arithmetic, and such that modified wNAFs are generated
11035 (which avoid length expansion in many cases).
11039 * Add a function EC_GROUP_check_discriminant() (defined via
11040 EC_METHOD) that verifies that the curve discriminant is non-zero.
11042 Add a function EC_GROUP_check() that makes some sanity tests
11043 on a EC_GROUP, its generator and order. This includes
11044 EC_GROUP_check_discriminant().
11046 *Nils Larsch <nla@trustcenter.de>*
11048 * Add ECDSA in new directory crypto/ecdsa/.
11050 Add applications 'openssl ecparam' and 'openssl ecdsa'
11051 (these are based on 'openssl dsaparam' and 'openssl dsa').
11053 ECDSA support is also included in various other files across the
11054 library. Most notably,
11055 - 'openssl req' now has a '-newkey ecdsa:file' option;
11056 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11057 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11058 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11059 them suitable for ECDSA where domain parameters must be
11060 extracted before the specific public key;
11061 - ECDSA engine support has been added.
11063 *Nils Larsch <nla@trustcenter.de>*
11065 * Include some named elliptic curves, and add OIDs from X9.62,
11066 SECG, and WAP/WTLS. Each curve can be obtained from the new
11068 EC_GROUP_new_by_curve_name(),
11069 and the list of available named curves can be obtained with
11070 EC_get_builtin_curves().
11071 Also add a 'curve_name' member to EC_GROUP objects, which can be
11073 EC_GROUP_set_curve_name()
11074 EC_GROUP_get_curve_name()
11076 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11078 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11079 was actually never needed) and in BN_mul(). The removal in BN_mul()
11080 required a small change in bn_mul_part_recursive() and the addition
11081 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11082 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11083 bn_sub_words() and bn_add_words() except they take arrays with
11088 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11090 * Cleanse PEM buffers before freeing them since they may contain
11093 *Benjamin Bennett <ben@psc.edu>*
11095 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11096 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11097 authentication-only ciphersuites.
11101 * Since AES128 and AES256 share a single mask bit in the logic of
11102 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11103 kludge to work properly if AES128 is available and AES256 isn't.
11107 * Expand security boundary to match 1.1.1 module.
11111 * Remove redundant features: hash file source, editing of test vectors
11112 modify fipsld to use external fips_premain.c signature.
11116 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11117 run algorithm test programs.
11121 * Make algorithm test programs more tolerant of whitespace.
11125 * Have SSL/TLS server implementation tolerate "mismatched" record
11126 protocol version while receiving ClientHello even if the
11127 ClientHello is fragmented. (The server can't insist on the
11128 particular protocol version it has chosen before the ServerHello
11129 message has informed the client about his choice.)
11133 * Load error codes if they are not already present instead of using a
11134 static variable. This allows them to be cleanly unloaded and reloaded.
11138 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11140 * Introduce limits to prevent malicious keys being able to
11141 cause a denial of service. ([CVE-2006-2940])
11143 *Steve Henson, Bodo Moeller*
11145 * Fix ASN.1 parsing of certain invalid structures that can result
11146 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11148 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11149 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11151 * Fix SSL client code which could crash if connecting to a
11152 malicious SSLv2 server. ([CVE-2006-4343])
11154 *Tavis Ormandy and Will Drewry, Google Security Team*
11156 * Change ciphersuite string processing so that an explicit
11157 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11158 will no longer include "AES128-SHA"), and any other similar
11159 ciphersuite (same bitmap) from *other* protocol versions (so that
11160 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11161 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11162 changes from 0.9.8b and 0.9.8d.
11166 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11168 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11169 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11171 * Change the Unix randomness entropy gathering to use poll() when
11172 possible instead of select(), since the latter has some
11173 undesirable limitations.
11175 *Darryl Miles via Richard Levitte and Bodo Moeller*
11177 * Disable rogue ciphersuites:
11179 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11180 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11181 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11183 The latter two were purportedly from
11184 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11187 Also deactivate the remaining ciphersuites from
11188 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11189 unofficial, and the ID has long expired.
11193 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11194 dual-core machines) and other potential thread-safety issues.
11198 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11200 * Adapt fipsld and the build system to link against the validated FIPS
11201 module in FIPS mode.
11205 * Fixes for VC++ 2005 build under Windows.
11209 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11210 from a Windows bash shell such as MSYS. It is autodetected from the
11211 "config" script when run from a VC++ environment. Modify standard VC++
11212 build to use fipscanister.o from the GNU make build.
11216 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11218 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11219 The value now differs depending on if you build for FIPS or not.
11220 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11221 safely run with a non-FIPSed libcrypto, as it may crash because of
11222 the difference induced by this change.
11226 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11228 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11229 (part of SSL_OP_ALL). This option used to disable the
11230 countermeasure against man-in-the-middle protocol-version
11231 rollback in the SSL 2.0 server implementation, which is a bad
11232 idea. ([CVE-2005-2969])
11234 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11235 for Information Security, National Institute of Advanced Industrial
11236 Science and Technology [AIST, Japan)]*
11238 * Minimal support for X9.31 signatures and PSS padding modes. This is
11239 mainly for FIPS compliance and not fully integrated at this stage.
11243 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11244 the exponentiation using a fixed-length exponent. (Otherwise,
11245 the information leaked through timing could expose the secret key
11246 after many signatures; cf. Bleichenbacher's attack on DSA with
11251 * Make a new fixed-window mod_exp implementation the default for
11252 RSA, DSA, and DH private-key operations so that the sequence of
11253 squares and multiplies and the memory access pattern are
11254 independent of the particular secret key. This will mitigate
11255 cache-timing and potential related attacks.
11257 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11258 and this is automatically used by BN_mod_exp_mont() if the new flag
11259 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11260 will use this BN flag for private exponents unless the flag
11261 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11262 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11264 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11266 * Change the client implementation for SSLv23_method() and
11267 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11268 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11269 (Previously, the SSL 2.0 backwards compatible Client Hello
11270 message format would be used even with SSL_OP_NO_SSLv2.)
11274 * Add support for smime-type MIME parameter in S/MIME messages which some
11279 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11280 a threadsafe manner. Modify rsa code to use new function and add calls
11281 to dsa and dh code (which had race conditions before).
11285 * Include the fixed error library code in the C error file definitions
11286 instead of fixing them up at runtime. This keeps the error code
11287 structures constant.
11291 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11293 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11296 * Fixes for newer kerberos headers. NB: the casts are needed because
11297 the 'length' field is signed on one version and unsigned on another
11298 with no (?) obvious way to tell the difference, without these VC++
11299 complains. Also the "definition" of FAR (blank) is no longer included
11300 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11301 some needed definitions.
11305 * Undo Cygwin change.
11309 * Added support for proxy certificates according to RFC 3820.
11310 Because they may be a security thread to unaware applications,
11311 they must be explicitly allowed in run-time. See
11312 docs/HOWTO/proxy_certificates.txt for further information.
11316 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11318 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11319 server and client random values. Previously
11320 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11321 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11323 This change has negligible security impact because:
11325 1. Server and client random values still have 24 bytes of pseudo random
11328 2. Server and client random values are sent in the clear in the initial
11331 3. The master secret is derived using the premaster secret (48 bytes in
11332 size for static RSA ciphersuites) as well as client server and random
11335 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11338 *Stephen Henson, reported by UK NISCC*
11340 * Use Windows randomness collection on Cygwin.
11344 * Fix hang in EGD/PRNGD query when communication socket is closed
11345 prematurely by EGD/PRNGD.
11347 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11349 * Prompt for pass phrases when appropriate for PKCS12 input format.
11353 * Back-port of selected performance improvements from development
11354 branch, as well as improved support for PowerPC platforms.
11358 * Add lots of checks for memory allocation failure, error codes to indicate
11359 failure and freeing up memory if a failure occurs.
11361 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11363 * Add new -passin argument to dgst.
11367 * Perform some character comparisons of different types in X509_NAME_cmp:
11368 this is needed for some certificates that re-encode DNs into UTF8Strings
11369 (in violation of RFC3280) and can't or won't issue name rollover
11374 * Make an explicit check during certificate validation to see that
11375 the CA setting in each certificate on the chain is correct. As a
11376 side effect always do the following basic checks on extensions,
11377 not just when there's an associated purpose to the check:
11379 - if there is an unhandled critical extension (unless the user
11380 has chosen to ignore this fault)
11381 - if the path length has been exceeded (if one is set at all)
11382 - that certain extensions fit the associated purpose (if one has
11387 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11389 * Avoid a race condition when CRLs are checked in a multi threaded
11390 environment. This would happen due to the reordering of the revoked
11391 entries during signature checking and serial number lookup. Now the
11392 encoding is cached and the serial number sort performed under a lock.
11393 Add new STACK function sk_is_sorted().
11397 * Add Delta CRL to the extension code.
11401 * Various fixes to s3_pkt.c so alerts are sent properly.
11403 *David Holmes <d.holmes@f5.com>*
11405 * Reduce the chances of duplicate issuer name and serial numbers (in
11406 violation of RFC3280) using the OpenSSL certificate creation utilities.
11407 This is done by creating a random 64 bit value for the initial serial
11408 number when a serial number file is created or when a self signed
11409 certificate is created using 'openssl req -x509'. The initial serial
11410 number file is created using 'openssl x509 -next_serial' in CA.pl
11411 rather than being initialized to 1.
11415 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11417 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11418 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11420 *Joe Orton, Steve Henson*
11422 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11425 *Joe Orton, Steve Henson*
11427 * Make it possible to have multiple active certificates with the same
11428 subject in the CA index file. This is done only if the keyword
11429 'unique_subject' is set to 'no' in the main CA section (default
11430 if 'CA_default') of the configuration file. The value is saved
11431 with the database itself in a separate index attribute file,
11432 named like the index file with '.attr' appended to the name.
11436 * X509 verify fixes. Disable broken certificate workarounds when
11437 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11438 keyUsage extension present. Don't accept CRLs with unhandled critical
11439 extensions: since verify currently doesn't process CRL extensions this
11440 rejects a CRL with *any* critical extensions. Add new verify error codes
11445 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11446 A clarification of RFC2560 will require the use of OCTET STRINGs and
11447 some implementations cannot handle the current raw format. Since OpenSSL
11448 copies and compares OCSP nonces as opaque blobs without any attempt at
11449 parsing them this should not create any compatibility issues.
11453 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11454 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11455 this HMAC (and other) operations are several times slower than OpenSSL
11460 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11462 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11464 * Use the correct content when signing type "other".
11468 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11470 * Fix various bugs revealed by running the NISCC test suite:
11472 Stop out of bounds reads in the ASN1 code when presented with
11473 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11475 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11477 If verify callback ignores invalid public key errors don't try to check
11478 certificate signature with the NULL public key.
11482 * New -ignore_err option in ocsp application to stop the server
11483 exiting on the first error in a request.
11487 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11488 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11493 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11494 extra data after the compression methods not only for TLS 1.0
11495 but also for SSL 3.0 (as required by the specification).
11497 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11499 * Change X509_certificate_type() to mark the key as exported/exportable
11500 when it's 512 *bits* long, not 512 bytes.
11504 * Change AES_cbc_encrypt() so it outputs exact multiple of
11505 blocks during encryption.
11509 * Various fixes to base64 BIO and non blocking I/O. On write
11510 flushes were not handled properly if the BIO retried. On read
11511 data was not being buffered properly and had various logic bugs.
11512 This also affects blocking I/O when the data being decoded is a
11517 * Various S/MIME bugfixes and compatibility changes:
11518 output correct application/pkcs7 MIME type if
11519 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11520 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11521 of files as .eml work). Correctly handle very long lines in MIME
11526 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11528 * Countermeasure against the Klima-Pokorny-Rosa extension of
11529 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11530 a protocol version number mismatch like a decryption error
11531 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11535 * Turn on RSA blinding by default in the default implementation
11536 to avoid a timing attack. Applications that don't want it can call
11537 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11538 They would be ill-advised to do so in most cases.
11540 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11542 * Change RSA blinding code so that it works when the PRNG is not
11543 seeded (in this case, the secret RSA exponent is abused as
11544 an unpredictable seed -- if it is not unpredictable, there
11545 is no point in blinding anyway). Make RSA blinding thread-safe
11546 by remembering the creator's thread ID in rsa->blinding and
11547 having all other threads use local one-time blinding factors
11548 (this requires more computation than sharing rsa->blinding, but
11549 avoids excessive locking; and if an RSA object is not shared
11550 between threads, blinding will still be very fast).
11554 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11555 ENGINE as defaults for all supported algorithms irrespective of
11556 the 'flags' parameter. 'flags' is now honoured, so applications
11557 should make sure they are passing it correctly.
11561 * Target "mingw" now allows native Windows code to be generated in
11562 the Cygwin environment as well as with the MinGW compiler.
11566 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11568 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11569 via timing by performing a MAC computation even if incorrect
11570 block cipher padding has been found. This is a countermeasure
11571 against active attacks where the attacker has to distinguish
11572 between bad padding and a MAC verification error. ([CVE-2003-0078])
11574 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11575 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11576 Martin Vuagnoux (EPFL, Ilion)*
11578 * Make the no-err option work as intended. The intention with no-err
11579 is not to have the whole error stack handling routines removed from
11580 libcrypto, it's only intended to remove all the function name and
11581 reason texts, thereby removing some of the footprint that may not
11582 be interesting if those errors aren't displayed anyway.
11584 NOTE: it's still possible for any application or module to have its
11585 own set of error texts inserted. The routines are there, just not
11586 used by default when no-err is given.
11590 * Add support for FreeBSD on IA64.
11592 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11594 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11595 Kerberos function mit_des_cbc_cksum(). Before this change,
11596 the value returned by DES_cbc_cksum() was like the one from
11597 mit_des_cbc_cksum(), except the bytes were swapped.
11599 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11601 * Allow an application to disable the automatic SSL chain building.
11602 Before this a rather primitive chain build was always performed in
11603 ssl3_output_cert_chain(): an application had no way to send the
11604 correct chain if the automatic operation produced an incorrect result.
11606 Now the chain builder is disabled if either:
11608 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11610 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11612 The reasoning behind this is that an application would not want the
11613 auto chain building to take place if extra chain certificates are
11614 present and it might also want a means of sending no additional
11615 certificates (for example the chain has two certificates and the
11620 * Add the possibility to build without the ENGINE framework.
11622 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11624 * Under Win32 gmtime() can return NULL: check return value in
11625 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11629 * DSA routines: under certain error conditions uninitialized BN objects
11630 could be freed. Solution: make sure initialization is performed early
11631 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11632 Nils Larsch <nla@trustcenter.de> via PR#459)
11636 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11637 checked on reconnect on the client side, therefore session resumption
11638 could still fail with a "ssl session id is different" error. This
11639 behaviour is masked when SSL_OP_ALL is used due to
11640 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11641 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11642 followup to PR #377.
11646 * IA-32 assembler support enhancements: unified ELF targets, support
11647 for SCO/Caldera platforms, fix for Cygwin shared build.
11651 * Add support for FreeBSD on sparc64. As a consequence, support for
11652 FreeBSD on non-x86 processors is separate from x86 processors on
11653 the config script, much like the NetBSD support.
11655 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11657 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11659 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11662 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11663 code (06) was taken as the first octet of the session ID and the last
11664 octet was ignored consequently. As a result SSLv2 client side session
11665 caching could not have worked due to the session ID mismatch between
11667 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11672 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11673 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11678 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11679 seems that in spite of existing for more than a year, many application
11680 author have done nothing to provide the necessary callbacks, which
11681 means that this particular engine will not work properly anywhere.
11682 This is a very unfortunate situation which forces us, in the name
11683 of usability, to give the hw_ncipher.c a static lock, which is part
11685 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11686 appear in 0.9.8 or later. We EXPECT application authors to have
11687 dealt properly with this when 0.9.8 is released (unless we actually
11688 make such changes in the libcrypto locking code that changes will
11689 have to be made anyway).
11693 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11694 octets have been read, EOF or an error occurs. Without this change
11695 some truncated ASN1 structures will not produce an error.
11699 * Disable Heimdal support, since it hasn't been fully implemented.
11700 Still give the possibility to force the use of Heimdal, but with
11701 warnings and a request that patches get sent to openssl-dev.
11705 * Add the VC-CE target, introduce the WINCE sysname, and add
11706 INSTALL.WCE and appropriate conditionals to make it build.
11708 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11710 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11711 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11712 edit numbers of the version.
11714 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11716 * Introduce safe string copy and catenation functions
11717 (BUF_strlcpy() and BUF_strlcat()).
11719 *Ben Laurie (CHATS) and Richard Levitte*
11721 * Avoid using fixed-size buffers for one-line DNs.
11723 *Ben Laurie (CHATS)*
11725 * Add BUF_MEM_grow_clean() to avoid information leakage when
11726 resizing buffers containing secrets, and use where appropriate.
11728 *Ben Laurie (CHATS)*
11730 * Avoid using fixed size buffers for configuration file location.
11732 *Ben Laurie (CHATS)*
11734 * Avoid filename truncation for various CA files.
11736 *Ben Laurie (CHATS)*
11738 * Use sizeof in preference to magic numbers.
11740 *Ben Laurie (CHATS)*
11742 * Avoid filename truncation in cert requests.
11744 *Ben Laurie (CHATS)*
11746 * Add assertions to check for (supposedly impossible) buffer
11749 *Ben Laurie (CHATS)*
11751 * Don't cache truncated DNS entries in the local cache (this could
11752 potentially lead to a spoofing attack).
11754 *Ben Laurie (CHATS)*
11756 * Fix various buffers to be large enough for hex/decimal
11757 representations in a platform independent manner.
11759 *Ben Laurie (CHATS)*
11761 * Add CRYPTO_realloc_clean() to avoid information leakage when
11762 resizing buffers containing secrets, and use where appropriate.
11764 *Ben Laurie (CHATS)*
11766 * Add BIO_indent() to avoid much slightly worrying code to do
11769 *Ben Laurie (CHATS)*
11771 * Convert sprintf()/BIO_puts() to BIO_printf().
11773 *Ben Laurie (CHATS)*
11775 * buffer_gets() could terminate with the buffer only half
11778 *Ben Laurie (CHATS)*
11780 * Add assertions to prevent user-supplied crypto functions from
11781 overflowing internal buffers by having large block sizes, etc.
11783 *Ben Laurie (CHATS)*
11785 * New OPENSSL_assert() macro (similar to assert(), but enabled
11788 *Ben Laurie (CHATS)*
11790 * Eliminate unused copy of key in RC4.
11792 *Ben Laurie (CHATS)*
11794 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11796 *Ben Laurie (CHATS)*
11798 * Fix off-by-one error in EGD path.
11800 *Ben Laurie (CHATS)*
11802 * If RANDFILE path is too long, ignore instead of truncating.
11804 *Ben Laurie (CHATS)*
11806 * Eliminate unused and incorrectly sized X.509 structure
11809 *Ben Laurie (CHATS)*
11811 * Eliminate unused and dangerous function knumber().
11813 *Ben Laurie (CHATS)*
11815 * Eliminate unused and dangerous structure, KSSL_ERR.
11817 *Ben Laurie (CHATS)*
11819 * Protect against overlong session ID context length in an encoded
11820 session object. Since these are local, this does not appear to be
11823 *Ben Laurie (CHATS)*
11825 * Change from security patch (see 0.9.6e below) that did not affect
11826 the 0.9.6 release series:
11828 Remote buffer overflow in SSL3 protocol - an attacker could
11829 supply an oversized master key in Kerberos-enabled versions.
11832 *Ben Laurie (CHATS)*
11834 * Change the SSL kerb5 codes to match RFC 2712.
11838 * Make -nameopt work fully for req and add -reqopt switch.
11840 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11842 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11844 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11846 * Make sure tests can be performed even if the corresponding algorithms
11847 have been removed entirely. This was also the last step to make
11848 OpenSSL compilable with DJGPP under all reasonable conditions.
11850 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11852 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11853 to allow version independent disabling of normally unselected ciphers,
11854 which may be activated as a side-effect of selecting a single cipher.
11856 (E.g., cipher list string "RSA" enables ciphersuites that are left
11857 out of "ALL" because they do not provide symmetric encryption.
11858 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11860 *Lutz Jaenicke, Bodo Moeller*
11862 * Add appropriate support for separate platform-dependent build
11863 directories. The recommended way to make a platform-dependent
11864 build directory is the following (tested on Linux), maybe with
11867 # Place yourself outside of the OpenSSL source tree. In
11868 # this example, the environment variable OPENSSL_SOURCE
11869 # is assumed to contain the absolute OpenSSL source directory.
11870 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11871 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11872 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11873 mkdir -p `dirname $F`
11874 ln -s $OPENSSL_SOURCE/$F $F
11877 To be absolutely sure not to disturb the source tree, a "make clean"
11878 is a good thing. If it isn't successful, don't worry about it,
11879 it probably means the source directory is very clean.
11883 * Make sure any ENGINE control commands make local copies of string
11884 pointers passed to them whenever necessary. Otherwise it is possible
11885 the caller may have overwritten (or deallocated) the original string
11886 data when a later ENGINE operation tries to use the stored values.
11888 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11890 * Improve diagnostics in file reading and command-line digests.
11892 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11894 * Add AES modes CFB and OFB to the object database. Correct an
11895 error in AES-CFB decryption.
11899 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11900 allows existing EVP_CIPHER_CTX structures to be reused after
11901 calling `EVP_*Final()`. This behaviour is used by encryption
11902 BIOs and some applications. This has the side effect that
11903 applications must explicitly clean up cipher contexts with
11904 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11908 * Check the values of dna and dnb in bn_mul_recursive before calling
11909 bn_mul_comba (a non zero value means the a or b arrays do not contain
11910 n2 elements) and fallback to bn_mul_normal if either is not zero.
11914 * Fix escaping of non-ASCII characters when using the -subj option
11915 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11919 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11920 form for "surname", serialNumber has no short form.
11921 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11922 therefore remove "mail" short name for "internet 7".
11923 The OID for unique identifiers in X509 certificates is
11924 x500UniqueIdentifier, not uniqueIdentifier.
11925 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11929 * Add an "init" command to the ENGINE config module and auto initialize
11930 ENGINEs. Without any "init" command the ENGINE will be initialized
11931 after all ctrl commands have been executed on it. If init=1 the
11932 ENGINE is initialized at that point (ctrls before that point are run
11933 on the uninitialized ENGINE and after on the initialized one). If
11934 init=0 then the ENGINE will not be initialized at all.
11938 * Fix the 'app_verify_callback' interface so that the user-defined
11939 argument is actually passed to the callback: In the
11940 SSL_CTX_set_cert_verify_callback() prototype, the callback
11941 declaration has been changed from
11944 int (*cb)(X509_STORE_CTX *,void *);
11945 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11946 i=s->ctx->app_verify_callback(&ctx)
11947 has been changed into
11948 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11950 To update applications using SSL_CTX_set_cert_verify_callback(),
11951 a dummy argument can be added to their callback functions.
11953 *D. K. Smetters <smetters@parc.xerox.com>*
11955 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11957 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11959 * Add and OPENSSL_LOAD_CONF define which will cause
11960 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11961 This allows older applications to transparently support certain
11962 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11963 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11964 load the config file and OPENSSL_add_all_algorithms_conf() which will
11965 always load it have also been added.
11969 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11970 Adjust NIDs and EVP layer.
11972 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11974 * Config modules support in openssl utility.
11976 Most commands now load modules from the config file,
11977 though in a few (such as version) this isn't done
11978 because it couldn't be used for anything.
11980 In the case of ca and req the config file used is
11981 the same as the utility itself: that is the -config
11982 command line option can be used to specify an
11987 * Move default behaviour from OPENSSL_config(). If appname is NULL
11988 use "openssl_conf" if filename is NULL use default openssl config file.
11992 * Add an argument to OPENSSL_config() to allow the use of an alternative
11993 config section name. Add a new flag to tolerate a missing config file
11994 and move code to CONF_modules_load_file().
11998 * Support for crypto accelerator cards from Accelerated Encryption
11999 Processing, www.aep.ie. (Use engine 'aep')
12000 The support was copied from 0.9.6c [engine] and adapted/corrected
12001 to work with the new engine framework.
12003 *AEP Inc. and Richard Levitte*
12005 * Support for SureWare crypto accelerator cards from Baltimore
12006 Technologies. (Use engine 'sureware')
12007 The support was copied from 0.9.6c [engine] and adapted
12008 to work with the new engine framework.
12012 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12013 make the newer ENGINE framework commands for the CHIL engine work.
12015 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12017 * Make it possible to produce shared libraries on ReliantUNIX.
12019 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12021 * Add the configuration target debug-linux-ppro.
