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 * Added a new BIO_s_dgram_mem() to read/write datagrams to memory
31 * Add a mac salt length option for the pkcs12 command.
35 * Add more SRTP protection profiles from RFC8723 and RFC8269.
39 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
41 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
43 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
44 supported and enabled.
48 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
49 to the list of ciphersuites providing Perfect Forward Secrecy as
50 required by SECLEVEL >= 3.
52 *Dmitry Belyavskiy, Nicola Tuveri*
54 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
55 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
56 SSL_get0_iana_groups() function-like macro, retrieves the list of
57 supported groups sent by the peer.
58 The function SSL_client_hello_get_extension_order() populates
59 a caller-supplied array with the list of extension types present in the
60 ClientHello, in order of appearance.
64 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
65 to make it possible to use empty passphrase strings.
69 * RNDR and RNDRRS support in provider functions to provide
70 random number generation for Arm CPUs (aarch64).
74 * s_client and s_server apps now explicitly say when the TLS version
75 does not include the renegotiation mechanism. This avoids confusion
76 between that scenario versus when the TLS version includes secure
77 renegotiation but the peer lacks support for it.
81 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
83 *Tomasz Kantecki, Andrey Matyukov*
85 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
86 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
87 of 160 bits and above and less than 224 bits were previously accepted by
88 default but are now no longer allowed. By default TLS compression was
89 already disabled in previous OpenSSL versions. At security level 2 it cannot
94 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
99 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
100 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
101 will need to load the legacy crypto provider.
105 * The various OBJ_* functions have been made thread safe.
109 * CCM8 cipher suites in TLS have been downgraded to security level zero
110 because they use a short authentication tag which lowers their strength.
114 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
119 * Parallel dual-prime 1536/2048-bit modular exponentiation for
120 AVX512_IFMA capable processors.
122 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
124 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
125 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
126 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
127 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
128 `OPENSSL_NO_DEPRECATED_3_1`.
130 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
131 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
132 definitions for these functions regardless of whether
133 `OPENSSL_NO_DEPRECATED_3_1` is defined.
135 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
136 functions regardless of whether they are using them. It is recommended that
137 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
141 * When generating safe-prime DH parameters set the recommended private key
142 length equivalent to minimum key lengths as in RFC 7919.
146 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
147 such as adding a trace facility for debugging certificate chain building.
151 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
152 in particular supporting requests for central key generation, generalized
153 polling, and various types of genm/genp exchanges defined in CMP Updates.
157 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
158 like correcting the TLS and proxy support and adding tracing for debugging.
162 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
166 * Fixed and extended `util/check-format.pl` for checking adherence to the
167 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
168 The checks are meanwhile more complete and yield fewer false positives.
175 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
176 listed here are only a brief description.
177 The migration guide contains more detailed information related to new features,
178 breaking changes, and mappings for the large list of deprecated functions.
180 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
182 ### Changes between 3.0.3 and 3.0.4 [21 June 2022]
184 * In addition to the c_rehash shell command injection identified in
185 CVE-2022-1292, further bugs where the c_rehash script does not
186 properly sanitise shell metacharacters to prevent command injection have been
189 When the CVE-2022-1292 was fixed it was not discovered that there
190 are other places in the script where the file names of certificates
191 being hashed were possibly passed to a command executed through the shell.
193 This script is distributed by some operating systems in a manner where
194 it is automatically executed. On such operating systems, an attacker
195 could execute arbitrary commands with the privileges of the script.
197 Use of the c_rehash script is considered obsolete and should be replaced
198 by the OpenSSL rehash command line tool.
201 *Daniel Fiala, Tomáš Mráz*
203 * Case insensitive string comparison no longer uses locales. It has instead
204 been directly implemented.
208 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
210 * Case insensitive string comparison is reimplemented via new locale-agnostic
211 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
212 comparison. The previous implementation had problems when the Turkish locale
217 * Fixed a bug in the c_rehash script which was not properly sanitising shell
218 metacharacters to prevent command injection. This script is distributed by
219 some operating systems in a manner where it is automatically executed. On
220 such operating systems, an attacker could execute arbitrary commands with the
221 privileges of the script.
223 Use of the c_rehash script is considered obsolete and should be replaced
224 by the OpenSSL rehash command line tool.
229 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
230 certificate on an OCSP response. The bug caused the function in the case
231 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
232 response (meaning a successful verification) even in the case where the
233 response signing certificate fails to verify.
235 It is anticipated that most users of `OCSP_basic_verify` will not use the
236 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
237 a negative value (indicating a fatal error) in the case of a certificate
238 verification failure. The normal expected return value in this case would be
241 This issue also impacts the command line OpenSSL "ocsp" application. When
242 verifying an ocsp response with the "-no_cert_checks" option the command line
243 application will report that the verification is successful even though it
244 has in fact failed. In this case the incorrect successful response will also
245 be accompanied by error messages showing the failure and contradicting the
246 apparently successful result.
251 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
252 AAD data as the MAC key. This made the MAC key trivially predictable.
254 An attacker could exploit this issue by performing a man-in-the-middle attack
255 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
256 that the modified data would still pass the MAC integrity check.
258 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
259 endpoint will always be rejected by the recipient and the connection will
260 fail at that point. Many application protocols require data to be sent from
261 the client to the server first. Therefore, in such a case, only an OpenSSL
262 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
264 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
265 sent in both directions. In this case both clients and servers could be
266 affected, regardless of the application protocol.
268 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
269 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
270 the handshake when using this ciphersuite.
272 The confidentiality of data is not impacted by this issue, i.e. an attacker
273 cannot decrypt data that has been encrypted using this ciphersuite - they can
276 In order for this attack to work both endpoints must legitimately negotiate
277 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
278 OpenSSL 3.0, and is not available within the default provider or the default
279 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
280 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
281 following must have occurred:
283 1) OpenSSL must have been compiled with the (non-default) compile time option
284 enable-weak-ssl-ciphers
286 2) OpenSSL must have had the legacy provider explicitly loaded (either
287 through application code or via configuration)
289 3) The ciphersuite must have been explicitly added to the ciphersuite list
291 4) The libssl security level must have been set to 0 (default is 1)
293 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
295 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
296 others that both endpoints have in common
301 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
302 occuppied by the removed hash table entries.
304 This function is used when decoding certificates or keys. If a long lived
305 process periodically decodes certificates or keys its memory usage will
306 expand without bounds and the process might be terminated by the operating
307 system causing a denial of service. Also traversing the empty hash table
308 entries will take increasingly more time.
310 Typically such long lived processes might be TLS clients or TLS servers
311 configured to accept client certificate authentication.
314 *Hugo Landau, Aliaksei Levin*
316 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
317 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
318 statistics are no longer supported. For compatibility, these statistics are
319 still listed in the output but are now always reported as zero.
323 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
325 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
326 for non-prime moduli.
328 Internally this function is used when parsing certificates that contain
329 elliptic curve public keys in compressed form or explicit elliptic curve
330 parameters with a base point encoded in compressed form.
332 It is possible to trigger the infinite loop by crafting a certificate that
333 has invalid explicit curve parameters.
335 Since certificate parsing happens prior to verification of the certificate
336 signature, any process that parses an externally supplied certificate may thus
337 be subject to a denial of service attack. The infinite loop can also be
338 reached when parsing crafted private keys as they can contain explicit
339 elliptic curve parameters.
341 Thus vulnerable situations include:
343 - TLS clients consuming server certificates
344 - TLS servers consuming client certificates
345 - Hosting providers taking certificates or private keys from customers
346 - Certificate authorities parsing certification requests from subscribers
347 - Anything else which parses ASN.1 elliptic curve parameters
349 Also any other applications that use the BN_mod_sqrt() where the attacker
350 can control the parameter values are vulnerable to this DoS issue.
355 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
356 to the list of ciphersuites providing Perfect Forward Secrecy as
357 required by SECLEVEL >= 3.
359 *Dmitry Belyavskiy, Nicola Tuveri*
361 * Made the AES constant time code for no-asm configurations
362 optional due to the resulting 95% performance degradation.
363 The AES constant time code can be enabled, for no assembly
364 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
368 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
373 * The negative return value handling of the certificate verification callback
374 was reverted. The replacement is to set the verification retry state with
375 the SSL_set_retry_verify() function.
379 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
381 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
382 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
383 verify a certificate supplied by a server. That function may return a
384 negative return value to indicate an internal error (for example out of
385 memory). Such a negative return value is mishandled by OpenSSL and will cause
386 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
387 success and a subsequent call to SSL_get_error() to return the value
388 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
389 returned by OpenSSL if the application has previously called
390 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
391 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
392 totally unexpected and applications may not behave correctly as a result. The
393 exact behaviour will depend on the application but it could result in
394 crashes, infinite loops or other similar incorrect responses.
396 This issue is made more serious in combination with a separate bug in OpenSSL
397 3.0 that will cause X509_verify_cert() to indicate an internal error when
398 processing a certificate chain. This will occur where a certificate does not
399 include the Subject Alternative Name extension but where a Certificate
400 Authority has enforced name constraints. This issue can occur even with valid
406 * Corrected a few file name and file reference bugs in the build,
407 installation and setup scripts, which lead to installation verification
408 failures. Slightly enhanced the installation verification script.
412 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
417 * Fixed PVK encoder to properly query for the passphrase.
421 * Multiple fixes in the OSSL_HTTP API functions.
425 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
426 OSSL_PARAM_INTEGER data type and return error on negative numbers
427 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
428 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
432 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
436 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
440 * Multiple threading fixes.
444 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
448 * Allow fetching an operation from the provider that owns an unexportable key
449 as a fallback if that is still allowed by the property query.
453 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
455 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
460 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
461 S390X capability vector to zero. This simplifies testing of different code
462 paths on S390X architecture.
466 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
467 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
468 SP 800-38D". The communication will fail at this point.
472 * The EC_GROUP_clear_free() function is deprecated as there is nothing
473 confidential in EC_GROUP data.
477 * The byte order mark (BOM) character is ignored if encountered at the
478 beginning of a PEM-formatted file.
482 * Added CMS support for the Russian GOST algorithms.
486 * Due to move of the implementation of cryptographic operations
487 to the providers, validation of various operation parameters can
488 be postponed until the actual operation is executed where previously
489 it happened immediately when an operation parameter was set.
491 For example when setting an unsupported curve with
492 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
493 fail but later keygen operations with the EVP_PKEY_CTX will fail.
495 *OpenSSL team members and many third party contributors*
497 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
498 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
499 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
500 instead to retrieve these algorithms from a provider.
504 * On build targets where the multilib postfix is set in the build
505 configuration the libdir directory was changing based on whether
506 the lib directory with the multilib postfix exists on the system
507 or not. This unpredictable behavior was removed and eventual
508 multilib postfix is now always added to the default libdir. Use
509 `--libdir=lib` to override the libdir if adding the postfix is
514 * The triple DES key wrap functionality now conforms to RFC 3217 but is
515 no longer interoperable with OpenSSL 1.1.1.
519 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
520 function codes, this function can only cause problems for calling
525 * Add a configurable flag to output date formats as ISO 8601. Does not
526 change the default date format.
530 * Version of MSVC earlier than 1300 could get link warnings, which could
531 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
532 Support for this flag has been removed.
536 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
537 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
538 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
539 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
540 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
544 * The signatures of the functions to get and set options on SSL and
545 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
546 Some source code changes may be required.
550 * The public definitions of conf_method_st and conf_st have been
551 deprecated. They will be made opaque in a future release.
553 *Rich Salz and Tomáš Mráz*
555 * Client-initiated renegotiation is disabled by default. To allow it, use
556 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
557 flag, or the "ClientRenegotiation" config parameter as appropriate.
561 * Add "abspath" and "includedir" pragma's to config files, to prevent,
562 or modify relative pathname inclusion.
566 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
567 validated. Please consult the README-FIPS and
568 README-PROVIDERS files, as well as the migration guide.
570 *OpenSSL team members and many third party contributors*
572 * For the key types DH and DHX the allowed settable parameters are now different.
576 * The openssl commands that read keys, certificates, and CRLs now
577 automatically detect the PEM or DER format of the input files.
579 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
581 * Added enhanced PKCS#12 APIs which accept a library context.
585 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
589 * Added support for Kernel TLS (KTLS).
591 *Boris Pismenny, John Baldwin and Andrew Gallatin*
593 * Support for RFC 5746 secure renegotiation is now required by default for
594 SSL or TLS connections to succeed.
598 * The signature of the `copy` functional parameter of the
599 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
600 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
601 the signature of the `pub_decode` functional parameter of the
602 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
603 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
607 * The error return values from some control calls (ctrl) have changed.
611 * A public key check is now performed during EVP_PKEY_derive_set_peer().
615 * Many functions in the EVP_ namespace that are getters of values from
616 implementations or contexts were renamed to include get or get0 in their
617 names. Old names are provided as macro aliases for compatibility and
622 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
623 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
624 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
629 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
632 * The output from the command line applications may have minor
637 * The output from numerous "printing" may have minor changes.
641 * Windows thread synchronization uses read/write primitives (SRWLock) when
642 supported by the OS, otherwise CriticalSection continues to be used.
646 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
647 work on read only BIO source/sinks that do not support these functions.
648 This allows piping or redirection of a file BIO using stdin to be buffered
649 into memory. This is used internally in OSSL_DECODER_from_bio().
653 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
654 this function would return one of the values OSSL_STORE_INFO_NAME,
655 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
656 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
657 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
658 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
659 using this function should be amended to handle the changed return value.
663 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
664 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
665 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
666 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
667 contains more than one certificate identifier: This means that all
668 certificates referenced there MUST be part of the validation chain.
672 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
673 RC5, DESX and DES have been moved to the legacy provider.
677 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
678 RIPEMD-160 have been moved to the legacy provider.
682 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
687 * The deprecated functions EVP_PKEY_get0_RSA(),
688 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
689 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
690 well as the similarly named "get1" functions behave differently in
695 * A number of functions handling low-level keys or engines were deprecated
696 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
697 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
698 EVP_PKEY_get0_siphash().
702 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
703 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
704 will need to load the legacy crypto provider. This includes these PBE
705 algorithms which use this KDF:
706 - NID_pbeWithMD2AndDES_CBC
707 - NID_pbeWithMD5AndDES_CBC
708 - NID_pbeWithSHA1AndRC2_CBC
709 - NID_pbeWithMD2AndRC2_CBC
710 - NID_pbeWithMD5AndRC2_CBC
711 - NID_pbeWithSHA1AndDES_CBC
715 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
716 BIO_debug_callback() functions.
720 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
721 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
725 * The RAND_METHOD APIs have been deprecated.
729 * The SRP APIs have been deprecated.
733 * Add a compile time option to prevent the caching of provider fetched
734 algorithms. This is enabled by including the no-cached-fetch option
735 at configuration time.
739 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
740 count of PKCS12_DEFAULT_ITER.
742 *Tomáš Mráz and Sahana Prasad*
744 * The openssl speed command does not use low-level API calls anymore.
748 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
751 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
753 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
757 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
758 providers may supply their own group implementations (using either the "key
759 exchange" or the "key encapsulation" methods) which will automatically be
760 detected and used by libssl.
762 *Matt Caswell, Nicola Tuveri*
764 * The undocumented function X509_certificate_type() has been deprecated;
768 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
772 * Removed RSA padding mode for SSLv23 (which was only used for
773 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
774 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
779 * Deprecated the obsolete X9.31 RSA key generation related functions.
781 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
782 is not allowed to return a value > 1, this is no more taken as failure.
784 *Viktor Dukhovni and David von Oheimb*
786 * Deprecated the obsolete X9.31 RSA key generation related functions
787 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
788 BN_X931_generate_prime_ex().
792 * The default key generation method for the regular 2-prime RSA keys was
793 changed to the FIPS 186-4 B.3.6 method.
797 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
801 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
805 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
806 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
808 *Rich Salz, Richard Levitte, and David von Oheimb*
810 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
814 * Deprecated `OCSP_parse_url()`.
818 * Validation of SM2 keys has been separated from the validation of regular EC
823 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
824 switches: a validation failure triggers an early exit, returning a failure
825 exit status to the parent process.
829 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
830 to ignore unknown ciphers.
834 * The `-cipher-commands` and `-digest-commands` options
835 of the command line utility `list` have been deprecated.
836 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
840 * Added convenience functions for generating asymmetric key pairs:
841 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
842 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
846 * All of the low level EC_KEY functions have been deprecated.
848 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
850 * Deprecated all the libcrypto and libssl error string loading
855 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
856 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
861 * The `-crypt` option to the `passwd` command line tool has been removed.
865 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
870 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
874 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
875 EVP_PKEY_get1_tls_encodedpoint().
879 * The security callback, which can be customised by application code, supports
880 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
881 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
885 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
886 interface. Their functionality remains unchanged.
890 * Added new option for 'openssl list', '-providers', which will display the
891 list of loaded providers, their names, version and status. It optionally
892 displays their gettable parameters.
896 * Removed EVP_PKEY_set_alias_type().
900 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
901 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
905 * Changed all "STACK" functions to be macros instead of inline functions. Macro
906 parameters are still checked for type safety at compile time via helper
911 * Remove the RAND_DRBG API
913 *Paul Dale and Matthias St. Pierre*
915 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
916 as well as actual hostnames.
920 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
921 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
922 conversely, silently ignore DTLS protocol version bounds when configuring
923 TLS-based contexts. The commands can be repeated to set bounds of both
924 types. The same applies with the corresponding "min_protocol" and
925 "max_protocol" command-line switches, in case some application uses both TLS
928 SSL_CTX instances that are created for a fixed protocol version (e.g.
929 `TLSv1_server_method()`) also silently ignore version bounds. Previously
930 attempts to apply bounds to these protocol versions would result in an
931 error. Now only the "version-flexible" SSL_CTX instances are subject to
932 limits in configuration files in command-line options.
936 * Deprecated the `ENGINE` API. Engines should be replaced with providers
941 * Reworked the recorded ERR codes to make better space for system errors.
942 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
943 given code is a system error (true) or an OpenSSL error (false).
947 * Reworked the test perl framework to better allow parallel testing.
949 *Nicola Tuveri and David von Oheimb*
951 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
952 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
956 * 'Configure' has been changed to figure out the configuration target if
957 none is given on the command line. Consequently, the 'config' script is
958 now only a mere wrapper. All documentation is changed to only mention
961 *Rich Salz and Richard Levitte*
963 * Added a library context `OSSL_LIB_CTX` that applications as well as
964 other libraries can use to form a separate context within which
965 libcrypto operations are performed.
969 * Added various `_ex` functions to the OpenSSL API that support using
970 a non-default `OSSL_LIB_CTX`.
974 * Handshake now fails if Extended Master Secret extension is dropped
979 * Dropped interactive mode from the `openssl` program.
983 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
985 *David von Oheimb and Shane Lontis*
987 * Deprecated `EC_METHOD_get_field_type()`.
991 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
992 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
993 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
997 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1001 * Add CAdES-BES signature verification support, mostly derived
1002 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1004 *Filipe Raimundo da Silva*
1006 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1010 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1011 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1015 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1019 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1020 EC_KEY_precompute_mult().
1024 * Deprecated EC_POINTs_mul().
1028 * Removed FIPS_mode() and FIPS_mode_set().
1032 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1036 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1037 EC_POINT_get_Jprojective_coordinates_GFp().
1041 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1042 arrays to be more easily constructed via a series of utility functions.
1043 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1044 the various push functions and finally convert to a passable OSSL_PARAM
1045 array using OSSL_PARAM_BLD_to_param().
1049 * The security strength of SHA1 and MD5 based signatures in TLS has been
1054 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1055 contain a provider side internal key.
1059 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1063 * Project text documents not yet having a proper file name extension
1064 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1065 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1069 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1070 have been converted to Markdown with the goal to produce documents
1071 which not only look pretty when viewed online in the browser, but
1072 remain well readable inside a plain text editor.
1074 To achieve this goal, a 'minimalistic' Markdown style has been applied
1075 which avoids formatting elements that interfere too much with the
1076 reading flow in the text file. For example, it
1078 * avoids [ATX headings][] and uses [setext headings][] instead
1079 (which works for `<h1>` and `<h2>` headings only).
1080 * avoids [inline links][] and uses [reference links][] instead.
1081 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1083 [ATX headings]: https://github.github.com/gfm/#atx-headings
1084 [setext headings]: https://github.github.com/gfm/#setext-headings
1085 [inline links]: https://github.github.com/gfm/#inline-link
1086 [reference links]: https://github.github.com/gfm/#reference-link
1087 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1088 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1090 *Matthias St. Pierre*
1092 * The test suite is changed to preserve results of each test recipe.
1093 A new directory test-runs/ with subdirectories named like the
1094 test recipes are created in the build tree for this purpose.
1098 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1099 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1100 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1102 *David von Oheimb, Martin Peylo*
1104 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1105 It supports arbitrary request and response content types, GET redirection,
1106 TLS, connections via HTTP(S) proxies, connections and exchange via
1107 user-defined BIOs (allowing implicit connections), persistent connections,
1108 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1109 The legacy OCSP-focused (and only partly documented) API
1110 is retained for backward compatibility, while most of it is deprecated.
1114 * Added `util/check-format.pl`, a tool for checking adherence to the
1115 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1116 The checks performed are incomplete and yield some false positives.
1117 Still the tool should be useful for detecting most typical glitches.
1121 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1122 If domain name resolution yields multiple IP addresses all of them are tried
1123 after `connect()` failures.