12022 Make 'openssl rsa' use the general key loading routines
12023 implemented in `apps.c`, and make those routines able to
12024 handle the key format FORMAT_NETSCAPE and the variant
12027 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12029 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12031 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12033 * Add -keyform to rsautl, and document -engine.
12035 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12037 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12038 BIO_R_NO_SUCH_FILE error code rather than the generic
12039 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12043 * Add new functions
12044 ERR_peek_last_error
12045 ERR_peek_last_error_line
12046 ERR_peek_last_error_line_data.
12047 These are similar to
12049 ERR_peek_error_line
12050 ERR_peek_error_line_data,
12051 but report on the latest error recorded rather than the first one
12052 still in the error queue.
12054 *Ben Laurie, Bodo Moeller*
12056 * default_algorithms option in ENGINE config module. This allows things
12058 default_algorithms = ALL
12059 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12063 * Preliminary ENGINE config module.
12067 * New experimental application configuration code.
12071 * Change the AES code to follow the same name structure as all other
12072 symmetric ciphers, and behave the same way. Move everything to
12073 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12075 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12077 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12079 *Ben Laurie and Theo de Raadt*
12081 * Add option to output public keys in req command.
12083 *Massimiliano Pala madwolf@openca.org*
12085 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12086 (up to about 10% better than before for P-192 and P-224).
12090 * New functions/macros
12092 SSL_CTX_set_msg_callback(ctx, cb)
12093 SSL_CTX_set_msg_callback_arg(ctx, arg)
12094 SSL_set_msg_callback(ssl, cb)
12095 SSL_set_msg_callback_arg(ssl, arg)
12097 to request calling a callback function
12099 void cb(int write_p, int version, int content_type,
12100 const void *buf, size_t len, SSL *ssl, void *arg)
12102 whenever a protocol message has been completely received
12103 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12104 protocol version according to which the SSL library interprets
12105 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12106 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12107 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12108 specification (change_cipher_spec(20), alert(21), handshake(22)).
12109 'buf' and 'len' point to the actual message, 'ssl' to the
12110 SSL object, and 'arg' is the application-defined value set by
12111 SSL[_CTX]_set_msg_callback_arg().
12113 'openssl s_client' and 'openssl s_server' have new '-msg' options
12114 to enable a callback that displays all protocol messages.
12118 * Change the shared library support so shared libraries are built as
12119 soon as the corresponding static library is finished, and thereby get
12120 openssl and the test programs linked against the shared library.
12121 This still only happens when the keyword "shard" has been given to
12122 the configuration scripts.
12124 NOTE: shared library support is still an experimental thing, and
12125 backward binary compatibility is still not guaranteed.
12127 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12129 * Add support for Subject Information Access extension.
12131 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12133 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12134 additional bytes when new memory had to be allocated, not just
12135 when reusing an existing buffer.
12139 * New command line and configuration option 'utf8' for the req command.
12140 This allows field values to be specified as UTF8 strings.
12144 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12145 runs for the former and machine-readable output for the latter.
12149 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12150 of the e-mail address in the DN (i.e., it will go into a certificate
12151 extension only). The new configuration file option 'email_in_dn = no'
12152 has the same effect.
12154 *Massimiliano Pala madwolf@openca.org*
12156 * Change all functions with names starting with `des_` to be starting
12157 with `DES_` instead. Add wrappers that are compatible with libdes,
12158 but are named `_ossl_old_des_*`. Finally, add macros that map the
12159 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12160 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12161 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12164 Since we provide two compatibility mappings, the user needs to
12165 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12166 compatibility is desired. The default (i.e., when that macro
12167 isn't defined) is OpenSSL 0.9.6c compatibility.
12169 There are also macros that enable and disable the support of old
12170 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12171 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12172 are defined, the default will apply: to support the old des routines.
12174 In either case, one must include openssl/des.h to get the correct
12175 definitions. Do not try to just include openssl/des_old.h, that
12178 NOTE: This is a major break of an old API into a new one. Software
12179 authors are encouraged to switch to the `DES_` style functions. Some
12180 time in the future, des_old.h and the libdes compatibility functions
12181 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12182 default), and then completely removed.
12186 * Test for certificates which contain unsupported critical extensions.
12187 If such a certificate is found during a verify operation it is
12188 rejected by default: this behaviour can be overridden by either
12189 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12190 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12191 X509_supported_extension() has also been added which returns 1 if a
12192 particular extension is supported.
12196 * Modify the behaviour of EVP cipher functions in similar way to digests
12197 to retain compatibility with existing code.
12201 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12202 compatibility with existing code. In particular the 'ctx' parameter does
12203 not have to be to be initialized before the call to EVP_DigestInit() and
12204 it is tidied up after a call to EVP_DigestFinal(). New function
12205 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12206 EVP_MD_CTX_copy() changed to not require the destination to be
12207 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12208 requires the destination to be valid.
12210 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12211 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12215 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12216 so that complete 'Handshake' protocol structures are kept in memory
12217 instead of overwriting 'msg_type' and 'length' with 'body' data.
12221 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12223 *Massimo Santin via Richard Levitte*
12225 * Major restructuring to the underlying ENGINE code. This includes
12226 reduction of linker bloat, separation of pure "ENGINE" manipulation
12227 (initialisation, etc) from functionality dealing with implementations
12228 of specific crypto interfaces. This change also introduces integrated
12229 support for symmetric ciphers and digest implementations - so ENGINEs
12230 can now accelerate these by providing EVP_CIPHER and EVP_MD
12231 implementations of their own. This is detailed in
12232 [crypto/engine/README.md](crypto/engine/README.md)
12233 as it couldn't be adequately described here. However, there are a few
12234 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12235 were changed in the original introduction of ENGINE code have now
12236 reverted back - the hooking from this code to ENGINE is now a good
12237 deal more passive and at run-time, operations deal directly with
12238 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12239 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12240 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12241 they were not being used by the framework as there is no concept of a
12242 BIGNUM_METHOD and they could not be generalised to the new
12243 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12244 ENGINE_cpy() has been removed as it cannot be consistently defined in
12249 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12253 * Change mkdef.pl to sort symbols that get the same entry number,
12254 and make sure the automatically generated functions `ERR_load_*`
12255 become part of libeay.num as well.
12259 * New function SSL_renegotiate_pending(). This returns true once
12260 renegotiation has been requested (either SSL_renegotiate() call
12261 or HelloRequest/ClientHello received from the peer) and becomes
12262 false once a handshake has been completed.
12263 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12264 sends a HelloRequest, but does not ensure that a handshake takes
12265 place. SSL_renegotiate_pending() is useful for checking if the
12266 client has followed the request.)
12270 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12271 By default, clients may request session resumption even during
12272 renegotiation (if session ID contexts permit); with this option,
12273 session resumption is possible only in the first handshake.
12275 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12276 more bits available for options that should not be part of
12277 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12281 * Add some demos for certificate and certificate request creation.
12285 * Make maximum certificate chain size accepted from the peer application
12286 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12287 "Douglas E. Engert" <deengert@anl.gov>.
12291 * Add support for shared libraries for Unixware-7
12292 (Boyd Lynn Gerber <gerberb@zenez.com>).
12296 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12297 be done prior to destruction. Use this to unload error strings from
12298 ENGINEs that load their own error strings. NB: This adds two new API
12299 functions to "get" and "set" this destroy handler in an ENGINE.
12303 * Alter all existing ENGINE implementations (except "openssl" and
12304 "openbsd") to dynamically instantiate their own error strings. This
12305 makes them more flexible to be built both as statically-linked ENGINEs
12306 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12307 Also, add stub code to each that makes building them as self-contained
12308 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12312 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12313 implementations into applications that are completely implemented in
12314 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12315 commands that can be used to configure what shared-library to load and
12316 to control aspects of the way it is handled. Also, made an update to
12317 the [README-Engine.md](README-Engine.md) file
12318 that brings its information up-to-date and
12319 provides some information and instructions on the "dynamic" ENGINE
12320 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12324 * Make it possible to unload ranges of ERR strings with a new
12325 "ERR_unload_strings" function.
12329 * Add a copy() function to EVP_MD.
12333 * Make EVP_MD routines take a context pointer instead of just the
12334 md_data void pointer.
12338 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12339 that the digest can only process a single chunk of data
12340 (typically because it is provided by a piece of
12341 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12342 is only going to provide a single chunk of data, and hence the
12343 framework needn't accumulate the data for oneshot drivers.
12347 * As with "ERR", make it possible to replace the underlying "ex_data"
12348 functions. This change also alters the storage and management of global
12349 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12350 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12351 index counters. The API functions that use this state have been changed
12352 to take a "class_index" rather than pointers to the class's local STACK
12353 and counter, and there is now an API function to dynamically create new
12354 classes. This centralisation allows us to (a) plug a lot of the
12355 thread-safety problems that existed, and (b) makes it possible to clean
12356 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12357 such data would previously have always leaked in application code and
12358 workarounds were in place to make the memory debugging turn a blind eye
12359 to it. Application code that doesn't use this new function will still
12360 leak as before, but their memory debugging output will announce it now
12361 rather than letting it slide.
12363 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12364 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12365 has a return value to indicate success or failure.
12369 * Make it possible to replace the underlying "ERR" functions such that the
12370 global state (2 LHASH tables and 2 locks) is only used by the "default"
12371 implementation. This change also adds two functions to "get" and "set"
12372 the implementation prior to it being automatically set the first time
12373 any other ERR function takes place. Ie. an application can call "get",
12374 pass the return value to a module it has just loaded, and that module
12375 can call its own "set" function using that value. This means the
12376 module's "ERR" operations will use (and modify) the error state in the
12377 application and not in its own statically linked copy of OpenSSL code.
12381 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12382 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12383 the operation, and provides a more encapsulated way for external code
12384 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12385 to use these functions rather than manually incrementing the counts.
12387 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12391 * Add EVP test program.
12395 * Add symmetric cipher support to ENGINE. Expect the API to change!
12399 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12400 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12401 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12402 These allow a CRL to be built without having to access X509_CRL fields
12403 directly. Modify 'ca' application to use new functions.
12407 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12408 bug workarounds. Rollback attack detection is a security feature.
12409 The problem will only arise on OpenSSL servers when TLSv1 is not
12410 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12411 Software authors not wanting to support TLSv1 will have special reasons
12412 for their choice and can explicitly enable this option.
12414 *Bodo Moeller, Lutz Jaenicke*
12416 * Rationalise EVP so it can be extended: don't include a union of
12417 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12418 (similar to those existing for EVP_CIPHER_CTX).
12423 EVP_MD_CTX_init(&md); /* new function call */
12424 EVP_DigestInit(&md, EVP_sha1());
12425 EVP_DigestUpdate(&md, in, len);
12426 EVP_DigestFinal(&md, out, NULL);
12427 EVP_MD_CTX_cleanup(&md); /* new function call */
12431 * Make DES key schedule conform to the usual scheme, as well as
12432 correcting its structure. This means that calls to DES functions
12433 now have to pass a pointer to a des_key_schedule instead of a
12434 plain des_key_schedule (which was actually always a pointer
12437 des_key_schedule ks;
12439 des_set_key_checked(..., &ks);
12440 des_ncbc_encrypt(..., &ks, ...);
12442 (Note that a later change renames 'des_...' into 'DES_...'.)
12446 * Initial reduction of linker bloat: the use of some functions, such as
12447 PEM causes large amounts of unused functions to be linked in due to
12448 poor organisation. For example pem_all.c contains every PEM function
12449 which has a knock on effect of linking in large amounts of (unused)
12450 ASN1 code. Grouping together similar functions and splitting unrelated
12451 functions prevents this.
12455 * Cleanup of EVP macros.
12459 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12460 correct `_ecb suffix`.
12464 * Add initial OCSP responder support to ocsp application. The
12465 revocation information is handled using the text based index
12466 use by the ca application. The responder can either handle
12467 requests generated internally, supplied in files (for example
12468 via a CGI script) or using an internal minimal server.
12472 * Add configuration choices to get zlib compression for TLS.
12476 * Changes to Kerberos SSL for RFC 2712 compliance:
12477 1. Implemented real KerberosWrapper, instead of just using
12478 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12479 2. Implemented optional authenticator field of KerberosWrapper.
12481 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12482 and authenticator structs; see crypto/krb5/.
12484 Generalized Kerberos calls to support multiple Kerberos libraries.
12485 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12486 via Richard Levitte*
12488 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12489 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12490 values for each of the key sizes rather than having just
12491 parameters (and 'speed' generating keys each time).
12495 * Speed up EVP routines.
12498 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12499 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12500 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12501 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12503 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12504 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12505 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12508 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12510 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12514 * Added the OS2-EMX target.
12516 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12518 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12519 New functions to support `NCONF` routines in extension code.
12520 New function `CONF_set_nconf()`
12521 to allow functions which take an `NCONF` to also handle the old `LHASH`
12522 structure: this means that the old `CONF` compatible routines can be
12523 retained (in particular w.rt. extensions) without having to duplicate the
12524 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12528 * Enhance the general user interface with mechanisms for inner control
12529 and with possibilities to have yes/no kind of prompts.
12533 * Change all calls to low-level digest routines in the library and
12534 applications to use EVP. Add missing calls to HMAC_cleanup() and
12535 don't assume HMAC_CTX can be copied using memcpy().
12537 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12539 * Add the possibility to control engines through control names but with
12540 arbitrary arguments instead of just a string.
12541 Change the key loaders to take a UI_METHOD instead of a callback
12542 function pointer. NOTE: this breaks binary compatibility with earlier
12543 versions of OpenSSL [engine].
12544 Adapt the nCipher code for these new conditions and add a card insertion
12549 * Enhance the general user interface with mechanisms to better support
12550 dialog box interfaces, application-defined prompts, the possibility
12551 to use defaults (for example default passwords from somewhere else)
12552 and interrupts/cancellations.
12556 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12557 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12561 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12562 tidy up some unnecessarily weird code in 'sk_new()').
12564 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12566 * Change the key loading routines for ENGINEs to use the same kind
12567 callback (pem_password_cb) as all other routines that need this
12572 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12573 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12574 than this minimum value is recommended.
12578 * New random seeder for OpenVMS, using the system process statistics
12579 that are easily reachable.
12583 * Windows apparently can't transparently handle global
12584 variables defined in DLLs. Initialisations such as:
12586 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12588 won't compile. This is used by the any applications that need to
12589 declare their own ASN1 modules. This was fixed by adding the option
12590 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12591 needed for static libraries under Win32.
12595 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12596 setting of purpose and trust fields. New X509_STORE trust and
12597 purpose functions and tidy up setting in other SSL functions.
12601 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12602 structure. These are inherited by X509_STORE_CTX when it is
12603 initialised. This allows various defaults to be set in the
12604 X509_STORE structure (such as flags for CRL checking and custom
12605 purpose or trust settings) for functions which only use X509_STORE_CTX
12606 internally such as S/MIME.
12608 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12609 trust settings if they are not set in X509_STORE. This allows X509_STORE
12610 purposes and trust (in S/MIME for example) to override any set by default.
12612 Add command line options for CRL checking to smime, s_client and s_server
12617 * Initial CRL based revocation checking. If the CRL checking flag(s)
12618 are set then the CRL is looked up in the X509_STORE structure and
12619 its validity and signature checked, then if the certificate is found
12620 in the CRL the verify fails with a revoked error.
12622 Various new CRL related callbacks added to X509_STORE_CTX structure.
12624 Command line options added to 'verify' application to support this.
12626 This needs some additional work, such as being able to handle multiple
12627 CRLs with different times, extension based lookup (rather than just
12628 by subject name) and ultimately more complete V2 CRL extension
12633 * Add a general user interface API (crypto/ui/). This is designed
12634 to replace things like des_read_password and friends (backward
12635 compatibility functions using this new API are provided).
12636 The purpose is to remove prompting functions from the DES code
12637 section as well as provide for prompting through dialog boxes in
12638 a window system and the like.
12642 * Add "ex_data" support to ENGINE so implementations can add state at a
12643 per-structure level rather than having to store it globally.
12647 * Make it possible for ENGINE structures to be copied when retrieved by
12648 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12649 This causes the "original" ENGINE structure to act like a template,
12650 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12651 operational state can be localised to each ENGINE structure, despite the
12652 fact they all share the same "methods". New ENGINE structures returned in
12653 this case have no functional references and the return value is the single
12654 structural reference. This matches the single structural reference returned
12655 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12660 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12661 needs to match any other type at all we need to manually clear the
12666 * Changes to the "openssl engine" utility to include;
12667 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12668 about an ENGINE's available control commands.
12669 - executing control commands from command line arguments using the
12670 '-pre' and '-post' switches. '-post' is only used if '-t' is
12671 specified and the ENGINE is successfully initialised. The syntax for
12672 the individual commands are colon-separated, for example;
12673 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12677 * New dynamic control command support for ENGINEs. ENGINEs can now
12678 declare their own commands (numbers), names (strings), descriptions,
12679 and input types for run-time discovery by calling applications. A
12680 subset of these commands are implicitly classed as "executable"
12681 depending on their input type, and only these can be invoked through
12682 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12683 can be based on user input, config files, etc). The distinction is
12684 that "executable" commands cannot return anything other than a boolean
12685 result and can only support numeric or string input, whereas some
12686 discoverable commands may only be for direct use through
12687 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12688 pointers, or other custom uses. The "executable" commands are to
12689 support parameterisations of ENGINE behaviour that can be
12690 unambiguously defined by ENGINEs and used consistently across any
12691 OpenSSL-based application. Commands have been added to all the
12692 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12693 control over shared-library paths without source code alterations.
12697 * Changed all ENGINE implementations to dynamically allocate their
12698 ENGINEs rather than declaring them statically. Apart from this being
12699 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12700 this also allows the implementations to compile without using the
12701 internal engine_int.h header.
12705 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12706 'const' value. Any code that should be able to modify a RAND_METHOD
12707 should already have non-const pointers to it (ie. they should only
12708 modify their own ones).
12712 * Made a variety of little tweaks to the ENGINE code.
12713 - "atalla" and "ubsec" string definitions were moved from header files
12714 to C code. "nuron" string definitions were placed in variables
12715 rather than hard-coded - allowing parameterisation of these values
12716 later on via ctrl() commands.
12717 - Removed unused "#if 0"'d code.
12718 - Fixed engine list iteration code so it uses ENGINE_free() to release
12719 structural references.
12720 - Constified the RAND_METHOD element of ENGINE structures.
12721 - Constified various get/set functions as appropriate and added
12722 missing functions (including a catch-all ENGINE_cpy that duplicates
12723 all ENGINE values onto a new ENGINE except reference counts/state).
12724 - Removed NULL parameter checks in get/set functions. Setting a method
12725 or function to NULL is a way of cancelling out a previously set
12726 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12727 and doesn't justify the extra error symbols and code.
12728 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12729 flags from engine_int.h to engine.h.
12730 - Changed prototypes for ENGINE handler functions (init(), finish(),
12731 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12735 * Implement binary inversion algorithm for BN_mod_inverse in addition
12736 to the algorithm using long division. The binary algorithm can be
12737 used only if the modulus is odd. On 32-bit systems, it is faster
12738 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12739 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12740 up to 450 bits. In 64-bit environments, the binary algorithm
12741 appears to be advantageous for much longer moduli; here we use it
12742 for moduli up to 2048 bits.
12746 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12747 could not support the combine flag in choice fields.
12751 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12752 extensions from a certificate request to the certificate.
12756 * Allow multiple 'certopt' and 'nameopt' options to be separated
12757 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12758 file: this allows the display of the certificate about to be
12759 signed to be customised, to allow certain fields to be included
12760 or excluded and extension details. The old system didn't display
12761 multicharacter strings properly, omitted fields not in the policy
12762 and couldn't display additional details such as extensions.
12766 * Function EC_POINTs_mul for multiple scalar multiplication
12767 of an arbitrary number of elliptic curve points
12768 \sum scalars[i]*points[i],
12769 optionally including the generator defined for the EC_GROUP:
12770 scalar*generator + \sum scalars[i]*points[i].
12772 EC_POINT_mul is a simple wrapper function for the typical case
12773 that the point list has just one item (besides the optional
12778 * First EC_METHODs for curves over GF(p):
12780 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12781 operations and provides various method functions that can also
12782 operate with faster implementations of modular arithmetic.
12784 EC_GFp_mont_method() reuses most functions that are part of
12785 EC_GFp_simple_method, but uses Montgomery arithmetic.
12787 *Bodo Moeller; point addition and point doubling
12788 implementation directly derived from source code provided by
12789 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12791 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12792 crypto/ec/ec_lib.c):
12794 Curves are EC_GROUP objects (with an optional group generator)
12795 based on EC_METHODs that are built into the library.
12797 Points are EC_POINT objects based on EC_GROUP objects.
12799 Most of the framework would be able to handle curves over arbitrary
12800 finite fields, but as there are no obvious types for fields other
12801 than GF(p), some functions are limited to that for now.
12805 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12806 that the file contains a complete HTTP response.
12810 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12811 change the def and num file printf format specifier from "%-40sXXX"
12812 to "%-39s XXX". The latter will always guarantee a space after the
12813 field while the former will cause them to run together if the field
12814 is 40 of more characters long.
12818 * Constify the cipher and digest 'method' functions and structures
12819 and modify related functions to take constant EVP_MD and EVP_CIPHER
12824 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12825 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12829 * Modify `EVP_Digest*()` routines so they now return values. Although the
12830 internal software routines can never fail additional hardware versions
12835 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12837 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12838 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12841 ERR_R_NESTED_ASN1_ERROR
12843 ERR_R_MISSING_ASN1_EOS
12844 were 4 .. 9, conflicting with
12845 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12847 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12848 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12850 Add new error code 'ERR_R_INTERNAL_ERROR'.
12854 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12859 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12860 sets the subject name for a new request or supersedes the
12861 subject name in a given request. Formats that can be parsed are
12862 'CN=Some Name, OU=myOU, C=IT'
12864 'CN=Some Name/OU=myOU/C=IT'.
12866 Add options '-batch' and '-verbose' to 'openssl req'.
12868 *Massimiliano Pala <madwolf@hackmasters.net>*
12870 * Introduce the possibility to access global variables through
12871 functions on platform were that's the best way to handle exporting
12872 global variables in shared libraries. To enable this functionality,
12873 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12874 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12875 is normally done by Configure or something similar).
12877 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12878 in the source file (foo.c) like this:
12880 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12881 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12883 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12884 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12886 OPENSSL_DECLARE_GLOBAL(int,foo);
12887 #define foo OPENSSL_GLOBAL_REF(foo)
12888 OPENSSL_DECLARE_GLOBAL(double,bar);
12889 #define bar OPENSSL_GLOBAL_REF(bar)
12891 The #defines are very important, and therefore so is including the
12892 header file everywhere where the defined globals are used.
12894 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12895 of ASN.1 items, but that structure is a bit different.
12897 The largest change is in util/mkdef.pl which has been enhanced with
12898 better and easier to understand logic to choose which symbols should
12899 go into the Windows .def files as well as a number of fixes and code
12900 cleanup (among others, algorithm keywords are now sorted
12901 lexicographically to avoid constant rewrites).
12905 * In BN_div() keep a copy of the sign of 'num' before writing the
12906 result to 'rm' because if rm==num the value will be overwritten
12907 and produce the wrong result if 'num' is negative: this caused
12908 problems with BN_mod() and BN_nnmod().
12912 * Function OCSP_request_verify(). This checks the signature on an
12913 OCSP request and verifies the signer certificate. The signer
12914 certificate is just checked for a generic purpose and OCSP request
12919 * Add OCSP_check_validity() function to check the validity of OCSP
12920 responses. OCSP responses are prepared in real time and may only
12921 be a few seconds old. Simply checking that the current time lies
12922 between thisUpdate and nextUpdate max reject otherwise valid responses
12923 caused by either OCSP responder or client clock inaccuracy. Instead
12924 we allow thisUpdate and nextUpdate to fall within a certain period of
12925 the current time. The age of the response can also optionally be
12926 checked. Two new options -validity_period and -status_age added to
12931 * If signature or public key algorithm is unrecognized print out its
12932 OID rather that just UNKNOWN.
12936 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12937 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12938 ID to be generated from the issuer certificate alone which can then be
12939 passed to OCSP_id_issuer_cmp().
12943 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12944 ASN1 modules to export functions returning ASN1_ITEM pointers
12945 instead of the ASN1_ITEM structures themselves. This adds several
12946 new macros which allow the underlying ASN1 function/structure to
12947 be accessed transparently. As a result code should not use ASN1_ITEM
12948 references directly (such as &X509_it) but instead use the relevant
12949 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12950 use of the new ASN1 code on platforms where exporting structures
12951 is problematical (for example in shared libraries) but exporting
12952 functions returning pointers to structures is not.
12956 * Add support for overriding the generation of SSL/TLS session IDs.
12957 These callbacks can be registered either in an SSL_CTX or per SSL.
12958 The purpose of this is to allow applications to control, if they wish,
12959 the arbitrary values chosen for use as session IDs, particularly as it
12960 can be useful for session caching in multiple-server environments. A
12961 command-line switch for testing this (and any client code that wishes
12962 to use such a feature) has been added to "s_server".
12964 *Geoff Thorpe, Lutz Jaenicke*
12966 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12967 of the form `#if defined(...) || defined(...) || ...` and
12968 `#if !defined(...) && !defined(...) && ...`. This also avoids
12969 the growing number of special cases it was previously handling.
12973 * Make all configuration macros available for application by making
12974 sure they are available in opensslconf.h, by giving them names starting
12975 with `OPENSSL_` to avoid conflicts with other packages and by making
12976 sure e_os2.h will cover all platform-specific cases together with
12978 Additionally, it is now possible to define configuration/platform-
12979 specific names (called "system identities"). In the C code, these
12980 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12981 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12982 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12987 * New option -set_serial to 'req' and 'x509' this allows the serial
12988 number to use to be specified on the command line. Previously self
12989 signed certificates were hard coded with serial number 0 and the
12990 CA options of 'x509' had to use a serial number in a file which was
12995 * New options to 'ca' utility to support V2 CRL entry extensions.
12996 Currently CRL reason, invalidity date and hold instruction are
12997 supported. Add new CRL extensions to V3 code and some new objects.
13001 * New function EVP_CIPHER_CTX_set_padding() this is used to
13002 disable standard block padding (aka PKCS#5 padding) in the EVP
13003 API, which was previously mandatory. This means that the data is
13004 not padded in any way and so the total length much be a multiple
13005 of the block size, otherwise an error occurs.
13009 * Initial (incomplete) OCSP SSL support.
13013 * New function OCSP_parse_url(). This splits up a URL into its host,
13014 port and path components: primarily to parse OCSP URLs. New -url
13015 option to ocsp utility.
13019 * New nonce behavior. The return value of OCSP_check_nonce() now
13020 reflects the various checks performed. Applications can decide
13021 whether to tolerate certain situations such as an absent nonce
13022 in a response when one was present in a request: the ocsp application
13023 just prints out a warning. New function OCSP_add1_basic_nonce()
13024 this is to allow responders to include a nonce in a response even if
13025 the request is nonce-less.
13029 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13030 skipped when using openssl x509 multiple times on a single input file,
13031 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13035 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13036 set string type: to handle setting ASN1_TIME structures. Fix ca
13037 utility to correctly initialize revocation date of CRLs.
13041 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13042 the clients preferred ciphersuites and rather use its own preferences.
13043 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13044 Internet Explorer by ensuring unchanged hash method during stepup.
13045 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13049 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13050 to aes and add a new 'exist' option to print out symbols that don't
13055 * Additional options to ocsp utility to allow flags to be set and
13056 additional certificates supplied.
13060 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13061 OCSP client a number of certificate to only verify the response
13066 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13067 handle the new API. Currently only ECB, CBC modes supported. Add new
13070 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13071 Encryption Standard (AES) Ciphersuites for Transport Layer
13072 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13073 not enabled by default and were not part of the "ALL" ciphersuite
13074 alias because they were not yet official; they could be
13075 explicitly requested by specifying the "AESdraft" ciphersuite
13076 group alias. In the final release of OpenSSL 0.9.7, the group
13077 alias is called "AES" and is part of "ALL".)
13079 *Ben Laurie, Steve Henson, Bodo Moeller*
13081 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13082 request to response.
13086 * Functions for OCSP responders. OCSP_request_onereq_count(),
13087 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13088 extract information from a certificate request. OCSP_response_create()
13089 creates a response and optionally adds a basic response structure.
13090 OCSP_basic_add1_status() adds a complete single response to a basic
13091 response and returns the OCSP_SINGLERESP structure just added (to allow
13092 extensions to be included for example). OCSP_basic_add1_cert() adds a
13093 certificate to a basic response and OCSP_basic_sign() signs a basic
13094 response with various flags. New helper functions ASN1_TIME_check()
13095 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13096 (converts ASN1_TIME to GeneralizedTime).
13100 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13101 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13102 structure from a certificate. X509_pubkey_digest() digests the public_key
13103 contents: this is used in various key identifiers.
13107 * Make sk_sort() tolerate a NULL argument.
13109 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13111 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13112 passed by the function are trusted implicitly. If any of them signed the
13113 response then it is assumed to be valid and is not verified.
13117 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13118 to data. This was previously part of the PKCS7 ASN1 code. This
13119 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13120 *Steve Henson, reported by Kenneth R. Robinette
13121 <support@securenetterm.com>*
13123 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13124 routines: without these tracing memory leaks is very painful.
13125 Fix leaks in PKCS12 and PKCS7 routines.
13129 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13130 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13131 effectively meant GeneralizedTime would never be used. Now it
13132 is initialised to -1 but X509_time_adj() now has to check the value
13133 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13134 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13135 *Steve Henson, reported by Kenneth R. Robinette
13136 <support@securenetterm.com>*
13138 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13139 result in a zero length in the ASN1_INTEGER structure which was
13140 not consistent with the structure when d2i_ASN1_INTEGER() was used
13141 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13142 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13143 where it did not print out a minus for negative ASN1_INTEGER.
13147 * Add summary printout to ocsp utility. The various functions which
13148 convert status values to strings have been renamed to:
13149 OCSP_response_status_str(), OCSP_cert_status_str() and
13150 OCSP_crl_reason_str() and are no longer static. New options
13151 to verify nonce values and to disable verification. OCSP response
13152 printout format cleaned up.
13156 * Add additional OCSP certificate checks. These are those specified
13157 in RFC2560. This consists of two separate checks: the CA of the
13158 certificate being checked must either be the OCSP signer certificate
13159 or the issuer of the OCSP signer certificate. In the latter case the
13160 OCSP signer certificate must contain the OCSP signing extended key
13161 usage. This check is performed by attempting to match the OCSP
13162 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13163 in the OCSP_CERTID structures of the response.
13167 * Initial OCSP certificate verification added to OCSP_basic_verify()
13168 and related routines. This uses the standard OpenSSL certificate
13169 verify routines to perform initial checks (just CA validity) and
13170 to obtain the certificate chain. Then additional checks will be
13171 performed on the chain. Currently the root CA is checked to see
13172 if it is explicitly trusted for OCSP signing. This is used to set
13173 a root CA as a global signing root: that is any certificate that
13174 chains to that CA is an acceptable OCSP signing certificate.
13178 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13179 extensions from a separate configuration file.
13180 As when reading extensions from the main configuration file,
13181 the '-extensions ...' option may be used for specifying the
13184 *Massimiliano Pala <madwolf@comune.modena.it>*
13186 * New OCSP utility. Allows OCSP requests to be generated or
13187 read. The request can be sent to a responder and the output
13188 parsed, outputted or printed in text form. Not complete yet:
13189 still needs to check the OCSP response validity.
13193 * New subcommands for 'openssl ca':
13194 `openssl ca -status <serial>` prints the status of the cert with
13195 the given serial number (according to the index file).
13196 `openssl ca -updatedb` updates the expiry status of certificates
13199 *Massimiliano Pala <madwolf@comune.modena.it>*
13201 * New '-newreq-nodes' command option to CA.pl. This is like
13202 '-newreq', but calls 'openssl req' with the '-nodes' option
13203 so that the resulting key is not encrypted.
13205 *Damien Miller <djm@mindrot.org>*
13207 * New configuration for the GNU Hurd.
13209 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13211 * Initial code to implement OCSP basic response verify. This
13212 is currently incomplete. Currently just finds the signer's
13213 certificate and verifies the signature on the response.
13217 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13218 value of OPENSSLDIR. This is available via the new '-d' option
13219 to 'openssl version', and is also included in 'openssl version -a'.
13223 * Allowing defining memory allocation callbacks that will be given
13224 file name and line number information in additional arguments
13225 (a `const char*` and an int). The basic functionality remains, as
13226 well as the original possibility to just replace malloc(),
13227 realloc() and free() by functions that do not know about these
13228 additional arguments. To register and find out the current
13229 settings for extended allocation functions, the following
13230 functions are provided:
13232 CRYPTO_set_mem_ex_functions
13233 CRYPTO_set_locked_mem_ex_functions
13234 CRYPTO_get_mem_ex_functions
13235 CRYPTO_get_locked_mem_ex_functions
13237 These work the same way as CRYPTO_set_mem_functions and friends.
13238 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13239 extended allocation function is enabled.
13240 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13241 a conventional allocation function is enabled.
13243 *Richard Levitte, Bodo Moeller*
13245 * Finish off removing the remaining LHASH function pointer casts.
13246 There should no longer be any prototype-casting required when using
13247 the LHASH abstraction, and any casts that remain are "bugs". See
13248 the callback types and macros at the head of lhash.h for details
13249 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13253 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13254 If /dev/[u]random devices are not available or do not return enough
13255 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13257 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13258 /etc/entropy will be queried once each in this sequence, querying stops
13259 when enough entropy was collected without querying more sockets.
13263 * Change the Unix RAND_poll() variant to be able to poll several
13264 random devices, as specified by DEVRANDOM, until a sufficient amount
13265 of data has been collected. We spend at most 10 ms on each file
13266 (select timeout) and read in non-blocking mode. DEVRANDOM now
13267 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13268 (previously it was just the string "/dev/urandom"), so on typical
13269 platforms the 10 ms delay will never occur.
13270 Also separate out the Unix variant to its own file, rand_unix.c.
13271 For VMS, there's a currently-empty rand_vms.c.
13275 * Move OCSP client related routines to ocsp_cl.c. These
13276 provide utility functions which an application needing
13277 to issue a request to an OCSP responder and analyse the
13278 response will typically need: as opposed to those which an
13279 OCSP responder itself would need which will be added later.
13281 OCSP_request_sign() signs an OCSP request with an API similar
13282 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13283 response. OCSP_response_get1_basic() extracts basic response
13284 from response. OCSP_resp_find_status(): finds and extracts status
13285 information from an OCSP_CERTID structure (which will be created
13286 when the request structure is built). These are built from lower
13287 level functions which work on OCSP_SINGLERESP structures but
13288 won't normally be used unless the application wishes to examine
13289 extensions in the OCSP response for example.
13291 Replace nonce routines with a pair of functions.
13292 OCSP_request_add1_nonce() adds a nonce value and optionally
13293 generates a random value. OCSP_check_nonce() checks the
13294 validity of the nonce in an OCSP response.
13298 * Change function OCSP_request_add() to OCSP_request_add0_id().
13299 This doesn't copy the supplied OCSP_CERTID and avoids the
13300 need to free up the newly created id. Change return type
13301 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13302 This can then be used to add extensions to the request.
13303 Deleted OCSP_request_new(), since most of its functionality
13304 is now in OCSP_REQUEST_new() (and the case insensitive name
13305 clash) apart from the ability to set the request name which
13306 will be added elsewhere.
13310 * Update OCSP API. Remove obsolete extensions argument from
13311 various functions. Extensions are now handled using the new
13312 OCSP extension code. New simple OCSP HTTP function which
13313 can be used to send requests and parse the response.
13317 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13318 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13319 uses the special reorder version of SET OF to sort the attributes
13320 and reorder them to match the encoded order. This resolves a long
13321 standing problem: a verify on a PKCS7 structure just after signing
13322 it used to fail because the attribute order did not match the
13323 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13324 it uses the received order. This is necessary to tolerate some broken
13325 software that does not order SET OF. This is handled by encoding
13326 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13327 to produce the required SET OF.
13331 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13332 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13333 files to get correct declarations of the ASN.1 item variables.
13337 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13338 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13339 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13340 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13341 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13342 ASN1_ITEM and no wrapper functions.
13346 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13347 replace the old function pointer based I/O routines. Change most of
13348 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13352 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13353 lines, recognize more "algorithms" that can be deselected, and make
13354 it complain about algorithm deselection that isn't recognised.
13358 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13359 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13360 to use new functions. Add NO_ASN1_OLD which can be set to remove
13361 some old style ASN1 functions: this can be used to determine if old
13362 code will still work when these eventually go away.
13366 * New extension functions for OCSP structures, these follow the
13367 same conventions as certificates and CRLs.
13371 * New function X509V3_add1_i2d(). This automatically encodes and
13372 adds an extension. Its behaviour can be customised with various
13373 flags to append, replace or delete. Various wrappers added for
13374 certificates and CRLs.
13378 * Fix to avoid calling the underlying ASN1 print routine when
13379 an extension cannot be parsed. Correct a typo in the
13380 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13384 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13385 entries for variables.
13389 * Add functionality to `apps/openssl.c` for detecting locking
13390 problems: As the program is single-threaded, all we have
13391 to do is register a locking callback using an array for
13392 storing which locks are currently held by the program.
13396 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13397 SSL_get_ex_data_X509_STORE_idx(), which is used in
13398 ssl_verify_cert_chain() and thus can be called at any time
13399 during TLS/SSL handshakes so that thread-safety is essential.
13400 Unfortunately, the ex_data design is not at all suited
13401 for multi-threaded use, so it probably should be abolished.
13405 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13407 *Broadcom, tweaked and integrated by Geoff Thorpe*
13409 * Move common extension printing code to new function
13410 X509V3_print_extensions(). Reorganise OCSP print routines and
13411 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13415 * New function X509_signature_print() to remove duplication in some
13420 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13421 set (this was treated exactly the same as SET OF previously). This
13422 is used to reorder the STACK representing the structure to match the
13423 encoding. This will be used to get round a problem where a PKCS7
13424 structure which was signed could not be verified because the STACK
13425 order did not reflect the encoded order.
13429 * Reimplement the OCSP ASN1 module using the new code.
13433 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13434 for its ASN1 operations. The old style function pointers still exist
13435 for now but they will eventually go away.
13439 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13440 completely replaces the old ASN1 functionality with a table driven
13441 encoder and decoder which interprets an ASN1_ITEM structure describing
13442 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13443 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13444 has also been converted to the new form.
13448 * Change BN_mod_exp_recp so that negative moduli are tolerated
13449 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13450 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13451 for negative moduli.
13455 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13456 of not touching the result's sign bit.
13460 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13465 * Changed the LHASH code to use prototypes for callbacks, and created
13466 macros to declare and implement thin (optionally static) functions
13467 that provide type-safety and avoid function pointer casting for the
13468 type-specific callbacks.
13472 * Added Kerberos Cipher Suites to be used with TLS, as written in
13474 *Veers Staats <staatsvr@asc.hpc.mil>,
13475 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13477 * Reformat the FAQ so the different questions and answers can be divided
13478 in sections depending on the subject.
13482 * Have the zlib compression code load ZLIB.DLL dynamically under
13487 * New function BN_mod_sqrt for computing square roots modulo a prime
13488 (using the probabilistic Tonelli-Shanks algorithm unless
13489 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13490 be handled deterministically).
13492 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13494 * Make BN_mod_inverse faster by explicitly handling small quotients
13495 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13496 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13500 * New function BN_kronecker.
13504 * Fix BN_gcd so that it works on negative inputs; the result is
13505 positive unless both parameters are zero.
13506 Previously something reasonably close to an infinite loop was
13507 possible because numbers could be growing instead of shrinking
13508 in the implementation of Euclid's algorithm.
13512 * Fix BN_is_word() and BN_is_one() macros to take into account the
13513 sign of the number in question.
13515 Fix BN_is_word(a,w) to work correctly for w == 0.
13517 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13518 because its test if the absolute value of 'a' equals 'w'.
13519 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13520 it exists mostly for use in the implementations of BN_is_zero(),
13521 BN_is_one(), and BN_is_word().
13525 * New function BN_swap.
13529 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13530 the exponentiation functions are more likely to produce reasonable
13531 results on negative inputs.
13535 * Change BN_mod_mul so that the result is always non-negative.
13536 Previously, it could be negative if one of the factors was negative;
13537 I don't think anyone really wanted that behaviour.
13541 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13542 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13543 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13544 and add new functions:
13553 BN_mod_lshift1_quick
13555 BN_mod_lshift_quick
13557 These functions always generate non-negative results.
13559 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13560 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13562 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13563 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13564 be reduced modulo `m`.
13566 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13569 The following entry accidentally appeared in the CHANGES file
13570 distributed with OpenSSL 0.9.7. The modifications described in
13571 it do *not* apply to OpenSSL 0.9.7.
13573 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13574 was actually never needed) and in BN_mul(). The removal in BN_mul()
13575 required a small change in bn_mul_part_recursive() and the addition
13576 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13577 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13578 bn_sub_words() and bn_add_words() except they take arrays with
13584 * In 'openssl passwd', verify passwords read from the terminal
13585 unless the '-salt' option is used (which usually means that
13586 verification would just waste user's time since the resulting
13587 hash is going to be compared with some given password hash)
13588 or the new '-noverify' option is used.
13590 This is an incompatible change, but it does not affect
13591 non-interactive use of 'openssl passwd' (passwords on the command
13592 line, '-stdin' option, '-in ...' option) and thus should not
13593 cause any problems.
13597 * Remove all references to RSAref, since there's no more need for it.
13601 * Make DSO load along a path given through an environment variable
13602 (SHLIB_PATH) with shl_load().
13606 * Constify the ENGINE code as a result of BIGNUM constification.
13607 Also constify the RSA code and most things related to it. In a
13608 few places, most notable in the depth of the ASN.1 code, ugly
13609 casts back to non-const were required (to be solved at a later
13614 * Make it so the openssl application has all engines loaded by default.
13618 * Constify the BIGNUM routines a little more.
13622 * Add the following functions:
13624 ENGINE_load_cswift()
13626 ENGINE_load_atalla()
13627 ENGINE_load_nuron()
13628 ENGINE_load_builtin_engines()
13630 That way, an application can itself choose if external engines that
13631 are built-in in OpenSSL shall ever be used or not. The benefit is
13632 that applications won't have to be linked with libdl or other dso
13633 libraries unless it's really needed.
13635 Changed 'openssl engine' to load all engines on demand.
13636 Changed the engine header files to avoid the duplication of some
13637 declarations (they differed!).
13641 * 'openssl engine' can now list capabilities.
13645 * Better error reporting in 'openssl engine'.
13649 * Never call load_dh_param(NULL) in s_server.
13653 * Add engine application. It can currently list engines by name and
13654 identity, and test if they are actually available.
13658 * Improve RPM specification file by forcing symbolic linking and making
13659 sure the installed documentation is also owned by root.root.
13661 *Damien Miller <djm@mindrot.org>*
13663 * Give the OpenSSL applications more possibilities to make use of
13664 keys (public as well as private) handled by engines.
13668 * Add OCSP code that comes from CertCo.
13672 * Add VMS support for the Rijndael code.
13676 * Added untested support for Nuron crypto accelerator.
13680 * Add support for external cryptographic devices. This code was
13681 previously distributed separately as the "engine" branch.
13683 *Geoff Thorpe, Richard Levitte*
13685 * Rework the filename-translation in the DSO code. It is now possible to
13686 have far greater control over how a "name" is turned into a filename
13687 depending on the operating environment and any oddities about the
13688 different shared library filenames on each system.
13692 * Support threads on FreeBSD-elf in Configure.
13696 * Fix for SHA1 assembly problem with MASM: it produces
13697 warnings about corrupt line number information when assembling
13698 with debugging information. This is caused by the overlapping
13701 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13704 NCONF_get_number() has no error checking at all. As a replacement,
13705 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13706 promoted strongly. The old NCONF_get_number is kept around for
13707 binary backward compatibility.
13708 Make it possible for methods to load from something other than a BIO,
13709 by providing a function pointer that is given a name instead of a BIO.
13710 For example, this could be used to load configuration data from an
13715 * Fix for non blocking accept BIOs. Added new I/O special reason
13716 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13717 with non blocking I/O was not possible because no retry code was
13718 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13723 * Added the beginnings of Rijndael support.
13727 * Fix for bug in DirectoryString mask setting. Add support for
13728 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13729 to allow certificate printing to more controllable, additional
13730 'certopt' option to 'x509' to allow new printing options to be
13735 * Clean old EAY MD5 hack from e_os.h.
13739 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13741 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13742 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13744 *Joe Orton, Steve Henson*
13746 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13748 * Fix additional bug revealed by the NISCC test suite:
13750 Stop bug triggering large recursion when presented with
13751 certain ASN.1 tags ([CVE-2003-0851])
13755 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13757 * Fix various bugs revealed by running the NISCC test suite:
13759 Stop out of bounds reads in the ASN1 code when presented with
13760 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13762 If verify callback ignores invalid public key errors don't try to check
13763 certificate signature with the NULL public key.
13767 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13768 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13773 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13774 extra data after the compression methods not only for TLS 1.0
13775 but also for SSL 3.0 (as required by the specification).
13777 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13779 * Change X509_certificate_type() to mark the key as exported/exportable
13780 when it's 512 *bits* long, not 512 bytes.
13784 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13786 * Countermeasure against the Klima-Pokorny-Rosa extension of
13787 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13788 a protocol version number mismatch like a decryption error
13789 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13793 * Turn on RSA blinding by default in the default implementation
13794 to avoid a timing attack. Applications that don't want it can call
13795 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13796 They would be ill-advised to do so in most cases.
13798 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13800 * Change RSA blinding code so that it works when the PRNG is not
13801 seeded (in this case, the secret RSA exponent is abused as
13802 an unpredictable seed -- if it is not unpredictable, there
13803 is no point in blinding anyway). Make RSA blinding thread-safe
13804 by remembering the creator's thread ID in rsa->blinding and
13805 having all other threads use local one-time blinding factors
13806 (this requires more computation than sharing rsa->blinding, but
13807 avoids excessive locking; and if an RSA object is not shared
13808 between threads, blinding will still be very fast).
13812 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13814 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13815 via timing by performing a MAC computation even if incorrect
13816 block cipher padding has been found. This is a countermeasure
13817 against active attacks where the attacker has to distinguish
13818 between bad padding and a MAC verification error. ([CVE-2003-0078])
13820 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13821 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13822 Martin Vuagnoux (EPFL, Ilion)*
13824 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13826 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13827 memory from its contents. This is done with a counter that will
13828 place alternating values in each byte. This can be used to solve
13829 two issues: 1) the removal of calls to memset() by highly optimizing
13830 compilers, and 2) cleansing with other values than 0, since those can
13831 be read through on certain media, for example a swap space on disk.
13835 * Bugfix: client side session caching did not work with external caching,
13836 because the session->cipher setting was not restored when reloading
13837 from the external cache. This problem was masked, when
13838 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13839 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13843 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13844 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13846 *Zeev Lieber <zeev-l@yahoo.com>*
13848 * Undo an undocumented change introduced in 0.9.6e which caused
13849 repeated calls to OpenSSL_add_all_ciphers() and
13850 OpenSSL_add_all_digests() to be ignored, even after calling
13855 * Change the default configuration reader to deal with last line not
13856 being properly terminated.
13860 * Change X509_NAME_cmp() so it applies the special rules on handling
13861 DN values that are of type PrintableString, as well as RDNs of type
13862 emailAddress where the value has the type ia5String.
13864 *stefank@valicert.com via Richard Levitte*
13866 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13867 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13868 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13869 the bitwise-OR of the two for use by the majority of applications
13870 wanting this behaviour, and update the docs. The documented
13871 behaviour and actual behaviour were inconsistent and had been
13872 changing anyway, so this is more a bug-fix than a behavioural
13875 *Geoff Thorpe, diagnosed by Nadav Har'El*
13877 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13878 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13882 * Fix initialization code race conditions in
13883 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13884 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13885 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13886 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13887 ssl2_get_cipher_by_char(),
13888 ssl3_get_cipher_by_char().
13890 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13892 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13893 the cached sessions are flushed, as the remove_cb() might use ex_data
13894 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13895 (see [openssl.org #212]).
13897 *Geoff Thorpe, Lutz Jaenicke*
13899 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13900 length, instead of the encoding length to d2i_ASN1_OBJECT.
13904 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13906 * [In 0.9.6g-engine release:]
13907 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13909 *Lynn Gazis <lgazis@rainbow.com>*
13911 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13913 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13914 and get fix the header length calculation.
13915 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13916 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13918 * Use proper error handling instead of 'assertions' in buffer
13919 overflow checks added in 0.9.6e. This prevents DoS (the
13920 assertions could call abort()).
13922 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13924 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13926 * Add various sanity checks to asn1_get_length() to reject
13927 the ASN1 length bytes if they exceed sizeof(long), will appear
13928 negative or the content length exceeds the length of the
13931 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13933 * Fix cipher selection routines: ciphers without encryption had no flags
13934 for the cipher strength set and where therefore not handled correctly
13935 by the selection routines (PR #130).