1127 * All of the low level RSA functions have been deprecated.
1131 * X509 certificates signed using SHA1 are no longer allowed at security
1136 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1137 modified to use PKEY APIs. These commands are now in maintenance mode
1138 and no new features will be added to them.
1142 * The command line utility rsautl has been deprecated.
1146 * The command line utilities genrsa and rsa have been modified to use PKEY
1147 APIs. They now write PKCS#8 keys by default. These commands are now in
1148 maintenance mode and no new features will be added to them.
1152 * All of the low level DH functions have been deprecated.
1154 *Paul Dale and Matt Caswell*
1156 * All of the low level DSA functions have been deprecated.
1160 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1161 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1165 * Deprecated low level ECDH and ECDSA functions.
1169 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1173 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1174 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1175 a new formulation to include all the things it can be used for,
1176 as well as words of caution.
1180 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1184 * All of the low level HMAC functions have been deprecated.
1186 *Paul Dale and David von Oheimb*
1188 * Over two thousand fixes were made to the documentation, including:
1189 - Common options (such as -rand/-writerand, TLS version control, etc)
1190 were refactored and point to newly-enhanced descriptions in openssl.pod.
1191 - Added style conformance for all options (with help from Richard Levitte),
1192 documented all reported missing options, added a CI build to check
1193 that all options are documented and that no unimplemented options
1195 - Documented some internals, such as all use of environment variables.
1196 - Addressed all internal broken L<> references.
1200 * All of the low level CMAC functions have been deprecated.
1204 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1205 functions have been deprecated.
1207 *Paul Dale and David von Oheimb*
1209 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1210 set of functions. The documentation mentioned negative values for some
1211 errors, but this was never the case, so the mention of negative values
1214 Code that followed the documentation and thereby check with something
1215 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1219 * All of the low level cipher functions have been deprecated.
1221 *Matt Caswell and Paul Dale*
1223 * Removed include/openssl/opensslconf.h.in and replaced it with
1224 include/openssl/configuration.h.in, which differs in not including
1225 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1226 was added to include both.
1228 This allows internal hacks where one might need to modify the set
1229 of configured macros, for example this if deprecated symbols are
1230 still supposed to be available internally:
1232 #include <openssl/configuration.h>
1234 #undef OPENSSL_NO_DEPRECATED
1235 #define OPENSSL_SUPPRESS_DEPRECATED
1237 #include <openssl/macros.h>
1239 This should not be used by applications that use the exported
1240 symbols, as that will lead to linking errors.
1244 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1245 used in exponentiation with 512-bit moduli. No EC algorithms are
1246 affected. Analysis suggests that attacks against 2-prime RSA1024,
1247 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1248 difficult to perform and are not believed likely. Attacks against DH512
1249 are considered just feasible. However, for an attack the target would
1250 have to re-use the DH512 private key, which is not recommended anyway.
1251 Also applications directly using the low-level API BN_mod_exp may be
1252 affected if they use BN_FLG_CONSTTIME.
1257 * Most memory-debug features have been deprecated, and the functionality
1258 replaced with no-ops.
1262 * Added documentation for the STACK API.
1266 * Introduced a new method type and API, OSSL_ENCODER, to represent
1267 generic encoders. These do the same sort of job that PEM writers
1268 and d2i functions do, but with support for methods supplied by
1269 providers, and the possibility for providers to support other
1274 * Introduced a new method type and API, OSSL_DECODER, to represent
1275 generic decoders. These do the same sort of job that PEM readers
1276 and i2d functions do, but with support for methods supplied by
1277 providers, and the possibility for providers to support other
1282 * Added a .pragma directive to the syntax of configuration files, to
1283 allow varying behavior in a supported and predictable manner.
1284 Currently added pragma:
1288 This allows dollar signs to be a keyword character unless it's
1289 followed by a opening brace or parenthesis. This is useful for
1290 platforms where dollar signs are commonly used in names, such as
1291 volume names and system directory names on VMS.
1295 * Added functionality to create an EVP_PKEY from user data.
1299 * Change the interpretation of the '--api' configuration option to
1300 mean that this is a desired API compatibility level with no
1301 further meaning. The previous interpretation, that this would
1302 also mean to remove all deprecated symbols up to and including
1303 the given version, no requires that 'no-deprecated' is also used
1304 in the configuration.
1306 When building applications, the desired API compatibility level
1307 can be set with the OPENSSL_API_COMPAT macro like before. For
1308 API compatibility version below 3.0, the old style numerical
1309 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1310 For version 3.0 and on, the value is expected to be the decimal
1311 value calculated from the major and minor version like this:
1313 MAJOR * 10000 + MINOR * 100
1317 -DOPENSSL_API_COMPAT=30000 For 3.0
1318 -DOPENSSL_API_COMPAT=30200 For 3.2
1320 To hide declarations that are deprecated up to and including the
1321 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1322 given when building the application as well.
1326 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1327 access to certificate and CRL stores via URIs and OSSL_STORE
1330 This adds the following functions:
1332 - X509_LOOKUP_store()
1333 - X509_STORE_load_file()
1334 - X509_STORE_load_path()
1335 - X509_STORE_load_store()
1336 - SSL_add_store_cert_subjects_to_stack()
1337 - SSL_CTX_set_default_verify_store()
1338 - SSL_CTX_load_verify_file()
1339 - SSL_CTX_load_verify_dir()
1340 - SSL_CTX_load_verify_store()
1344 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1345 The presence of this system service is determined at run-time.
1349 * Added functionality to create an EVP_PKEY context based on data
1350 for methods from providers. This takes an algorithm name and a
1351 property query string and simply stores them, with the intent
1352 that any operation that uses this context will use those strings
1353 to fetch the needed methods implicitly, thereby making the port
1354 of application written for pre-3.0 OpenSSL easier.
1358 * The undocumented function NCONF_WIN32() has been deprecated; for
1359 conversion details see the HISTORY section of doc/man5/config.pod
1363 * Introduced the new functions EVP_DigestSignInit_ex() and
1364 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1365 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1366 pages for further details.
1370 * Over two thousand fixes were made to the documentation, including:
1371 adding missing command flags, better style conformance, documentation
1374 *Rich Salz, Richard Levitte*
1376 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1377 X25519, X448, Ed25519 and Ed448.
1381 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1386 * Deprecated the public definition of `ERR_STATE` as well as the function
1387 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1392 * Added ERR functionality to give callers access to the stored function
1393 names that have replaced the older function code based functions.
1395 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1396 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1397 ERR_peek_error_all() and ERR_peek_last_error_all().
1399 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1400 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1401 ERR_func_error_string().
1405 * Extended testing to be verbose for failing tests only. The make variables
1406 VERBOSE_FAILURE or VF can be used to enable this:
1408 $ make VF=1 test # Unix
1409 $ mms /macro=(VF=1) test ! OpenVMS
1410 $ nmake VF=1 test # Windows
1414 * Added the `-copy_extensions` option to the `x509` command for use with
1415 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1416 all extensions in the request are copied to the certificate or vice versa.
1418 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1420 * Added the `-copy_extensions` option to the `req` command for use with
1421 `-x509`. When given with the `copy` or `copyall` argument,
1422 all extensions in the certification request are copied to the certificate.
1426 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1427 they generate are by default RFC 5280 compliant in the following sense:
1428 There is a subjectKeyIdentifier extension with a hash value of the public key
1429 and for not self-signed certs there is an authorityKeyIdentifier extension
1430 with a keyIdentifier field or issuer information identifying the signing key.
1431 This is done unless some configuration overrides the new default behavior,
1432 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1436 * Added several checks to `X509_verify_cert()` according to requirements in
1437 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1438 (which may be done by using the CLI option `-x509_strict`):
1439 * The basicConstraints of CA certificates must be marked critical.
1440 * CA certificates must explicitly include the keyUsage extension.
1441 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1442 * The issuer name of any certificate must not be empty.
1443 * The subject name of CA certs, certs with keyUsage crlSign,
1444 and certs without subjectAlternativeName must not be empty.
1445 * If a subjectAlternativeName extension is given it must not be empty.
1446 * The signatureAlgorithm field and the cert signature must be consistent.
1447 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1448 must not be marked critical.
1449 * The authorityKeyIdentifier must be given for X.509v3 certs
1450 unless they are self-signed.
1451 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1455 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1456 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1460 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1461 used even when parsing explicit parameters, when loading a encoded key
1462 or calling `EC_GROUP_new_from_ecpkparameters()`/
1463 `EC_GROUP_new_from_ecparameters()`.
1464 This prevents bypass of security hardening and performance gains,
1465 especially for curves with specialized EC_METHODs.
1466 By default, if a key encoded with explicit parameters is loaded and later
1467 encoded, the output is still encoded with explicit parameters, even if
1468 internally a "named" EC_GROUP is used for computation.
1472 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1473 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1474 NULL. After this change, only the cofactor parameter can be NULL. It also
1475 does some minimal sanity checks on the passed order.
1480 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1481 An attack is simple, if the first CMS_recipientInfo is valid but the
1482 second CMS_recipientInfo is chosen ciphertext. If the second
1483 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1484 encryption key will be replaced by garbage, and the message cannot be
1485 decoded, but if the RSA decryption fails, the correct encryption key is
1486 used and the recipient will not notice the attack.
1487 As a work around for this potential attack the length of the decrypted
1488 key must be equal to the cipher default key length, in case the
1489 certifiate is not given and all recipientInfo are tried out.
1490 The old behaviour can be re-enabled in the CMS code by setting the
1491 CMS_DEBUG_DECRYPT flag.
1495 * Early start up entropy quality from the DEVRANDOM seed source has been
1496 improved for older Linux systems. The RAND subsystem will wait for
1497 /dev/random to be producing output before seeding from /dev/urandom.
1498 The seeded state is stored for future library initialisations using
1499 a system global shared memory segment. The shared memory identifier
1500 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1501 the desired value. The default identifier is 114.
1505 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1506 when primes for RSA keys are computed.
1507 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1508 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1509 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1510 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1511 This avoids possible fingerprinting of newly generated RSA modules.
1515 * Correct the extended master secret constant on EBCDIC systems. Without this
1516 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1517 negotiate EMS will fail. Unfortunately this also means that TLS connections
1518 between EBCDIC systems with this fix, and EBCDIC systems without this
1519 fix will fail if they negotiate EMS.
1523 * Changed the library initialisation so that the config file is now loaded
1524 by default. This was already the case for libssl. It now occurs for both
1525 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1526 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1530 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1531 where the former acts as a replacement for `ERR_put_error()`, and the
1532 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1533 `ERR_raise_data()` adds more flexibility by taking a format string and
1534 an arbitrary number of arguments following it, to be processed with
1539 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1540 to check if a named provider is loaded and available. When called, it
1541 will also activate all fallback providers if such are still present.
1545 * Enforce a minimum DH modulus size of 512 bits.
1549 * Changed DH parameters to generate the order q subgroup instead of 2q.
1550 Previously generated DH parameters are still accepted by DH_check
1551 but DH_generate_key works around that by clearing bit 0 of the
1552 private key for those. This avoids leaking bit 0 of the private key.
1556 * Significantly reduce secure memory usage by the randomness pools.
1560 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1565 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1566 algorithms. An implementation of a key exchange algorithm can be obtained
1567 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1568 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1569 the older EVP_PKEY_derive_init() function. See the man pages for the new
1570 functions for further details.
1574 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1578 * Removed the function names from error messages and deprecated the
1583 * Removed NextStep support and the macro OPENSSL_UNISTD
1587 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1588 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1589 Also removed "export var as function" capability; we do not export
1590 variables, only functions.
1594 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1595 an error and 1 indicating success. In previous versions of OpenSSL this
1596 was a void type. If a key was set longer than the maximum possible this
1601 * Support SM2 signing and verification schemes with X509 certificate.
1605 * Use SHA256 as the default digest for TS query in the `ts` app.
1609 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1613 * Default cipher lists/suites are now available via a function, the
1614 #defines are deprecated.
1618 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1619 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1620 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1624 * Join the directories crypto/x509 and crypto/x509v3
1628 * Added command 'openssl kdf' that uses the EVP_KDF API.
1632 * Added command 'openssl mac' that uses the EVP_MAC API.
1636 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1637 as default directories. Also added the command 'openssl info'
1638 for scripting purposes.
1642 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1647 * Add prediction resistance to the DRBG reseeding process.
1651 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1652 mandated by IEEE Std 1619-2018.
1656 * Added newline escaping functionality to a filename when using openssl dgst.
1657 This output format is to replicate the output format found in the `*sum`
1658 checksum programs. This aims to preserve backward compatibility.
1660 *Matt Eaton, Richard Levitte, and Paul Dale*
1662 * Removed the heartbeat message in DTLS feature, as it has very
1663 little usage and doesn't seem to fulfill a valuable purpose.
1664 The configuration option is now deprecated.
1668 * Changed the output of 'openssl {digestname} < file' to display the
1669 digest name in its output.
1673 * Added a new generic trace API which provides support for enabling
1674 instrumentation through trace output.
1676 *Richard Levitte & Matthias St. Pierre*
1678 * Added build tests for C++. These are generated files that only do one
1679 thing, to include one public OpenSSL head file each. This tests that
1680 the public header files can be usefully included in a C++ application.
1682 This test isn't enabled by default. It can be enabled with the option
1683 'enable-buildtest-c++'.
1687 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1691 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1695 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1699 * Added KMAC to EVP_MAC.
1703 * Added property based algorithm implementation selection framework to
1708 * Added SCA hardening for modular field inversion in EC_GROUP through
1709 a new dedicated field_inv() pointer in EC_METHOD.
1710 This also addresses a leakage affecting conversions from projective
1711 to affine coordinates.
1713 *Billy Bob Brumley, Nicola Tuveri*
1715 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1716 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1717 those algorithms that were already supported through the EVP_PKEY API
1718 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1719 and scrypt are now wrappers that call EVP_KDF.
1723 * Build devcrypto engine as a dynamic engine.
1725 *Eneas U de Queiroz*
1727 * Add keyed BLAKE2 to EVP_MAC.
1731 * Fix a bug in the computation of the endpoint-pair shared secret used
1732 by DTLS over SCTP. This breaks interoperability with older versions
1733 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1734 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1735 interoperability with such broken implementations. However, enabling
1736 this switch breaks interoperability with correct implementations.
1738 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1739 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1743 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1747 * Changed the license to the Apache License v2.0.
1751 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1753 - Major releases (indicated by incrementing the MAJOR release number)
1754 may introduce incompatible API/ABI changes.
1755 - Minor releases (indicated by incrementing the MINOR release number)
1756 may introduce new features but retain API/ABI compatibility.
1757 - Patch releases (indicated by incrementing the PATCH number)
1758 are intended for bug fixes and other improvements of existing
1759 features only (like improving performance or adding documentation)
1760 and retain API/ABI compatibility.
1764 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1768 * Remove the 'dist' target and add a tarball building script. The
1769 'dist' target has fallen out of use, and it shouldn't be
1770 necessary to configure just to create a source distribution.
1774 * Recreate the OS390-Unix config target. It no longer relies on a
1775 special script like it did for OpenSSL pre-1.1.0.
1779 * Instead of having the source directories listed in Configure, add
1780 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1785 * Add GMAC to EVP_MAC.
1789 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1793 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1794 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1795 to facilitate the continued use of MACs through raw private keys in
1796 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1800 * Deprecate ECDH_KDF_X9_62().
1804 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1805 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1806 are retained for backwards compatibility.
1810 * AES-XTS mode now enforces that its two keys are different to mitigate
1811 the attacked described in "Efficient Instantiations of Tweakable
1812 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1813 Details of this attack can be obtained from:
1814 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1818 * Rename the object files, i.e. give them other names than in previous
1819 versions. Their names now include the name of the final product, as
1820 well as its type mnemonic (bin, lib, shlib).
1824 * Added new option for 'openssl list', '-objects', which will display the
1825 list of built in objects, i.e. OIDs with names.
1829 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1830 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1835 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1836 improves application performance by removing data copies and providing
1837 applications with zero-copy system calls such as sendfile and splice.
1841 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1845 * `PKCS12_parse` now maintains the order of the parsed certificates
1846 when outputting them via `*ca` (rather than reversing it).
1850 * Deprecated pthread fork support methods.
1854 * Added support for FFDHE key exchange in TLS 1.3.
1858 * Added a new concept for OpenSSL plugability: providers. This
1859 functionality is designed to replace the ENGINE API and ENGINE
1860 implementations, and to be much more dynamic, allowing provider
1861 authors to introduce new algorithms among other things, as long as
1862 there's an API that supports the algorithm type.
1864 With this concept comes a new core API for interaction between
1865 libcrypto and provider implementations. Public libcrypto functions
1866 that want to use providers do so through this core API.
1868 The main documentation for this core API is found in
1869 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1870 refer to other manuals describing the API specific for supported
1871 algorithm types (also called operations).
1878 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
1880 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
1882 * Avoid loading of a dynamic engine twice.
1886 * Prioritise DANE TLSA issuer certs over peer certs
1890 * Fixed random API for MacOS prior to 10.12
1892 These MacOS versions don't support the CommonCrypto APIs
1896 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1898 * Fixed an SM2 Decryption Buffer Overflow.
1900 In order to decrypt SM2 encrypted data an application is expected to
1901 call the API function EVP_PKEY_decrypt(). Typically an application will
1902 call this function twice. The first time, on entry, the "out" parameter
1903 can be NULL and, on exit, the "outlen" parameter is populated with the
1904 buffer size required to hold the decrypted plaintext. The application
1905 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1906 again, but this time passing a non-NULL value for the "out" parameter.
1908 A bug in the implementation of the SM2 decryption code means that the
1909 calculation of the buffer size required to hold the plaintext returned
1910 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1911 size required by the second call. This can lead to a buffer overflow
1912 when EVP_PKEY_decrypt() is called by the application a second time with
1913 a buffer that is too small.
1915 A malicious attacker who is able present SM2 content for decryption to
1916 an application could cause attacker chosen data to overflow the buffer
1917 by up to a maximum of 62 bytes altering the contents of other data held
1918 after the buffer, possibly changing application behaviour or causing
1919 the application to crash. The location of the buffer is application
1920 dependent but is typically heap allocated.
1925 * Fixed various read buffer overruns processing ASN.1 strings
1927 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1928 structure which contains a buffer holding the string data and a field
1929 holding the buffer length. This contrasts with normal C strings which
1930 are repesented as a buffer for the string data which is terminated
1931 with a NUL (0) byte.
1933 Although not a strict requirement, ASN.1 strings that are parsed using
1934 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1935 well as any string whose value has been set with the ASN1_STRING_set()
1936 function will additionally NUL terminate the byte array in the
1937 ASN1_STRING structure.
1939 However, it is possible for applications to directly construct valid
1940 ASN1_STRING structures which do not NUL terminate the byte array by
1941 directly setting the "data" and "length" fields in the ASN1_STRING
1942 array. This can also happen by using the ASN1_STRING_set0() function.
1944 Numerous OpenSSL functions that print ASN.1 data have been found to
1945 assume that the ASN1_STRING byte array will be NUL terminated, even
1946 though this is not guaranteed for strings that have been directly
1947 constructed. Where an application requests an ASN.1 structure to be
1948 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1949 been directly constructed by the application without NUL terminating
1950 the "data" field, then a read buffer overrun can occur.
1952 The same thing can also occur during name constraints processing
1953 of certificates (for example if a certificate has been directly
1954 constructed by the application instead of loading it via the OpenSSL
1955 parsing functions, and the certificate contains non NUL terminated
1956 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1957 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1959 If a malicious actor can cause an application to directly construct an
1960 ASN1_STRING and then process it through one of the affected OpenSSL
1961 functions then this issue could be hit. This might result in a crash
1962 (causing a Denial of Service attack). It could also result in the
1963 disclosure of private memory contents (such as private keys, or
1964 sensitive plaintext).
1969 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1971 * Fixed a problem with verifying a certificate chain when using the
1972 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1973 the certificates present in a certificate chain. It is not set by default.
1975 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1976 the chain that have explicitly encoded elliptic curve parameters was added
1977 as an additional strict check.
1979 An error in the implementation of this check meant that the result of a
1980 previous check to confirm that certificates in the chain are valid CA
1981 certificates was overwritten. This effectively bypasses the check
1982 that non-CA certificates must not be able to issue other certificates.
1984 If a "purpose" has been configured then there is a subsequent opportunity
1985 for checks that the certificate is a valid CA. All of the named "purpose"
1986 values implemented in libcrypto perform this check. Therefore, where
1987 a purpose is set the certificate chain will still be rejected even when the
1988 strict flag has been used. A purpose is set by default in libssl client and
1989 server certificate verification routines, but it can be overridden or
1990 removed by an application.
1992 In order to be affected, an application must explicitly set the
1993 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1994 for the certificate verification or, in the case of TLS client or server
1995 applications, override the default purpose.
2000 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2001 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2002 renegotiation ClientHello omits the signature_algorithms extension (where it
2003 was present in the initial ClientHello), but includes a
2004 signature_algorithms_cert extension then a NULL pointer dereference will
2005 result, leading to a crash and a denial of service attack.