13939 * Fix EVP_dsa_sha macro.
13944 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13945 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13946 that was added in OpenSSL 0.9.6d.
13948 As the countermeasure turned out to be incompatible with some
13949 broken SSL implementations, the new option is part of SSL_OP_ALL.
13950 SSL_OP_ALL is usually employed when compatibility with weird SSL
13951 implementations is desired (e.g. '-bugs' option to 's_client' and
13952 's_server'), so the new option is automatically set in many
13957 * Changes in security patch:
13959 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13960 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13961 Air Force Materiel Command, USAF, under agreement number
13964 * Add various sanity checks to asn1_get_length() to reject
13965 the ASN1 length bytes if they exceed sizeof(long), will appear
13966 negative or the content length exceeds the length of the
13967 supplied buffer. ([CVE-2002-0659])
13969 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13971 * Assertions for various potential buffer overflows, not known to
13972 happen in practice.
13974 *Ben Laurie (CHATS)*
13976 * Various temporary buffers to hold ASCII versions of integers were
13977 too small for 64 bit platforms. ([CVE-2002-0655])
13978 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13980 * Remote buffer overflow in SSL3 protocol - an attacker could
13981 supply an oversized session ID to a client. ([CVE-2002-0656])
13983 *Ben Laurie (CHATS)*
13985 * Remote buffer overflow in SSL2 protocol - an attacker could
13986 supply an oversized client master key. ([CVE-2002-0656])
13988 *Ben Laurie (CHATS)*
13990 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13992 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13993 encoded as NULL) with id-dsa-with-sha1.
13995 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13997 * Check various `X509_...()` return values in `apps/req.c`.
13999 *Nils Larsch <nla@trustcenter.de>*
14001 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14002 an end-of-file condition would erroneously be flagged, when the CRLF
14003 was just at the end of a processed block. The bug was discovered when
14004 processing data through a buffering memory BIO handing the data to a
14005 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14006 <ptsekov@syntrex.com> and Nedelcho Stanev.
14010 * Implement a countermeasure against a vulnerability recently found
14011 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14012 before application data chunks to avoid the use of known IVs
14013 with data potentially chosen by the attacker.
14017 * Fix length checks in ssl3_get_client_hello().
14021 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14022 to prevent ssl3_read_internal() from incorrectly assuming that
14023 ssl3_read_bytes() found application data while handshake
14024 processing was enabled when in fact s->s3->in_read_app_data was
14025 merely automatically cleared during the initial handshake.
14027 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14029 * Fix object definitions for Private and Enterprise: they were not
14030 recognized in their shortname (=lowercase) representation. Extend
14031 obj_dat.pl to issue an error when using undefined keywords instead
14032 of silently ignoring the problem (Svenning Sorensen
14033 <sss@sss.dnsalias.net>).
14037 * Fix DH_generate_parameters() so that it works for 'non-standard'
14038 generators, i.e. generators other than 2 and 5. (Previously, the
14039 code did not properly initialise the 'add' and 'rem' values to
14040 BN_generate_prime().)
14042 In the new general case, we do not insist that 'generator' is
14043 actually a primitive root: This requirement is rather pointless;
14044 a generator of the order-q subgroup is just as good, if not
14049 * Map new X509 verification errors to alerts. Discovered and submitted by
14050 Tom Wu <tom@arcot.com>.
14054 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14055 returning non-zero before the data has been completely received
14056 when using non-blocking I/O.
14058 *Bodo Moeller; problem pointed out by John Hughes*
14060 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14062 *Ben Laurie, Lutz Jaenicke*
14064 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14065 Yoram Zahavi <YoramZ@gilian.com>).
14069 * Add information about CygWin 1.3 and on, and preserve proper
14070 configuration for the versions before that.
14072 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14074 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14075 check whether we deal with a copy of a session and do not delete from
14076 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14077 <izhar@checkpoint.com>.
14081 * Do not store session data into the internal session cache, if it
14082 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14083 flag is set). Proposed by Aslam <aslam@funk.com>.
14087 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14092 * [In 0.9.6d-engine release:]
14093 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14095 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14097 * Add the configuration target linux-s390x.
14099 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14101 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14102 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14103 variable as an indication that a ClientHello message has been
14104 received. As the flag value will be lost between multiple
14105 invocations of ssl3_accept when using non-blocking I/O, the
14106 function may not be aware that a handshake has actually taken
14107 place, thus preventing a new session from being added to the
14110 To avoid this problem, we now set s->new_session to 2 instead of
14111 using a local variable.
14113 *Lutz Jaenicke, Bodo Moeller*
14115 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14116 if the SSL_R_LENGTH_MISMATCH error is detected.
14118 *Geoff Thorpe, Bodo Moeller*
14120 * New 'shared_ldflag' column in Configure platform table.
14124 * Fix EVP_CIPHER_mode macro.
14126 *"Dan S. Camper" <dan@bti.net>*
14128 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14129 type, we must throw them away by setting rr->length to 0.
14131 *D P Chang <dpc@qualys.com>*
14133 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14135 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14136 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14137 worked incorrectly for those cases where range = `10..._2` and
14138 `3*range` is two bits longer than range.)
14142 * Only add signing time to PKCS7 structures if it is not already
14147 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14148 OBJ_ld_ce should be OBJ_id_ce.
14149 Also some ip-pda OIDs in crypto/objects/objects.txt were
14150 incorrect (cf. RFC 3039).
14152 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14154 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14155 returns early because it has nothing to do.
14157 *Andy Schneider <andy.schneider@bjss.co.uk>*
14159 * [In 0.9.6c-engine release:]
14160 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14162 *Andy Schneider <andy.schneider@bjss.co.uk>*
14164 * [In 0.9.6c-engine release:]
14165 Add support for Cryptographic Appliance's keyserver technology.
14166 (Use engine 'keyclient')
14168 *Cryptographic Appliances and Geoff Thorpe*
14170 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14171 is called via tools/c89.sh because arguments have to be
14172 rearranged (all '-L' options must appear before the first object
14175 *Richard Shapiro <rshapiro@abinitio.com>*
14177 * [In 0.9.6c-engine release:]
14178 Add support for Broadcom crypto accelerator cards, backported
14181 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14183 * [In 0.9.6c-engine release:]
14184 Add support for SureWare crypto accelerator cards from
14185 Baltimore Technologies. (Use engine 'sureware')
14187 *Baltimore Technologies and Mark Cox*
14189 * [In 0.9.6c-engine release:]
14190 Add support for crypto accelerator cards from Accelerated
14191 Encryption Processing, www.aep.ie. (Use engine 'aep')
14193 *AEP Inc. and Mark Cox*
14195 * Add a configuration entry for gcc on UnixWare.
14197 *Gary Benson <gbenson@redhat.com>*
14199 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14200 messages are stored in a single piece (fixed-length part and
14201 variable-length part combined) and fix various bugs found on the way.
14205 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14206 instead. BIO_gethostbyname() does not know what timeouts are
14207 appropriate, so entries would stay in cache even when they have
14209 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14211 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14212 faced with a pathologically small ClientHello fragment that does
14213 not contain client_version: Instead of aborting with an error,
14214 simply choose the highest available protocol version (i.e.,
14215 TLS 1.0 unless it is disabled). In practice, ClientHello
14216 messages are never sent like this, but this change gives us
14217 strictly correct behaviour at least for TLS.
14221 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14222 never resets s->method to s->ctx->method when called from within
14223 one of the SSL handshake functions.
14225 *Bodo Moeller; problem pointed out by Niko Baric*
14227 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14228 (sent using the client's version number) if client_version is
14229 smaller than the protocol version in use. Also change
14230 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14231 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14232 the client will at least see that alert.
14236 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14241 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14242 client receives HelloRequest while in a handshake.
14244 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14246 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14247 should end in 'break', not 'goto end' which circumvents various
14248 cleanups done in state SSL_ST_OK. But session related stuff
14249 must be disabled for SSL_ST_OK in the case that we just sent a
14252 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14253 before just sending a HelloRequest.
14255 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14257 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14258 reveal whether illegal block cipher padding was found or a MAC
14259 verification error occurred. (Neither SSLerr() codes nor alerts
14260 are directly visible to potential attackers, but the information
14261 may leak via logfiles.)
14263 Similar changes are not required for the SSL 2.0 implementation
14264 because the number of padding bytes is sent in clear for SSL 2.0,
14265 and the extra bytes are just ignored. However ssl/s2_pkt.c
14266 failed to verify that the purported number of padding bytes is in
14271 * Add OpenUNIX-8 support including shared libraries
14272 (Boyd Lynn Gerber <gerberb@zenez.com>).
14276 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14277 'wristwatch attack' using huge encoding parameters (cf.
14278 James H. Manger's CRYPTO 2001 paper). Note that the
14279 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14280 encoding parameters and hence was not vulnerable.
14284 * BN_sqr() bug fix.
14286 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14288 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14289 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14290 followed by modular reduction.
14292 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14294 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14295 equivalent based on BN_pseudo_rand() instead of BN_rand().
14299 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14300 This function was broken, as the check for a new client hello message
14301 to handle SGC did not allow these large messages.
14302 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14306 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14310 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14311 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14315 * Rework the configuration and shared library support for Tru64 Unix.
14316 The configuration part makes use of modern compiler features and
14317 still retains old compiler behavior for those that run older versions
14318 of the OS. The shared library support part includes a variant that
14319 uses the RPATH feature, and is available through the special
14320 configuration target "alpha-cc-rpath", which will never be selected
14323 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14325 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14326 with the same message size as in ssl3_get_certificate_request().
14327 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14328 messages might inadvertently be reject as too long.
14330 *Petr Lampa <lampa@fee.vutbr.cz>*
14332 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14336 * Modified SSL library such that the verify_callback that has been set
14337 specifically for an SSL object with SSL_set_verify() is actually being
14338 used. Before the change, a verify_callback set with this function was
14339 ignored and the verify_callback() set in the SSL_CTX at the time of
14340 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14341 to allow the necessary settings.
14345 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14346 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14347 done automatically (in contradiction to the requirements of the C
14348 standard). This made problems when used from OpenSSH.
14352 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14353 dh->length and always used
14355 BN_rand_range(priv_key, dh->p).
14357 BN_rand_range() is not necessary for Diffie-Hellman, and this
14358 specific range makes Diffie-Hellman unnecessarily inefficient if
14359 dh->length (recommended exponent length) is much smaller than the
14360 length of dh->p. We could use BN_rand_range() if the order of
14361 the subgroup was stored in the DH structure, but we only have
14366 BN_rand(priv_key, l, ...)
14368 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14375 RSA_eay_public_encrypt
14376 RSA_eay_private_decrypt
14377 RSA_eay_private_encrypt (signing)
14378 RSA_eay_public_decrypt (signature verification)
14380 (default implementations for RSA_public_encrypt,
14381 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14382 always reject numbers >= n.
14386 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14387 to synchronize access to 'locking_thread'. This is necessary on
14388 systems where access to 'locking_thread' (an 'unsigned long'
14389 variable) is not atomic.
14393 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14394 *before* setting the 'crypto_lock_rand' flag. The previous code had
14395 a race condition if 0 is a valid thread ID.
14397 *Travis Vitek <vitek@roguewave.com>*
14399 * Add support for shared libraries under Irix.
14401 *Albert Chin-A-Young <china@thewrittenword.com>*
14403 * Add configuration option to build on Linux on both big-endian and
14404 little-endian MIPS.
14406 *Ralf Baechle <ralf@uni-koblenz.de>*
14408 * Add the possibility to create shared libraries on HP-UX.
14412 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14414 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14415 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14416 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14417 PRNG state recovery was possible based on the output of
14418 one PRNG request appropriately sized to gain knowledge on
14419 'md' followed by enough consecutive 1-byte PRNG requests
14420 to traverse all of 'state'.
14422 1. When updating 'md_local' (the current thread's copy of 'md')
14423 during PRNG output generation, hash all of the previous
14424 'md_local' value, not just the half used for PRNG output.
14426 2. Make the number of bytes from 'state' included into the hash
14427 independent from the number of PRNG bytes requested.
14429 The first measure alone would be sufficient to avoid
14430 Markku-Juhani's attack. (Actually it had never occurred
14431 to me that the half of 'md_local' used for chaining was the
14432 half from which PRNG output bytes were taken -- I had always
14433 assumed that the secret half would be used.) The second
14434 measure makes sure that additional data from 'state' is never
14435 mixed into 'md_local' in small portions; this heuristically
14436 further strengthens the PRNG.
14440 * Fix crypto/bn/asm/mips3.s.
14444 * When only the key is given to "enc", the IV is undefined. Print out
14445 an error message in this case.
14449 * Handle special case when X509_NAME is empty in X509 printing routines.
14453 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14454 positive and less than q.
14458 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14459 used: it isn't thread safe and the add_lock_callback should handle
14462 *Paul Rose <Paul.Rose@bridge.com>*
14464 * Verify that incoming data obeys the block size in
14465 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14471 *Ulf Möller, Bodo Möller*
14473 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14474 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14475 when fixing the server behaviour for backwards-compatible 'client
14476 hello' messages. (Note that the attack is impractical against
14477 SSL 3.0 and TLS 1.0 anyway because length and version checking
14478 means that the probability of guessing a valid ciphertext is
14479 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14482 Before 0.9.5, the countermeasure (hide the error by generating a
14483 random 'decryption result') did not work properly because
14484 ERR_clear_error() was missing, meaning that SSL_get_error() would
14485 detect the supposedly ignored error.
14487 Both problems are now fixed.
14491 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14492 (previously it was 1024).
14496 * Fix for compatibility mode trust settings: ignore trust settings
14497 unless some valid trust or reject settings are present.
14501 * Fix for blowfish EVP: its a variable length cipher.
14505 * Fix various bugs related to DSA S/MIME verification. Handle missing
14506 parameters in DSA public key structures and return an error in the
14507 DSA routines if parameters are absent.
14511 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14512 in the current directory if neither $RANDFILE nor $HOME was set.
14513 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14514 caused some confusion to Windows users who haven't defined $HOME.
14515 Thus RAND_file_name() is changed again: e_os.h can define a
14516 DEFAULT_HOME, which will be used if $HOME is not set.
14517 For Windows, we use "C:"; on other platforms, we still require
14518 environment variables.
14520 * Move 'if (!initialized) RAND_poll()' into regions protected by
14521 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14522 having multiple threads call RAND_poll() concurrently.
14526 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14527 combination of a flag and a thread ID variable.
14528 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14529 flag), *other* threads can enter ssleay_add_bytes without obeying
14530 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14531 that they do not hold after the first thread unsets add_do_not_lock).
14535 * Change bctest again: '-x' expressions are not available in all
14536 versions of 'test'.
14540 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14542 * Fix a couple of memory leaks in PKCS7_dataDecode()
14544 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14546 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14547 the default extension for executables, if any. Also, make the perl
14548 scripts that use symlink() to test if it really exists and use "cp"
14549 if it doesn't. All this made OpenSSL compilable and installable in
14554 * Fix for asn1_GetSequence() for indefinite length constructed data.
14555 If SEQUENCE is length is indefinite just set c->slen to the total
14556 amount of data available.
14558 *Steve Henson, reported by shige@FreeBSD.org*
14560 *This change does not apply to 0.9.7.*
14562 * Change bctest to avoid here-documents inside command substitution
14563 (workaround for FreeBSD /bin/sh bug).
14564 For compatibility with Ultrix, avoid shell functions (introduced
14565 in the bctest version that searches along $PATH).
14569 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14570 with des_encrypt() defined on some operating systems, like Solaris
14575 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14576 On the Importance of Eliminating Errors in Cryptographic
14577 Computations, J. Cryptology 14 (2001) 2, 101-119,
14578 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14582 * MIPS assembler BIGNUM division bug fix.
14586 * Disabled incorrect Alpha assembler code.
14590 * Fix PKCS#7 decode routines so they correctly update the length
14591 after reading an EOC for the EXPLICIT tag.
14595 *This change does not apply to 0.9.7.*
14597 * Fix bug in PKCS#12 key generation routines. This was triggered
14598 if a 3DES key was generated with a 0 initial byte. Include
14599 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14600 (but broken) behaviour.
14604 * Enhance bctest to search for a working bc along $PATH and print
14607 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14609 * Fix memory leaks in err.c: free err_data string if necessary;
14610 don't write to the wrong index in ERR_set_error_data.
14614 * Implement ssl23_peek (analogous to ssl23_read), which previously
14619 * Replace rdtsc with `_emit` statements for VC++ version 5.
14621 *Jeremy Cooper <jeremy@baymoo.org>*
14623 * Make it possible to reuse SSLv2 sessions.
14627 * In copy_email() check for >= 0 as a return value for
14628 X509_NAME_get_index_by_NID() since 0 is a valid index.
14630 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14632 * Avoid coredump with unsupported or invalid public keys by checking if
14633 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14634 PKCS7_verify() fails with non detached data.
14638 * Don't use getenv in library functions when run as setuid/setgid.
14639 New function OPENSSL_issetugid().
14643 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14644 due to incorrect handling of multi-threading:
14646 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14648 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14650 3. Count how many times MemCheck_off() has been called so that
14651 nested use can be treated correctly. This also avoids
14652 inband-signalling in the previous code (which relied on the
14653 assumption that thread ID 0 is impossible).
14657 * Add "-rand" option also to s_client and s_server.
14661 * Fix CPU detection on Irix 6.x.
14662 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14663 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14665 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14670 *This change does not apply to 0.9.7.*
14672 * Use the cached encoding of an X509_NAME structure rather than
14673 copying it. This is apparently the reason for the libsafe "errors"
14674 but the code is actually correct.
14678 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14679 Bleichenbacher's DSA attack.
14680 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14681 to be set and top=0 forces the highest bit to be set; top=-1 is new
14682 and leaves the highest bit random.
14684 *Ulf Moeller, Bodo Moeller*
14686 * In the `NCONF_...`-based implementations for `CONF_...` queries
14687 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14688 a temporary CONF structure with the data component set to NULL
14689 (which gives segmentation faults in lh_retrieve).
14690 Instead, use NULL for the CONF pointer in CONF_get_string and
14691 CONF_get_number (which may use environment variables) and directly
14692 return NULL from CONF_get_section.
14696 * Fix potential buffer overrun for EBCDIC.
14700 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14701 keyUsage if basicConstraints absent for a CA.
14705 * Make SMIME_write_PKCS7() write mail header values with a format that
14706 is more generally accepted (no spaces before the semicolon), since
14707 some programs can't parse those values properly otherwise. Also make
14708 sure BIO's that break lines after each write do not create invalid
14713 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14714 macros previously used would not encode an empty SEQUENCE OF
14715 and break the signature.
14719 *This change does not apply to 0.9.7.*
14721 * Zero the premaster secret after deriving the master secret in
14726 * Add some EVP_add_digest_alias registrations (as found in
14727 OpenSSL_add_all_digests()) to SSL_library_init()
14728 aka OpenSSL_add_ssl_algorithms(). This provides improved
14729 compatibility with peers using X.509 certificates
14730 with unconventional AlgorithmIdentifier OIDs.
14734 * Fix for Irix with NO_ASM.
14736 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14738 * ./config script fixes.
14740 *Ulf Moeller, Richard Levitte*
14742 * Fix 'openssl passwd -1'.
14746 * Change PKCS12_key_gen_asc() so it can cope with non null
14747 terminated strings whose length is passed in the passlen
14748 parameter, for example from PEM callbacks. This was done
14749 by adding an extra length parameter to asc2uni().
14751 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14753 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14754 call failed, free the DSA structure.
14758 * Fix to uni2asc() to cope with zero length Unicode strings.
14759 These are present in some PKCS#12 files.
14763 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14764 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14765 when writing a 32767 byte record.
14767 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14769 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14770 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14772 (RSA objects have a reference count access to which is protected
14773 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14774 so they are meant to be shared between threads.)
14775 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14776 "Reddie, Steven" <Steven.Reddie@ca.com>*
14778 * Fix a deadlock in CRYPTO_mem_leaks().
14782 * Use better test patterns in bntest.
14786 * rand_win.c fix for Borland C.
14790 * BN_rshift bugfix for n == 0.
14794 * Add a 'bctest' script that checks for some known 'bc' bugs
14795 so that 'make test' does not abort just because 'bc' is broken.
14799 * Store verify_result within SSL_SESSION also for client side to
14800 avoid potential security hole. (Re-used sessions on the client side
14801 always resulted in verify_result==X509_V_OK, not using the original
14802 result of the server certificate verification.)
14806 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14807 SSL3_RT_APPLICATION_DATA, return 0.
14808 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14813 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14814 releases, have been re-implemented by renaming the previous
14815 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14816 and ssl3_read_internal, respectively, and adding 'peek' parameters
14817 to them. The new ssl[23]_{read,peek} functions are calls to
14818 ssl[23]_read_internal with the 'peek' flag set appropriately.
14819 A 'peek' parameter has also been added to ssl3_read_bytes, which
14820 does the actual work for ssl3_read_internal.
14824 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14825 the method-specific "init()" handler. Also clean up ex_data after
14826 calling the method-specific "finish()" handler. Previously, this was
14827 happening the other way round.
14831 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14832 The previous value, 12, was not always sufficient for BN_mod_exp().
14836 * Make sure that shared libraries get the internal name engine with
14837 the full version number and not just 0. This should mark the
14838 shared libraries as not backward compatible. Of course, this should
14839 be changed again when we can guarantee backward binary compatibility.
14843 * Fix typo in get_cert_by_subject() in by_dir.c
14845 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14847 * Rework the system to generate shared libraries:
14849 - Make note of the expected extension for the shared libraries and
14850 if there is a need for symbolic links from for example libcrypto.so.0
14851 to libcrypto.so.0.9.7. There is extended info in Configure for
14854 - Make as few rebuilds of the shared libraries as possible.
14856 - Still avoid linking the OpenSSL programs with the shared libraries.
14858 - When installing, install the shared libraries separately from the
14863 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14865 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14866 and not in SSL_clear because the latter is also used by the
14867 accept/connect functions; previously, the settings made by
14868 SSL_set_read_ahead would be lost during the handshake.
14870 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14872 * Correct util/mkdef.pl to be selective about disabled algorithms.
14873 Previously, it would create entries for disabled algorithms no
14878 * Added several new manual pages for SSL_* function.
14882 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14884 * In ssl23_get_client_hello, generate an error message when faced
14885 with an initial SSL 3.0/TLS record that is too small to contain the
14886 first two bytes of the ClientHello message, i.e. client_version.
14887 (Note that this is a pathologic case that probably has never happened
14888 in real life.) The previous approach was to use the version number
14889 from the record header as a substitute; but our protocol choice
14890 should not depend on that one because it is not authenticated
14891 by the Finished messages.
14895 * More robust randomness gathering functions for Windows.
14897 *Jeffrey Altman <jaltman@columbia.edu>*
14899 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14900 not set then we don't setup the error code for issuer check errors
14901 to avoid possibly overwriting other errors which the callback does
14902 handle. If an application does set the flag then we assume it knows
14903 what it is doing and can handle the new informational codes
14908 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14909 a general "ANY" type, as such it should be able to decode anything
14910 including tagged types. However it didn't check the class so it would
14911 wrongly interpret tagged types in the same way as their universal
14912 counterpart and unknown types were just rejected. Changed so that the
14913 tagged and unknown types are handled in the same way as a SEQUENCE:
14914 that is the encoding is stored intact. There is also a new type
14915 "V_ASN1_OTHER" which is used when the class is not universal, in this
14916 case we have no idea what the actual type is so we just lump them all
14921 * On VMS, stdout may very well lead to a file that is written to
14922 in a record-oriented fashion. That means that every write() will
14923 write a separate record, which will be read separately by the
14924 programs trying to read from it. This can be very confusing.
14926 The solution is to put a BIO filter in the way that will buffer
14927 text until a linefeed is reached, and then write everything a
14928 line at a time, so every record written will be an actual line,
14929 not chunks of lines and not (usually doesn't happen, but I've
14930 seen it once) several lines in one record. BIO_f_linebuffer() is
14933 Currently, it's a VMS-only method, because that's where it has
14934 been tested well enough.
14938 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14939 it can return incorrect results.
14940 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14941 but it was in 0.9.6-beta[12].)
14945 * Disable the check for content being present when verifying detached
14946 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14947 include zero length content when signing messages.
14951 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14952 BIO_ctrl (for BIO pairs).
14956 * Add DSO method for VMS.
14960 * Bug fix: Montgomery multiplication could produce results with the
14965 * Add RPM specification openssl.spec and modify it to build three
14966 packages. The default package contains applications, application
14967 documentation and run-time libraries. The devel package contains
14968 include files, static libraries and function documentation. The
14969 doc package contains the contents of the doc directory. The original
14970 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14974 * Add a large number of documentation files for many SSL routines.