2007 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2008 (which is the default configuration). OpenSSL TLS clients are not impacted by
2012 *Peter Kästle and Samuel Sapalski*
2014 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2016 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2017 create a unique hash value based on the issuer and serial number data
2018 contained within an X509 certificate. However it was failing to correctly
2019 handle any errors that may occur while parsing the issuer field (which might
2020 occur if the issuer field is maliciously constructed). This may subsequently
2021 result in a NULL pointer deref and a crash leading to a potential denial of
2027 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2028 padding mode to correctly check for rollback attacks. This is considered a
2029 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2034 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2035 functions. Previously they could overflow the output length argument in some
2036 cases where the input length is close to the maximum permissable length for
2037 an integer on the platform. In such cases the return value from the function
2038 call would be 1 (indicating success), but the output length value would be
2039 negative. This could cause applications to behave incorrectly or crash.
2044 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2045 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2046 could be exploited in a side channel attack to recover the password. Since
2047 the attack is local host only this is outside of the current OpenSSL
2048 threat model and therefore no CVE is assigned.
2050 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2055 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2057 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2058 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2059 If an attacker can control both items being compared then this could lead
2060 to a possible denial of service attack. OpenSSL itself uses the
2061 GENERAL_NAME_cmp function for two purposes:
2062 1) Comparing CRL distribution point names between an available CRL and a
2063 CRL distribution point embedded in an X509 certificate
2064 2) When verifying that a timestamp response token signer matches the
2065 timestamp authority name (exposed via the API functions
2066 TS_RESP_verify_response and TS_RESP_verify_token)
2071 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2073 * Certificates with explicit curve parameters are now disallowed in
2074 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2078 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2079 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2080 conversely, silently ignore DTLS protocol version bounds when configuring
2081 TLS-based contexts. The commands can be repeated to set bounds of both
2082 types. The same applies with the corresponding "min_protocol" and
2083 "max_protocol" command-line switches, in case some application uses both TLS
2086 SSL_CTX instances that are created for a fixed protocol version (e.g.
2087 TLSv1_server_method()) also silently ignore version bounds. Previously
2088 attempts to apply bounds to these protocol versions would result in an
2089 error. Now only the "version-flexible" SSL_CTX instances are subject to
2090 limits in configuration files in command-line options.
2094 * Handshake now fails if Extended Master Secret extension is dropped
2099 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2101 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2103 * Fixed segmentation fault in SSL_check_chain()
2104 Server or client applications that call the SSL_check_chain() function
2105 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2106 dereference as a result of incorrect handling of the
2107 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2108 or unrecognised signature algorithm is received from the peer. This could
2109 be exploited by a malicious peer in a Denial of Service attack.
2114 * Added AES consttime code for no-asm configurations
2115 an optional constant time support for AES was added
2116 when building openssl for no-asm.
2117 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2118 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2119 At this time this feature is by default disabled.
2120 It will be enabled by default in 3.0.
2124 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2126 * Revert the change of EOF detection while reading in libssl to avoid
2127 regressions in applications depending on the current way of reporting
2128 the EOF. As the existing method is not fully accurate the change to
2129 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2130 branch and will be present in the 3.0 release.
2134 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2135 when primes for RSA keys are computed.
2136 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2137 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2138 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
2139 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2140 This avoids possible fingerprinting of newly generated RSA modules.
2144 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2146 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2147 while reading in libssl then we would report an error back to the
2148 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2149 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2150 therefore give a hint as to what went wrong.
2154 * Check that ed25519 and ed448 are allowed by the security level. Previously
2155 signature algorithms not using an MD were not being checked that they were
2156 allowed by the security level.
2160 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2161 was not quite right. The behaviour was not consistent between resumption
2162 and normal handshakes, and also not quite consistent with historical
2163 behaviour. The behaviour in various scenarios has been clarified and
2164 it has been updated to make it match historical behaviour as closely as
2169 * *[VMS only]* The header files that the VMS compilers include automatically,
2170 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2171 that the C++ compiler doesn't understand. This is a shortcoming in the
2172 compiler, but can be worked around with `__cplusplus` guards.
2174 C++ applications that use OpenSSL libraries must be compiled using the
2175 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2176 functions. Otherwise, only functions with symbols of less than 31
2177 characters can be used, as the linker will not be able to successfully
2178 resolve symbols with longer names.
2182 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2183 The presence of this system service is determined at run-time.
2187 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2192 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2194 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2195 number generator (RNG). This was intended to include protection in the
2196 event of a fork() system call in order to ensure that the parent and child
2197 processes did not share the same RNG state. However this protection was not
2198 being used in the default case.
2200 A partial mitigation for this issue is that the output from a high
2201 precision timer is mixed into the RNG state so the likelihood of a parent
2202 and child process sharing state is significantly reduced.
2204 If an application already calls OPENSSL_init_crypto() explicitly using
2205 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2208 *Matthias St. Pierre*
2210 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2211 used even when parsing explicit parameters, when loading a encoded key
2212 or calling `EC_GROUP_new_from_ecpkparameters()`/
2213 `EC_GROUP_new_from_ecparameters()`.
2214 This prevents bypass of security hardening and performance gains,
2215 especially for curves with specialized EC_METHODs.
2216 By default, if a key encoded with explicit parameters is loaded and later
2217 encoded, the output is still encoded with explicit parameters, even if
2218 internally a "named" EC_GROUP is used for computation.
2222 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2223 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2224 NULL. After this change, only the cofactor parameter can be NULL. It also
2225 does some minimal sanity checks on the passed order.
2230 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2231 An attack is simple, if the first CMS_recipientInfo is valid but the
2232 second CMS_recipientInfo is chosen ciphertext. If the second
2233 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2234 encryption key will be replaced by garbage, and the message cannot be
2235 decoded, but if the RSA decryption fails, the correct encryption key is
2236 used and the recipient will not notice the attack.
2237 As a work around for this potential attack the length of the decrypted
2238 key must be equal to the cipher default key length, in case the
2239 certifiate is not given and all recipientInfo are tried out.
2240 The old behaviour can be re-enabled in the CMS code by setting the
2241 CMS_DEBUG_DECRYPT flag.
2246 * Early start up entropy quality from the DEVRANDOM seed source has been
2247 improved for older Linux systems. The RAND subsystem will wait for
2248 /dev/random to be producing output before seeding from /dev/urandom.
2249 The seeded state is stored for future library initialisations using
2250 a system global shared memory segment. The shared memory identifier
2251 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2252 the desired value. The default identifier is 114.
2256 * Correct the extended master secret constant on EBCDIC systems. Without this
2257 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2258 negotiate EMS will fail. Unfortunately this also means that TLS connections
2259 between EBCDIC systems with this fix, and EBCDIC systems without this
2260 fix will fail if they negotiate EMS.
2264 * Use Windows installation paths in the mingw builds
2266 Mingw isn't a POSIX environment per se, which means that Windows
2267 paths should be used for installation.
2272 * Changed DH_check to accept parameters with order q and 2q subgroups.
2273 With order 2q subgroups the bit 0 of the private key is not secret
2274 but DH_generate_key works around that by clearing bit 0 of the
2275 private key for those. This avoids leaking bit 0 of the private key.
2279 * Significantly reduce secure memory usage by the randomness pools.
2283 * Revert the DEVRANDOM_WAIT feature for Linux systems
2285 The DEVRANDOM_WAIT feature added a select() call to wait for the
2286 /dev/random device to become readable before reading from the
2287 /dev/urandom device.
2289 It turned out that this change had negative side effects on
2290 performance which were not acceptable. After some discussion it
2291 was decided to revert this feature and leave it up to the OS
2292 resp. the platform maintainer to ensure a proper initialization
2293 during early boot time.
2295 *Matthias St. Pierre*
2297 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2299 * Add build tests for C++. These are generated files that only do one
2300 thing, to include one public OpenSSL head file each. This tests that
2301 the public header files can be usefully included in a C++ application.
2303 This test isn't enabled by default. It can be enabled with the option
2304 'enable-buildtest-c++'.
2308 * Enable SHA3 pre-hashing for ECDSA and DSA.
2312 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2313 This changes the size when using the `genpkey` command when no size is given.
2314 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2315 generation commands to use 2048 bits by default.
2319 * Reorganize the manual pages to consistently have RETURN VALUES,
2320 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2321 util/fix-doc-nits accordingly.
2323 *Paul Yang, Joshua Lock*
2325 * Add the missing accessor EVP_PKEY_get0_engine()
2329 * Have commands like `s_client` and `s_server` output the signature scheme
2330 along with other cipher suite parameters when debugging.
2334 * Make OPENSSL_config() error agnostic again.
2338 * Do the error handling in RSA decryption constant time.
2342 * Prevent over long nonces in ChaCha20-Poly1305.
2344 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2345 for every encryption operation. RFC 7539 specifies that the nonce value
2346 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2347 and front pads the nonce with 0 bytes if it is less than 12
2348 bytes. However it also incorrectly allows a nonce to be set of up to 16
2349 bytes. In this case only the last 12 bytes are significant and any
2350 additional leading bytes are ignored.
2352 It is a requirement of using this cipher that nonce values are
2353 unique. Messages encrypted using a reused nonce value are susceptible to
2354 serious confidentiality and integrity attacks. If an application changes
2355 the default nonce length to be longer than 12 bytes and then makes a
2356 change to the leading bytes of the nonce expecting the new value to be a
2357 new unique nonce then such an application could inadvertently encrypt
2358 messages with a reused nonce.
2360 Additionally the ignored bytes in a long nonce are not covered by the
2361 integrity guarantee of this cipher. Any application that relies on the
2362 integrity of these ignored leading bytes of a long nonce may be further
2363 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2364 is safe because no such use sets such a long nonce value. However user
2365 applications that use this cipher directly and set a non-default nonce
2366 length to be longer than 12 bytes may be vulnerable.
2368 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2374 * Add DEVRANDOM_WAIT feature for Linux systems
2376 On older Linux systems where the getrandom() system call is not available,
2377 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2378 Contrary to getrandom(), the /dev/urandom device will not block during
2379 early boot when the kernel CSPRNG has not been seeded yet.
2381 To mitigate this known weakness, use select() to wait for /dev/random to
2382 become readable before reading from /dev/urandom.
2384 * Ensure that SM2 only uses SM3 as digest algorithm
2388 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2390 * Change the info callback signals for the start and end of a post-handshake
2391 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2392 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2393 confused by this and assume that a TLSv1.2 renegotiation has started. This
2394 can break KeyUpdate handling. Instead we no longer signal the start and end
2395 of a post handshake message exchange (although the messages themselves are
2396 still signalled). This could break some applications that were expecting
2397 the old signals. However without this KeyUpdate is not usable for many
2402 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2404 * Timing vulnerability in DSA signature generation
2406 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2407 timing side channel attack. An attacker could use variations in the signing
2408 algorithm to recover the private key.
2410 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2415 * Timing vulnerability in ECDSA signature generation
2417 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2418 timing side channel attack. An attacker could use variations in the signing
2419 algorithm to recover the private key.
2421 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2426 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2427 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2428 of two gigabytes and the error handling improved.
2430 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2431 categorized as a normal bug, not a security issue, because the DRBG reseeds
2432 automatically and is fully functional even without additional randomness
2433 provided by the application.
2435 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2437 * Add a new ClientHello callback. Provides a callback interface that gives
2438 the application the ability to adjust the nascent SSL object at the
2439 earliest stage of ClientHello processing, immediately after extensions have
2440 been collected but before they have been processed. In particular, this
2441 callback can adjust the supported TLS versions in response to the contents
2446 * Add SM2 base algorithm support.
2450 * s390x assembly pack: add (improved) hardware-support for the following
2451 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2452 aes-cfb/cfb8, aes-ecb.
2456 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2457 parameter is no longer accepted, as it leads to a corrupt table. NULL
2458 pem_str is reserved for alias entries only.
2462 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2463 step for prime curves. The new implementation is based on formulae from
2464 differential addition-and-doubling in homogeneous projective coordinates
2465 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2466 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2467 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2468 to work in projective coordinates.
2470 *Billy Bob Brumley, Nicola Tuveri*
2472 * Change generating and checking of primes so that the error rate of not
2473 being prime depends on the intended use based on the size of the input.
2474 For larger primes this will result in more rounds of Miller-Rabin.
2475 The maximal error rate for primes with more than 1080 bits is lowered
2478 *Kurt Roeckx, Annie Yousar*
2480 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2484 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2485 moving between systems, and to avoid confusion when a Windows build is
2486 done with mingw vs with MSVC. For POSIX installs, there's still a
2487 symlink or copy named 'tsget' to avoid that confusion as well.
2491 * Revert blinding in ECDSA sign and instead make problematic addition
2492 length-invariant. Switch even to fixed-length Montgomery multiplication.
2496 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2497 step for binary curves. The new implementation is based on formulae from
2498 differential addition-and-doubling in mixed Lopez-Dahab projective
2499 coordinates, modified to independently blind the operands.
2501 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2503 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2504 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2505 EC_METHODs to implement their own specialized "ladder step", to take
2506 advantage of more favorable coordinate systems or more efficient
2507 differential addition-and-doubling algorithms.
2509 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2511 * Modified the random device based seed sources to keep the relevant
2512 file descriptors open rather than reopening them on each access.
2513 This allows such sources to operate in a chroot() jail without
2514 the associated device nodes being available. This behaviour can be
2515 controlled using RAND_keep_random_devices_open().
2519 * Numerous side-channel attack mitigations have been applied. This may have
2520 performance impacts for some algorithms for the benefit of improved
2521 security. Specific changes are noted in this change log by their respective
2526 * AIX shared library support overhaul. Switch to AIX "natural" way of
2527 handling shared libraries, which means collecting shared objects of
2528 different versions and bitnesses in one common archive. This allows to
2529 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2530 doesn't affect the way 3rd party applications are linked, only how
2531 multi-version installation is managed.
2535 * Make ec_group_do_inverse_ord() more robust and available to other
2536 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2537 mitigations are applied to the fallback BN_mod_inverse().
2538 When using this function rather than BN_mod_inverse() directly, new
2539 EC cryptosystem implementations are then safer-by-default.
2543 * Add coordinate blinding for EC_POINT and implement projective
2544 coordinate blinding for generic prime curves as a countermeasure to
2545 chosen point SCA attacks.
2547 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2549 * Add blinding to ECDSA and DSA signatures to protect against side channel
2550 attacks discovered by Keegan Ryan (NCC Group).
2554 * Enforce checking in the `pkeyutl` command to ensure that the input
2555 length does not exceed the maximum supported digest length when performing
2556 a sign, verify or verifyrecover operation.
2560 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2561 I/O in combination with something like select() or poll() will hang. This
2562 can be turned off again using SSL_CTX_clear_mode().
2563 Many applications do not properly handle non-application data records, and
2564 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2565 around the problems in those applications, but can also break some.
2566 It's recommended to read the manpages about SSL_read(), SSL_write(),
2567 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2568 SSL_CTX_set_read_ahead() again.
2572 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2573 now allow empty (zero character) pass phrases.
2577 * Apply blinding to binary field modular inversion and remove patent
2578 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2582 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2583 binary and prime elliptic curves.
2587 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2588 constant time fixed point multiplication.
2592 * Revise elliptic curve scalar multiplication with timing attack
2593 defenses: ec_wNAF_mul redirects to a constant time implementation
2594 when computing fixed point and variable point multiplication (which
2595 in OpenSSL are mostly used with secret scalars in keygen, sign,
2596 ECDH derive operations).
2597 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2600 * Updated CONTRIBUTING
2604 * Updated DRBG / RAND to request nonce and additional low entropy
2605 randomness from the system.
2607 *Matthias St. Pierre*
2609 * Updated 'openssl rehash' to use OpenSSL consistent default.
2613 * Moved the load of the ssl_conf module to libcrypto, which helps
2614 loading engines that libssl uses before libssl is initialised.
2618 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2622 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2624 *Ingo Schwarze, Rich Salz*
2626 * Added output of accepting IP address and port for 'openssl s_server'
2630 * Added a new API for TLSv1.3 ciphersuites:
2631 SSL_CTX_set_ciphersuites()
2632 SSL_set_ciphersuites()
2636 * Memory allocation failures consistently add an error to the error
2641 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2642 in libcrypto when run as setuid/setgid.
2646 * Load any config file by default when libssl is used.
2650 * Added new public header file <openssl/rand_drbg.h> and documentation
2651 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2653 *Matthias St. Pierre*
2655 * QNX support removed (cannot find contributors to get their approval
2656 for the license change).
2660 * TLSv1.3 replay protection for early data has been implemented. See the
2661 SSL_read_early_data() man page for further details.
2665 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2666 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2667 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2668 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2669 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2670 configuration has been separated out. See the ciphers man page or the
2671 SSL_CTX_set_ciphersuites() man page for more information.
2675 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2676 in responder mode now supports the new "-multi" option, which
2677 spawns the specified number of child processes to handle OCSP
2678 requests. The "-timeout" option now also limits the OCSP
2679 responder's patience to wait to receive the full client request
2680 on a newly accepted connection. Child processes are respawned
2681 as needed, and the CA index file is automatically reloaded
2682 when changed. This makes it possible to run the "ocsp" responder
2683 as a long-running service, making the OpenSSL CA somewhat more
2684 feature-complete. In this mode, most diagnostic messages logged
2685 after entering the event loop are logged via syslog(3) rather than
2690 * Added support for X448 and Ed448. Heavily based on original work by
2695 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2696 objects loaded. This adds the functions OSSL_STORE_expect() and
2697 OSSL_STORE_find() as well as needed tools to construct searches and
2698 get the search data out of them.
2702 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2703 version of OpenSSL should review their configuration settings to ensure
2704 that they are still appropriate for TLSv1.3. For further information see:
2705 <https://wiki.openssl.org/index.php/TLS1.3>
2709 * Grand redesign of the OpenSSL random generator
2711 The default RAND method now utilizes an AES-CTR DRBG according to
2712 NIST standard SP 800-90Ar1. The new random generator is essentially
2713 a port of the default random generator from the OpenSSL FIPS 2.0
2714 object module. It is a hybrid deterministic random bit generator
2715 using an AES-CTR bit stream and which seeds and reseeds itself
2716 automatically using trusted system entropy sources.
2718 Some of its new features are:
2719 - Support for multiple DRBG instances with seed chaining.
2720 - The default RAND method makes use of a DRBG.
2721 - There is a public and private DRBG instance.
2722 - The DRBG instances are fork-safe.
2723 - Keep all global DRBG instances on the secure heap if it is enabled.
2724 - The public and private DRBG instance are per thread for lock free
2727 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2729 * Changed Configure so it only says what it does and doesn't dump
2730 so much data. Instead, ./configdata.pm should be used as a script
2731 to display all sorts of configuration data.
2735 * Added processing of "make variables" to Configure.
2739 * Added SHA512/224 and SHA512/256 algorithm support.
2743 * The last traces of Netware support, first removed in 1.1.0, have
2748 * Get rid of Makefile.shared, and in the process, make the processing
2749 of certain files (rc.obj, or the .def/.map/.opt files produced from
2750 the ordinal files) more visible and hopefully easier to trace and
2751 debug (or make silent).
2755 * Make it possible to have environment variable assignments as
2756 arguments to config / Configure.
2760 * Add multi-prime RSA (RFC 8017) support.
2764 * Add SM3 implemented according to GB/T 32905-2016
2765 *Jack Lloyd <jack.lloyd@ribose.com>,*
2766 *Ronald Tse <ronald.tse@ribose.com>,*
2767 *Erick Borsboom <erick.borsboom@ribose.com>*
2769 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2770 as documented in RFC6066.
2771 Based on a patch from Tomasz Moń
2773 *Filipe Raimundo da Silva*
2775 * Add SM4 implemented according to GB/T 32907-2016.
2776 *Jack Lloyd <jack.lloyd@ribose.com>,*
2777 *Ronald Tse <ronald.tse@ribose.com>,*
2778 *Erick Borsboom <erick.borsboom@ribose.com>*
2780 * Reimplement -newreq-nodes and ERR_error_string_n; the
2781 original author does not agree with the license change.
2785 * Add ARIA AEAD TLS support.
2789 * Some macro definitions to support VS6 have been removed. Visual
2790 Studio 6 has not worked since 1.1.0
2794 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2795 without clearing the errors.
2799 * Add "atfork" functions. If building on a system that without
2800 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2801 requirements. The RAND facility now uses/requires this.
2809 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2810 not possible to disable entirely. However, it's still possible to
2811 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2814 To disable, configure with 'no-ui-console'. 'no-ui' is still
2815 possible to use as an alias. Check at compile time with the
2816 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2817 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2821 * Add a STORE module, which implements a uniform and URI based reader of
2822 stores that can contain keys, certificates, CRLs and numerous other
2823 objects. The main API is loosely based on a few stdio functions,
2824 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2825 OSSL_STORE_error and OSSL_STORE_close.
2826 The implementation uses backends called "loaders" to implement arbitrary
2827 URI schemes. There is one built in "loader" for the 'file' scheme.
2831 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2832 then adjusted to work on FreeBSD 8.4 as well.
2833 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2834 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2838 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2839 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2840 error code calls like this:
2842 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2844 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2845 that can be encoded in C. For the foreseeable future, this will only
2848 *Richard Levitte and Tim Hudson*
2850 * Removed BSD cryptodev engine.
2854 * Add a build target 'build_all_generated', to build all generated files
2855 and only that. This can be used to prepare everything that requires
2856 things like perl for a system that lacks perl and then move everything
2857 to that system and do the rest of the build there.
2861 * In the UI interface, make it possible to duplicate the user data. This
2862 can be used by engines that need to retain the data for a longer time
2863 than just the call where this user data is passed.
2867 * Ignore the '-named_curve auto' value for compatibility of applications
2870 *Tomáš Mráz <tmraz@fedoraproject.org>*
2872 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2873 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2874 alerts across multiple records (some of which could be empty). In practice
2875 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2876 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2877 support this at all. Supporting it adds significant complexity to the
2878 record layer, and its removal is unlikely to cause interoperability
2883 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2884 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2885 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2890 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2891 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2893 *Richard Levitte, Andy Polyakov*
2895 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2900 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2901 platform rather than 'mingw'.
2905 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2906 success if they are asked to add an object which already exists
2907 in the store. This change cascades to other functions which load
2908 certificates and CRLs.
2912 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2913 facilitate stack unwinding even from assembly subroutines.
2917 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2918 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2922 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2923 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2924 which is the minimum version we support.
2928 * Certificate time validation (X509_cmp_time) enforces stricter
2929 compliance with RFC 5280. Fractional seconds and timezone offsets
2930 are no longer allowed.
2934 * Add support for ARIA
2938 * s_client will now send the Server Name Indication (SNI) extension by
2939 default unless the new "-noservername" option is used. The server name is
2940 based on the host provided to the "-connect" option unless overridden by
2941 using "-servername".
2945 * Add support for SipHash
2949 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2950 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2951 prevent issues where no progress is being made and the peer continually
2952 sends unrecognised record types, using up resources processing them.
2956 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2957 using the algorithm defined in
2958 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2962 * Heartbeat support has been removed; the ABI is changed for now.
2964 *Richard Levitte, Rich Salz*
2966 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2970 * The RSA "null" method, which was partially supported to avoid patent
2971 issues, has been replaced to always returns NULL.
2978 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2980 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2981 used even when parsing explicit parameters, when loading a encoded key
2982 or calling `EC_GROUP_new_from_ecpkparameters()`/
2983 `EC_GROUP_new_from_ecparameters()`.
2984 This prevents bypass of security hardening and performance gains,
2985 especially for curves with specialized EC_METHODs.
2986 By default, if a key encoded with explicit parameters is loaded and later
2987 encoded, the output is still encoded with explicit parameters, even if
2988 internally a "named" EC_GROUP is used for computation.
2992 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2993 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2994 NULL. After this change, only the cofactor parameter can be NULL. It also
2995 does some minimal sanity checks on the passed order.
3000 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3001 An attack is simple, if the first CMS_recipientInfo is valid but the
3002 second CMS_recipientInfo is chosen ciphertext. If the second
3003 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3004 encryption key will be replaced by garbage, and the message cannot be
3005 decoded, but if the RSA decryption fails, the correct encryption key is
3006 used and the recipient will not notice the attack.
3007 As a work around for this potential attack the length of the decrypted
3008 key must be equal to the cipher default key length, in case the
3009 certifiate is not given and all recipientInfo are tried out.
3010 The old behaviour can be re-enabled in the CMS code by setting the
3011 CMS_DEBUG_DECRYPT flag.
3016 * Use Windows installation paths in the mingw builds
3018 Mingw isn't a POSIX environment per se, which means that Windows
3019 paths should be used for installation.
3024 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3026 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3027 This changes the size when using the `genpkey` command when no size is given.
3028 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3029 generation commands to use 2048 bits by default.
3033 * Prevent over long nonces in ChaCha20-Poly1305.
3035 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3036 for every encryption operation. RFC 7539 specifies that the nonce value
3037 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3038 and front pads the nonce with 0 bytes if it is less than 12
3039 bytes. However it also incorrectly allows a nonce to be set of up to 16
3040 bytes. In this case only the last 12 bytes are significant and any
3041 additional leading bytes are ignored.
3043 It is a requirement of using this cipher that nonce values are
3044 unique. Messages encrypted using a reused nonce value are susceptible to
3045 serious confidentiality and integrity attacks. If an application changes
3046 the default nonce length to be longer than 12 bytes and then makes a
3047 change to the leading bytes of the nonce expecting the new value to be a
3048 new unique nonce then such an application could inadvertently encrypt
3049 messages with a reused nonce.
3051 Additionally the ignored bytes in a long nonce are not covered by the
3052 integrity guarantee of this cipher. Any application that relies on the
3053 integrity of these ignored leading bytes of a long nonce may be further
3054 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3055 is safe because no such use sets such a long nonce value. However user
3056 applications that use this cipher directly and set a non-default nonce
3057 length to be longer than 12 bytes may be vulnerable.
3059 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3065 * Added SCA hardening for modular field inversion in EC_GROUP through
3066 a new dedicated field_inv() pointer in EC_METHOD.
3067 This also addresses a leakage affecting conversions from projective
3068 to affine coordinates.
3070 *Billy Bob Brumley, Nicola Tuveri*
3072 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3073 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3077 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3081 * Remove the 'dist' target and add a tarball building script. The
3082 'dist' target has fallen out of use, and it shouldn't be
3083 necessary to configure just to create a source distribution.
3087 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3089 * Timing vulnerability in DSA signature generation
3091 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3092 timing side channel attack. An attacker could use variations in the signing
3093 algorithm to recover the private key.
3095 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3100 * Timing vulnerability in ECDSA signature generation
3102 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3103 timing side channel attack. An attacker could use variations in the signing
3104 algorithm to recover the private key.
3106 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3111 * Add coordinate blinding for EC_POINT and implement projective
3112 coordinate blinding for generic prime curves as a countermeasure to
3113 chosen point SCA attacks.
3115 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3117 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3119 * Client DoS due to large DH parameter
3121 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3122 malicious server can send a very large prime value to the client. This will
3123 cause the client to spend an unreasonably long period of time generating a
3124 key for this prime resulting in a hang until the client has finished. This
3125 could be exploited in a Denial Of Service attack.
3127 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3132 * Cache timing vulnerability in RSA Key Generation
3134 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3135 a cache timing side channel attack. An attacker with sufficient access to
3136 mount cache timing attacks during the RSA key generation process could
3137 recover the private key.
3139 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3140 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3145 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3146 parameter is no longer accepted, as it leads to a corrupt table. NULL
3147 pem_str is reserved for alias entries only.
3151 * Revert blinding in ECDSA sign and instead make problematic addition
3152 length-invariant. Switch even to fixed-length Montgomery multiplication.
3156 * Change generating and checking of primes so that the error rate of not
3157 being prime depends on the intended use based on the size of the input.
3158 For larger primes this will result in more rounds of Miller-Rabin.
3159 The maximal error rate for primes with more than 1080 bits is lowered
3162 *Kurt Roeckx, Annie Yousar*
3164 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3168 * Add blinding to ECDSA and DSA signatures to protect against side channel
3169 attacks discovered by Keegan Ryan (NCC Group).
3173 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3174 now allow empty (zero character) pass phrases.
3178 * Certificate time validation (X509_cmp_time) enforces stricter
3179 compliance with RFC 5280. Fractional seconds and timezone offsets
3180 are no longer allowed.
3184 * Fixed a text canonicalisation bug in CMS
3186 Where a CMS detached signature is used with text content the text goes
3187 through a canonicalisation process first prior to signing or verifying a
3188 signature. This process strips trailing space at the end of lines, converts
3189 line terminators to CRLF and removes additional trailing line terminators
3190 at the end of a file. A bug in the canonicalisation process meant that
3191 some characters, such as form-feed, were incorrectly treated as whitespace
3192 and removed. This is contrary to the specification (RFC5485). This fix
3193 could mean that detached text data signed with an earlier version of
3194 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3195 signed with a fixed OpenSSL may fail to verify with an earlier version of
3196 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3197 and use the "-binary" flag (for the "cms" command line application) or set
3198 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3202 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3204 * Constructed ASN.1 types with a recursive definition could exceed the stack
3206 Constructed ASN.1 types with a recursive definition (such as can be found
3207 in PKCS7) could eventually exceed the stack given malicious input with
3208 excessive recursion. This could result in a Denial Of Service attack. There
3209 are no such structures used within SSL/TLS that come from untrusted sources
3210 so this is considered safe.
3212 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3218 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3220 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3221 effectively reduced to only comparing the least significant bit of each
3222 byte. This allows an attacker to forge messages that would be considered as
3223 authenticated in an amount of tries lower than that guaranteed by the
3224 security claims of the scheme. The module can only be compiled by the
3225 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3227 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3233 * Add a build target 'build_all_generated', to build all generated files
3234 and only that. This can be used to prepare everything that requires
3235 things like perl for a system that lacks perl and then move everything
3236 to that system and do the rest of the build there.
3240 * Backport SSL_OP_NO_RENGOTIATION
3242 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3243 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3244 changes this is no longer possible in 1.1.0. Therefore the new
3245 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3246 1.1.0 to provide equivalent functionality.
3248 Note that if an application built against 1.1.0h headers (or above) is run
3249 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3250 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3254 * Removed the OS390-Unix config target. It relied on a script that doesn't
3259 * rsaz_1024_mul_avx2 overflow bug on x86_64
3261 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3262 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3263 Analysis suggests that attacks against RSA and DSA as a result of this
3264 defect would be very difficult to perform and are not believed likely.
3265 Attacks against DH1024 are considered just feasible, because most of the
3266 work necessary to deduce information about a private key may be performed
3267 offline. The amount of resources required for such an attack would be
3268 significant. However, for an attack on TLS to be meaningful, the server
3269 would have to share the DH1024 private key among multiple clients, which is
3270 no longer an option since CVE-2016-0701.
3272 This only affects processors that support the AVX2 but not ADX extensions
3273 like Intel Haswell (4th generation).
3275 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3276 was originally found via the OSS-Fuzz project.
3281 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3283 * bn_sqrx8x_internal carry bug on x86_64
3285 There is a carry propagating bug in the x86_64 Montgomery squaring
3286 procedure. No EC algorithms are affected. Analysis suggests that attacks
3287 against RSA and DSA as a result of this defect would be very difficult to
3288 perform and are not believed likely. Attacks against DH are considered just
3289 feasible (although very difficult) because most of the work necessary to
3290 deduce information about a private key may be performed offline. The amount
3291 of resources required for such an attack would be very significant and
3292 likely only accessible to a limited number of attackers. An attacker would
3293 additionally need online access to an unpatched system using the target
3294 private key in a scenario with persistent DH parameters and a private
3295 key that is shared between multiple clients.
3297 This only affects processors that support the BMI1, BMI2 and ADX extensions
3298 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3300 This issue was reported to OpenSSL by the OSS-Fuzz project.
3305 * Malformed X.509 IPAddressFamily could cause OOB read
3307 If an X.509 certificate has a malformed IPAddressFamily extension,
3308 OpenSSL could do a one-byte buffer overread. The most likely result
3309 would be an erroneous display of the certificate in text format.
3311 This issue was reported to OpenSSL by the OSS-Fuzz project.
3316 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3318 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3319 platform rather than 'mingw'.
3323 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3324 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3325 which is the minimum version we support.
3329 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3331 * Encrypt-Then-Mac renegotiation crash
3333 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3334 negotiated where it was not in the original handshake (or vice-versa) then
3335 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3336 and servers are affected.
3338 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3343 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3345 * Truncated packet could crash via OOB read
3347 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3348 cipher is being used, then a truncated packet can cause that host to
3349 perform an out-of-bounds read, usually resulting in a crash.
3351 This issue was reported to OpenSSL by Robert Święcki of Google.
3356 * Bad (EC)DHE parameters cause a client crash
3358 If a malicious server supplies bad parameters for a DHE or ECDHE key
3359 exchange then this can result in the client attempting to dereference a
3360 NULL pointer leading to a client crash. This could be exploited in a Denial
3363 This issue was reported to OpenSSL by Guido Vranken.
3368 * BN_mod_exp may produce incorrect results on x86_64
3370 There is a carry propagating bug in the x86_64 Montgomery squaring
3371 procedure. No EC algorithms are affected. Analysis suggests that attacks
3372 against RSA and DSA as a result of this defect would be very difficult to
3373 perform and are not believed likely. Attacks against DH are considered just
3374 feasible (although very difficult) because most of the work necessary to
3375 deduce information about a private key may be performed offline. The amount
3376 of resources required for such an attack would be very significant and
3377 likely only accessible to a limited number of attackers. An attacker would
3378 additionally need online access to an unpatched system using the target
3379 private key in a scenario with persistent DH parameters and a private
3380 key that is shared between multiple clients. For example this can occur by
3381 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3382 similar to CVE-2015-3193 but must be treated as a separate problem.
3384 This issue was reported to OpenSSL by the OSS-Fuzz project.
3389 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3391 * ChaCha20/Poly1305 heap-buffer-overflow
3393 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3394 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3395 crash. This issue is not considered to be exploitable beyond a DoS.
3397 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3402 * CMS Null dereference
3404 Applications parsing invalid CMS structures can crash with a NULL pointer
3405 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3406 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3407 structure callback if an attempt is made to free certain invalid encodings.
3408 Only CHOICE structures using a callback which do not handle NULL value are
3411 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3416 * Montgomery multiplication may produce incorrect results
3418 There is a carry propagating bug in the Broadwell-specific Montgomery
3419 multiplication procedure that handles input lengths divisible by, but
3420 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3421 and DH private keys are impossible. This is because the subroutine in
3422 question is not used in operations with the private key itself and an input
3423 of the attacker's direct choice. Otherwise the bug can manifest itself as
3424 transient authentication and key negotiation failures or reproducible
3425 erroneous outcome of public-key operations with specially crafted input.
3426 Among EC algorithms only Brainpool P-512 curves are affected and one
3427 presumably can attack ECDH key negotiation. Impact was not analyzed in
3428 detail, because pre-requisites for attack are considered unlikely. Namely
3429 multiple clients have to choose the curve in question and the server has to
3430 share the private key among them, neither of which is default behaviour.
3431 Even then only clients that chose the curve will be affected.
3433 This issue was publicly reported as transient failures and was not
3434 initially recognized as a security issue. Thanks to Richard Morgan for
3435 providing reproducible case.
3440 * Removed automatic addition of RPATH in shared libraries and executables,
3441 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3445 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3447 * Fix Use After Free for large message sizes
3449 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3450 message larger than approx 16k is received then the underlying buffer to
3451 store the incoming message is reallocated and moved. Unfortunately a
3452 dangling pointer to the old location is left which results in an attempt to
3453 write to the previously freed location. This is likely to result in a
3454 crash, however it could potentially lead to execution of arbitrary code.
3456 This issue only affects OpenSSL 1.1.0a.
3458 This issue was reported to OpenSSL by Robert Święcki.
3463 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3465 * OCSP Status Request extension unbounded memory growth
3467 A malicious client can send an excessively large OCSP Status Request
3468 extension. If that client continually requests renegotiation, sending a
3469 large OCSP Status Request extension each time, then there will be unbounded
3470 memory growth on the server. This will eventually lead to a Denial Of
3471 Service attack through memory exhaustion. Servers with a default
3472 configuration are vulnerable even if they do not support OCSP. Builds using
3473 the "no-ocsp" build time option are not affected.
3475 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3480 * SSL_peek() hang on empty record
3482 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3483 sends an empty record. This could be exploited by a malicious peer in a
3484 Denial Of Service attack.
3486 This issue was reported to OpenSSL by Alex Gaynor.
3491 * Excessive allocation of memory in tls_get_message_header() and
3492 dtls1_preprocess_fragment()
3494 A (D)TLS message includes 3 bytes for its length in the header for the
3495 message. This would allow for messages up to 16Mb in length. Messages of
3496 this length are excessive and OpenSSL includes a check to ensure that a
3497 peer is sending reasonably sized messages in order to avoid too much memory
3498 being consumed to service a connection. A flaw in the logic of version
3499 1.1.0 means that memory for the message is allocated too early, prior to
3500 the excessive message length check. Due to way memory is allocated in
3501 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3502 to service a connection. This could lead to a Denial of Service through
3503 memory exhaustion. However, the excessive message length check still takes
3504 place, and this would cause the connection to immediately fail. Assuming
3505 that the application calls SSL_free() on the failed connection in a timely
3506 manner then the 21Mb of allocated memory will then be immediately freed
3507 again. Therefore the excessive memory allocation will be transitory in
3508 nature. This then means that there is only a security impact if:
3510 1) The application does not call SSL_free() in a timely manner in the event
3511 that the connection fails
3513 2) The application is working in a constrained environment where there is
3514 very little free memory
3516 3) The attacker initiates multiple connection attempts such that there are
3517 multiple connections in a state where memory has been allocated for the
3518 connection; SSL_free() has not yet been called; and there is insufficient
3519 memory to service the multiple requests.
3521 Except in the instance of (1) above any Denial Of Service is likely to be
3522 transitory because as soon as the connection fails the memory is
3523 subsequently freed again in the SSL_free() call. However there is an
3524 increased risk during this period of application crashes due to the lack of
3525 memory - which would then mean a more serious Denial of Service.
3527 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3528 (CVE-2016-6307 and CVE-2016-6308)
3532 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3533 had to be removed. Primary reason is that vendor assembler can't
3534 assemble our modules with -KPIC flag. As result it, assembly
3535 support, was not even available as option. But its lack means
3536 lack of side-channel resistant code, which is incompatible with
3537 security by todays standards. Fortunately gcc is readily available
3538 prepackaged option, which we firmly point at...
3542 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3544 * Windows command-line tool supports UTF-8 opt-in option for arguments
3545 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3546 (to any value) allows Windows user to access PKCS#12 file generated
3547 with Windows CryptoAPI and protected with non-ASCII password, as well
3548 as files generated under UTF-8 locale on Linux also protected with
3553 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3554 have been disabled by default and removed from DEFAULT, just like RC4.
3555 See the RC4 item below to re-enable both.
3559 * The method for finding the storage location for the Windows RAND seed file
3560 has changed. First we check %RANDFILE%. If that is not set then we check
3561 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3562 all else fails we fall back to C:\.
3566 * The EVP_EncryptUpdate() function has had its return type changed from void
3567 to int. A return of 0 indicates and error while a return of 1 indicates
3572 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3573 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3574 off the constant time implementation for RSA, DSA and DH have been made
3575 no-ops and deprecated.
3579 * Windows RAND implementation was simplified to only get entropy by
3580 calling CryptGenRandom(). Various other RAND-related tickets
3583 *Joseph Wylie Yandle, Rich Salz*
3585 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3586 and `OPENSSL_LH_`, respectively. The old names are available
3587 with API compatibility. They new names are now completely documented.
3591 * Unify TYPE_up_ref(obj) methods signature.
3592 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3593 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3594 int (instead of void) like all others TYPE_up_ref() methods.
3595 So now these methods also check the return value of CRYPTO_atomic_add(),
3596 and the validity of object reference counter.
3598 *fdasilvayy@gmail.com*
3600 * With Windows Visual Studio builds, the .pdb files are installed
3601 alongside the installed libraries and executables. For a static
3602 library installation, ossl_static.pdb is the associate compiler
3603 generated .pdb file to be used when linking programs.
3607 * Remove openssl.spec. Packaging files belong with the packagers.
3611 * Automatic Darwin/OSX configuration has had a refresh, it will now
3612 recognise x86_64 architectures automatically. You can still decide
3613 to build for a different bitness with the environment variable
3614 KERNEL_BITS (can be 32 or 64), for example:
3616 KERNEL_BITS=32 ./config
3620 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3621 256 bit AES and HMAC with SHA256.
3625 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3629 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3633 * To enable users to have their own config files and build file templates,
3634 Configure looks in the directory indicated by the environment variable
3635 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3636 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3637 name and is used as is.
3641 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3642 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3643 X509_CERT_FILE_CTX was removed.
3647 * "shared" builds are now the default. To create only static libraries use
3648 the "no-shared" Configure option.
3652 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3653 All of these option have not worked for some while and are fundamental
3658 * Make various cleanup routines no-ops and mark them as deprecated. Most
3659 global cleanup functions are no longer required because they are handled
3660 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3661 Explicitly de-initing can cause problems (e.g. where a library that uses
3662 OpenSSL de-inits, but an application is still using it). The affected
3663 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3664 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3665 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3666 COMP_zlib_cleanup().
3670 * --strict-warnings no longer enables runtime debugging options
3671 such as REF_DEBUG. Instead, debug options are automatically
3672 enabled with '--debug' builds.
3674 *Andy Polyakov, Emilia Käsper*
3676 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3677 have been moved out of the public header files. New functions for managing
3678 these have been added.
3682 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3683 objects have been moved out of the public header files. New
3684 functions for managing these have been added.
3688 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3689 have been moved out of the public header files. New functions for managing
3690 these have been added.
3694 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3695 moved out of the public header files. New functions for managing these
3700 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3704 * Removed the mk1mf build scripts.
3708 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3709 it is always safe to #include a header now.
3713 * Removed the aged BC-32 config and all its supporting scripts
3717 * Removed support for Ultrix, Netware, and OS/2.
3721 * Add support for HKDF.
3723 *Alessandro Ghedini*
3725 * Add support for blake2b and blake2s
3729 * Added support for "pipelining". Ciphers that have the
3730 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3731 encryptions/decryptions simultaneously. There are currently no built-in
3732 ciphers with this property but the expectation is that engines will be able
3733 to offer it to significantly improve throughput. Support has been extended
3734 into libssl so that multiple records for a single connection can be
3735 processed in one go (for >=TLS 1.1).
3739 * Added the AFALG engine. This is an async capable engine which is able to
3740 offload work to the Linux kernel. In this initial version it only supports
3741 AES128-CBC. The kernel must be version 4.1.0 or greater.
3745 * OpenSSL now uses a new threading API. It is no longer necessary to
3746 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3747 are two supported threading models: pthreads and windows threads. It is
3748 also possible to configure OpenSSL at compile time for "no-threads". The
3749 old threading API should no longer be used. The functions have been
3750 replaced with "no-op" compatibility macros.