14976 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14978 * Add a configuration entry for Sony News 4.
14980 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14982 * Don't set the two most significant bits to one when generating a
14983 random number < q in the DSA library.
14987 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14988 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14989 the underlying transport is blocking) if a handshake took place.
14990 (The default behaviour is needed by applications such as s_client
14991 and s_server that use select() to determine when to use SSL_read;
14992 but for applications that know in advance when to expect data, it
14993 just makes things more complicated.)
14997 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15002 * Add a few more EBCDIC conditionals that make `req` and `x509`
15003 work better on such systems.
15005 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15007 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15008 Update PKCS12_parse() so it copies the friendlyName and the
15009 keyid to the certificates aux info.
15013 * Fix bug in PKCS7_verify() which caused an infinite loop
15014 if there was more than one signature.
15016 *Sven Uszpelkat <su@celocom.de>*
15018 * Major change in util/mkdef.pl to include extra information
15019 about each symbol, as well as presenting variables as well
15020 as functions. This change means that there's n more need
15021 to rebuild the .num files when some algorithms are excluded.
15025 * Allow the verify time to be set by an application,
15026 rather than always using the current time.
15030 * Phase 2 verify code reorganisation. The certificate
15031 verify code now looks up an issuer certificate by a
15032 number of criteria: subject name, authority key id
15033 and key usage. It also verifies self signed certificates
15034 by the same criteria. The main comparison function is
15035 X509_check_issued() which performs these checks.
15037 Lot of changes were necessary in order to support this
15038 without completely rewriting the lookup code.
15040 Authority and subject key identifier are now cached.
15042 The LHASH 'certs' is X509_STORE has now been replaced
15043 by a STACK_OF(X509_OBJECT). This is mainly because an
15044 LHASH can't store or retrieve multiple objects with
15045 the same hash value.
15047 As a result various functions (which were all internal
15048 use only) have changed to handle the new X509_STORE
15049 structure. This will break anything that messed round
15050 with X509_STORE internally.
15052 The functions X509_STORE_add_cert() now checks for an
15053 exact match, rather than just subject name.
15055 The X509_STORE API doesn't directly support the retrieval
15056 of multiple certificates matching a given criteria, however
15057 this can be worked round by performing a lookup first
15058 (which will fill the cache with candidate certificates)
15059 and then examining the cache for matches. This is probably
15060 the best we can do without throwing out X509_LOOKUP
15061 entirely (maybe later...).
15063 The X509_VERIFY_CTX structure has been enhanced considerably.
15065 All certificate lookup operations now go via a get_issuer()
15066 callback. Although this currently uses an X509_STORE it
15067 can be replaced by custom lookups. This is a simple way
15068 to bypass the X509_STORE hackery necessary to make this
15069 work and makes it possible to use more efficient techniques
15070 in future. A very simple version which uses a simple
15071 STACK for its trusted certificate store is also provided
15072 using X509_STORE_CTX_trusted_stack().
15074 The verify_cb() and verify() callbacks now have equivalents
15075 in the X509_STORE_CTX structure.
15077 X509_STORE_CTX also has a 'flags' field which can be used
15078 to customise the verify behaviour.
15082 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15083 excludes S/MIME capabilities.
15087 * When a certificate request is read in keep a copy of the
15088 original encoding of the signed data and use it when outputting
15089 again. Signatures then use the original encoding rather than
15090 a decoded, encoded version which may cause problems if the
15091 request is improperly encoded.
15095 * For consistency with other BIO_puts implementations, call
15096 buffer_write(b, ...) directly in buffer_puts instead of calling
15099 In BIO_puts, increment b->num_write as in BIO_write.
15101 *Peter.Sylvester@EdelWeb.fr*
15103 * Fix BN_mul_word for the case where the word is 0. (We have to use
15104 BN_zero, we may not return a BIGNUM with an array consisting of
15105 words set to zero.)
15109 * Avoid calling abort() from within the library when problems are
15110 detected, except if preprocessor symbols have been defined
15111 (such as REF_CHECK, BN_DEBUG etc.).
15115 * New openssl application 'rsautl'. This utility can be
15116 used for low-level RSA operations. DER public key
15117 BIO/fp routines also added.
15121 * New Configure entry and patches for compiling on QNX 4.
15123 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15125 * A demo state-machine implementation was sponsored by
15126 Nuron (<http://www.nuron.com/>) and is now available in
15127 demos/state_machine.
15131 * New options added to the 'dgst' utility for signature
15132 generation and verification.
15136 * Unrecognized PKCS#7 content types are now handled via a
15137 catch all ASN1_TYPE structure. This allows unsupported
15138 types to be stored as a "blob" and an application can
15139 encode and decode it manually.
15143 * Fix various signed/unsigned issues to make a_strex.c
15144 compile under VC++.
15146 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15148 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15149 length if passed a buffer. ASN1_INTEGER_to_BN failed
15150 if passed a NULL BN and its argument was negative.
15152 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15154 * Modification to PKCS#7 encoding routines to output definite
15155 length encoding. Since currently the whole structures are in
15156 memory there's not real point in using indefinite length
15157 constructed encoding. However if OpenSSL is compiled with
15158 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15162 * Added BIO_vprintf() and BIO_vsnprintf().
15166 * Added more prefixes to parse for in the strings written
15167 through a logging bio, to cover all the levels that are available
15168 through syslog. The prefixes are now:
15170 PANIC, EMERG, EMR => LOG_EMERG
15171 ALERT, ALR => LOG_ALERT
15172 CRIT, CRI => LOG_CRIT
15173 ERROR, ERR => LOG_ERR
15174 WARNING, WARN, WAR => LOG_WARNING
15175 NOTICE, NOTE, NOT => LOG_NOTICE
15176 INFO, INF => LOG_INFO
15177 DEBUG, DBG => LOG_DEBUG
15179 and as before, if none of those prefixes are present at the
15180 beginning of the string, LOG_ERR is chosen.
15182 On Win32, the `LOG_*` levels are mapped according to this:
15184 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15185 LOG_WARNING => EVENTLOG_WARNING_TYPE
15186 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15190 * Made it possible to reconfigure with just the configuration
15191 argument "reconf" or "reconfigure". The command line arguments
15192 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15193 and are retrieved from there when reconfiguring.
15199 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15201 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15205 * The obj_dat.pl script was messing up the sorting of object
15206 names. The reason was that it compared the quoted version
15207 of strings as a result "OCSP" > "OCSP Signing" because
15208 " > SPACE. Changed script to store unquoted versions of
15209 names and add quotes on output. It was also omitting some
15210 names from the lookup table if they were given a default
15211 value (that is if SN is missing it is given the same
15212 value as LN and vice versa), these are now added on the
15213 grounds that if an object has a name we should be able to
15214 look it up. Finally added warning output when duplicate
15215 short or long names are found.
15219 * Changes needed for Tandem NSK.
15221 *Scott Uroff <scott@xypro.com>*
15223 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15224 RSA_padding_check_SSLv23(), special padding was never detected
15225 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15226 version rollback attacks was not effective.
15228 In s23_clnt.c, don't use special rollback-attack detection padding
15229 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15230 client; similarly, in s23_srvr.c, don't do the rollback check if
15231 SSL 2.0 is the only protocol enabled in the server.
15235 * Make it possible to get hexdumps of unprintable data with 'openssl
15236 asn1parse'. By implication, the functions ASN1_parse_dump() and
15237 BIO_dump_indent() are added.
15241 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15242 these print out strings and name structures based on various
15243 flags including RFC2253 support and proper handling of
15244 multibyte characters. Added options to the 'x509' utility
15245 to allow the various flags to be set.
15249 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15250 Also change the functions X509_cmp_current_time() and
15251 X509_gmtime_adj() work with an ASN1_TIME structure,
15252 this will enable certificates using GeneralizedTime in validity
15253 dates to be checked.
15257 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15258 negative public key encodings) on by default,
15259 NO_NEG_PUBKEY_BUG can be set to disable it.
15263 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15264 content octets. An i2c_ASN1_OBJECT is unnecessary because
15265 the encoding can be trivially obtained from the structure.
15269 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15270 not read locks (`CRYPTO_r_[un]lock`).
15274 * A first attempt at creating official support for shared
15275 libraries through configuration. I've kept it so the
15276 default is static libraries only, and the OpenSSL programs
15277 are always statically linked for now, but there are
15278 preparations for dynamic linking in place.
15279 This has been tested on Linux and Tru64.
15283 * Randomness polling function for Win9x, as described in:
15284 Peter Gutmann, Software Generation of Practically Strong
15289 * Fix so PRNG is seeded in req if using an already existing
15294 * New options to smime application. -inform and -outform
15295 allow alternative formats for the S/MIME message including
15296 PEM and DER. The -content option allows the content to be
15297 specified separately. This should allow things like Netscape
15298 form signing output easier to verify.
15302 * Fix the ASN1 encoding of tags using the 'long form'.
15306 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15307 STRING types. These convert content octets to and from the
15308 underlying type. The actual tag and length octets are
15309 already assumed to have been read in and checked. These
15310 are needed because all other string types have virtually
15311 identical handling apart from the tag. By having versions
15312 of the ASN1 functions that just operate on content octets
15313 IMPLICIT tagging can be handled properly. It also allows
15314 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15315 and ASN1_INTEGER are identical apart from the tag.
15319 * Change the handling of OID objects as follows:
15321 - New object identifiers are inserted in objects.txt, following
15322 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15323 - objects.pl is used to process obj_mac.num and create a new
15325 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15328 This is currently kind of a hack, and the perl code in objects.pl
15329 isn't very elegant, but it works as I intended. The simplest way
15330 to check that it worked correctly is to look in obj_dat.h and
15331 check the array nid_objs and make sure the objects haven't moved
15332 around (this is important!). Additions are OK, as well as
15333 consistent name changes.
15337 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15341 * Addition of the command line parameter '-rand file' to 'openssl req'.
15342 The given file adds to whatever has already been seeded into the
15343 random pool through the RANDFILE configuration file option or
15344 environment variable, or the default random state file.
15348 * mkstack.pl now sorts each macro group into lexical order.
15349 Previously the output order depended on the order the files
15350 appeared in the directory, resulting in needless rewriting
15355 * Patches to make OpenSSL compile under Win32 again. Mostly
15356 work arounds for the VC++ problem that it treats func() as
15357 func(void). Also stripped out the parts of mkdef.pl that
15358 added extra typesafe functions: these no longer exist.
15362 * Reorganisation of the stack code. The macros are now all
15363 collected in safestack.h . Each macro is defined in terms of
15364 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15365 DEBUG_SAFESTACK is now handled in terms of function casts,
15366 this has the advantage of retaining type safety without the
15367 use of additional functions. If DEBUG_SAFESTACK is not defined
15368 then the non typesafe macros are used instead. Also modified the
15369 mkstack.pl script to handle the new form. Needs testing to see
15370 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15371 the default if no major problems. Similar behaviour for ASN1_SET_OF
15372 and PKCS12_STACK_OF.
15376 * When some versions of IIS use the 'NET' form of private key the
15377 key derivation algorithm is different. Normally MD5(password) is
15378 used as a 128 bit RC4 key. In the modified case
15379 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15380 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15381 as the old Netscape_RSA functions except they have an additional
15382 'sgckey' parameter which uses the modified algorithm. Also added
15383 an -sgckey command line option to the rsa utility. Thanks to
15384 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15385 algorithm to openssl-dev.
15389 * The evp_local.h macros were using 'c.##kname' which resulted in
15390 invalid expansion on some systems (SCO 5.0.5 for example).
15391 Corrected to 'c.kname'.
15393 *Phillip Porch <root@theporch.com>*
15395 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15396 a STACK of email addresses from a certificate or request, these look
15397 in the subject name and the subject alternative name extensions and
15398 omit any duplicate addresses.
15402 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15403 This makes DSA verification about 2 % faster.
15407 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15408 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15409 plus overhead for 1024 bit moduli).
15410 This makes exponentiations about 0.5 % faster for 1024 bit
15411 exponents (as measured by "openssl speed rsa2048").
15415 * Rename memory handling macros to avoid conflicts with other
15417 Malloc => OPENSSL_malloc
15418 Malloc_locked => OPENSSL_malloc_locked
15419 Realloc => OPENSSL_realloc
15420 Free => OPENSSL_free
15424 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15425 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15429 * CygWin32 support.
15431 *John Jarvie <jjarvie@newsguy.com>*
15433 * The type-safe stack code has been rejigged. It is now only compiled
15434 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15435 by default all type-specific stack functions are "#define"d back to
15436 standard stack functions. This results in more streamlined output
15437 but retains the type-safety checking possibilities of the original
15442 * The STACK code has been cleaned up, and certain type declarations
15443 that didn't make a lot of sense have been brought in line. This has
15444 also involved a cleanup of sorts in safestack.h to more correctly
15445 map type-safe stack functions onto their plain stack counterparts.
15446 This work has also resulted in a variety of "const"ifications of
15447 lots of the code, especially `_cmp` operations which should normally
15448 be prototyped with "const" parameters anyway.
15452 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15453 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15454 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15455 where all of 'md' is used each time the PRNG is used, but 'state'
15456 is used only indexed by a cyclic counter. As entropy may not be
15457 well distributed from the beginning, 'md' is important as a
15458 chaining variable. However, the output function chains only half
15459 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15460 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15461 in all of 'state' being rewritten, with the new values depending
15462 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15466 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15467 the handshake is continued after ssl_verify_cert_chain();
15468 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15469 can lead to 'unexplainable' connection aborts later.
15471 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15473 * Major EVP API cipher revision.
15474 Add hooks for extra EVP features. This allows various cipher
15475 parameters to be set in the EVP interface. Support added for variable
15476 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15477 setting of RC2 and RC5 parameters.
15479 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15482 Remove lots of duplicated code from the EVP library. For example *every*
15483 cipher init() function handles the 'iv' in the same way according to the
15484 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15485 for CFB and OFB modes they zero ctx->num.
15487 New functionality allows removal of S/MIME code RC2 hack.
15489 Most of the routines have the same form and so can be declared in terms
15492 By shifting this to the top level EVP_CipherInit() it can be removed from
15493 all individual ciphers. If the cipher wants to handle IVs or keys
15494 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15497 Change lots of functions like EVP_EncryptUpdate() to now return a
15498 value: although software versions of the algorithms cannot fail
15499 any installed hardware versions can.
15503 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15504 this option is set, tolerate broken clients that send the negotiated
15505 protocol version number instead of the requested protocol version
15510 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15511 i.e. non-zero for export ciphersuites, zero otherwise.
15512 Previous versions had this flag inverted, inconsistent with
15513 rsa_tmp_cb (..._TMP_RSA_CB).
15515 *Bodo Moeller; problem reported by Amit Chopra*
15517 * Add missing DSA library text string. Work around for some IIS
15518 key files with invalid SEQUENCE encoding.
15522 * Add a document (doc/standards.txt) that list all kinds of standards
15523 and so on that are implemented in OpenSSL.
15527 * Enhance c_rehash script. Old version would mishandle certificates
15528 with the same subject name hash and wouldn't handle CRLs at all.
15529 Added -fingerprint option to crl utility, to support new c_rehash
15534 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15538 * Fix for SSL server purpose checking. Server checking was
15539 rejecting certificates which had extended key usage present
15540 but no ssl client purpose.
15542 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15544 * Make PKCS#12 code work with no password. The PKCS#12 spec
15545 is a little unclear about how a blank password is handled.
15546 Since the password in encoded as a BMPString with terminating
15547 double NULL a zero length password would end up as just the
15548 double NULL. However no password at all is different and is
15549 handled differently in the PKCS#12 key generation code. NS
15550 treats a blank password as zero length. MSIE treats it as no
15551 password on export: but it will try both on import. We now do
15552 the same: PKCS12_parse() tries zero length and no password if
15553 the password is set to "" or NULL (NULL is now a valid password:
15554 it wasn't before) as does the pkcs12 application.
15558 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15559 perror when PEM_read_bio_X509_REQ fails, the error message must
15560 be obtained from the error queue.
15564 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15565 it in ERR_remove_state if appropriate, and change ERR_get_state
15566 accordingly to avoid race conditions (this is necessary because
15567 thread_hash is no longer constant once set).
15571 * Bugfix for linux-elf makefile.one.
15575 * RSA_get_default_method() will now cause a default
15576 RSA_METHOD to be chosen if one doesn't exist already.
15577 Previously this was only set during a call to RSA_new()
15578 or RSA_new_method(NULL) meaning it was possible for
15579 RSA_get_default_method() to return NULL.
15583 * Added native name translation to the existing DSO code
15584 that will convert (if the flag to do so is set) filenames
15585 that are sufficiently small and have no path information
15586 into a canonical native form. Eg. "blah" converted to
15587 "libblah.so" or "blah.dll" etc.
15591 * New function ERR_error_string_n(e, buf, len) which is like
15592 ERR_error_string(e, buf), but writes at most 'len' bytes
15593 including the 0 terminator. For ERR_error_string_n, 'buf'
15596 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15598 * CONF library reworked to become more general. A new CONF
15599 configuration file reader "class" is implemented as well as a
15600 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15601 old `CONF_*` functions are still there, but are reimplemented to
15602 work in terms of the new functions. Also, a set of functions
15603 to handle the internal storage of the configuration data is
15604 provided to make it easier to write new configuration file
15605 reader "classes" (I can definitely see something reading a
15606 configuration file in XML format, for example), called `_CONF_*`,
15607 or "the configuration storage API"...
15609 The new configuration file reading functions are:
15611 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15612 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15614 NCONF_default, NCONF_WIN32
15616 NCONF_dump_fp, NCONF_dump_bio
15618 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15619 NCONF_new creates a new CONF object. This works in the same way
15620 as other interfaces in OpenSSL, like the BIO interface.
15621 `NCONF_dump_*` dump the internal storage of the configuration file,
15622 which is useful for debugging. All other functions take the same
15623 arguments as the old `CONF_*` functions with the exception of the
15624 first that must be a `CONF *` instead of a `LHASH *`.
15626 To make it easier to use the new classes with the old `CONF_*` functions,
15627 the function CONF_set_default_method is provided.
15631 * Add '-tls1' option to 'openssl ciphers', which was already
15632 mentioned in the documentation but had not been implemented.
15633 (This option is not yet really useful because even the additional
15634 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15638 * Initial DSO code added into libcrypto for letting OpenSSL (and
15639 OpenSSL-based applications) load shared libraries and bind to
15640 them in a portable way.
15642 *Geoff Thorpe, with contributions from Richard Levitte*
15644 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15646 * Make sure _lrotl and _lrotr are only used with MSVC.
15648 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15649 (the default implementation of RAND_status).
15651 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15652 to '-clrext' (= clear extensions), as intended and documented.
15653 *Bodo Moeller; inconsistency pointed out by Michael Attili
15654 <attili@amaxo.com>*
15656 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15657 was larger than the MD block size.
15659 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15661 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15662 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15663 using the passed key: if the passed key was a private key the result
15664 of X509_print(), for example, would be to print out all the private key
15669 * des_quad_cksum() byte order bug fix.
15670 *Ulf Möller, using the problem description in krb4-0.9.7, where
15671 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15673 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15676 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15678 * For easily testing in shell scripts whether some command
15679 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15680 returns with exit code 0 iff no command of the given name is available.
15681 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15682 the output goes to stdout and nothing is printed to stderr.
15683 Additional arguments are always ignored.
15685 Since for each cipher there is a command of the same name,
15686 the 'no-cipher' compilation switches can be tested this way.
15688 ('openssl no-XXX' is not able to detect pseudo-commands such
15689 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15693 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15697 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15698 is set; it will be thrown away anyway because each handshake creates
15700 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15701 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15702 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15703 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15707 * New s_client option -ign_eof: EOF at stdin is ignored, and
15708 'Q' and 'R' lose their special meanings (quit/renegotiate).
15709 This is part of what -quiet does; unlike -quiet, -ign_eof
15710 does not suppress any output.
15714 * Add compatibility options to the purpose and trust code. The
15715 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15716 accepts a certificate or CA, this was the previous behaviour,
15717 with all the associated security issues.
15719 X509_TRUST_COMPAT is the old trust behaviour: only and
15720 automatically trust self signed roots in certificate store. A
15721 new trust setting X509_TRUST_DEFAULT is used to specify that
15722 a purpose has no associated trust setting and it should instead
15723 use the value in the default purpose.
15727 * Fix the PKCS#8 DSA private key code so it decodes keys again
15728 and fix a memory leak.
15732 * In util/mkerr.pl (which implements 'make errors'), preserve
15733 reason strings from the previous version of the .c file, as
15734 the default to have only downcase letters (and digits) in
15735 automatically generated reasons codes is not always appropriate.
15739 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15740 using strerror. Previously, ERR_reason_error_string() returned
15741 library names as reason strings for SYSerr; but SYSerr is a special
15742 case where small numbers are errno values, not library numbers.
15746 * Add '-dsaparam' option to 'openssl dhparam' application. This
15747 converts DSA parameters into DH parameters. (When creating parameters,
15748 DSA_generate_parameters is used.)
15752 * Include 'length' (recommended exponent length) in C code generated
15753 by 'openssl dhparam -C'.
15757 * The second argument to set_label in perlasm was already being used
15758 so couldn't be used as a "file scope" flag. Moved to third argument
15763 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15764 instead of RAND_bytes for encryption IVs and salts.
15768 * Include RAND_status() into RAND_METHOD instead of implementing
15769 it only for md_rand.c Otherwise replacing the PRNG by calling
15770 RAND_set_rand_method would be impossible.
15774 * Don't let DSA_generate_key() enter an infinite loop if the random
15775 number generation fails.
15779 * New 'rand' application for creating pseudo-random output.
15783 * Added configuration support for Linux/IA64
15785 *Rolf Haberrecker <rolf@suse.de>*
15787 * Assembler module support for Mingw32.
15791 * Shared library support for HPUX (in shlib/).
15793 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15795 * Shared library support for Solaris gcc.
15797 *Lutz Behnke <behnke@trustcenter.de>*
15799 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15801 * PKCS7_encrypt() was adding text MIME headers twice because they
15802 were added manually and by SMIME_crlf_copy().
15806 * In bntest.c don't call BN_rand with zero bits argument.
15808 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15810 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15811 case was implemented. This caused BN_div_recp() to fail occasionally.
15815 * Add an optional second argument to the set_label() in the perl
15816 assembly language builder. If this argument exists and is set
15817 to 1 it signals that the assembler should use a symbol whose
15818 scope is the entire file, not just the current function. This
15819 is needed with MASM which uses the format label:: for this scope.
15821 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15823 * Change the ASN1 types so they are typedefs by default. Before
15824 almost all types were #define'd to ASN1_STRING which was causing
15825 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15830 * Change names of new functions to the new get1/get0 naming
15831 convention: After 'get1', the caller owns a reference count
15832 and has to call `..._free`; 'get0' returns a pointer to some
15833 data structure without incrementing reference counters.
15834 (Some of the existing 'get' functions increment a reference
15835 counter, some don't.)
15836 Similarly, 'set1' and 'add1' functions increase reference
15837 counters or duplicate objects.
15841 * Allow for the possibility of temp RSA key generation failure:
15842 the code used to assume it always worked and crashed on failure.
15846 * Fix potential buffer overrun problem in BIO_printf().
15847 *Ulf Möller, using public domain code by Patrick Powell; problem
15848 pointed out by David Sacerdote <das33@cornell.edu>*
15850 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15851 RAND_egd() and RAND_status(). In the command line application,
15852 the EGD socket can be specified like a seed file using RANDFILE
15857 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15858 Some CAs (e.g. Verisign) distribute certificates in this form.
15862 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15863 list to exclude them. This means that no special compilation option
15864 is needed to use anonymous DH: it just needs to be included in the
15869 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15870 EVP_MD_type. The old functionality is available in a new macro called
15871 EVP_MD_md(). Change code that uses it and update docs.
15875 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15876 where the `void *` argument is replaced by a function pointer argument.
15877 Previously `void *` was abused to point to functions, which works on
15878 many platforms, but is not correct. As these functions are usually
15879 called by macros defined in OpenSSL header files, most source code
15880 should work without changes.
15884 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15885 sections with information on -D... compiler switches used for
15886 compiling the library so that applications can see them. To enable
15887 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15888 must be defined. E.g.,
15889 #define OPENSSL_ALGORITHM_DEFINES
15890 #include <openssl/opensslconf.h>
15891 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15893 *Richard Levitte, Ulf and Bodo Möller*
15895 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15900 * Change the 'other' type in certificate aux info to a STACK_OF
15901 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15902 the required ASN1 format: arbitrary types determined by an OID.
15906 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15907 argument to 'req'. This is not because the function is newer or
15908 better than others it just uses the work 'NEW' in the certificate
15909 request header lines. Some software needs this.
15913 * Reorganise password command line arguments: now passwords can be
15914 obtained from various sources. Delete the PEM_cb function and make
15915 it the default behaviour: i.e. if the callback is NULL and the
15916 usrdata argument is not NULL interpret it as a null terminated pass
15917 phrase. If usrdata and the callback are NULL then the pass phrase
15918 is prompted for as usual.
15922 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15923 the support is automatically enabled. The resulting binaries will
15924 autodetect the card and use it if present.
15926 *Ben Laurie and Compaq Inc.*
15928 * Work around for Netscape hang bug. This sends certificate request
15929 and server done in one record. Since this is perfectly legal in the
15930 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15931 the bugs/SSLv3 entry for more info.
15935 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15939 * Add -rand argument to smime and pkcs12 applications and read/write
15944 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15948 * Add command line password options to the remaining applications.
15952 * Bug fix for BN_div_recp() for numerators with an even number of
15957 * More tests in bntest.c, and changed test_bn output.
15961 * ./config recognizes MacOS X now.
15965 * Bug fix for BN_div() when the first words of num and divisor are
15966 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15970 * Add support for various broken PKCS#8 formats, and command line
15971 options to produce them.
15975 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15976 get temporary BIGNUMs from a BN_CTX.
15980 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15985 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15986 include a #define from the old name to the new. The original intent
15987 was that statically linked binaries could for example just call
15988 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15989 link with digests. This never worked because SSLeay_add_all_digests()
15990 and SSLeay_add_all_ciphers() were in the same source file so calling
15991 one would link with the other. They are now in separate source files.
15995 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15999 * Use a less unusual form of the Miller-Rabin primality test (it used
16000 a binary algorithm for exponentiation integrated into the Miller-Rabin
16001 loop, our standard modexp algorithms are faster).
16005 * Support for the EBCDIC character set completed.
16007 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16009 * Source code cleanups: use const where appropriate, eliminate casts,
16010 use `void *` instead of `char *` in lhash.
16014 * Bugfix: ssl3_send_server_key_exchange was not restartable
16015 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16016 this the server could overwrite ephemeral keys that the client
16021 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16022 using 50 iterations of the Rabin-Miller test.
16024 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16025 iterations of the Rabin-Miller test as required by the appendix
16026 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16027 As BN_is_prime_fasttest includes trial division, DSA parameter
16028 generation becomes much faster.
16030 This implies a change for the callback functions in DSA_is_prime
16031 and DSA_generate_parameters: The callback function is called once
16032 for each positive witness in the Rabin-Miller test, not just
16033 occasionally in the inner loop; and the parameters to the
16034 callback function now provide an iteration count for the outer
16035 loop rather than for the current invocation of the inner loop.
16036 DSA_generate_parameters additionally can call the callback
16037 function with an 'iteration count' of -1, meaning that a
16038 candidate has passed the trial division test (when q is generated
16039 from an application-provided seed, trial division is skipped).
16043 * New function BN_is_prime_fasttest that optionally does trial
16044 division before starting the Rabin-Miller test and has
16045 an additional BN_CTX * argument (whereas BN_is_prime always
16046 has to allocate at least one BN_CTX).
16047 'callback(1, -1, cb_arg)' is called when a number has passed the
16048 trial division stage.
16052 * Fix for bug in CRL encoding. The validity dates weren't being handled
16057 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16061 * New function BN_pseudo_rand().
16065 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16066 bignum version of BN_from_montgomery() with the working code from
16067 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16072 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16073 made it impossible to use the same SSL_SESSION data structure in
16074 SSL2 clients in multiple threads.
16078 * The return value of RAND_load_file() no longer counts bytes obtained
16079 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16080 to seed the PRNG (previously an explicit byte count was required).
16082 *Ulf Möller, Bodo Möller*
16084 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16085 used `char *` instead of `void *` and had casts all over the place.
16089 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16093 * Retain source code compatibility for BN_prime_checks macro:
16094 BN_is_prime(..., BN_prime_checks, ...) now uses
16095 BN_prime_checks_for_size to determine the appropriate number of
16096 Rabin-Miller iterations.
16100 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16101 DH_CHECK_P_NOT_SAFE_PRIME.
16102 (Check if this is true? OpenPGP calls them "strong".)
16106 * Merge the functionality of "dh" and "gendh" programs into a new program
16107 "dhparam". The old programs are retained for now but will handle DH keys
16108 (instead of parameters) in future.
16112 * Make the ciphers, s_server and s_client programs check the return values
16113 when a new cipher list is set.
16117 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16118 ciphers. Before when the 56bit ciphers were enabled the sorting was
16121 The syntax for the cipher sorting has been extended to support sorting by
16122 cipher-strength (using the strength_bits hard coded in the tables).
16123 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16125 Fix a bug in the cipher-command parser: when supplying a cipher command
16126 string with an "undefined" symbol (neither command nor alphanumeric
16127 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16128 an error is flagged.
16130 Due to the strength-sorting extension, the code of the
16131 ssl_create_cipher_list() function was completely rearranged. I hope that
16132 the readability was also increased :-)
16134 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16136 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16137 for the first serial number and places 2 in the serial number file. This
16138 avoids problems when the root CA is created with serial number zero and
16139 the first user certificate has the same issuer name and serial number
16144 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16145 the new code. Add documentation for this stuff.
16149 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16150 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16151 structures and behave in an analogous way to the X509v3 functions:
16152 they shouldn't be called directly but wrapper functions should be used
16155 So we also now have some wrapper functions that call the X509at functions
16156 when passed certificate requests. (TO DO: similar things can be done with
16157 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16158 things. Some of these need some d2i or i2d and print functionality
16159 because they handle more complex structures.)
16163 * Add missing #ifndefs that caused missing symbols when building libssl
16164 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16165 NO_RSA in `ssl/s2*.c`.
16167 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16169 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16170 has a return value which indicates the quality of the random data
16171 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16172 error queue. New function RAND_pseudo_bytes() generates output that is
16173 guaranteed to be unique but not unpredictable. RAND_add is like
16174 RAND_seed, but takes an extra argument for an entropy estimate
16175 (RAND_seed always assumes full entropy).
16179 * Do more iterations of Rabin-Miller probable prime test (specifically,
16180 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16181 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16182 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16183 false-positive rate of at most 2^-80 for random input.
16187 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16191 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16192 in the 0.9.5 release), this returns the chain
16193 from an X509_CTX structure with a dup of the stack and all
16194 the X509 reference counts upped: so the stack will exist
16195 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16198 Also make SSL_SESSION_print() print out the verify return
16203 * Add manpage for the pkcs12 command. Also change the default
16204 behaviour so MAC iteration counts are used unless the new
16205 -nomaciter option is used. This improves file security and
16206 only older versions of MSIE (4.0 for example) need it.
16210 * Honor the no-xxx Configure options when creating .DEF files.
16214 * Add PKCS#10 attributes to field table: challengePassword,
16215 unstructuredName and unstructuredAddress. These are taken from
16216 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16217 international characters are used.
16219 More changes to X509_ATTRIBUTE code: allow the setting of types
16220 based on strings. Remove the 'loc' parameter when adding
16221 attributes because these will be a SET OF encoding which is sorted
16226 * Initial changes to the 'req' utility to allow request generation
16227 automation. This will allow an application to just generate a template
16228 file containing all the field values and have req construct the
16231 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16232 used all over the place including certificate requests and PKCS#7
16233 structures. They are currently handled manually where necessary with
16234 some primitive wrappers for PKCS#7. The new functions behave in a
16235 manner analogous to the X509 extension functions: they allow
16236 attributes to be looked up by NID and added.
16238 Later something similar to the X509V3 code would be desirable to
16239 automatically handle the encoding, decoding and printing of the
16240 more complex types. The string types like challengePassword can
16241 be handled by the string table functions.
16243 Also modified the multi byte string table handling. Now there is
16244 a 'global mask' which masks out certain types. The table itself
16245 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16246 is useful when for example there is only one permissible type
16247 (as in countryName) and using the mask might result in no valid
16252 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16253 SSL_get_peer_finished to allow applications to obtain the latest
16254 Finished messages sent to the peer or expected from the peer,
16255 respectively. (SSL_get_peer_finished is usually the Finished message
16256 actually received from the peer, otherwise the protocol will be aborted.)
16258 As the Finished message are message digests of the complete handshake
16259 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16260 be used for external authentication procedures when the authentication
16261 provided by SSL/TLS is not desired or is not enough.
16265 * Enhanced support for Alpha Linux is added. Now ./config checks if
16266 the host supports BWX extension and if Compaq C is present on the
16267 $PATH. Just exploiting of the BWX extension results in 20-30%
16268 performance kick for some algorithms, e.g. DES and RC4 to mention
16269 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16274 * Add support for MS "fast SGC". This is arguably a violation of the
16275 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16276 weak crypto and after checking the certificate is SGC a second one
16277 with strong crypto. MS SGC stops the first handshake after receiving
16278 the server certificate message and sends a second client hello. Since
16279 a server will typically do all the time consuming operations before
16280 expecting any further messages from the client (server key exchange
16281 is the most expensive) there is little difference between the two.
16283 To get OpenSSL to support MS SGC we have to permit a second client
16284 hello message after we have sent server done. In addition we have to
16285 reset the MAC if we do get this second client hello.
16289 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16290 if a DER encoded private key is RSA or DSA traditional format. Changed
16291 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16292 format DER encoded private key. Newer code should use PKCS#8 format which
16293 has the key type encoded in the ASN1 structure. Added DER private key
16294 support to pkcs8 application.
16298 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16299 ciphersuites has been selected (as required by the SSL 3/TLS 1
16300 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16301 is set, we interpret this as a request to violate the specification
16302 (the worst that can happen is a handshake failure, and 'correct'
16303 behaviour would result in a handshake failure anyway).
16307 * In SSL_CTX_add_session, take into account that there might be multiple
16308 SSL_SESSION structures with the same session ID (e.g. when two threads
16309 concurrently obtain them from an external cache).
16310 The internal cache can handle only one SSL_SESSION with a given ID,
16311 so if there's a conflict, we now throw out the old one to achieve
16316 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16317 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16318 some routines that use cipher OIDs: some ciphers do not have OIDs
16319 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16324 * Simplify the trust setting structure and code. Now we just have
16325 two sequences of OIDs for trusted and rejected settings. These will
16326 typically have values the same as the extended key usage extension
16327 and any application specific purposes.
16329 The trust checking code now has a default behaviour: it will just
16330 check for an object with the same NID as the passed id. Functions can
16331 be provided to override either the default behaviour or the behaviour
16332 for a given id. SSL client, server and email already have functions
16333 in place for compatibility: they check the NID and also return "trusted"
16334 if the certificate is self signed.
16338 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16339 traditional format into an EVP_PKEY structure.
16343 * Add a password callback function PEM_cb() which either prompts for
16344 a password if usr_data is NULL or otherwise assumes it is a null
16345 terminated password. Allow passwords to be passed on command line
16346 environment or config files in a few more utilities.
16350 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16351 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16352 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16353 Update documentation.
16357 * Support for ASN1 "NULL" type. This could be handled before by using
16358 ASN1_TYPE but there wasn't any function that would try to read a NULL
16359 and produce an error if it couldn't. For compatibility we also have
16360 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16361 don't allocate anything because they don't need to.
16365 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16368 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16370 * Rebuild of the memory allocation routines used by OpenSSL code and
16371 possibly others as well. The purpose is to make an interface that
16372 provide hooks so anyone can build a separate set of allocation and
16373 deallocation routines to be used by OpenSSL, for example memory
16374 pool implementations, or something else, which was previously hard
16375 since Malloc(), Realloc() and Free() were defined as macros having
16376 the values malloc, realloc and free, respectively (except for Win32
16377 compilations). The same is provided for memory debugging code.
16378 OpenSSL already comes with functionality to find memory leaks, but
16379 this gives people a chance to debug other memory problems.
16381 With these changes, a new set of functions and macros have appeared:
16383 CRYPTO_set_mem_debug_functions() [F]
16384 CRYPTO_get_mem_debug_functions() [F]
16385 CRYPTO_dbg_set_options() [F]
16386 CRYPTO_dbg_get_options() [F]
16387 CRYPTO_malloc_debug_init() [M]
16389 The memory debug functions are NULL by default, unless the library
16390 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16391 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16392 gives the standard debugging functions that come with OpenSSL) or
16393 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16394 provided by the library user) must be used. When the standard
16395 debugging functions are used, CRYPTO_dbg_set_options can be used to
16396 request additional information:
16397 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16398 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16400 Also, things like CRYPTO_set_mem_functions will always give the
16401 expected result (the new set of functions is used for allocation
16402 and deallocation) at all times, regardless of platform and compiler
16405 To finish it up, some functions that were never use in any other
16406 way than through macros have a new API and new semantic:
16408 CRYPTO_dbg_malloc()
16409 CRYPTO_dbg_realloc()
16412 All macros of value have retained their old syntax.
16414 *Richard Levitte and Bodo Moeller*
16416 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16417 ordering of SMIMECapabilities wasn't in "strength order" and there
16418 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16423 * Some ASN1 types with illegal zero length encoding (INTEGER,
16424 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16426 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16428 * Merge in my S/MIME library for OpenSSL. This provides a simple
16429 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16430 functionality to handle multipart/signed properly) and a utility
16431 called 'smime' to call all this stuff. This is based on code I
16432 originally wrote for Celo who have kindly allowed it to be
16433 included in OpenSSL.
16437 * Add variants des_set_key_checked and des_set_key_unchecked of
16438 des_set_key (aka des_key_sched). Global variable des_check_key
16439 decides which of these is called by des_set_key; this way
16440 des_check_key behaves as it always did, but applications and
16441 the library itself, which was buggy for des_check_key == 1,
16442 have a cleaner way to pick the version they need.
16446 * New function PKCS12_newpass() which changes the password of a
16451 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16452 dynamic mix. In both cases the ids can be used as an index into the
16453 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16454 functions so they accept a list of the field values and the
16455 application doesn't need to directly manipulate the X509_TRUST
16460 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16465 * Modify the way the V3 extension code looks up extensions. This now
16466 works in a similar way to the object code: we have some "standard"
16467 extensions in a static table which is searched with OBJ_bsearch()
16468 and the application can add dynamic ones if needed. The file
16469 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16470 updated whenever a new extension is added to the core code and kept
16471 in ext_nid order. There is a simple program 'tabtest.c' which checks
16472 this. New extensions are not added too often so this file can readily
16473 be maintained manually.
16475 There are two big advantages in doing things this way. The extensions
16476 can be looked up immediately and no longer need to be "added" using
16477 X509V3_add_standard_extensions(): this function now does nothing.
16478 Side note: I get *lots* of email saying the extension code doesn't
16479 work because people forget to call this function.
16480 Also no dynamic allocation is done unless new extensions are added:
16481 so if we don't add custom extensions there is no need to call
16482 X509V3_EXT_cleanup().
16486 * Modify enc utility's salting as follows: make salting the default. Add a
16487 magic header, so unsalted files fail gracefully instead of just decrypting
16488 to garbage. This is because not salting is a big security hole, so people
16489 should be discouraged from doing it.
16493 * Fixes and enhancements to the 'x509' utility. It allowed a message
16494 digest to be passed on the command line but it only used this
16495 parameter when signing a certificate. Modified so all relevant
16496 operations are affected by the digest parameter including the
16497 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16498 DSA key was used because it didn't fix the digest.
16502 * Initial certificate chain verify code. Currently tests the untrusted
16503 certificates for consistency with the verify purpose (which is set
16504 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16506 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16507 this is because it will reject chains with invalid extensions whereas
16508 every previous version of OpenSSL and SSLeay made no checks at all.
16510 Trust code: checks the root CA for the relevant trust settings. Trust
16511 settings have an initial value consistent with the verify purpose: e.g.
16512 if the verify purpose is for SSL client use it expects the CA to be
16513 trusted for SSL client use. However the default value can be changed to
16514 permit custom trust settings: one example of this would be to only trust
16515 certificates from a specific "secure" set of CAs.
16517 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16518 which should be used for version portability: especially since the
16519 verify structure is likely to change more often now.
16521 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16522 to set them. If not set then assume SSL clients will verify SSL servers
16525 Two new options to the verify program: -untrusted allows a set of
16526 untrusted certificates to be passed in and -purpose which sets the
16527 intended purpose of the certificate. If a purpose is set then the
16528 new chain verify code is used to check extension consistency.
16532 * Support for the authority information access extension.
16536 * Modify RSA and DSA PEM read routines to transparently handle
16537 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16538 public keys in a format compatible with certificate
16539 SubjectPublicKeyInfo structures. Unfortunately there were already
16540 functions called *_PublicKey_* which used various odd formats so
16541 these are retained for compatibility: however the DSA variants were
16542 never in a public release so they have been deleted. Changed dsa/rsa
16543 utilities to handle the new format: note no releases ever handled public
16544 keys so we should be OK.
16546 The primary motivation for this change is to avoid the same fiasco
16547 that dogs private keys: there are several incompatible private key
16548 formats some of which are standard and some OpenSSL specific and
16549 require various evil hacks to allow partial transparent handling and
16550 even then it doesn't work with DER formats. Given the option anything
16551 other than PKCS#8 should be dumped: but the other formats have to
16552 stay in the name of compatibility.
16554 With public keys and the benefit of hindsight one standard format
16555 is used which works with EVP_PKEY, RSA or DSA structures: though
16556 it clearly returns an error if you try to read the wrong kind of key.
16558 Added a -pubkey option to the 'x509' utility to output the public key.
16559 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16560 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16561 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16562 that do the same as the `EVP_PKEY_assign_*()` except they up the
16563 reference count of the added key (they don't "swallow" the
16568 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16569 CRLs would fail if the file contained no certificates or no CRLs:
16570 added a new function to read in both types and return the number
16571 read: this means that if none are read it will be an error. The
16572 DER versions of the certificate and CRL reader would always fail
16573 because it isn't possible to mix certificates and CRLs in DER format
16574 without choking one or the other routine. Changed this to just read
16575 a certificate: this is the best we can do. Also modified the code
16576 in `apps/verify.c` to take notice of return codes: it was previously
16577 attempting to read in certificates from NULL pointers and ignoring
16578 any errors: this is one reason why the cert and CRL reader seemed
16579 to work. It doesn't check return codes from the default certificate
16580 routines: these may well fail if the certificates aren't installed.
16584 * Code to support otherName option in GeneralName.
16588 * First update to verify code. Change the verify utility
16589 so it warns if it is passed a self signed certificate:
16590 for consistency with the normal behaviour. X509_verify
16591 has been modified to it will now verify a self signed
16592 certificate if *exactly* the same certificate appears
16593 in the store: it was previously impossible to trust a
16594 single self signed certificate. This means that:
16595 openssl verify ss.pem
16596 now gives a warning about a self signed certificate but
16597 openssl verify -CAfile ss.pem ss.pem
16602 * For servers, store verify_result in SSL_SESSION data structure
16603 (and add it to external session representation).
16604 This is needed when client certificate verifications fails,
16605 but an application-provided verification callback (set by
16606 SSL_CTX_set_cert_verify_callback) allows accepting the session
16607 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16608 but returns 1): When the session is reused, we have to set
16609 ssl->verify_result to the appropriate error code to avoid
16612 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16614 * Fix a bug in the new PKCS#7 code: it didn't consider the
16615 case in PKCS7_dataInit() where the signed PKCS7 structure
16616 didn't contain any existing data because it was being created.
16618 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16620 * Add a salt to the key derivation routines in enc.c. This
16621 forms the first 8 bytes of the encrypted file. Also add a
16622 -S option to allow a salt to be input on the command line.
16626 * New function X509_cmp(). Oddly enough there wasn't a function
16627 to compare two certificates. We do this by working out the SHA1
16628 hash and comparing that. X509_cmp() will be needed by the trust
16633 * SSL_get1_session() is like SSL_get_session(), but increments
16634 the reference count in the SSL_SESSION returned.
16636 *Geoff Thorpe <geoff@eu.c2.net>*
16638 * Fix for 'req': it was adding a null to request attributes.
16639 Also change the X509_LOOKUP and X509_INFO code to handle
16640 certificate auxiliary information.
16644 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16649 * Add the possibility to add extra information to the memory leak
16650 detecting output, to form tracebacks, showing from where each
16651 allocation was originated: CRYPTO_push_info("constant string") adds
16652 the string plus current file name and line number to a per-thread
16653 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16654 is like calling CYRPTO_pop_info() until the stack is empty.
16655 Also updated memory leak detection code to be multi-thread-safe.
16659 * Add options -text and -noout to pkcs7 utility and delete the
16660 encryption options which never did anything. Update docs.
16664 * Add options to some of the utilities to allow the pass phrase
16665 to be included on either the command line (not recommended on
16666 OSes like Unix) or read from the environment. Update the
16667 manpages and fix a few bugs.
16671 * Add a few manpages for some of the openssl commands.
16675 * Fix the -revoke option in ca. It was freeing up memory twice,
16676 leaking and not finding already revoked certificates.
16680 * Extensive changes to support certificate auxiliary information.
16681 This involves the use of X509_CERT_AUX structure and X509_AUX
16682 functions. An X509_AUX function such as PEM_read_X509_AUX()
16683 can still read in a certificate file in the usual way but it
16684 will also read in any additional "auxiliary information". By
16685 doing things this way a fair degree of compatibility can be
16686 retained: existing certificates can have this information added
16687 using the new 'x509' options.
16689 Current auxiliary information includes an "alias" and some trust
16690 settings. The trust settings will ultimately be used in enhanced
16691 certificate chain verification routines: currently a certificate
16692 can only be trusted if it is self signed and then it is trusted
16697 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16698 The problem was that one of the replacement routines had not been working
16699 since SSLeay releases. For now the offending routine has been replaced
16700 with non-optimised assembler. Even so, this now gives around 95%
16701 performance improvement for 1024 bit RSA signs.
16705 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16706 handling. Most clients have the effective key size in bits equal to
16707 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16708 A few however don't do this and instead use the size of the decrypted key
16709 to determine the RC2 key length and the AlgorithmIdentifier to determine
16710 the effective key length. In this case the effective key length can still
16711 be 40 bits but the key length can be 168 bits for example. This is fixed
16712 by manually forcing an RC2 key into the EVP_PKEY structure because the
16713 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16714 the key length and effective key length are equal.
16718 * Add a bunch of functions that should simplify the creation of
16719 X509_NAME structures. Now you should be able to do:
16720 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16721 and have it automatically work out the correct field type and fill in
16722 the structures. The more adventurous can try:
16723 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16724 and it will (hopefully) work out the correct multibyte encoding.
16728 * Change the 'req' utility to use the new field handling and multibyte
16729 copy routines. Before the DN field creation was handled in an ad hoc
16730 way in req, ca, and x509 which was rather broken and didn't support
16731 BMPStrings or UTF8Strings. Since some software doesn't implement
16732 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16733 using the dirstring_type option. See the new comment in the default
16734 openssl.cnf for more info.
16738 * Make crypto/rand/md_rand.c more robust:
16739 - Assure unique random numbers after fork().
16740 - Make sure that concurrent threads access the global counter and
16741 md serializably so that we never lose entropy in them
16742 or use exactly the same state in multiple threads.
16743 Access to the large state is not always serializable because
16744 the additional locking could be a performance killer, and
16745 md should be large enough anyway.
16749 * New file `apps/app_rand.c` with commonly needed functionality
16750 for handling the random seed file.
16752 Use the random seed file in some applications that previously did not:
16754 dsaparam -genkey (which also ignored its '-rand' option),
16757 x509 (when signing).
16758 Except on systems with /dev/urandom, it is crucial to have a random
16759 seed file at least for key creation, DSA signing, and for DH exchanges;
16760 for RSA signatures we could do without one.
16762 gendh and gendsa (unlike genrsa) used to read only the first byte
16763 of each file listed in the '-rand' option. The function as previously
16764 found in genrsa is now in app_rand.c and is used by all programs
16765 that support '-rand'.
16769 * In RAND_write_file, use mode 0600 for creating files;
16770 don't just chmod when it may be too late.
16774 * Report an error from X509_STORE_load_locations
16775 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16779 * New function ASN1_mbstring_copy() this copies a string in either
16780 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16781 into an ASN1_STRING type. A mask of permissible types is passed
16782 and it chooses the "minimal" type to use or an error if not type
16787 * Add function equivalents to the various macros in asn1.h. The old
16788 macros are retained with an `M_` prefix. Code inside the library can
16789 use the `M_` macros. External code (including the openssl utility)
16790 should *NOT* in order to be "shared library friendly".