3752 *Alessandro Ghedini, Matt Caswell*
3754 * Modify behavior of ALPN to invoke callback after SNI/servername
3755 callback, such that updates to the SSL_CTX affect ALPN.
3759 * Add SSL_CIPHER queries for authentication and key-exchange.
3763 * Changes to the DEFAULT cipherlist:
3764 - Prefer (EC)DHE handshakes over plain RSA.
3765 - Prefer AEAD ciphers over legacy ciphers.
3766 - Prefer ECDSA over RSA when both certificates are available.
3767 - Prefer TLSv1.2 ciphers/PRF.
3768 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3773 * Change the ECC default curve list to be this, in order: x25519,
3774 secp256r1, secp521r1, secp384r1.
3778 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3779 disabled by default. They can be re-enabled using the
3780 enable-weak-ssl-ciphers option to Configure.
3784 * If the server has ALPN configured, but supports no protocols that the
3785 client advertises, send a fatal "no_application_protocol" alert.
3786 This behaviour is SHALL in RFC 7301, though it isn't universally
3787 implemented by other servers.
3791 * Add X25519 support.
3792 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3793 for public and private key encoding using the format documented in
3794 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3795 key generation and key derivation.
3797 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3802 * Deprecate SRP_VBASE_get_by_user.
3803 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3804 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3805 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3806 seed, even if the seed is configured.
3808 Users should use SRP_VBASE_get1_by_user instead. Note that in
3809 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3810 also that even though configuring the SRP seed attempts to hide
3811 invalid usernames by continuing the handshake with fake
3812 credentials, this behaviour is not constant time and no strong
3813 guarantees are made that the handshake is indistinguishable from
3814 that of a valid user.
3818 * Configuration change; it's now possible to build dynamic engines
3819 without having to build shared libraries and vice versa. This
3820 only applies to the engines in `engines/`, those in `crypto/engine/`
3821 will always be built into libcrypto (i.e. "static").
3823 Building dynamic engines is enabled by default; to disable, use
3824 the configuration option "disable-dynamic-engine".
3826 The only requirements for building dynamic engines are the
3827 presence of the DSO module and building with position independent
3828 code, so they will also automatically be disabled if configuring
3829 with "disable-dso" or "disable-pic".
3831 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3832 are also taken away from openssl/opensslconf.h, as they are
3837 * Configuration change; if there is a known flag to compile
3838 position independent code, it will always be applied on the
3839 libcrypto and libssl object files, and never on the application
3840 object files. This means other libraries that use routines from
3841 libcrypto / libssl can be made into shared libraries regardless
3842 of how OpenSSL was configured.
3844 If this isn't desirable, the configuration options "disable-pic"
3845 or "no-pic" can be used to disable the use of PIC. This will
3846 also disable building shared libraries and dynamic engines.
3850 * Removed JPAKE code. It was experimental and has no wide use.
3854 * The INSTALL_PREFIX Makefile variable has been renamed to
3855 DESTDIR. That makes for less confusion on what this variable
3856 is for. Also, the configuration option --install_prefix is
3861 * Heartbeat for TLS has been removed and is disabled by default
3862 for DTLS; configure with enable-heartbeats. Code that uses the
3863 old #define's might need to be updated.
3865 *Emilia Käsper, Rich Salz*
3867 * Rename REF_CHECK to REF_DEBUG.
3871 * New "unified" build system
3873 The "unified" build system is aimed to be a common system for all
3874 platforms we support. With it comes new support for VMS.
3876 This system builds supports building in a different directory tree
3877 than the source tree. It produces one Makefile (for unix family
3878 or lookalikes), or one descrip.mms (for VMS).
3880 The source of information to make the Makefile / descrip.mms is
3881 small files called 'build.info', holding the necessary
3882 information for each directory with source to compile, and a
3883 template in Configurations, like unix-Makefile.tmpl or
3886 With this change, the library names were also renamed on Windows
3887 and on VMS. They now have names that are closer to the standard
3888 on Unix, and include the major version number, and in certain
3889 cases, the architecture they are built for. See "Notes on shared
3890 libraries" in INSTALL.
3892 We rely heavily on the perl module Text::Template.
3896 * Added support for auto-initialisation and de-initialisation of the library.
3897 OpenSSL no longer requires explicit init or deinit routines to be called,
3898 except in certain circumstances. See the OPENSSL_init_crypto() and
3899 OPENSSL_init_ssl() man pages for further information.
3903 * The arguments to the DTLSv1_listen function have changed. Specifically the
3904 "peer" argument is now expected to be a BIO_ADDR object.
3906 * Rewrite of BIO networking library. The BIO library lacked consistent
3907 support of IPv6, and adding it required some more extensive
3908 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3909 which hold all types of addresses and chains of address information.
3910 It also introduces a new API, with functions like BIO_socket,
3911 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3912 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3913 have been adapted accordingly.
3917 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3922 * CRIME protection: disable compression by default, even if OpenSSL is
3923 compiled with zlib enabled. Applications can still enable compression
3924 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3925 using the SSL_CONF library to configure compression.
3929 * The signature of the session callback configured with
3930 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3931 was explicitly marked as `const unsigned char*` instead of
3936 * Always DPURIFY. Remove the use of uninitialized memory in the
3937 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3941 * Removed many obsolete configuration items, including
3942 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3943 MD2_CHAR, MD2_INT, MD2_LONG
3945 IDEA_SHORT, IDEA_LONG
3946 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3948 *Rich Salz, with advice from Andy Polyakov*
3950 * Many BN internals have been moved to an internal header file.
3952 *Rich Salz with help from Andy Polyakov*
3954 * Configuration and writing out the results from it has changed.
3955 Files such as Makefile include/openssl/opensslconf.h and are now
3956 produced through general templates, such as Makefile.in and
3957 crypto/opensslconf.h.in and some help from the perl module
3960 Also, the center of configuration information is no longer
3961 Makefile. Instead, Configure produces a perl module in
3962 configdata.pm which holds most of the config data (in the hash
3963 table %config), the target data that comes from the target
3964 configuration in one of the `Configurations/*.conf` files (in
3969 * To clarify their intended purposes, the Configure options
3970 --prefix and --openssldir change their semantics, and become more
3971 straightforward and less interdependent.
3973 --prefix shall be used exclusively to give the location INSTALLTOP
3974 where programs, scripts, libraries, include files and manuals are
3975 going to be installed. The default is now /usr/local.
3977 --openssldir shall be used exclusively to give the default
3978 location OPENSSLDIR where certificates, private keys, CRLs are
3979 managed. This is also where the default openssl.cnf gets
3981 If the directory given with this option is a relative path, the
3982 values of both the --prefix value and the --openssldir value will
3983 be combined to become OPENSSLDIR.
3984 The default for --openssldir is INSTALLTOP/ssl.
3986 Anyone who uses --openssldir to specify where OpenSSL is to be
3987 installed MUST change to use --prefix instead.
3991 * The GOST engine was out of date and therefore it has been removed. An up
3992 to date GOST engine is now being maintained in an external repository.
3993 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3994 support for GOST ciphersuites (these are only activated if a GOST engine
3999 * EGD is no longer supported by default; use enable-egd when
4002 *Ben Kaduk and Rich Salz*
4004 * The distribution now has Makefile.in files, which are used to
4005 create Makefile's when Configure is run. *Configure must be run
4006 before trying to build now.*
4010 * The return value for SSL_CIPHER_description() for error conditions
4015 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4017 Obtaining and performing DNSSEC validation of TLSA records is
4018 the application's responsibility. The application provides
4019 the TLSA records of its choice to OpenSSL, and these are then
4020 used to authenticate the peer.
4022 The TLSA records need not even come from DNS. They can, for
4023 example, be used to implement local end-entity certificate or
4024 trust-anchor "pinning", where the "pin" data takes the form
4025 of TLSA records, which can augment or replace verification
4026 based on the usual WebPKI public certification authorities.
4030 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4031 continues to support deprecated interfaces in default builds.
4032 However, applications are strongly advised to compile their
4033 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4034 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4035 or the 1.1.0 releases.
4037 In environments in which all applications have been ported to
4038 not use any deprecated interfaces OpenSSL's Configure script
4039 should be used with the --api=1.1.0 option to entirely remove
4040 support for the deprecated features from the library and
4041 unconditionally disable them in the installed headers.
4042 Essentially the same effect can be achieved with the "no-deprecated"
4043 argument to Configure, except that this will always restrict
4044 the build to just the latest API, rather than a fixed API
4047 As applications are ported to future revisions of the API,
4048 they should update their compile-time OPENSSL_API_COMPAT define
4049 accordingly, but in most cases should be able to continue to
4050 compile with later releases.
4052 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4053 0x10000000L and 0x00908000L, respectively. However those
4054 versions did not support the OPENSSL_API_COMPAT feature, and
4055 so applications are not typically tested for explicit support
4056 of just the undeprecated features of either release.
4060 * Add support for setting the minimum and maximum supported protocol.
4061 It can bet set via the SSL_set_min_proto_version() and
4062 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4063 MaxProtocol. It's recommended to use the new APIs to disable
4064 protocols instead of disabling individual protocols using
4065 SSL_set_options() or SSL_CONF's Protocol. This change also
4066 removes support for disabling TLS 1.2 in the OpenSSL TLS
4067 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4071 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4075 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4076 and integrates ECDSA and ECDH functionality into EC. Implementations can
4077 now redirect key generation and no longer need to convert to or from
4080 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4081 include the ec.h header file instead.
4085 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4086 ciphers who are no longer supported and drops support the ephemeral RSA key
4087 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4091 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4092 opaque. For HMAC_CTX, the following constructors and destructors
4095 HMAC_CTX *HMAC_CTX_new(void);
4096 void HMAC_CTX_free(HMAC_CTX *ctx);
4098 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4099 destroy such methods has been added. See EVP_MD_meth_new(3) and
4100 EVP_CIPHER_meth_new(3) for documentation.
4103 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4104 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4105 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4106 an already created structure.
4107 2) For consistency with the majority of our object creators and
4108 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4109 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4110 for deprecated builds.
4114 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4115 cryptographic operations to be performed asynchronously as long as an
4116 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4117 further details. Libssl has also had this capability integrated with the
4118 introduction of the new mode SSL_MODE_ASYNC and associated error
4119 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4120 pages. This work was developed in partnership with Intel Corp.
4124 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4125 always enabled now. If you want to disable the support you should
4126 exclude it using the list of supported ciphers. This also means that the
4127 "-no_ecdhe" option has been removed from s_server.
4131 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4132 SSL_{CTX_}set1_curves() which can set a list.
4136 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4137 curve you want to support using SSL_{CTX_}set1_curves().
4141 * State machine rewrite. The state machine code has been significantly
4142 refactored in order to remove much duplication of code and solve issues
4143 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4144 further details). This change does have some associated API changes.
4145 Notably the SSL_state() function has been removed and replaced by
4146 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4147 SSL_set_state() has been removed altogether. The previous handshake states
4148 defined in ssl.h and ssl3.h have also been removed.
4152 * All instances of the string "ssleay" in the public API were replaced
4153 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4154 Some error codes related to internal RSA_eay API's were renamed.
4158 * The demo files in crypto/threads were moved to demo/threads.
4162 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4165 *Matt Caswell, Rich Salz*
4167 * New ASN.1 embed macro.
4169 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4170 structure is not allocated: it is part of the parent. That is instead of
4178 This reduces memory fragmentation and make it impossible to accidentally
4179 set a mandatory field to NULL.
4181 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4182 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4183 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4188 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4192 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4193 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4194 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4195 DES and RC4 ciphersuites.
4199 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4200 This changes the decoding behaviour for some invalid messages,
4201 though the change is mostly in the more lenient direction, and
4202 legacy behaviour is preserved as much as possible.
4206 * Fix no-stdio build.
4207 *David Woodhouse <David.Woodhouse@intel.com> and also*
4208 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4210 * New testing framework
4211 The testing framework has been largely rewritten and is now using
4212 perl and the perl modules Test::Harness and an extended variant of
4213 Test::More called OpenSSL::Test to do its work. All test scripts in
4214 test/ have been rewritten into test recipes, and all direct calls to
4215 executables in test/Makefile have become individual recipes using the
4216 simplified testing OpenSSL::Test::Simple.
4218 For documentation on our testing modules, do:
4220 perldoc test/testlib/OpenSSL/Test/Simple.pm
4221 perldoc test/testlib/OpenSSL/Test.pm
4225 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4226 are used; the latter aborts on memory leaks (usually checked on exit).
4227 Some undocumented "set malloc, etc., hooks" functions were removed
4228 and others were changed. All are now documented.
4232 * In DSA_generate_parameters_ex, if the provided seed is too short,
4235 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4237 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4238 from RFC4279, RFC4785, RFC5487, RFC5489.
4240 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4241 original RSA_PSK patch.
4245 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4246 era flag was never set throughout the codebase (only read). Also removed
4247 SSL3_FLAGS_POP_BUFFER which was only used if
4248 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4252 * Changed the default name options in the "ca", "crl", "req" and "x509"
4253 to be "oneline" instead of "compat".
4257 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4258 not aware of clients that still exhibit this bug, and the workaround
4259 hasn't been working properly for a while.
4263 * The return type of BIO_number_read() and BIO_number_written() as well as
4264 the corresponding num_read and num_write members in the BIO structure has
4265 changed from unsigned long to uint64_t. On platforms where an unsigned
4266 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4271 * Given the pervasive nature of TLS extensions it is inadvisable to run
4272 OpenSSL without support for them. It also means that maintaining
4273 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4274 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4278 * Removed support for the two export grade static DH ciphersuites
4279 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4280 were newly added (along with a number of other static DH ciphersuites) to
4281 1.0.2. However the two export ones have *never* worked since they were
4282 introduced. It seems strange in any case to be adding new export
4283 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4287 * Version negotiation has been rewritten. In particular SSLv23_method(),
4288 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4289 and turned into macros which simply call the new preferred function names
4290 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4291 should use the new names instead. Also as part of this change the ssl23.h
4292 header file has been removed.
4296 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4297 code and the associated standard is no longer considered fit-for-purpose.
4301 * RT2547 was closed. When generating a private key, try to make the
4302 output file readable only by the owner. This behavior change might
4303 be noticeable when interacting with other software.
4305 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4310 * Added HTTP GET support to the ocsp command.
4314 * Changed default digest for the dgst and enc commands from MD5 to
4319 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4323 * Added support for TLS extended master secret from
4324 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4325 initial patch which was a great help during development.
4329 * All libssl internal structures have been removed from the public header
4330 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4331 now redundant). Users should not attempt to access internal structures
4332 directly. Instead they should use the provided API functions.
4336 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4337 Access to deprecated functions can be re-enabled by running config with
4338 "enable-deprecated". In addition applications wishing to use deprecated
4339 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4340 will, by default, disable some transitive includes that previously existed
4341 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4345 * Added support for OCB mode. OpenSSL has been granted a patent license
4346 compatible with the OpenSSL license for use of OCB. Details are available
4347 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4348 for OCB can be removed by calling config with no-ocb.
4352 * SSLv2 support has been removed. It still supports receiving a SSLv2
4353 compatible client hello.
4357 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4358 done while fixing the error code for the key-too-small case.
4360 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4362 * CA.sh has been removed; use CA.pl instead.
4366 * Removed old DES API.
4370 * Remove various unsupported platforms:
4376 Sinix/ReliantUNIX RM400
4381 16-bit platforms such as WIN16
4385 * Clean up OPENSSL_NO_xxx #define's
4386 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4387 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4388 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4389 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4390 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4391 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4392 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4393 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4394 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4395 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4399 * Cleaned up dead code
4400 Remove all but one '#ifdef undef' which is to be looked at.
4404 * Clean up calling of xxx_free routines.
4405 Just like free(), fix most of the xxx_free routines to accept
4406 NULL. Remove the non-null checks from callers. Save much code.
4410 * Add secure heap for storage of private keys (when possible).
4411 Add BIO_s_secmem(), CBIGNUM, etc.
4412 Contributed by Akamai Technologies under our Corporate CLA.
4416 * Experimental support for a new, fast, unbiased prime candidate generator,
4417 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4419 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4421 * New output format NSS in the sess_id command line tool. This allows
4422 exporting the session id and the master key in NSS keylog format.
4424 *Martin Kaiser <martin@kaiser.cx>*
4426 * Harmonize version and its documentation. -f flag is used to display
4429 *mancha <mancha1@zoho.com>*
4431 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4432 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4434 *mancha <mancha1@zoho.com>*
4436 * Fix some double frees. These are not thought to be exploitable.
4438 *mancha <mancha1@zoho.com>*
4440 * A missing bounds check in the handling of the TLS heartbeat extension
4441 can be used to reveal up to 64k of memory to a connected client or
4444 Thanks for Neel Mehta of Google Security for discovering this bug and to
4445 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4446 preparing the fix ([CVE-2014-0160])
4448 *Adam Langley, Bodo Moeller*
4450 * Fix for the attack described in the paper "Recovering OpenSSL
4451 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4452 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4453 <http://eprint.iacr.org/2014/140>
4455 Thanks to Yuval Yarom and Naomi Benger for discovering this
4456 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4458 *Yuval Yarom and Naomi Benger*
4460 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4461 this fixes a limitation in previous versions of OpenSSL.
4465 * Experimental encrypt-then-mac support.
4467 Experimental support for encrypt then mac from
4468 draft-gutmann-tls-encrypt-then-mac-02.txt
4470 To enable it set the appropriate extension number (0x42 for the test
4471 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4473 For non-compliant peers (i.e. just about everything) this should have no
4476 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4480 * Add EVP support for key wrapping algorithms, to avoid problems with
4481 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4482 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4483 algorithms and include tests cases.
4487 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4492 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4493 MGF1 digest and OAEP label.
4497 * Make openssl verify return errors.
4499 *Chris Palmer <palmer@google.com> and Ben Laurie*
4501 * New function ASN1_TIME_diff to calculate the difference between two
4502 ASN1_TIME structures or one structure and the current time.
4506 * Update fips_test_suite to support multiple command line options. New
4507 test to induce all self test errors in sequence and check expected
4512 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4513 sign or verify all in one operation.
4517 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4518 test programs and fips_test_suite. Includes functionality to parse
4519 the minimal script output of fipsalgest.pl directly.
4523 * Add authorisation parameter to FIPS_module_mode_set().
4527 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4531 * Use separate DRBG fields for internal and external flags. New function
4532 FIPS_drbg_health_check() to perform on demand health checking. Add
4533 generation tests to fips_test_suite with reduced health check interval to
4534 demonstrate periodic health checking. Add "nodh" option to
4535 fips_test_suite to skip very slow DH test.
4539 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4544 * More extensive health check for DRBG checking many more failure modes.
4545 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4546 combination: call this in fips_test_suite.
4550 * Add support for canonical generation of DSA parameter 'g'. See
4553 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4554 POST to handle HMAC cases.
4558 * Add functions FIPS_module_version() and FIPS_module_version_text()
4559 to return numerical and string versions of the FIPS module number.
4563 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4564 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4565 outside the validated module in the FIPS capable OpenSSL.
4569 * Minor change to DRBG entropy callback semantics. In some cases
4570 there is no multiple of the block length between min_len and
4571 max_len. Allow the callback to return more than max_len bytes
4572 of entropy but discard any extra: it is the callback's responsibility
4573 to ensure that the extra data discarded does not impact the
4574 requested amount of entropy.
4578 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4579 information in FIPS186-3, SP800-57 and SP800-131A.
4583 * CCM support via EVP. Interface is very similar to GCM case except we
4584 must supply all data in one chunk (i.e. no update, final) and the
4585 message length must be supplied if AAD is used. Add algorithm test
4590 * Initial version of POST overhaul. Add POST callback to allow the status
4591 of POST to be monitored and/or failures induced. Modify fips_test_suite
4592 to use callback. Always run all selftests even if one fails.
4596 * XTS support including algorithm test driver in the fips_gcmtest program.
4597 Note: this does increase the maximum key length from 32 to 64 bytes but
4598 there should be no binary compatibility issues as existing applications
4599 will never use XTS mode.
4603 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4604 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4605 performs algorithm blocking for unapproved PRNG types. Also do not
4606 set PRNG type in FIPS_mode_set(): leave this to the application.
4607 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4608 the standard OpenSSL PRNG: set additional data to a date time vector.
4612 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4613 This shouldn't present any incompatibility problems because applications
4614 shouldn't be using these directly and any that are will need to rethink
4615 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4619 * Extensive self tests and health checking required by SP800-90 DRBG.
4620 Remove strength parameter from FIPS_drbg_instantiate and always
4621 instantiate at maximum supported strength.
4625 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4629 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4633 * New function DH_compute_key_padded() to compute a DH key and pad with
4634 leading zeroes if needed: this complies with SP800-56A et al.
4638 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4639 anything, incomplete, subject to change and largely untested at present.
4643 * Modify fipscanisteronly build option to only build the necessary object
4644 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4648 * Add experimental option FIPSSYMS to give all symbols in
4649 fipscanister.o and FIPS or fips prefix. This will avoid
4650 conflicts with future versions of OpenSSL. Add perl script
4651 util/fipsas.pl to preprocess assembly language source files
4652 and rename any affected symbols.
4656 * Add selftest checks and algorithm block of non-fips algorithms in
4657 FIPS mode. Remove DES2 from selftests.
4661 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4662 return internal method without any ENGINE dependencies. Add new
4663 tiny fips sign and verify functions.