16794 * Add various functions that can check a certificate's extensions
16795 to see if it usable for various purposes such as SSL client,
16796 server or S/MIME and CAs of these types. This is currently
16797 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16798 verification. Also added a -purpose flag to x509 utility to
16799 print out all the purposes.
16803 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16808 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16809 for, obtain and decode and extension and obtain its critical flag.
16810 This allows all the necessary extension code to be handled in a
16811 single function call.
16815 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16816 platforms. See crypto/rc4/rc4_enc.c for further details.
16820 * New -noout option to asn1parse. This causes no output to be produced
16821 its main use is when combined with -strparse and -out to extract data
16822 from a file (which may not be in ASN.1 format).
16826 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16827 when producing the local key id.
16829 *Richard Levitte <levitte@stacken.kth.se>*
16831 * New option -dhparam in s_server. This allows a DH parameter file to be
16832 stated explicitly. If it is not stated then it tries the first server
16833 certificate file. The previous behaviour hard coded the filename
16838 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16839 a public key to be input or output. For example:
16840 openssl rsa -in key.pem -pubout -out pubkey.pem
16841 Also added necessary DSA public key functions to handle this.
16845 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16846 in the message. This was handled by allowing
16847 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16849 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16851 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16852 to the end of the strings whereas this didn't. This would cause problems
16853 if strings read with d2i_ASN1_bytes() were later modified.
16855 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16857 * Fix for base64 decode bug. When a base64 bio reads only one line of
16858 data and it contains EOF it will end up returning an error. This is
16859 caused by input 46 bytes long. The cause is due to the way base64
16860 BIOs find the start of base64 encoded data. They do this by trying a
16861 trial decode on each line until they find one that works. When they
16862 do a flag is set and it starts again knowing it can pass all the
16863 data directly through the decoder. Unfortunately it doesn't reset
16864 the context it uses. This means that if EOF is reached an attempt
16865 is made to pass two EOFs through the context and this causes the
16866 resulting error. This can also cause other problems as well. As is
16867 usual with these problems it takes *ages* to find and the fix is
16868 trivial: move one line.
16870 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16872 * Ugly workaround to get s_client and s_server working under Windows. The
16873 old code wouldn't work because it needed to select() on sockets and the
16874 tty (for keypresses and to see if data could be written). Win32 only
16875 supports select() on sockets so we select() with a 1s timeout on the
16876 sockets and then see if any characters are waiting to be read, if none
16877 are present then we retry, we also assume we can always write data to
16878 the tty. This isn't nice because the code then blocks until we've
16879 received a complete line of data and it is effectively polling the
16880 keyboard at 1s intervals: however it's quite a bit better than not
16881 working at all :-) A dedicated Windows application might handle this
16882 with an event loop for example.
16886 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16887 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16888 will be called when RSA_sign() and RSA_verify() are used. This is useful
16889 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16890 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16891 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16892 This necessitated the support of an extra signature type NID_md5_sha1
16893 for SSL signatures and modifications to the SSL library to use it instead
16894 of calling RSA_public_decrypt() and RSA_private_encrypt().
16898 * Add new -verify -CAfile and -CApath options to the crl program, these
16899 will lookup a CRL issuers certificate and verify the signature in a
16900 similar way to the verify program. Tidy up the crl program so it
16901 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16902 less strict. It will now permit CRL extensions even if it is not
16903 a V2 CRL: this will allow it to tolerate some broken CRLs.
16907 * Initialize all non-automatic variables each time one of the openssl
16908 sub-programs is started (this is necessary as they may be started
16909 multiple times from the "OpenSSL>" prompt).
16911 *Lennart Bang, Bodo Moeller*
16913 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16914 removing all other RSA functionality (this is what NO_RSA does). This
16915 is so (for example) those in the US can disable those operations covered
16916 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16921 * Non-copying interface to BIO pairs.
16922 (still largely untested)
16926 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16927 ASCII string. This was handled independently in various places before.
16931 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16932 UTF8 strings a character at a time.
16936 * Use client_version from client hello to select the protocol
16937 (s23_srvr.c) and for RSA client key exchange verification
16938 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16942 * Add various utility functions to handle SPKACs, these were previously
16943 handled by poking round in the structure internals. Added new function
16944 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16945 print, verify and generate SPKACs. Based on an original idea from
16946 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16950 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16954 * Allow the config file extension section to be overwritten on the
16955 command line. Based on an original idea from Massimiliano Pala
16956 <madwolf@comune.modena.it>. The new option is called -extensions
16957 and can be applied to ca, req and x509. Also -reqexts to override
16958 the request extensions in req and -crlexts to override the crl extensions
16963 * Add new feature to the SPKAC handling in ca. Now you can include
16964 the same field multiple times by preceding it by "XXXX." for example:
16967 this is the same syntax as used in the req config file.
16971 * Allow certificate extensions to be added to certificate requests. These
16972 are specified in a 'req_extensions' option of the req section of the
16973 config file. They can be printed out with the -text option to req but
16974 are otherwise ignored at present.
16978 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16979 data read consists of only the final block it would not decrypted because
16980 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16981 A misplaced 'break' also meant the decrypted final block might not be
16982 copied until the next read.
16986 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16987 a few extra parameters to the DH structure: these will be useful if
16988 for example we want the value of 'q' or implement X9.42 DH.
16992 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16993 provides hooks that allow the default DSA functions or functions on a
16994 "per key" basis to be replaced. This allows hardware acceleration and
16995 hardware key storage to be handled without major modification to the
16996 library. Also added low-level modexp hooks and CRYPTO_EX structure and
16997 associated functions.
17001 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17002 as "read only": it can't be written to and the buffer it points to will
17003 not be freed. Reading from a read only BIO is much more efficient than
17004 a normal memory BIO. This was added because there are several times when
17005 an area of memory needs to be read from a BIO. The previous method was
17006 to create a memory BIO and write the data to it, this results in two
17007 copies of the data and an O(n^2) reading algorithm. There is a new
17008 function BIO_new_mem_buf() which creates a read only memory BIO from
17009 an area of memory. Also modified the PKCS#7 routines to use read only
17014 * Bugfix: ssl23_get_client_hello did not work properly when called in
17015 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17016 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17017 but a retry condition occurred while trying to read the rest.
17021 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17022 NID_pkcs7_encrypted by default: this was wrong since this should almost
17023 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17024 the encrypted data type: this is a more sensible place to put it and it
17025 allows the PKCS#12 code to be tidied up that duplicated this
17030 * Changed obj_dat.pl script so it takes its input and output files on
17031 the command line. This should avoid shell escape redirection problems
17036 * Initial support for certificate extension requests, these are included
17037 in things like Xenroll certificate requests. Included functions to allow
17038 extensions to be obtained and added.
17042 * -crlf option to s_client and s_server for sending newlines as
17043 CRLF (as required by many protocols).
17047 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17049 * Install libRSAglue.a when OpenSSL is built with RSAref.
17051 *Ralf S. Engelschall*
17053 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17055 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17057 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17062 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17063 DH parameters/keys (q is lost during that conversion, but the resulting
17064 DH parameters contain its length).
17066 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17067 much faster than DH_generate_parameters (which creates parameters
17068 where `p = 2*q + 1`), and also the smaller q makes DH computations
17069 much more efficient (160-bit exponentiation instead of 1024-bit
17070 exponentiation); so this provides a convenient way to support DHE
17071 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17072 utter importance to use
17073 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17075 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17076 when such DH parameters are used, because otherwise small subgroup
17077 attacks may become possible!
17081 * Avoid memory leak in i2d_DHparams.
17085 * Allow the -k option to be used more than once in the enc program:
17086 this allows the same encrypted message to be read by multiple recipients.
17090 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17091 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17092 it will always use the numerical form of the OID, even if it has a short
17097 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17098 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17099 otherwise bn_mod_exp was called. In the case of hardware keys for example
17100 no private key components need be present and it might store extra data
17101 in the RSA structure, which cannot be accessed from bn_mod_exp.
17102 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17103 private key operations.
17107 * Added support for SPARC Linux.
17111 * pem_password_cb function type incompatibly changed from
17112 typedef int pem_password_cb(char *buf, int size, int rwflag);
17114 ....(char *buf, int size, int rwflag, void *userdata);
17115 so that applications can pass data to their callbacks:
17116 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17117 additional void * argument, which is just handed through whenever
17118 the password callback is called.
17120 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17122 New function SSL_CTX_set_default_passwd_cb_userdata.
17124 Compatibility note: As many C implementations push function arguments
17125 onto the stack in reverse order, the new library version is likely to
17126 interoperate with programs that have been compiled with the old
17127 pem_password_cb definition (PEM_whatever takes some data that
17128 happens to be on the stack as its last argument, and the callback
17129 just ignores this garbage); but there is no guarantee whatsoever that
17132 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17133 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17134 problems not only on Windows, but also on some Unix platforms.
17135 To avoid problematic command lines, these definitions are now in an
17136 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17137 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17141 * MIPS III/IV assembler module is reimplemented.
17145 * More DES library cleanups: remove references to srand/rand and
17146 delete an unused file.
17150 * Add support for the free Netwide assembler (NASM) under Win32,
17151 since not many people have MASM (ml) and it can be hard to obtain.
17152 This is currently experimental but it seems to work OK and pass all
17153 the tests. Check out INSTALL.W32 for info.
17157 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17158 without temporary keys kept an extra copy of the server key,
17159 and connections with temporary keys did not free everything in case
17164 * New function RSA_check_key and new openssl rsa option -check
17165 for verifying the consistency of RSA keys.
17167 *Ulf Moeller, Bodo Moeller*
17169 * Various changes to make Win32 compile work:
17170 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17171 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17172 comparison" warnings.
17173 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17177 * Add a debugging option to PKCS#5 v2 key generation function: when
17178 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17179 derived keys are printed to stderr.
17183 * Copy the flags in ASN1_STRING_dup().
17185 *Roman E. Pavlov <pre@mo.msk.ru>*
17187 * The x509 application mishandled signing requests containing DSA
17188 keys when the signing key was also DSA and the parameters didn't match.
17190 It was supposed to omit the parameters when they matched the signing key:
17191 the verifying software was then supposed to automatically use the CA's
17192 parameters if they were absent from the end user certificate.
17194 Omitting parameters is no longer recommended. The test was also
17195 the wrong way round! This was probably due to unusual behaviour in
17196 EVP_cmp_parameters() which returns 1 if the parameters match.
17197 This meant that parameters were omitted when they *didn't* match and
17198 the certificate was useless. Certificates signed with 'ca' didn't have
17201 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17203 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17204 The interface is as follows:
17205 Applications can use
17206 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17207 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17208 "off" is now the default.
17209 The library internally uses
17210 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17211 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17212 to disable memory-checking temporarily.
17214 Some inconsistent states that previously were possible (and were
17215 even the default) are now avoided.
17217 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17218 with each memory chunk allocated; this is occasionally more helpful
17219 than just having a counter.
17221 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17223 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17228 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17229 which largely parallels "options", but is for changing API behaviour,
17230 whereas "options" are about protocol behaviour.
17231 Initial "mode" flags are:
17233 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17234 a single record has been written.
17235 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17236 retries use the same buffer location.
17237 (But all of the contents must be
17242 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17245 * Fix problems with no-hmac etc.
17247 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17249 * New functions RSA_get_default_method(), RSA_set_method() and
17250 RSA_get_method(). These allows replacement of RSA_METHODs without having
17251 to mess around with the internals of an RSA structure.
17255 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17256 Also really enable memory leak checks in openssl.c and in some
17259 *Chad C. Mulligan, Bodo Moeller*
17261 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17262 up the length of negative integers. This has now been simplified to just
17263 store the length when it is first determined and use it later, rather
17264 than trying to keep track of where data is copied and updating it to
17266 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17268 * Add a new function PKCS7_signatureVerify. This allows the verification
17269 of a PKCS#7 signature but with the signing certificate passed to the
17270 function itself. This contrasts with PKCS7_dataVerify which assumes the
17271 certificate is present in the PKCS#7 structure. This isn't always the
17272 case: certificates can be omitted from a PKCS#7 structure and be
17273 distributed by "out of band" means (such as a certificate database).
17277 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17278 function prototypes in pem.h, also change util/mkdef.pl to add the
17279 necessary function names.
17283 * mk1mf.pl (used by Windows builds) did not properly read the
17284 options set by Configure in the top level Makefile, and Configure
17285 was not even able to write more than one option correctly.
17286 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17290 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17291 file to be loaded from a BIO or FILE pointer. The BIO version will
17292 for example allow memory BIOs to contain config info.
17296 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17297 Whoever hopes to achieve shared-library compatibility across versions
17298 must use this, not the compile-time macro.
17299 (Exercise 0.9.4: Which is the minimum library version required by
17301 Note: All this applies only to multi-threaded programs, others don't
17306 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17307 through a BIO pair triggered the default case, i.e.
17308 SSLerr(...,SSL_R_UNKNOWN_STATE).
17312 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17313 can use the SSL library even if none of the specific BIOs is
17318 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17319 for the encoded length.
17321 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17323 * Add initial documentation of the X509V3 functions.
17327 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17328 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17329 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17330 secure PKCS#8 private key format with a high iteration count.
17334 * Fix determination of Perl interpreter: A perl or perl5
17335 *directory* in $PATH was also accepted as the interpreter.
17337 *Ralf S. Engelschall*
17339 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17340 wrong with it but it was very old and did things like calling
17341 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17342 unusual formatting.
17346 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17347 to use the new extension code.
17351 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17352 with macros. This should make it easier to change their form, add extra
17353 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17358 * Add to configuration table a new entry that can specify an alternative
17359 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17360 according to Mark Crispin <MRC@Panda.COM>.
17364 * DES CBC did not update the IV. Weird.
17368 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17369 Changing the behaviour of the former might break existing programs --
17370 where IV updating is needed, des_ncbc_encrypt can be used.
17373 * When bntest is run from "make test" it drives bc to check its
17374 calculations, as well as internally checking them. If an internal check
17375 fails, it needs to cause bc to give a non-zero result or make test carries
17376 on without noticing the failure. Fixed.
17380 * DES library cleanups.
17384 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17385 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17386 ciphers. NOTE: although the key derivation function has been verified
17387 against some published test vectors it has not been extensively tested
17388 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17393 * Instead of "mkdir -p", which is not fully portable, use new
17394 Perl script "util/mkdir-p.pl".
17398 * Rewrite the way password based encryption (PBE) is handled. It used to
17399 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17400 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17401 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17402 the 'parameter' field of the AlgorithmIdentifier is passed to the
17403 underlying key generation function so it must do its own ASN1 parsing.
17404 This has also changed the EVP_PBE_CipherInit() function which now has a
17405 'parameter' argument instead of literal salt and iteration count values
17406 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17410 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17411 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17412 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17413 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17414 value was just used as a "magic string" and not used directly its
17415 value doesn't matter.
17419 * Introduce some semblance of const correctness to BN. Shame C doesn't
17424 * "linux-sparc64" configuration (ultrapenguin).
17426 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17427 "linux-sparc" configuration.
17429 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17431 * config now generates no-xxx options for missing ciphers.
17435 * Support the EBCDIC character set (work in progress).
17436 File ebcdic.c not yet included because it has a different license.
17438 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17440 * Support BS2000/OSD-POSIX.
17442 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17444 * Make callbacks for key generation use `void *` instead of `char *`.
17448 * Make S/MIME samples compile (not yet tested).
17452 * Additional typesafe stacks.
17456 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17460 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17462 * New configuration variant "sco5-gcc".
17464 * Updated some demos.
17466 *Sean O Riordain, Wade Scholine*
17468 * Add missing BIO_free at exit of pkcs12 application.
17472 * Fix memory leak in conf.c.
17476 * Updates for Win32 to assembler version of MD5.
17480 * Set #! path to perl in `apps/der_chop` to where we found it
17481 instead of using a fixed path.
17485 * SHA library changes for irix64-mips4-cc.
17489 * Improvements for VMS support.
17493 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17495 * Bignum library bug fix. IRIX 6 passes "make test" now!
17496 This also avoids the problems with SC4.2 and unpatched SC5.
17498 *Andy Polyakov <appro@fy.chalmers.se>*
17500 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17501 These are required because of the typesafe stack would otherwise break
17502 existing code. If old code used a structure member which used to be STACK
17503 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17504 sk_num or sk_value it would produce an error because the num, data members
17505 are not present in STACK_OF. Now it just produces a warning. sk_set
17506 replaces the old method of assigning a value to sk_value
17507 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17508 that does this will no longer work (and should use sk_set instead) but
17509 this could be regarded as a "questionable" behaviour anyway.
17513 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17514 correctly handle encrypted S/MIME data.
17518 * Change type of various DES function arguments from des_cblock
17519 (which means, in function argument declarations, pointer to char)
17520 to des_cblock * (meaning pointer to array with 8 char elements),
17521 which allows the compiler to do more typechecking; it was like
17522 that back in SSLeay, but with lots of ugly casts.
17524 Introduce new type const_des_cblock.
17528 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17529 problems: find RecipientInfo structure that matches recipient certificate
17530 and initialise the ASN1 structures properly based on passed cipher.
17534 * Belatedly make the BN tests actually check the results.
17538 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17539 to and from BNs: it was completely broken. New compilation option
17540 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17541 key elements as negative integers.
17545 * Reorganize and speed up MD5.
17547 *Andy Polyakov <appro@fy.chalmers.se>*
17551 *Richard Levitte <richard@levitte.org>*
17553 * New option -out to asn1parse to allow the parsed structure to be
17554 output to a file. This is most useful when combined with the -strparse
17555 option to examine the output of things like OCTET STRINGS.
17559 * Make SSL library a little more fool-proof by not requiring any longer
17560 that `SSL_set_{accept,connect}_state` be called before
17561 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17562 in many applications because usually everything *appeared* to work as
17563 intended anyway -- now it really works as intended).
17567 * Move openssl.cnf out of lib/.
17571 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17572 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17573 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17575 *Ralf S. Engelschall*
17577 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17578 handle PKCS#7 enveloped data properly.
17580 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17582 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17583 copying pointers. The cert_st handling is changed by this in
17584 various ways (and thus what used to be known as ctx->default_cert
17585 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17586 any longer when s->cert does not give us what we need).
17587 ssl_cert_instantiate becomes obsolete by this change.
17588 As soon as we've got the new code right (possibly it already is?),
17589 we have solved a couple of bugs of the earlier code where s->cert
17590 was used as if it could not have been shared with other SSL structures.
17592 Note that using the SSL API in certain dirty ways now will result
17593 in different behaviour than observed with earlier library versions:
17594 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17595 does not influence s as it used to.
17597 In order to clean up things more thoroughly, inside SSL_SESSION
17598 we don't use CERT any longer, but a new structure SESS_CERT
17599 that holds per-session data (if available); currently, this is
17600 the peer's certificate chain and, for clients, the server's certificate
17601 and temporary key. CERT holds only those values that can have
17602 meaningful defaults in an SSL_CTX.
17606 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17607 from the internal representation. Various PKCS#7 fixes: remove some
17608 evil casts and set the enc_dig_alg field properly based on the signing
17613 * Allow PKCS#12 password to be set from the command line or the
17614 environment. Let 'ca' get its config file name from the environment
17615 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17620 * Allow certificate policies extension to use an IA5STRING for the
17621 organization field. This is contrary to the PKIX definition but
17622 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17627 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17628 without disallowing inline assembler and the like for non-pedantic builds.
17632 * Support Borland C++ builder.
17634 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17640 * SHA-1 cleanups and performance enhancements.
17642 *Andy Polyakov <appro@fy.chalmers.se>*
17644 * Sparc v8plus assembler for the bignum library.
17646 *Andy Polyakov <appro@fy.chalmers.se>*
17648 * Accept any -xxx and +xxx compiler options in Configure.
17652 * Update HPUX configuration.
17656 * Add missing `sk_<type>_unshift()` function to safestack.h
17658 *Ralf S. Engelschall*
17660 * New function SSL_CTX_use_certificate_chain_file that sets the
17661 "extra_cert"s in addition to the certificate. (This makes sense
17662 only for "PEM" format files, as chains as a whole are not
17667 * Support verify_depth from the SSL API.
17668 x509_vfy.c had what can be considered an off-by-one-error:
17669 Its depth (which was not part of the external interface)
17670 was actually counting the number of certificates in a chain;
17671 now it really counts the depth.
17675 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17676 instead of X509err, which often resulted in confusing error
17677 messages since the error codes are not globally unique
17678 (e.g. an alleged error in ssl3_accept when a certificate
17679 didn't match the private key).
17681 * New function SSL_CTX_set_session_id_context that allows to set a default
17682 value (so that you don't need SSL_set_session_id_context for each
17683 connection using the SSL_CTX).
17687 * OAEP decoding bug fix.
17691 * Support INSTALL_PREFIX for package builders, as proposed by
17696 * New Configure options "threads" and "no-threads". For systems
17697 where the proper compiler options are known (currently Solaris
17698 and Linux), "threads" is the default.
17702 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17706 * Install various scripts to $(OPENSSLDIR)/misc, not to
17707 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17708 such as /usr/local/bin.
17712 * "make linux-shared" to build shared libraries.
17714 *Niels Poppe <niels@netbox.org>*
17716 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17720 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17721 extension adding in x509 utility.
17725 * Remove NOPROTO sections and error code comments.
17729 * Partial rewrite of the DEF file generator to now parse the ANSI
17734 * New Configure options --prefix=DIR and --openssldir=DIR.
17738 * Complete rewrite of the error code script(s). It is all now handled
17739 by one script at the top level which handles error code gathering,
17740 header rewriting and C source file generation. It should be much better
17741 than the old method: it now uses a modified version of Ulf's parser to
17742 read the ANSI prototypes in all header files (thus the old K&R definitions
17743 aren't needed for error creation any more) and do a better job of
17744 translating function codes into names. The old 'ASN1 error code embedded
17745 in a comment' is no longer necessary and it doesn't use .err files which
17746 have now been deleted. Also the error code call doesn't have to appear all
17747 on one line (which resulted in some large lines...).
17751 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17755 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17756 0 (which usually indicates a closed connection), but continue reading.
17760 * Fix some race conditions.
17764 * Add support for CRL distribution points extension. Add Certificate
17765 Policies and CRL distribution points documentation.
17769 * Move the autogenerated header file parts to crypto/opensslconf.h.
17773 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17774 8 of keying material. Merlin has also confirmed interop with this fix
17775 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17777 *Merlin Hughes <merlin@baltimore.ie>*
17779 * Fix lots of warnings.
17781 *Richard Levitte <levitte@stacken.kth.se>*
17783 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17784 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17786 *Richard Levitte <levitte@stacken.kth.se>*
17788 * Fix problems with sizeof(long) == 8.
17790 *Andy Polyakov <appro@fy.chalmers.se>*
17792 * Change functions to ANSI C.
17796 * Fix typos in error codes.
17798 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17800 * Remove defunct assembler files from Configure.
17804 * SPARC v8 assembler BIGNUM implementation.
17806 *Andy Polyakov <appro@fy.chalmers.se>*
17808 * Support for Certificate Policies extension: both print and set.
17809 Various additions to support the r2i method this uses.
17813 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17814 return a const string when you are expecting an allocated buffer.
17818 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17819 types DirectoryString and DisplayText.
17823 * Add code to allow r2i extensions to access the configuration database,
17824 add an LHASH database driver and add several ctx helper functions.
17828 * Fix an evil bug in bn_expand2() which caused various BN functions to
17829 fail when they extended the size of a BIGNUM.
17833 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17834 support typesafe stack.
17838 * Fix typo in SSL_[gs]et_options().
17840 *Nils Frostberg <nils@medcom.se>*
17842 * Delete various functions and files that belonged to the (now obsolete)
17843 old X509V3 handling code.
17847 * New Configure option "rsaref".
17851 * Don't auto-generate pem.h.
17855 * Introduce type-safe ASN.1 SETs.
17859 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17861 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17863 * Introduce type-safe STACKs. This will almost certainly break lots of code
17864 that links with OpenSSL (well at least cause lots of warnings), but fear
17865 not: the conversion is trivial, and it eliminates loads of evil casts. A
17866 few STACKed things have been converted already. Feel free to convert more.
17867 In the fullness of time, I'll do away with the STACK type altogether.
17871 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17872 specified in `<certfile>` by updating the entry in the index.txt file.
17873 This way one no longer has to edit the index.txt file manually for
17874 revoking a certificate. The -revoke option does the gory details now.
17876 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17878 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17879 `-text` option at all and this way the `-noout -text` combination was
17880 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17882 *Ralf S. Engelschall*
17884 * Make sure a corresponding plain text error message exists for the
17885 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17886 verify callback function determined that a certificate was revoked.
17888 *Ralf S. Engelschall*
17890 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17891 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17892 all available ciphers including rc5, which was forgotten until now.