4667 * New build option no-ec2m to disable characteristic 2 code.
4671 * New build option "fipscanisteronly". This only builds fipscanister.o
4672 and (currently) associated fips utilities. Uses the file Makefile.fips
4673 instead of Makefile.org as the prototype.
4677 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4678 Update fips_gcmtest to use IV generator.
4682 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4683 setting output buffer to NULL. The `*Final` function must be
4684 called although it will not retrieve any additional data. The tag
4685 can be set or retrieved with a ctrl. The IV length is by default 12
4686 bytes (96 bits) but can be set to an alternative value. If the IV
4687 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4692 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4693 underlying do_cipher function handles all cipher semantics itself
4694 including padding and finalisation. This is useful if (for example)
4695 an ENGINE cipher handles block padding itself. The behaviour of
4696 do_cipher is subtly changed if this flag is set: the return value
4697 is the number of characters written to the output buffer (zero is
4698 no longer an error code) or a negative error code. Also if the
4699 input buffer is NULL and length 0 finalisation should be performed.
4703 * If a candidate issuer certificate is already part of the constructed
4704 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4708 * Improve forward-security support: add functions
4710 void SSL_CTX_set_not_resumable_session_callback(
4711 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4712 void SSL_set_not_resumable_session_callback(
4713 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4715 for use by SSL/TLS servers; the callback function will be called whenever a
4716 new session is created, and gets to decide whether the session may be
4717 cached to make it resumable (return 0) or not (return 1). (As by the
4718 SSL/TLS protocol specifications, the session_id sent by the server will be
4719 empty to indicate that the session is not resumable; also, the server will
4720 not generate RFC 4507 (RFC 5077) session tickets.)
4722 A simple reasonable callback implementation is to return is_forward_secure.
4723 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4724 by the SSL/TLS server library, indicating whether it can provide forward
4727 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4729 * New -verify_name option in command line utilities to set verification
4734 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4735 Add CMAC pkey methods.
4739 * Experimental renegotiation in s_server -www mode. If the client
4740 browses /reneg connection is renegotiated. If /renegcert it is
4741 renegotiated requesting a certificate.
4745 * Add an "external" session cache for debugging purposes to s_server. This
4746 should help trace issues which normally are only apparent in deployed
4747 multi-process servers.
4751 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4752 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4753 BIO_set_cipher() and some obscure PEM functions were changed so they
4754 can now return an error. The RAND changes required a change to the
4755 RAND_METHOD structure.
4759 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4760 a gcc attribute to warn if the result of a function is ignored. This
4761 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4762 whose return value is often ignored.
4766 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4767 These allow SCTs (signed certificate timestamps) to be requested and
4768 validated when establishing a connection.
4770 *Rob Percival <robpercival@google.com>*
4775 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4777 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4778 used even when parsing explicit parameters, when loading a encoded key
4779 or calling `EC_GROUP_new_from_ecpkparameters()`/
4780 `EC_GROUP_new_from_ecparameters()`.
4781 This prevents bypass of security hardening and performance gains,
4782 especially for curves with specialized EC_METHODs.
4783 By default, if a key encoded with explicit parameters is loaded and later
4784 encoded, the output is still encoded with explicit parameters, even if
4785 internally a "named" EC_GROUP is used for computation.
4789 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4790 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4791 NULL. After this change, only the cofactor parameter can be NULL. It also
4792 does some minimal sanity checks on the passed order.
4797 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4798 An attack is simple, if the first CMS_recipientInfo is valid but the
4799 second CMS_recipientInfo is chosen ciphertext. If the second
4800 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4801 encryption key will be replaced by garbage, and the message cannot be
4802 decoded, but if the RSA decryption fails, the correct encryption key is
4803 used and the recipient will not notice the attack.
4804 As a work around for this potential attack the length of the decrypted
4805 key must be equal to the cipher default key length, in case the
4806 certifiate is not given and all recipientInfo are tried out.
4807 The old behaviour can be re-enabled in the CMS code by setting the
4808 CMS_DEBUG_DECRYPT flag.
4813 * Document issue with installation paths in diverse Windows builds
4815 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4816 binaries and run-time config file.
4821 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4823 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4824 This changes the size when using the `genpkey` command when no size is given.
4825 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4826 generation commands to use 2048 bits by default.
4830 * Add FIPS support for Android Arm 64-bit
4832 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4833 Module in Version 2.0.10. For some reason, the corresponding target
4834 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4835 built with FIPS support on Android Arm 64-bit. This omission has been
4838 *Matthias St. Pierre*
4840 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4842 * 0-byte record padding oracle
4844 If an application encounters a fatal protocol error and then calls
4845 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4846 then OpenSSL can respond differently to the calling application if a 0 byte
4847 record is received with invalid padding compared to if a 0 byte record is
4848 received with an invalid MAC. If the application then behaves differently
4849 based on that in a way that is detectable to the remote peer, then this
4850 amounts to a padding oracle that could be used to decrypt data.
4852 In order for this to be exploitable "non-stitched" ciphersuites must be in
4853 use. Stitched ciphersuites are optimised implementations of certain
4854 commonly used ciphersuites. Also the application must call SSL_shutdown()
4855 twice even if a protocol error has occurred (applications should not do
4856 this but some do anyway).
4858 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4859 Aviram, with additional investigation by Steven Collison and Andrew
4860 Hourselt. It was reported to OpenSSL on 10th December 2018.
4865 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4869 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4871 * Microarchitecture timing vulnerability in ECC scalar multiplication
4873 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4874 shown to be vulnerable to a microarchitecture timing side channel attack.
4875 An attacker with sufficient access to mount local timing attacks during
4876 ECDSA signature generation could recover the private key.
4878 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4879 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4885 * Timing vulnerability in DSA signature generation
4887 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4888 timing side channel attack. An attacker could use variations in the signing
4889 algorithm to recover the private key.
4891 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4896 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4897 Module, accidentally introduced while backporting security fixes from the
4898 development branch and hindering the use of ECC in FIPS mode.
4902 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4904 * Client DoS due to large DH parameter
4906 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4907 malicious server can send a very large prime value to the client. This will
4908 cause the client to spend an unreasonably long period of time generating a
4909 key for this prime resulting in a hang until the client has finished. This
4910 could be exploited in a Denial Of Service attack.
4912 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4917 * Cache timing vulnerability in RSA Key Generation
4919 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4920 a cache timing side channel attack. An attacker with sufficient access to
4921 mount cache timing attacks during the RSA key generation process could
4922 recover the private key.
4924 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4925 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4930 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4931 parameter is no longer accepted, as it leads to a corrupt table. NULL
4932 pem_str is reserved for alias entries only.
4936 * Revert blinding in ECDSA sign and instead make problematic addition
4937 length-invariant. Switch even to fixed-length Montgomery multiplication.
4941 * Change generating and checking of primes so that the error rate of not
4942 being prime depends on the intended use based on the size of the input.
4943 For larger primes this will result in more rounds of Miller-Rabin.
4944 The maximal error rate for primes with more than 1080 bits is lowered
4947 *Kurt Roeckx, Annie Yousar*
4949 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4953 * Add blinding to ECDSA and DSA signatures to protect against side channel
4954 attacks discovered by Keegan Ryan (NCC Group).
4958 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4959 now allow empty (zero character) pass phrases.
4963 * Certificate time validation (X509_cmp_time) enforces stricter
4964 compliance with RFC 5280. Fractional seconds and timezone offsets
4965 are no longer allowed.
4969 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4971 * Constructed ASN.1 types with a recursive definition could exceed the stack
4973 Constructed ASN.1 types with a recursive definition (such as can be found
4974 in PKCS7) could eventually exceed the stack given malicious input with
4975 excessive recursion. This could result in a Denial Of Service attack. There
4976 are no such structures used within SSL/TLS that come from untrusted sources
4977 so this is considered safe.
4979 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4985 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4987 * Read/write after SSL object in error state
4989 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4990 mechanism. The intent was that if a fatal error occurred during a handshake
4991 then OpenSSL would move into the error state and would immediately fail if
4992 you attempted to continue the handshake. This works as designed for the
4993 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4994 SSL_connect()), however due to a bug it does not work correctly if
4995 SSL_read() or SSL_write() is called directly. In that scenario, if the
4996 handshake fails then a fatal error will be returned in the initial function
4997 call. If SSL_read()/SSL_write() is subsequently called by the application
4998 for the same SSL object then it will succeed and the data is passed without
4999 being decrypted/encrypted directly from the SSL/TLS record layer.
5001 In order to exploit this issue an application bug would have to be present
5002 that resulted in a call to SSL_read()/SSL_write() being issued after having
5003 already received a fatal error.
5005 This issue was reported to OpenSSL by David Benjamin (Google).
5010 * rsaz_1024_mul_avx2 overflow bug on x86_64
5012 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5013 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5014 Analysis suggests that attacks against RSA and DSA as a result of this
5015 defect would be very difficult to perform and are not believed likely.
5016 Attacks against DH1024 are considered just feasible, because most of the
5017 work necessary to deduce information about a private key may be performed
5018 offline. The amount of resources required for such an attack would be
5019 significant. However, for an attack on TLS to be meaningful, the server
5020 would have to share the DH1024 private key among multiple clients, which is
5021 no longer an option since CVE-2016-0701.
5023 This only affects processors that support the AVX2 but not ADX extensions
5024 like Intel Haswell (4th generation).
5026 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5027 was originally found via the OSS-Fuzz project.
5032 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5034 * bn_sqrx8x_internal carry bug on x86_64
5036 There is a carry propagating bug in the x86_64 Montgomery squaring
5037 procedure. No EC algorithms are affected. Analysis suggests that attacks
5038 against RSA and DSA as a result of this defect would be very difficult to
5039 perform and are not believed likely. Attacks against DH are considered just
5040 feasible (although very difficult) because most of the work necessary to
5041 deduce information about a private key may be performed offline. The amount
5042 of resources required for such an attack would be very significant and
5043 likely only accessible to a limited number of attackers. An attacker would
5044 additionally need online access to an unpatched system using the target
5045 private key in a scenario with persistent DH parameters and a private
5046 key that is shared between multiple clients.
5048 This only affects processors that support the BMI1, BMI2 and ADX extensions
5049 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5051 This issue was reported to OpenSSL by the OSS-Fuzz project.
5056 * Malformed X.509 IPAddressFamily could cause OOB read
5058 If an X.509 certificate has a malformed IPAddressFamily extension,
5059 OpenSSL could do a one-byte buffer overread. The most likely result
5060 would be an erroneous display of the certificate in text format.
5062 This issue was reported to OpenSSL by the OSS-Fuzz project.
5066 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5068 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5069 platform rather than 'mingw'.
5073 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5075 * Truncated packet could crash via OOB read
5077 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5078 cipher is being used, then a truncated packet can cause that host to
5079 perform an out-of-bounds read, usually resulting in a crash.
5081 This issue was reported to OpenSSL by Robert Święcki of Google.
5086 * BN_mod_exp may produce incorrect results on x86_64
5088 There is a carry propagating bug in the x86_64 Montgomery squaring
5089 procedure. No EC algorithms are affected. Analysis suggests that attacks
5090 against RSA and DSA as a result of this defect would be very difficult to
5091 perform and are not believed likely. Attacks against DH are considered just
5092 feasible (although very difficult) because most of the work necessary to
5093 deduce information about a private key may be performed offline. The amount
5094 of resources required for such an attack would be very significant and
5095 likely only accessible to a limited number of attackers. An attacker would
5096 additionally need online access to an unpatched system using the target
5097 private key in a scenario with persistent DH parameters and a private
5098 key that is shared between multiple clients. For example this can occur by
5099 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5100 similar to CVE-2015-3193 but must be treated as a separate problem.
5102 This issue was reported to OpenSSL by the OSS-Fuzz project.
5107 * Montgomery multiplication may produce incorrect results
5109 There is a carry propagating bug in the Broadwell-specific Montgomery
5110 multiplication procedure that handles input lengths divisible by, but
5111 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5112 and DH private keys are impossible. This is because the subroutine in
5113 question is not used in operations with the private key itself and an input
5114 of the attacker's direct choice. Otherwise the bug can manifest itself as
5115 transient authentication and key negotiation failures or reproducible
5116 erroneous outcome of public-key operations with specially crafted input.
5117 Among EC algorithms only Brainpool P-512 curves are affected and one
5118 presumably can attack ECDH key negotiation. Impact was not analyzed in
5119 detail, because pre-requisites for attack are considered unlikely. Namely
5120 multiple clients have to choose the curve in question and the server has to
5121 share the private key among them, neither of which is default behaviour.
5122 Even then only clients that chose the curve will be affected.
5124 This issue was publicly reported as transient failures and was not
5125 initially recognized as a security issue. Thanks to Richard Morgan for
5126 providing reproducible case.
5131 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5132 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5133 prevent issues where no progress is being made and the peer continually
5134 sends unrecognised record types, using up resources processing them.
5138 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5140 * Missing CRL sanity check
5142 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5143 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5144 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5146 This issue only affects the OpenSSL 1.0.2i
5151 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5153 * OCSP Status Request extension unbounded memory growth
5155 A malicious client can send an excessively large OCSP Status Request
5156 extension. If that client continually requests renegotiation, sending a
5157 large OCSP Status Request extension each time, then there will be unbounded
5158 memory growth on the server. This will eventually lead to a Denial Of
5159 Service attack through memory exhaustion. Servers with a default
5160 configuration are vulnerable even if they do not support OCSP. Builds using
5161 the "no-ocsp" build time option are not affected.
5163 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5168 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5171 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5177 * OOB write in MDC2_Update()
5179 An overflow can occur in MDC2_Update() either if called directly or
5180 through the EVP_DigestUpdate() function using MDC2. If an attacker
5181 is able to supply very large amounts of input data after a previous
5182 call to EVP_EncryptUpdate() with a partial block then a length check
5183 can overflow resulting in a heap corruption.
5185 The amount of data needed is comparable to SIZE_MAX which is impractical
5188 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5193 * Malformed SHA512 ticket DoS
5195 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5196 DoS attack where a malformed ticket will result in an OOB read which will
5199 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5200 a custom server callback and ticket lookup mechanism.
5202 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5207 * OOB write in BN_bn2dec()
5209 The function BN_bn2dec() does not check the return value of BN_div_word().
5210 This can cause an OOB write if an application uses this function with an
5211 overly large BIGNUM. This could be a problem if an overly large certificate
5212 or CRL is printed out from an untrusted source. TLS is not affected because
5213 record limits will reject an oversized certificate before it is parsed.
5215 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5220 * OOB read in TS_OBJ_print_bio()
5222 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5223 the total length the OID text representation would use and not the amount
5224 of data written. This will result in OOB reads when large OIDs are
5227 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5232 * Pointer arithmetic undefined behaviour
5234 Avoid some undefined pointer arithmetic
5236 A common idiom in the codebase is to check limits in the following manner:
5239 Where "p" points to some malloc'd data of SIZE bytes and
5242 "len" here could be from some externally supplied data (e.g. from a TLS
5245 The rules of C pointer arithmetic are such that "p + len" is only well
5246 defined where len <= SIZE. Therefore the above idiom is actually
5247 undefined behaviour.
5249 For example this could cause problems if some malloc implementation
5250 provides an address for "p" such that "p + len" actually overflows for
5251 values of len that are too big and therefore p + len < limit.
5253 This issue was reported to OpenSSL by Guido Vranken
5258 * Constant time flag not preserved in DSA signing
5260 Operations in the DSA signing algorithm should run in constant time in
5261 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5262 implementation means that a non-constant time codepath is followed for
5263 certain operations. This has been demonstrated through a cache-timing
5264 attack to be sufficient for an attacker to recover the private DSA key.
5266 This issue was reported by César Pereida (Aalto University), Billy Brumley
5267 (Tampere University of Technology), and Yuval Yarom (The University of
5268 Adelaide and NICTA).
5273 * DTLS buffered message DoS
5275 In a DTLS connection where handshake messages are delivered out-of-order
5276 those messages that OpenSSL is not yet ready to process will be buffered
5277 for later use. Under certain circumstances, a flaw in the logic means that
5278 those messages do not get removed from the buffer even though the handshake
5279 has been completed. An attacker could force up to approx. 15 messages to
5280 remain in the buffer when they are no longer required. These messages will
5281 be cleared when the DTLS connection is closed. The default maximum size for
5282 a message is 100k. Therefore the attacker could force an additional 1500k
5283 to be consumed per connection. By opening many simulataneous connections an
5284 attacker could cause a DoS attack through memory exhaustion.
5286 This issue was reported to OpenSSL by Quan Luo.
5291 * DTLS replay protection DoS
5293 A flaw in the DTLS replay attack protection mechanism means that records
5294 that arrive for future epochs update the replay protection "window" before
5295 the MAC for the record has been validated. This could be exploited by an
5296 attacker by sending a record for the next epoch (which does not have to
5297 decrypt or have a valid MAC), with a very large sequence number. This means
5298 that all subsequent legitimate packets are dropped causing a denial of
5299 service for a specific DTLS connection.
5301 This issue was reported to OpenSSL by the OCAP audit team.
5306 * Certificate message OOB reads
5308 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5309 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5310 theoretical DoS risk but this has not been observed in practice on common
5313 The messages affected are client certificate, client certificate request
5314 and server certificate. As a result the attack can only be performed
5315 against a client or a server which enables client authentication.
5317 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5322 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5324 * Prevent padding oracle in AES-NI CBC MAC check
5326 A MITM attacker can use a padding oracle attack to decrypt traffic
5327 when the connection uses an AES CBC cipher and the server support
5330 This issue was introduced as part of the fix for Lucky 13 padding
5331 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5332 constant time by making sure that always the same bytes are read and
5333 compared against either the MAC or padding bytes. But it no longer
5334 checked that there was enough data to have both the MAC and padding
5337 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5341 * Fix EVP_EncodeUpdate overflow
5343 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5344 Base64 encoding of binary data. If an attacker is able to supply very large
5345 amounts of input data then a length check can overflow resulting in a heap
5348 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5349 the `PEM_write_bio*` family of functions. These are mainly used within the
5350 OpenSSL command line applications, so any application which processes data
5351 from an untrusted source and outputs it as a PEM file should be considered
5352 vulnerable to this issue. User applications that call these APIs directly
5353 with large amounts of untrusted data may also be vulnerable.
5355 This issue was reported by Guido Vranken.
5360 * Fix EVP_EncryptUpdate overflow
5362 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5363 is able to supply very large amounts of input data after a previous call to
5364 EVP_EncryptUpdate() with a partial block then a length check can overflow
5365 resulting in a heap corruption. Following an analysis of all OpenSSL
5366 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5367 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5368 the first called function after an EVP_EncryptInit(), and therefore that
5369 specific call must be safe. The second form is where the length passed to
5370 EVP_EncryptUpdate() can be seen from the code to be some small value and
5371 therefore there is no possibility of an overflow. Since all instances are
5372 one of these two forms, it is believed that there can be no overflows in
5373 internal code due to this problem. It should be noted that
5374 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5375 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5376 of these calls have also been analysed too and it is believed there are no
5377 instances in internal usage where an overflow could occur.
5379 This issue was reported by Guido Vranken.
5384 * Prevent ASN.1 BIO excessive memory allocation
5386 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5387 a short invalid encoding can cause allocation of large amounts of memory
5388 potentially consuming excessive resources or exhausting memory.
5390 Any application parsing untrusted data through d2i BIO functions is
5391 affected. The memory based functions such as d2i_X509() are *not* affected.
5392 Since the memory based functions are used by the TLS library, TLS
5393 applications are not affected.
5395 This issue was reported by Brian Carpenter.
5402 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5403 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5404 in arbitrary stack data being returned in the buffer.
5406 This issue was reported by Guido Vranken.
5411 * Modify behavior of ALPN to invoke callback after SNI/servername
5412 callback, such that updates to the SSL_CTX affect ALPN.
5416 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5421 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5422 methods are enabled and ssl2 is disabled the methods return NULL.
5426 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5428 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5429 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5430 provide any "EXPORT" or "LOW" strength ciphers.
5434 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5435 is by default disabled at build-time. Builds that are not configured with
5436 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5437 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5438 will need to explicitly call either of:
5440 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5442 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5444 as appropriate. Even if either of those is used, or the application
5445 explicitly uses the version-specific SSLv2_method() or its client and
5446 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5447 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5448 ciphers, and SSLv2 56-bit DES are no longer available.
5453 * Fix a double-free in DSA code
5455 A double free bug was discovered when OpenSSL parses malformed DSA private
5456 keys and could lead to a DoS attack or memory corruption for applications
5457 that receive DSA private keys from untrusted sources. This scenario is
5460 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5466 * Disable SRP fake user seed to address a server memory leak.
5468 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5470 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5471 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5472 was changed to ignore the "fake user" SRP seed, even if the seed
5475 Users should use SRP_VBASE_get1_by_user instead. Note that in
5476 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5477 also that even though configuring the SRP seed attempts to hide
5478 invalid usernames by continuing the handshake with fake
5479 credentials, this behaviour is not constant time and no strong
5480 guarantees are made that the handshake is indistinguishable from
5481 that of a valid user.
5486 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5488 In the BN_hex2bn function the number of hex digits is calculated using an
5489 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5490 large values of `i` this can result in `bn_expand` not allocating any
5491 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5492 field as NULL leading to a subsequent NULL ptr deref. For very large values
5493 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5494 In this case memory is allocated to the internal BIGNUM data field, but it
5495 is insufficiently sized leading to heap corruption. A similar issue exists
5496 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5497 is ever called by user applications with very large untrusted hex/dec data.