17893 In order to let the testing shell script know which algorithms
17894 are available, a new (up to now undocumented) command
17895 `openssl list-cipher-commands` is used.
17899 * Bugfix: s_client occasionally would sleep in select() when
17900 it should have checked SSL_pending() first.
17904 * New functions DSA_do_sign and DSA_do_verify to provide access to
17905 the raw DSA values prior to ASN.1 encoding.
17909 * Tweaks to Configure
17911 *Niels Poppe <niels@netbox.org>*
17913 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17918 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17922 * New config option to avoid instructions that are illegal on the 80386.
17923 The default code is faster, but requires at least a 486.
17927 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17928 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17929 same as SSL2_VERSION anyway.
17933 * New "-showcerts" option for s_client.
17937 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17938 application. Various cleanups and fixes.
17942 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17943 modify error routines to work internally. Add error codes and PBE init
17944 to library startup routines.
17948 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17949 packing functions to asn1 and evp. Changed function names and error
17950 codes along the way.
17954 * PKCS12 integration: and so it begins... First of several patches to
17955 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17956 objects to objects.h
17960 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17961 and display support for Thawte strong extranet extension.
17965 * Add LinuxPPC support.
17967 *Jeff Dubrule <igor@pobox.org>*
17969 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17970 bn_div_words in alpha.s.
17972 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17974 * Make sure the RSA OAEP test is skipped under -DRSAref because
17975 OAEP isn't supported when OpenSSL is built with RSAref.
17977 *Ulf Moeller <ulf@fitug.de>*
17979 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17980 so they no longer are missing under -DNOPROTO.
17982 *Soren S. Jorvang <soren@t.dk>*
17984 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17986 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17987 doesn't work when the session is reused. Coming soon!
17991 * Fix a security hole, that allows sessions to be reused in the wrong
17992 context thus bypassing client cert protection! All software that uses
17993 client certs and session caches in multiple contexts NEEDS PATCHING to
17994 allow session reuse! A fuller solution is in the works.
17996 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17998 * Some more source tree cleanups (removed obsolete files
17999 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18000 permission on "config" script to be executable) and a fix for the INSTALL
18003 *Ulf Moeller <ulf@fitug.de>*
18005 * Remove some legacy and erroneous uses of malloc, free instead of
18008 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18010 * Make rsa_oaep_test return non-zero on error.
18012 *Ulf Moeller <ulf@fitug.de>*
18014 * Add support for native Solaris shared libraries. Configure
18015 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18016 if someone would make that last step automatic.
18018 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18020 * ctx_size was not built with the right compiler during "make links". Fixed.
18024 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18025 except NULL ciphers". This means the default cipher list will no longer
18026 enable NULL ciphers. They need to be specifically enabled e.g. with
18027 the string "DEFAULT:eNULL".
18031 * Fix to RSA private encryption routines: if p < q then it would
18032 occasionally produce an invalid result. This will only happen with
18033 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18037 * Be less restrictive and allow also `perl util/perlpath.pl
18038 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18039 because this way one can also use an interpreter named `perl5` (which is
18040 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18041 installed as `perl`).
18043 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18045 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18047 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18049 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18050 advapi32.lib to Win32 build and change the pem test comparison
18051 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18052 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18053 and crypto/des/ede_cbcm_enc.c.
18057 * DES quad checksum was broken on big-endian architectures. Fixed.
18061 * Comment out two functions in bio.h that aren't implemented. Fix up the
18062 Win32 test batch file so it (might) work again. The Win32 test batch file
18063 is horrible: I feel ill....
18067 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18068 in e_os.h. Audit of header files to check ANSI and non ANSI
18069 sections: 10 functions were absent from non ANSI section and not exported
18070 from Windows DLLs. Fixed up libeay.num for new functions.
18074 * Make `openssl version` output lines consistent.
18076 *Ralf S. Engelschall*
18078 * Fix Win32 symbol export lists for BIO functions: Added
18079 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18080 to ms/libeay{16,32}.def.
18082 *Ralf S. Engelschall*
18084 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18085 fine under Unix and passes some trivial tests I've now added. But the
18086 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18087 added to make sure no one expects that this stuff really works in the
18088 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18089 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18092 *Ralf S. Engelschall*
18094 * Fix the generation of two part addresses in perl.
18096 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18098 * Add config entry for Linux on MIPS.
18100 *John Tobey <jtobey@channel1.com>*
18102 * Make links whenever Configure is run, unless we are on Windoze.
18106 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18107 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18112 * Add a useful kludge to allow package maintainers to specify compiler and
18113 other platforms details on the command line without having to patch the
18114 Configure script every time: One now can use
18115 `perl Configure <id>:<details>`,
18116 i.e. platform ids are allowed to have details appended
18117 to them (separated by colons). This is treated as there would be a static
18118 pre-configured entry in Configure's %table under key `<id>` with value
18119 `<details>` and `perl Configure <id>` is called. So, when you want to
18120 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18121 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18122 now, which overrides the FreeBSD-elf entry on-the-fly.
18124 *Ralf S. Engelschall*
18126 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18130 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18131 on the `perl Configure ...` command line. This way one can compile
18132 OpenSSL libraries with Position Independent Code (PIC) which is needed
18133 for linking it into DSOs.
18135 *Ralf S. Engelschall*
18137 * Remarkably, export ciphers were totally broken and no-one had noticed!
18142 * Cleaned up the LICENSE document: The official contact for any license
18143 questions now is the OpenSSL core team under openssl-core@openssl.org.
18144 And add a paragraph about the dual-license situation to make sure people
18145 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18146 to the OpenSSL toolkit.
18148 *Ralf S. Engelschall*
18150 * General source tree makefile cleanups: Made `making xxx in yyy...`
18151 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18152 Additionally cleaned up the `make links` target: Remove unnecessary
18153 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18154 to speed processing and no longer clutter the display with confusing
18155 stuff. Instead only the actually done links are displayed.
18157 *Ralf S. Engelschall*
18159 * Permit null encryption ciphersuites, used for authentication only. It used
18160 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18161 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18166 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18167 signed attributes when verifying signatures (this would break them),
18168 the detached data encoding was wrong and public keys obtained using
18169 X509_get_pubkey() weren't freed.
18173 * Add text documentation for the BUFFER functions. Also added a work around
18174 to a Win95 console bug. This was triggered by the password read stuff: the
18175 last character typed gets carried over to the next fread(). If you were
18176 generating a new cert request using 'req' for example then the last
18177 character of the passphrase would be CR which would then enter the first
18182 * Added the new 'Includes OpenSSL Cryptography Software' button as
18183 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18184 button and can be used by applications based on OpenSSL to show the
18185 relationship to the OpenSSL project.
18187 *Ralf S. Engelschall*
18189 * Remove confusing variables in function signatures in files
18190 ssl/ssl_lib.c and ssl/ssl.h.
18192 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18194 * Don't install bss_file.c under PREFIX/include/
18196 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18198 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18199 functions that return function pointers and has support for NT specific
18200 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18201 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18202 unsigned to signed types: this was killing the Win32 compile.
18206 * Add new certificate file to stack functions,
18207 SSL_add_dir_cert_subjects_to_stack() and
18208 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18209 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18210 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18211 This means that Apache-SSL and similar packages don't have to mess around
18212 to add as many CAs as they want to the preferred list.
18216 * Experiment with doxygen documentation. Currently only partially applied to
18218 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18219 openssl.doxy as the configuration file.
18223 * Get rid of remaining C++-style comments which strict C compilers hate.
18225 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18227 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18228 compiled in by default: it has problems with large keys.
18232 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18233 DH private keys and/or callback functions which directly correspond to
18234 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18235 is needed for applications which have to configure certificates on a
18236 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18238 For the RSA certificate situation is makes no difference, but
18239 for the DSA certificate situation this fixes the "no shared cipher"
18240 problem where the OpenSSL cipher selection procedure failed because the
18241 temporary keys were not overtaken from the context and the API provided
18242 no way to reconfigure them.
18243 The new functions now let applications reconfigure the stuff and they
18244 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18245 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18246 non-public-API function ssl_cert_instantiate() is used as a helper
18247 function and also to reduce code redundancy inside ssl_rsa.c.
18249 *Ralf S. Engelschall*
18251 * Move s_server -dcert and -dkey options out of the undocumented feature
18252 area because they are useful for the DSA situation and should be
18253 recognized by the users.
18255 *Ralf S. Engelschall*
18257 * Fix the cipher decision scheme for export ciphers: the export bits are
18258 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18259 SSL_EXP_MASK. So, the original variable has to be used instead of the
18260 already masked variable.
18262 *Richard Levitte <levitte@stacken.kth.se>*
18264 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18266 *Richard Levitte <levitte@stacken.kth.se>*
18268 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18269 from `int` to `unsigned int` because it is a length and initialized by
18270 EVP_DigestFinal() which expects an `unsigned int *`.
18272 *Richard Levitte <levitte@stacken.kth.se>*
18274 * Don't hard-code path to Perl interpreter on shebang line of Configure
18275 script. Instead use the usual Shell->Perl transition trick.
18277 *Ralf S. Engelschall*
18279 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18280 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18281 -noout -modulus` as it's already the case for `openssl rsa -noout
18282 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18283 currently the public key is printed (a decision which was already done by
18284 `openssl dsa -modulus` in the past) which serves a similar purpose.
18285 Additionally the NO_RSA no longer completely removes the whole -modulus
18286 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18289 *Ralf S. Engelschall*
18291 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18292 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18294 *Arne Ansper <arne@ats.cyber.ee>*
18296 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18297 to be added. Now both 'req' and 'ca' can use new objects defined in the
18302 * Add cool BIO that does syslog (or event log on NT).
18304 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18306 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18307 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18308 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18309 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18313 * Add preliminary config info for new extension code.
18317 * Make RSA_NO_PADDING really use no padding.
18319 *Ulf Moeller <ulf@fitug.de>*
18321 * Generate errors when private/public key check is done.
18325 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18326 for some CRL extensions and new objects added.
18330 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18331 key usage extension and fuller support for authority key id.
18335 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18336 padding method for RSA, which is recommended for new applications in PKCS
18337 #1 v2.0 (RFC 2437, October 1998).
18338 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18339 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18340 against Bleichbacher's attack on RSA.
18341 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18344 * Updates to the new SSL compression code
18346 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18348 * Fix so that the version number in the master secret, when passed
18349 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18350 (because the server will not accept higher), that the version number
18351 is 0x03,0x01, not 0x03,0x00
18353 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18355 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18356 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18357 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18361 * Support for RAW extensions where an arbitrary extension can be
18362 created by including its DER encoding. See `apps/openssl.cnf` for
18367 * Make sure latest Perl versions don't interpret some generated C array
18368 code as Perl array code in the crypto/err/err_genc.pl script.
18370 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18372 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18373 not many people have the assembler. Various Win32 compilation fixes and
18374 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18375 build instructions.
18379 * Modify configure script 'Configure' to automatically create crypto/date.h
18380 file under Win32 and also build pem.h from pem.org. New script
18381 util/mkfiles.pl to create the MINFO file on environments that can't do a
18382 'make files': perl util/mkfiles.pl >MINFO should work.
18386 * Major rework of DES function declarations, in the pursuit of correctness
18387 and purity. As a result, many evil casts evaporated, and some weirdness,
18388 too. You may find this causes warnings in your code. Zapping your evil
18389 casts will probably fix them. Mostly.
18393 * Fix for a typo in asn1.h. Bug fix to object creation script
18394 obj_dat.pl. It considered a zero in an object definition to mean
18395 "end of object": none of the objects in objects.h have any zeros
18396 so it wasn't spotted.
18398 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18400 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18401 Masking (CBCM). In the absence of test vectors, the best I have been able
18402 to do is check that the decrypt undoes the encrypt, so far. Send me test
18403 vectors if you have them.
18407 * Correct calculation of key length for export ciphers (too much space was
18408 allocated for null ciphers). This has not been tested!
18412 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18413 message is now correct (it understands "crypto" and "ssl" on its
18414 command line). There is also now an "update" option. This will update
18415 the util/ssleay.num and util/libeay.num files with any new functions.
18417 perl util/mkdef.pl crypto ssl update
18418 it will update them.
18422 * Overhauled the Perl interface:
18423 - ported BN stuff to OpenSSL's different BN library
18424 - made the perl/ source tree CVS-aware
18425 - renamed the package from SSLeay to OpenSSL (the files still contain
18426 their history because I've copied them in the repository)
18427 - removed obsolete files (the test scripts will be replaced
18428 by better Test::Harness variants in the future)
18430 *Ralf S. Engelschall*
18432 * First cut for a very conservative source tree cleanup:
18433 1. merge various obsolete readme texts into doc/ssleay.txt
18434 where we collect the old documents and readme texts.
18435 2. remove the first part of files where I'm already sure that we no
18436 longer need them because of three reasons: either they are just temporary
18437 files which were left by Eric or they are preserved original files where
18438 I've verified that the diff is also available in the CVS via "cvs diff
18439 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18440 the crypto/md/ stuff).
18442 *Ralf S. Engelschall*
18444 * More extension code. Incomplete support for subject and issuer alt
18445 name, issuer and authority key id. Change the i2v function parameters
18446 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18447 what that's for :-) Fix to ASN1 macro which messed up
18448 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18452 * Preliminary support for ENUMERATED type. This is largely copied from the
18457 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18459 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18461 * Make sure `make rehash` target really finds the `openssl` program.
18463 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18465 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18466 like to hear about it if this slows down other processors.
18470 * Add CygWin32 platform information to Configure script.
18472 *Alan Batie <batie@aahz.jf.intel.com>*
18474 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18476 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18478 * New program nseq to manipulate netscape certificate sequences
18482 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18487 * Fixes to BN code. Previously the default was to define BN_RECURSION
18488 but the BN code had some problems that would cause failures when
18489 doing certificate verification and some other functions.
18491 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18493 * Add ASN1 and PEM code to support netscape certificate sequences.
18497 * Add ASN1 and PEM code to support netscape certificate sequences.
18501 * Add several PKIX and private extended key usage OIDs.
18505 * Modify the 'ca' program to handle the new extension code. Modify
18506 openssl.cnf for new extension format, add comments.
18510 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18511 and add a sample to openssl.cnf so req -x509 now adds appropriate
18516 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18517 error code, add initial support to X509_print() and x509 application.
18521 * Takes a deep breath and start adding X509 V3 extension support code. Add
18522 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18523 stuff is currently isolated and isn't even compiled yet.
18527 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18528 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18529 Removed the versions check from X509 routines when loading extensions:
18530 this allows certain broken certificates that don't set the version
18531 properly to be processed.
18535 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18536 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18537 can still be regenerated with "make depend".
18541 * Spelling mistake in C version of CAST-128.
18543 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18545 * Changes to the error generation code. The perl script err-code.pl
18546 now reads in the old error codes and retains the old numbers, only
18547 adding new ones if necessary. It also only changes the .err files if new
18548 codes are added. The makefiles have been modified to only insert errors
18549 when needed (to avoid needlessly modifying header files). This is done
18550 by only inserting errors if the .err file is newer than the auto generated
18551 C file. To rebuild all the error codes from scratch (the old behaviour)
18552 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18553 or delete all the .err files.
18557 * CAST-128 was incorrectly implemented for short keys. The C version has
18558 been fixed, but is untested. The assembler versions are also fixed, but
18559 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18560 to regenerate it if needed.
18561 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18562 Hagino <itojun@kame.net>*
18564 * File was opened incorrectly in randfile.c.
18566 *Ulf Möller <ulf@fitug.de>*
18568 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18569 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18570 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18571 al: it's just almost always a UTCTime. Note this patch adds new error
18572 codes so do a "make errors" if there are problems.
18576 * Correct Linux 1 recognition in config.
18578 *Ulf Möller <ulf@fitug.de>*
18580 * Remove pointless MD5 hash when using DSA keys in ca.
18582 *Anonymous <nobody@replay.com>*
18584 * Generate an error if given an empty string as a cert directory. Also
18585 generate an error if handed NULL (previously returned 0 to indicate an
18586 error, but didn't set one).
18588 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18590 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18594 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18595 parameters. This was causing a warning which killed off the Win32 compile.
18599 * Remove C++ style comments from crypto/bn/bn_local.h.
18601 *Neil Costigan <neil.costigan@celocom.com>*
18603 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18604 based on a text string, looking up short and long names and finally
18605 "dot" format. The "dot" format stuff didn't work. Added new function
18606 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18607 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18608 OID is not part of the table.
18612 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18613 X509_LOOKUP_by_alias().
18617 * Sort openssl functions by name.
18621 * Get the `gendsa` command working and add it to the `list` command. Remove
18622 encryption from sample DSA keys (in case anyone is interested the password
18627 * Make *all* `*_free` functions accept a NULL pointer.
18629 *Frans Heymans <fheymans@isaserver.be>*
18631 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18634 *Anonymous <nobody@replay.com>*
18636 * s_server should send the CAfile as acceptable CAs, not its own cert.
18638 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18640 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18642 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18644 * Temp key "for export" tests were wrong in s3_srvr.c.
18646 *Anonymous <nobody@replay.com>*
18648 * Add prototype for temp key callback functions
18649 SSL_CTX_set_tmp_{rsa,dh}_callback().
18653 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18654 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18658 * X509_name_add_entry() freed the wrong thing after an error.
18660 *Arne Ansper <arne@ats.cyber.ee>*
18662 * rsa_eay.c would attempt to free a NULL context.
18664 *Arne Ansper <arne@ats.cyber.ee>*
18666 * BIO_s_socket() had a broken should_retry() on Windoze.
18668 *Arne Ansper <arne@ats.cyber.ee>*
18670 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18672 *Arne Ansper <arne@ats.cyber.ee>*
18674 * Make sure the already existing X509_STORE->depth variable is initialized
18675 in X509_STORE_new(), but document the fact that this variable is still
18676 unused in the certificate verification process.
18678 *Ralf S. Engelschall*
18680 * Fix the various library and `apps/` files to free up pkeys obtained from
18681 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18685 * Fix reference counting in X509_PUBKEY_get(). This makes
18686 demos/maurice/example2.c work, amongst others, probably.
18688 *Steve Henson and Ben Laurie*
18690 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18691 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18692 are no longer created. This way we have a single and consistent command
18693 line interface `openssl <command>`, similar to `cvs <command>`.
18695 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18697 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18698 BIT STRING wrapper always have zero unused bits.
18702 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18706 * Make the top-level INSTALL documentation easier to understand.
18710 * Makefiles updated to exit if an error occurs in a sub-directory
18711 make (including if user presses ^C) [Paul Sutton]
18713 * Make Montgomery context stuff explicit in RSA data structure.
18717 * Fix build order of pem and err to allow for generated pem.h.
18721 * Fix renumbering bug in X509_NAME_delete_entry().
18725 * Enhanced the err-ins.pl script so it makes the error library number
18726 global and can add a library name. This is needed for external ASN1 and
18727 other error libraries.
18731 * Fixed sk_insert which never worked properly.
18735 * Fix ASN1 macros so they can handle indefinite length constructed
18736 EXPLICIT tags. Some non standard certificates use these: they can now
18741 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18742 into a single doc/ssleay.txt bundle. This way the information is still
18743 preserved but no longer messes up this directory. Now it's new room for
18744 the new set of documentation files.
18746 *Ralf S. Engelschall*
18748 * SETs were incorrectly DER encoded. This was a major pain, because they
18749 shared code with SEQUENCEs, which aren't coded the same. This means that
18750 almost everything to do with SETs or SEQUENCEs has either changed name or
18751 number of arguments.
18753 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18755 * Fix test data to work with the above.
18759 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18760 was already fixed by Eric for 0.9.1 it seems.
18762 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18764 * Autodetect FreeBSD3.
18768 * Fix various bugs in Configure. This affects the following platforms:
18772 unixware-2.0-pentium
18777 * Eliminate generated files from CVS. Reorder tests to regenerate files
18778 before they are needed.
18782 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18786 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18788 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18789 changed SSLeay to OpenSSL in version strings.
18791 *Ralf S. Engelschall*
18793 * Some fixups to the top-level documents.
18797 * Fixed the nasty bug where rsaref.h was not found under compile-time
18798 because the symlink to include/ was missing.
18800 *Ralf S. Engelschall*
18802 * Incorporated the popular no-RSA/DSA-only patches
18803 which allow to compile a RSA-free SSLeay.
18805 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18807 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18808 when "ssleay" is still not found.
18810 *Ralf S. Engelschall*
18812 * Added more platforms to Configure: Cray T3E, HPUX 11,
18814 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18816 * Updated the README file.
18818 *Ralf S. Engelschall*
18820 * Added various .cvsignore files in the CVS repository subdirs
18821 to make a "cvs update" really silent.
18823 *Ralf S. Engelschall*
18825 * Recompiled the error-definition header files and added
18826 missing symbols to the Win32 linker tables.
18828 *Ralf S. Engelschall*
18830 * Cleaned up the top-level documents;
18831 o new files: CHANGES and LICENSE
18832 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18833 o merged COPYRIGHT into LICENSE
18834 o removed obsolete TODO file
18835 o renamed MICROSOFT to INSTALL.W32
18837 *Ralf S. Engelschall*
18839 * Removed dummy files from the 0.9.1b source tree:
18840 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18841 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18842 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18843 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18844 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18846 *Ralf S. Engelschall*
18848 * Added various platform portability fixes.
18852 * The Genesis of the OpenSSL rpject:
18853 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18854 Young and Tim J. Hudson created while they were working for C2Net until
18857 *The OpenSSL Project*
18859 ### Changes between 0.9.0b and 0.9.1b [not released]
18861 * Updated a few CA certificates under certs/
18865 * Changed some BIGNUM api stuff.
18869 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18870 DGUX x86, Linux Alpha, etc.
18874 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18875 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18880 * Add -strparse option to asn1pars program which parses nested
18883 *Dr Stephen Henson <shenson@bigfoot.com>*
18885 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18889 * DSA fix for "ca" program.
18893 * Added "-genkey" option to "dsaparam" program.
18897 * Added RIPE MD160 (rmd160) message digest.
18901 * Added -a (all) option to "ssleay version" command.
18905 * Added PLATFORM define which is the id given to Configure.
18909 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18913 * Extended the ASN.1 parser routines.
18917 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18921 * Added a BN_CTX to the BN library.
18925 * Fixed the weak key values in DES library
18929 * Changed API in EVP library for cipher aliases.
18933 * Added support for RC2/64bit cipher.
18937 * Converted the lhash library to the crypto/mem.c functions.
18941 * Added more recognized ASN.1 object ids.
18945 * Added more RSA padding checks for SSL/TLS.
18949 * Added BIO proxy/filter functionality.
18953 * Added extra_certs to SSL_CTX which can be used
18954 send extra CA certificates to the client in the CA cert chain sending
18955 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18959 * Now Fortezza is denied in the authentication phase because
18960 this is key exchange mechanism is not supported by SSLeay at all.
18964 * Additional PKCS1 checks.
18968 * Support the string "TLSv1" for all TLS v1 ciphers.
18972 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18973 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18977 * Fixed a few memory leaks.
18981 * Fixed various code and comment typos.
18985 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18986 bytes sent in the client random.
18988 *Edward Bishop <ebishop@spyglass.com>*
18992 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18993 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18994 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18995 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18996 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18997 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18998 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18999 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19000 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19001 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19002 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19003 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19004 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19005 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19006 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19007 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19008 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19009 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19010 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19011 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19012 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19013 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19014 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19015 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19016 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19017 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19018 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19019 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19020 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19021 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19022 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19023 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19024 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19025 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19026 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19027 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19028 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19029 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19030 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19031 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19032 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19033 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19034 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19035 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19036 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19037 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19038 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19039 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19040 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19041 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19042 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19043 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19044 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19045 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19046 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19047 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19048 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19049 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19050 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19051 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19052 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19053 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19054 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19055 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19056 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19057 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19058 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19059 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19060 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19061 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19062 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19063 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19064 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19065 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19066 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19067 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19068 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19069 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19070 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19071 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19072 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19073 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19074 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19075 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19076 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19077 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19078 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19079 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19080 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19081 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19082 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19083 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19084 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19085 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19086 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19087 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19088 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19089 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19090 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19091 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19092 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19093 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19094 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19095 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19096 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19097 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19098 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19099 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19100 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19101 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19102 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19103 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19104 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19105 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19106 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19107 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19108 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19109 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19110 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19111 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19112 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19113 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19114 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19115 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19116 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19117 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19118 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19119 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19120 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19121 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19122 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19123 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19124 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19125 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19126 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19127 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19128 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19129 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19130 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19131 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19132 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19133 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19134 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19135 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19136 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19137 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19138 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19139 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19140 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19141 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19142 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19143 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19144 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19145 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19146 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19147 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19148 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19149 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19150 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19151 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19152 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19153 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655