5498 This is anticipated to be a rare occurrence.
5500 All OpenSSL internal usage of these functions use data that is not expected
5501 to be untrusted, e.g. config file data or application command line
5502 arguments. If user developed applications generate config file data based
5503 on untrusted data then it is possible that this could also lead to security
5504 consequences. This is also anticipated to be rare.
5506 This issue was reported to OpenSSL by Guido Vranken.
5511 * Fix memory issues in `BIO_*printf` functions
5513 The internal `fmtstr` function used in processing a "%s" format string in
5514 the `BIO_*printf` functions could overflow while calculating the length of a
5515 string and cause an OOB read when printing very long strings.
5517 Additionally the internal `doapr_outch` function can attempt to write to an
5518 OOB memory location (at an offset from the NULL pointer) in the event of a
5519 memory allocation failure. In 1.0.2 and below this could be caused where
5520 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5521 could be in processing a very long "%s" format string. Memory leaks can
5524 The first issue may mask the second issue dependent on compiler behaviour.
5525 These problems could enable attacks where large amounts of untrusted data
5526 is passed to the `BIO_*printf` functions. If applications use these functions
5527 in this way then they could be vulnerable. OpenSSL itself uses these
5528 functions when printing out human-readable dumps of ASN.1 data. Therefore
5529 applications that print this data could be vulnerable if the data is from
5530 untrusted sources. OpenSSL command line applications could also be
5531 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5532 as command line arguments.
5534 Libssl is not considered directly vulnerable. Additionally certificates etc
5535 received via remote connections via libssl are also unlikely to be able to
5536 trigger these issues because of message size limits enforced within libssl.
5538 This issue was reported to OpenSSL Guido Vranken.
5543 * Side channel attack on modular exponentiation
5545 A side-channel attack was found which makes use of cache-bank conflicts on
5546 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5547 of RSA keys. The ability to exploit this issue is limited as it relies on
5548 an attacker who has control of code in a thread running on the same
5549 hyper-threaded core as the victim thread which is performing decryptions.
5551 This issue was reported to OpenSSL by Yuval Yarom, The University of
5552 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5553 Nadia Heninger, University of Pennsylvania with more information at
5554 <http://cachebleed.info>.
5559 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5560 if no keysize is specified with default_bits. This fixes an
5561 omission in an earlier change that changed all RSA/DSA key generation
5562 commands to use 2048 bits by default.
5566 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5568 * DH small subgroups
5570 Historically OpenSSL only ever generated DH parameters based on "safe"
5571 primes. More recently (in version 1.0.2) support was provided for
5572 generating X9.42 style parameter files such as those required for RFC 5114
5573 support. The primes used in such files may not be "safe". Where an
5574 application is using DH configured with parameters based on primes that are
5575 not "safe" then an attacker could use this fact to find a peer's private
5576 DH exponent. This attack requires that the attacker complete multiple
5577 handshakes in which the peer uses the same private DH exponent. For example
5578 this could be used to discover a TLS server's private DH exponent if it's
5579 reusing the private DH exponent or it's using a static DH ciphersuite.
5581 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5582 TLS. It is not on by default. If the option is not set then the server
5583 reuses the same private DH exponent for the life of the server process and
5584 would be vulnerable to this attack. It is believed that many popular
5585 applications do set this option and would therefore not be at risk.
5587 The fix for this issue adds an additional check where a "q" parameter is
5588 available (as is the case in X9.42 based parameters). This detects the
5589 only known attack, and is the only possible defense for static DH
5590 ciphersuites. This could have some performance impact.
5592 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5593 default and cannot be disabled. This could have some performance impact.
5595 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5600 * SSLv2 doesn't block disabled ciphers
5602 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5603 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5604 been disabled, provided that the SSLv2 protocol was not also disabled via
5607 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5608 and Sebastian Schinzel.
5613 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5615 * BN_mod_exp may produce incorrect results on x86_64
5617 There is a carry propagating bug in the x86_64 Montgomery squaring
5618 procedure. No EC algorithms are affected. Analysis suggests that attacks
5619 against RSA and DSA as a result of this defect would be very difficult to
5620 perform and are not believed likely. Attacks against DH are considered just
5621 feasible (although very difficult) because most of the work necessary to
5622 deduce information about a private key may be performed offline. The amount
5623 of resources required for such an attack would be very significant and
5624 likely only accessible to a limited number of attackers. An attacker would
5625 additionally need online access to an unpatched system using the target
5626 private key in a scenario with persistent DH parameters and a private
5627 key that is shared between multiple clients. For example this can occur by
5628 default in OpenSSL DHE based SSL/TLS ciphersuites.
5630 This issue was reported to OpenSSL by Hanno Böck.
5635 * Certificate verify crash with missing PSS parameter
5637 The signature verification routines will crash with a NULL pointer
5638 dereference if presented with an ASN.1 signature using the RSA PSS
5639 algorithm and absent mask generation function parameter. Since these
5640 routines are used to verify certificate signature algorithms this can be
5641 used to crash any certificate verification operation and exploited in a
5642 DoS attack. Any application which performs certificate verification is
5643 vulnerable including OpenSSL clients and servers which enable client
5646 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5651 * X509_ATTRIBUTE memory leak
5653 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5654 memory. This structure is used by the PKCS#7 and CMS routines so any
5655 application which reads PKCS#7 or CMS data from untrusted sources is
5656 affected. SSL/TLS is not affected.
5658 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5664 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5665 This changes the decoding behaviour for some invalid messages,
5666 though the change is mostly in the more lenient direction, and
5667 legacy behaviour is preserved as much as possible.
5671 * In DSA_generate_parameters_ex, if the provided seed is too short,
5674 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5676 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5678 * Alternate chains certificate forgery
5680 During certificate verification, OpenSSL will attempt to find an
5681 alternative certificate chain if the first attempt to build such a chain
5682 fails. An error in the implementation of this logic can mean that an
5683 attacker could cause certain checks on untrusted certificates to be
5684 bypassed, such as the CA flag, enabling them to use a valid leaf
5685 certificate to act as a CA and "issue" an invalid certificate.
5687 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5692 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5694 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5695 incompatibility in the handling of HMAC. The previous ABI has now been
5700 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5702 * Malformed ECParameters causes infinite loop
5704 When processing an ECParameters structure OpenSSL enters an infinite loop
5705 if the curve specified is over a specially malformed binary polynomial
5708 This can be used to perform denial of service against any
5709 system which processes public keys, certificate requests or
5710 certificates. This includes TLS clients and TLS servers with
5711 client authentication enabled.
5713 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5718 * Exploitable out-of-bounds read in X509_cmp_time
5720 X509_cmp_time does not properly check the length of the ASN1_TIME
5721 string and can read a few bytes out of bounds. In addition,
5722 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5725 An attacker can use this to craft malformed certificates and CRLs of
5726 various sizes and potentially cause a segmentation fault, resulting in
5727 a DoS on applications that verify certificates or CRLs. TLS clients
5728 that verify CRLs are affected. TLS clients and servers with client
5729 authentication enabled may be affected if they use custom verification
5732 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5733 independently by Hanno Böck.
5738 * PKCS7 crash with missing EnvelopedContent
5740 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5741 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5742 with missing content and trigger a NULL pointer dereference on parsing.
5744 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5745 structures from untrusted sources are affected. OpenSSL clients and
5746 servers are not affected.
5748 This issue was reported to OpenSSL by Michal Zalewski (Google).
5753 * CMS verify infinite loop with unknown hash function
5755 When verifying a signedData message the CMS code can enter an infinite loop
5756 if presented with an unknown hash function OID. This can be used to perform
5757 denial of service against any system which verifies signedData messages using
5759 This issue was reported to OpenSSL by Johannes Bauer.
5764 * Race condition handling NewSessionTicket
5766 If a NewSessionTicket is received by a multi-threaded client when attempting to
5767 reuse a previous ticket then a race condition can occur potentially leading to
5768 a double free of the ticket data.
5773 * Only support 256-bit or stronger elliptic curves with the
5774 'ecdh_auto' setting (server) or by default (client). Of supported
5775 curves, prefer P-256 (both).
5779 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5781 * ClientHello sigalgs DoS fix
5783 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5784 invalid signature algorithms extension a NULL pointer dereference will
5785 occur. This can be exploited in a DoS attack against the server.
5787 This issue was was reported to OpenSSL by David Ramos of Stanford
5791 *Stephen Henson and Matt Caswell*
5793 * Multiblock corrupted pointer fix
5795 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5796 feature only applies on 64 bit x86 architecture platforms that support AES
5797 NI instructions. A defect in the implementation of "multiblock" can cause
5798 OpenSSL's internal write buffer to become incorrectly set to NULL when
5799 using non-blocking IO. Typically, when the user application is using a
5800 socket BIO for writing, this will only result in a failed connection.
5801 However if some other BIO is used then it is likely that a segmentation
5802 fault will be triggered, thus enabling a potential DoS attack.
5804 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5809 * Segmentation fault in DTLSv1_listen fix
5811 The DTLSv1_listen function is intended to be stateless and processes the
5812 initial ClientHello from many peers. It is common for user code to loop
5813 over the call to DTLSv1_listen until a valid ClientHello is received with
5814 an associated cookie. A defect in the implementation of DTLSv1_listen means
5815 that state is preserved in the SSL object from one invocation to the next
5816 that can lead to a segmentation fault. Errors processing the initial
5817 ClientHello can trigger this scenario. An example of such an error could be
5818 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5821 This issue was reported to OpenSSL by Per Allansson.
5826 * Segmentation fault in ASN1_TYPE_cmp fix
5828 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5829 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5830 certificate signature algorithm consistency this can be used to crash any
5831 certificate verification operation and exploited in a DoS attack. Any
5832 application which performs certificate verification is vulnerable including
5833 OpenSSL clients and servers which enable client authentication.
5838 * Segmentation fault for invalid PSS parameters fix
5840 The signature verification routines will crash with a NULL pointer
5841 dereference if presented with an ASN.1 signature using the RSA PSS
5842 algorithm and invalid parameters. Since these routines are used to verify
5843 certificate signature algorithms this can be used to crash any
5844 certificate verification operation and exploited in a DoS attack. Any
5845 application which performs certificate verification is vulnerable including
5846 OpenSSL clients and servers which enable client authentication.
5848 This issue was was reported to OpenSSL by Brian Carpenter.
5853 * ASN.1 structure reuse memory corruption fix
5855 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5856 memory corruption via an invalid write. Such reuse is and has been
5857 strongly discouraged and is believed to be rare.
5859 Applications that parse structures containing CHOICE or ANY DEFINED BY
5860 components may be affected. Certificate parsing (d2i_X509 and related
5861 functions) are however not affected. OpenSSL clients and servers are
5867 * PKCS7 NULL pointer dereferences fix
5869 The PKCS#7 parsing code does not handle missing outer ContentInfo
5870 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5871 missing content and trigger a NULL pointer dereference on parsing.
5873 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5874 otherwise parse PKCS#7 structures from untrusted sources are
5875 affected. OpenSSL clients and servers are not affected.
5877 This issue was reported to OpenSSL by Michal Zalewski (Google).
5882 * DoS via reachable assert in SSLv2 servers fix
5884 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5885 servers that both support SSLv2 and enable export cipher suites by sending
5886 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5888 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5889 (OpenSSL development team).
5894 * Empty CKE with client auth and DHE fix
5896 If client auth is used then a server can seg fault in the event of a DHE
5897 ciphersuite being selected and a zero length ClientKeyExchange message
5898 being sent by the client. This could be exploited in a DoS attack.
5903 * Handshake with unseeded PRNG fix
5905 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5906 with an unseeded PRNG. The conditions are:
5907 - The client is on a platform where the PRNG has not been seeded
5908 automatically, and the user has not seeded manually
5909 - A protocol specific client method version has been used (i.e. not
5910 SSL_client_methodv23)
5911 - A ciphersuite is used that does not require additional random data from
5912 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5914 If the handshake succeeds then the client random that has been used will
5915 have been generated from a PRNG with insufficient entropy and therefore the
5916 output may be predictable.
5918 For example using the following command with an unseeded openssl will
5919 succeed on an unpatched platform:
5921 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5926 * Use After Free following d2i_ECPrivatekey error fix
5928 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5929 could cause a use after free condition. This, in turn, could cause a double
5930 free in several private key parsing functions (such as d2i_PrivateKey
5931 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5932 for applications that receive EC private keys from untrusted
5933 sources. This scenario is considered rare.
5935 This issue was discovered by the BoringSSL project and fixed in their
5941 * X509_to_X509_REQ NULL pointer deref fix
5943 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5944 the certificate key is invalid. This function is rarely used in practice.
5946 This issue was discovered by Brian Carpenter.
5951 * Removed the export ciphers from the DEFAULT ciphers
5955 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5957 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5958 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5959 So far those who have to target multiple platforms would compromise
5960 and argue that binary targeting say ARMv5 would still execute on
5961 ARMv8. "Universal" build resolves this compromise by providing
5962 near-optimal performance even on newer platforms.
5966 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5967 (other platforms pending).
5969 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5971 * Add support for the SignedCertificateTimestampList certificate and
5972 OCSP response extensions from RFC6962.
5976 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5977 for corner cases. (Certain input points at infinity could lead to
5978 bogus results, with non-infinity inputs mapped to infinity too.)
5982 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5983 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5984 common cases are optimized and there still is room for further
5985 improvements. Vector Permutation AES for Altivec is also added.
5989 * Add support for little-endian ppc64 Linux target.
5991 *Marcelo Cerri (IBM)*
5993 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5994 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5995 are optimized and there still is room for further improvements.
5996 Both 32- and 64-bit modes are supported.
5998 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6000 * Improved ARMv7 NEON support.
6004 * Support for SPARC Architecture 2011 crypto extensions, first
6005 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6006 SHA256/512, MD5, GHASH and modular exponentiation.
6008 *Andy Polyakov, David Miller*
6010 * Accelerated modular exponentiation for Intel processors, a.k.a.
6013 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6015 * Support for new and upcoming Intel processors, including AVX2,
6016 BMI and SHA ISA extensions. This includes additional "stitched"
6017 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6020 This work was sponsored by Intel Corp.
6024 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6025 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6026 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6030 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6031 this fixes a limitation in previous versions of OpenSSL.
6035 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6036 MGF1 digest and OAEP label.
6040 * Add EVP support for key wrapping algorithms, to avoid problems with
6041 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6042 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6043 algorithms and include tests cases.
6047 * Add functions to allocate and set the fields of an ECDSA_METHOD
6050 *Douglas E. Engert, Steve Henson*
6052 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6053 difference in days and seconds between two tm or ASN1_TIME structures.
6057 * Add -rev test option to s_server to just reverse order of characters
6058 received by client and send back to server. Also prints an abbreviated
6059 summary of the connection parameters.
6063 * New option -brief for s_client and s_server to print out a brief summary
6064 of connection parameters.
6068 * Add callbacks for arbitrary TLS extensions.
6070 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6072 * New option -crl_download in several openssl utilities to download CRLs
6073 from CRLDP extension in certificates.
6077 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6081 * New function X509_CRL_diff to generate a delta CRL from the difference
6082 of two full CRLs. Add support to "crl" utility.
6086 * New functions to set lookup_crls function and to retrieve
6087 X509_STORE from X509_STORE_CTX.
6091 * Print out deprecated issuer and subject unique ID fields in
6096 * Extend OCSP I/O functions so they can be used for simple general purpose
6097 HTTP as well as OCSP. New wrapper function which can be used to download
6098 CRLs using the OCSP API.
6102 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6106 * `SSL_CONF*` functions. These provide a common framework for application
6107 configuration using configuration files or command lines.
6111 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6112 message callback and prints the results. Needs compile time option
6113 "enable-ssl-trace". New options to s_client and s_server to enable
6118 * New ctrl and macro to retrieve supported points extensions.
6119 Print out extension in s_server and s_client.
6123 * New functions to retrieve certificate signature and signature
6128 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6133 * New Suite B modes for TLS code. These use and enforce the requirements
6134 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6135 only use Suite B curves. The Suite B modes can be set by using the
6136 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6140 * New chain verification flags for Suite B levels of security. Check
6141 algorithms are acceptable when flags are set in X509_verify_cert.
6145 * Make tls1_check_chain return a set of flags indicating checks passed
6146 by a certificate chain. Add additional tests to handle client
6147 certificates: checks for matching certificate type and issuer name
6152 * If an attempt is made to use a signature algorithm not in the peer
6153 preference list abort the handshake. If client has no suitable
6154 signature algorithms in response to a certificate request do not
6155 use the certificate.
6159 * If server EC tmp key is not in client preference list abort handshake.
6163 * Add support for certificate stores in CERT structure. This makes it
6164 possible to have different stores per SSL structure or one store in
6165 the parent SSL_CTX. Include distinct stores for certificate chain
6166 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6167 to build and store a certificate chain in CERT structure: returning
6168 an error if the chain cannot be built: this will allow applications
6169 to test if a chain is correctly configured.
6171 Note: if the CERT based stores are not set then the parent SSL_CTX
6172 store is used to retain compatibility with existing behaviour.
6176 * New function ssl_set_client_disabled to set a ciphersuite disabled
6177 mask based on the current session, check mask when sending client
6178 hello and checking the requested ciphersuite.
6182 * New ctrls to retrieve and set certificate types in a certificate
6183 request message. Print out received values in s_client. If certificate
6184 types is not set with custom values set sensible values based on
6185 supported signature algorithms.
6189 * Support for distinct client and server supported signature algorithms.
6193 * Add certificate callback. If set this is called whenever a certificate
6194 is required by client or server. An application can decide which
6195 certificate chain to present based on arbitrary criteria: for example
6196 supported signature algorithms. Add very simple example to s_server.
6197 This fixes many of the problems and restrictions of the existing client
6198 certificate callback: for example you can now clear an existing
6199 certificate and specify the whole chain.
6203 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6204 the certificate can be used for (if anything). Set valid_flags field
6205 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6206 to have similar checks in it.
6208 Add new "cert_flags" field to CERT structure and include a "strict mode".
6209 This enforces some TLS certificate requirements (such as only permitting
6210 certificate signature algorithms contained in the supported algorithms
6211 extension) which some implementations ignore: this option should be used
6212 with caution as it could cause interoperability issues.
6216 * Update and tidy signature algorithm extension processing. Work out
6217 shared signature algorithms based on preferences and peer algorithms
6218 and print them out in s_client and s_server. Abort handshake if no
6219 shared signature algorithms.
6223 * Add new functions to allow customised supported signature algorithms
6224 for SSL and SSL_CTX structures. Add options to s_client and s_server
6229 * New function SSL_certs_clear() to delete all references to certificates
6230 from an SSL structure. Before this once a certificate had been added
6231 it couldn't be removed.
6235 * Integrate hostname, email address and IP address checking with certificate
6236 verification. New verify options supporting checking in openssl utility.
6240 * Fixes and wildcard matching support to hostname and email checking
6241 functions. Add manual page.
6243 *Florian Weimer (Red Hat Product Security Team)*
6245 * New functions to check a hostname email or IP address against a
6246 certificate. Add options x509 utility to print results of checks against
6251 * Fix OCSP checking.
6253 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6255 * Initial experimental support for explicitly trusted non-root CAs.
6256 OpenSSL still tries to build a complete chain to a root but if an
6257 intermediate CA has a trust setting included that is used. The first
6258 setting is used: whether to trust (e.g., -addtrust option to the x509
6263 * Add -trusted_first option which attempts to find certificates in the
6264 trusted store even if an untrusted chain is also supplied.
6268 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6269 platform support for Linux and Android.
6273 * Support for linux-x32, ILP32 environment in x86_64 framework.
6277 * Experimental multi-implementation support for FIPS capable OpenSSL.
6278 When in FIPS mode the approved implementations are used as normal,
6279 when not in FIPS mode the internal unapproved versions are used instead.
6280 This means that the FIPS capable OpenSSL isn't forced to use the
6281 (often lower performance) FIPS implementations outside FIPS mode.
6285 * Transparently support X9.42 DH parameters when calling
6286 PEM_read_bio_DHparameters. This means existing applications can handle
6287 the new parameter format automatically.
6291 * Initial experimental support for X9.42 DH parameter format: mainly
6292 to support use of 'q' parameter for RFC5114 parameters.
6296 * Add DH parameters from RFC5114 including test data to dhtest.
6300 * Support for automatic EC temporary key parameter selection. If enabled
6301 the most preferred EC parameters are automatically used instead of
6302 hardcoded fixed parameters. Now a server just has to call:
6303 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6304 support ECDH and use the most appropriate parameters.
6308 * Enhance and tidy EC curve and point format TLS extension code. Use
6309 static structures instead of allocation if default values are used.
6310 New ctrls to set curves we wish to support and to retrieve shared curves.
6311 Print out shared curves in s_server. New options to s_server and s_client
6312 to set list of supported curves.
6316 * New ctrls to retrieve supported signature algorithms and
6317 supported curve values as an array of NIDs. Extend openssl utility
6318 to print out received values.
6322 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6323 between NIDs and the more common NIST names such as "P-256". Enhance
6324 ecparam utility and ECC method to recognise the NIST names for curves.
6328 * Enhance SSL/TLS certificate chain handling to support different
6329 chains for each certificate instead of one chain in the parent SSL_CTX.
6333 * Support for fixed DH ciphersuite client authentication: where both
6334 server and client use DH certificates with common parameters.
6338 * Support for fixed DH ciphersuites: those requiring DH server
6343 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6345 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6346 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6347 X509_CINF_get_signature were reverted post internal team review.
6352 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6354 * OCSP Status Request extension unbounded memory growth
6356 A malicious client can send an excessively large OCSP Status Request
6357 extension. If that client continually requests renegotiation, sending a
6358 large OCSP Status Request extension each time, then there will be unbounded
6359 memory growth on the server. This will eventually lead to a Denial Of
6360 Service attack through memory exhaustion. Servers with a default
6361 configuration are vulnerable even if they do not support OCSP. Builds using
6362 the "no-ocsp" build time option are not affected.
6364 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6369 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6372 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6378 * OOB write in MDC2_Update()
6380 An overflow can occur in MDC2_Update() either if called directly or
6381 through the EVP_DigestUpdate() function using MDC2. If an attacker
6382 is able to supply very large amounts of input data after a previous
6383 call to EVP_EncryptUpdate() with a partial block then a length check
6384 can overflow resulting in a heap corruption.
6386 The amount of data needed is comparable to SIZE_MAX which is impractical
6389 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6394 * Malformed SHA512 ticket DoS
6396 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6397 DoS attack where a malformed ticket will result in an OOB read which will
6400 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6401 a custom server callback and ticket lookup mechanism.
6403 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6408 * OOB write in BN_bn2dec()
6410 The function BN_bn2dec() does not check the return value of BN_div_word().
6411 This can cause an OOB write if an application uses this function with an
6412 overly large BIGNUM. This could be a problem if an overly large certificate
6413 or CRL is printed out from an untrusted source. TLS is not affected because
6414 record limits will reject an oversized certificate before it is parsed.
6416 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6421 * OOB read in TS_OBJ_print_bio()
6423 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6424 the total length the OID text representation would use and not the amount
6425 of data written. This will result in OOB reads when large OIDs are
6428 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6433 * Pointer arithmetic undefined behaviour
6435 Avoid some undefined pointer arithmetic
6437 A common idiom in the codebase is to check limits in the following manner:
6440 Where "p" points to some malloc'd data of SIZE bytes and
6443 "len" here could be from some externally supplied data (e.g. from a TLS
6446 The rules of C pointer arithmetic are such that "p + len" is only well
6447 defined where len <= SIZE. Therefore the above idiom is actually
6448 undefined behaviour.
6450 For example this could cause problems if some malloc implementation
6451 provides an address for "p" such that "p + len" actually overflows for
6452 values of len that are too big and therefore p + len < limit.
6454 This issue was reported to OpenSSL by Guido Vranken
6459 * Constant time flag not preserved in DSA signing
6461 Operations in the DSA signing algorithm should run in constant time in
6462 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6463 implementation means that a non-constant time codepath is followed for
6464 certain operations. This has been demonstrated through a cache-timing
6465 attack to be sufficient for an attacker to recover the private DSA key.
6467 This issue was reported by César Pereida (Aalto University), Billy Brumley
6468 (Tampere University of Technology), and Yuval Yarom (The University of
6469 Adelaide and NICTA).
6474 * DTLS buffered message DoS
6476 In a DTLS connection where handshake messages are delivered out-of-order
6477 those messages that OpenSSL is not yet ready to process will be buffered
6478 for later use. Under certain circumstances, a flaw in the logic means that
6479 those messages do not get removed from the buffer even though the handshake
6480 has been completed. An attacker could force up to approx. 15 messages to
6481 remain in the buffer when they are no longer required. These messages will
6482 be cleared when the DTLS connection is closed. The default maximum size for
6483 a message is 100k. Therefore the attacker could force an additional 1500k
6484 to be consumed per connection. By opening many simulataneous connections an
6485 attacker could cause a DoS attack through memory exhaustion.
6487 This issue was reported to OpenSSL by Quan Luo.
6492 * DTLS replay protection DoS
6494 A flaw in the DTLS replay attack protection mechanism means that records
6495 that arrive for future epochs update the replay protection "window" before
6496 the MAC for the record has been validated. This could be exploited by an
6497 attacker by sending a record for the next epoch (which does not have to
6498 decrypt or have a valid MAC), with a very large sequence number. This means
6499 that all subsequent legitimate packets are dropped causing a denial of
6500 service for a specific DTLS connection.
6502 This issue was reported to OpenSSL by the OCAP audit team.
6507 * Certificate message OOB reads
6509 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6510 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6511 theoretical DoS risk but this has not been observed in practice on common
6514 The messages affected are client certificate, client certificate request
6515 and server certificate. As a result the attack can only be performed
6516 against a client or a server which enables client authentication.
6518 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6523 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6525 * Prevent padding oracle in AES-NI CBC MAC check
6527 A MITM attacker can use a padding oracle attack to decrypt traffic
6528 when the connection uses an AES CBC cipher and the server support
6531 This issue was introduced as part of the fix for Lucky 13 padding
6532 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6533 constant time by making sure that always the same bytes are read and
6534 compared against either the MAC or padding bytes. But it no longer
6535 checked that there was enough data to have both the MAC and padding
6538 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6543 * Fix EVP_EncodeUpdate overflow
6545 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6546 Base64 encoding of binary data. If an attacker is able to supply very large
6547 amounts of input data then a length check can overflow resulting in a heap
6550 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6551 the `PEM_write_bio*` family of functions. These are mainly used within the
6552 OpenSSL command line applications, so any application which processes data
6553 from an untrusted source and outputs it as a PEM file should be considered
6554 vulnerable to this issue. User applications that call these APIs directly
6555 with large amounts of untrusted data may also be vulnerable.
6557 This issue was reported by Guido Vranken.
6562 * Fix EVP_EncryptUpdate overflow
6564 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6565 is able to supply very large amounts of input data after a previous call to
6566 EVP_EncryptUpdate() with a partial block then a length check can overflow
6567 resulting in a heap corruption. Following an analysis of all OpenSSL
6568 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6569 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6570 the first called function after an EVP_EncryptInit(), and therefore that
6571 specific call must be safe. The second form is where the length passed to
6572 EVP_EncryptUpdate() can be seen from the code to be some small value and
6573 therefore there is no possibility of an overflow. Since all instances are
6574 one of these two forms, it is believed that there can be no overflows in
6575 internal code due to this problem. It should be noted that
6576 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6577 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6578 of these calls have also been analysed too and it is believed there are no
6579 instances in internal usage where an overflow could occur.
6581 This issue was reported by Guido Vranken.
6586 * Prevent ASN.1 BIO excessive memory allocation
6588 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6589 a short invalid encoding can casuse allocation of large amounts of memory
6590 potentially consuming excessive resources or exhausting memory.
6592 Any application parsing untrusted data through d2i BIO functions is
6593 affected. The memory based functions such as d2i_X509() are *not* affected.
6594 Since the memory based functions are used by the TLS library, TLS
6595 applications are not affected.
6597 This issue was reported by Brian Carpenter.
6604 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6605 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6606 in arbitrary stack data being returned in the buffer.
6608 This issue was reported by Guido Vranken.
6613 * Modify behavior of ALPN to invoke callback after SNI/servername
6614 callback, such that updates to the SSL_CTX affect ALPN.
6618 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6623 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6624 methods are enabled and ssl2 is disabled the methods return NULL.
6628 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6630 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6631 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6632 provide any "EXPORT" or "LOW" strength ciphers.
6636 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6637 is by default disabled at build-time. Builds that are not configured with
6638 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6639 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6640 will need to explicitly call either of:
6642 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6644 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6646 as appropriate. Even if either of those is used, or the application
6647 explicitly uses the version-specific SSLv2_method() or its client and
6648 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6649 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6650 ciphers, and SSLv2 56-bit DES are no longer available.
6655 * Fix a double-free in DSA code
6657 A double free bug was discovered when OpenSSL parses malformed DSA private
6658 keys and could lead to a DoS attack or memory corruption for applications
6659 that receive DSA private keys from untrusted sources. This scenario is
6662 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6668 * Disable SRP fake user seed to address a server memory leak.
6670 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6672 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6673 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6674 was changed to ignore the "fake user" SRP seed, even if the seed
6677 Users should use SRP_VBASE_get1_by_user instead. Note that in
6678 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6679 also that even though configuring the SRP seed attempts to hide
6680 invalid usernames by continuing the handshake with fake
6681 credentials, this behaviour is not constant time and no strong
6682 guarantees are made that the handshake is indistinguishable from
6683 that of a valid user.
6688 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6690 In the BN_hex2bn function the number of hex digits is calculated using an
6691 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6692 large values of `i` this can result in `bn_expand` not allocating any
6693 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6694 field as NULL leading to a subsequent NULL ptr deref. For very large values
6695 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6696 In this case memory is allocated to the internal BIGNUM data field, but it
6697 is insufficiently sized leading to heap corruption. A similar issue exists
6698 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6699 is ever called by user applications with very large untrusted hex/dec data.
6700 This is anticipated to be a rare occurrence.
6702 All OpenSSL internal usage of these functions use data that is not expected
6703 to be untrusted, e.g. config file data or application command line
6704 arguments. If user developed applications generate config file data based
6705 on untrusted data then it is possible that this could also lead to security
6706 consequences. This is also anticipated to be rare.
6708 This issue was reported to OpenSSL by Guido Vranken.
6713 * Fix memory issues in `BIO_*printf` functions
6715 The internal `fmtstr` function used in processing a "%s" format string in
6716 the `BIO_*printf` functions could overflow while calculating the length of a
6717 string and cause an OOB read when printing very long strings.
6719 Additionally the internal `doapr_outch` function can attempt to write to an
6720 OOB memory location (at an offset from the NULL pointer) in the event of a
6721 memory allocation failure. In 1.0.2 and below this could be caused where
6722 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6723 could be in processing a very long "%s" format string. Memory leaks can
6726 The first issue may mask the second issue dependent on compiler behaviour.
6727 These problems could enable attacks where large amounts of untrusted data
6728 is passed to the `BIO_*printf` functions. If applications use these functions
6729 in this way then they could be vulnerable. OpenSSL itself uses these
6730 functions when printing out human-readable dumps of ASN.1 data. Therefore
6731 applications that print this data could be vulnerable if the data is from
6732 untrusted sources. OpenSSL command line applications could also be
6733 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6734 as command line arguments.
6736 Libssl is not considered directly vulnerable. Additionally certificates etc
6737 received via remote connections via libssl are also unlikely to be able to
6738 trigger these issues because of message size limits enforced within libssl.
6740 This issue was reported to OpenSSL Guido Vranken.
6745 * Side channel attack on modular exponentiation
6747 A side-channel attack was found which makes use of cache-bank conflicts on
6748 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6749 of RSA keys. The ability to exploit this issue is limited as it relies on
6750 an attacker who has control of code in a thread running on the same
6751 hyper-threaded core as the victim thread which is performing decryptions.
6753 This issue was reported to OpenSSL by Yuval Yarom, The University of
6754 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6755 Nadia Heninger, University of Pennsylvania with more information at
6756 <http://cachebleed.info>.
6761 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6762 if no keysize is specified with default_bits. This fixes an
6763 omission in an earlier change that changed all RSA/DSA key generation
6764 commands to use 2048 bits by default.
6768 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6770 * Protection for DH small subgroup attacks
6772 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6773 switched on by default and cannot be disabled. This could have some
6778 * SSLv2 doesn't block disabled ciphers
6780 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6781 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6782 been disabled, provided that the SSLv2 protocol was not also disabled via
6785 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6786 and Sebastian Schinzel.
6791 * Reject DH handshakes with parameters shorter than 1024 bits.
6795 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6797 * Certificate verify crash with missing PSS parameter
6799 The signature verification routines will crash with a NULL pointer
6800 dereference if presented with an ASN.1 signature using the RSA PSS
6801 algorithm and absent mask generation function parameter. Since these
6802 routines are used to verify certificate signature algorithms this can be
6803 used to crash any certificate verification operation and exploited in a
6804 DoS attack. Any application which performs certificate verification is
6805 vulnerable including OpenSSL clients and servers which enable client
6808 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6813 * X509_ATTRIBUTE memory leak
6815 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6816 memory. This structure is used by the PKCS#7 and CMS routines so any
6817 application which reads PKCS#7 or CMS data from untrusted sources is
6818 affected. SSL/TLS is not affected.
6820 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6826 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6827 This changes the decoding behaviour for some invalid messages,
6828 though the change is mostly in the more lenient direction, and
6829 legacy behaviour is preserved as much as possible.
6833 * In DSA_generate_parameters_ex, if the provided seed is too short,
6834 use a random seed, as already documented.
6836 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6838 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6840 * Alternate chains certificate forgery
6842 During certificate verfification, OpenSSL will attempt to find an
6843 alternative certificate chain if the first attempt to build such a chain
6844 fails. An error in the implementation of this logic can mean that an
6845 attacker could cause certain checks on untrusted certificates to be
6846 bypassed, such as the CA flag, enabling them to use a valid leaf
6847 certificate to act as a CA and "issue" an invalid certificate.
6849 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6855 * Race condition handling PSK identify hint
6857 If PSK identity hints are received by a multi-threaded client then
6858 the values are wrongly updated in the parent SSL_CTX structure. This can
6859 result in a race condition potentially leading to a double free of the
6865 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6867 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6868 incompatibility in the handling of HMAC. The previous ABI has now been
6871 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6873 * Malformed ECParameters causes infinite loop
6875 When processing an ECParameters structure OpenSSL enters an infinite loop
6876 if the curve specified is over a specially malformed binary polynomial
6879 This can be used to perform denial of service against any
6880 system which processes public keys, certificate requests or
6881 certificates. This includes TLS clients and TLS servers with
6882 client authentication enabled.
6884 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6889 * Exploitable out-of-bounds read in X509_cmp_time
6891 X509_cmp_time does not properly check the length of the ASN1_TIME
6892 string and can read a few bytes out of bounds. In addition,
6893 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6896 An attacker can use this to craft malformed certificates and CRLs of
6897 various sizes and potentially cause a segmentation fault, resulting in
6898 a DoS on applications that verify certificates or CRLs. TLS clients
6899 that verify CRLs are affected. TLS clients and servers with client
6900 authentication enabled may be affected if they use custom verification
6903 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6904 independently by Hanno Böck.
6909 * PKCS7 crash with missing EnvelopedContent
6911 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6912 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6913 with missing content and trigger a NULL pointer dereference on parsing.
6915 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6916 structures from untrusted sources are affected. OpenSSL clients and
6917 servers are not affected.
6919 This issue was reported to OpenSSL by Michal Zalewski (Google).
6924 * CMS verify infinite loop with unknown hash function
6926 When verifying a signedData message the CMS code can enter an infinite loop
6927 if presented with an unknown hash function OID. This can be used to perform
6928 denial of service against any system which verifies signedData messages using
6930 This issue was reported to OpenSSL by Johannes Bauer.
6935 * Race condition handling NewSessionTicket
6937 If a NewSessionTicket is received by a multi-threaded client when attempting to
6938 reuse a previous ticket then a race condition can occur potentially leading to
6939 a double free of the ticket data.
6944 * Reject DH handshakes with parameters shorter than 768 bits.
6946 *Kurt Roeckx and Emilia Kasper*
6948 * dhparam: generate 2048-bit parameters by default.
6950 *Kurt Roeckx and Emilia Kasper*
6952 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6954 * Segmentation fault in ASN1_TYPE_cmp fix
6956 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6957 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6958 certificate signature algorithm consistency this can be used to crash any
6959 certificate verification operation and exploited in a DoS attack. Any
6960 application which performs certificate verification is vulnerable including
6961 OpenSSL clients and servers which enable client authentication.
6966 * ASN.1 structure reuse memory corruption fix
6968 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6969 memory corruption via an invalid write. Such reuse is and has been
6970 strongly discouraged and is believed to be rare.
6972 Applications that parse structures containing CHOICE or ANY DEFINED BY
6973 components may be affected. Certificate parsing (d2i_X509 and related
6974 functions) are however not affected. OpenSSL clients and servers are
6980 * PKCS7 NULL pointer dereferences fix
6982 The PKCS#7 parsing code does not handle missing outer ContentInfo
6983 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6984 missing content and trigger a NULL pointer dereference on parsing.
6986 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6987 otherwise parse PKCS#7 structures from untrusted sources are
6988 affected. OpenSSL clients and servers are not affected.
6990 This issue was reported to OpenSSL by Michal Zalewski (Google).
6995 * DoS via reachable assert in SSLv2 servers fix
6997 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6998 servers that both support SSLv2 and enable export cipher suites by sending
6999 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7001 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7002 (OpenSSL development team).
7007 * Use After Free following d2i_ECPrivatekey error fix
7009 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7010 could cause a use after free condition. This, in turn, could cause a double
7011 free in several private key parsing functions (such as d2i_PrivateKey
7012 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7013 for applications that receive EC private keys from untrusted
7014 sources. This scenario is considered rare.
7016 This issue was discovered by the BoringSSL project and fixed in their
7022 * X509_to_X509_REQ NULL pointer deref fix
7024 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7025 the certificate key is invalid. This function is rarely used in practice.
7027 This issue was discovered by Brian Carpenter.
7032 * Removed the export ciphers from the DEFAULT ciphers
7036 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7038 * Build fixes for the Windows and OpenVMS platforms
7040 *Matt Caswell and Richard Levitte*
7042 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7044 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7045 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7046 dereference. This could lead to a Denial Of Service attack. Thanks to
7047 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7052 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7053 dtls1_buffer_record function under certain conditions. In particular this
7054 could occur if an attacker sent repeated DTLS records with the same
7055 sequence number but for the next epoch. The memory leak could be exploited
7056 by an attacker in a Denial of Service attack through memory exhaustion.
7057 Thanks to Chris Mueller for reporting this issue.
7062 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7063 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7064 method would be set to NULL which could later result in a NULL pointer
7065 dereference. Thanks to Frank Schmirler for reporting this issue.
7070 * Abort handshake if server key exchange message is omitted for ephemeral
7073 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7074 reporting this issue.
7079 * Remove non-export ephemeral RSA code on client and server. This code
7080 violated the TLS standard by allowing the use of temporary RSA keys in
7081 non-export ciphersuites and could be used by a server to effectively
7082 downgrade the RSA key length used to a value smaller than the server
7083 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7084 INRIA or reporting this issue.
7089 * Fixed issue where DH client certificates are accepted without verification.
7090 An OpenSSL server will accept a DH certificate for client authentication
7091 without the certificate verify message. This effectively allows a client to
7092 authenticate without the use of a private key. This only affects servers
7093 which trust a client certificate authority which issues certificates
7094 containing DH keys: these are extremely rare and hardly ever encountered.
7095 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7101 * Ensure that the session ID context of an SSL is updated when its
7102 SSL_CTX is updated via SSL_set_SSL_CTX.
7104 The session ID context is typically set from the parent SSL_CTX,
7105 and can vary with the CTX.
7109 * Fix various certificate fingerprint issues.
7111 By using non-DER or invalid encodings outside the signed portion of a
7112 certificate the fingerprint can be changed without breaking the signature.
7113 Although no details of the signed portion of the certificate can be changed
7114 this can cause problems with some applications: e.g. those using the
7115 certificate fingerprint for blacklists.
7117 1. Reject signatures with non zero unused bits.
7119 If the BIT STRING containing the signature has non zero unused bits reject
7120 the signature. All current signature algorithms require zero unused bits.
7122 2. Check certificate algorithm consistency.
7124 Check the AlgorithmIdentifier inside TBS matches the one in the
7125 certificate signature. NB: this will result in signature failure
7126 errors for some broken certificates.
7128 Thanks to Konrad Kraszewski from Google for reporting this issue.
7130 3. Check DSA/ECDSA signatures use DER.
7132 Re-encode DSA/ECDSA signatures and compare with the original received
7133 signature. Return an error if there is a mismatch.
7135 This will reject various cases including garbage after signature
7136 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7137 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7138 (negative or with leading zeroes).
7140 Further analysis was conducted and fixes were developed by Stephen Henson
7141 of the OpenSSL core team.
7147 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7148 results on some platforms, including x86_64. This bug occurs at random
7149 with a very low probability, and is not known to be exploitable in any
7150 way, though its exact impact is difficult to determine. Thanks to Pieter
7151 Wuille (Blockstream) who reported this issue and also suggested an initial
7152 fix. Further analysis was conducted by the OpenSSL development team and
7153 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7154 the OpenSSL core team.
7159 * Do not resume sessions on the server if the negotiated protocol
7160 version does not match the session's version. Resuming with a different
7161 version, while not strictly forbidden by the RFC, is of questionable
7162 sanity and breaks all known clients.
7164 *David Benjamin, Emilia Käsper*
7166 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7167 early CCS messages during renegotiation. (Note that because
7168 renegotiation is encrypted, this early CCS was not exploitable.)
7172 * Tighten client-side session ticket handling during renegotiation:
7173 ensure that the client only accepts a session ticket if the server sends
7174 the extension anew in the ServerHello. Previously, a TLS client would
7175 reuse the old extension state and thus accept a session ticket if one was
7176 announced in the initial ServerHello.
7178 Similarly, ensure that the client requires a session ticket if one
7179 was advertised in the ServerHello. Previously, a TLS client would
7180 ignore a missing NewSessionTicket message.
7184 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7188 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7189 sends a carefully crafted handshake message, to cause OpenSSL to fail
7190 to free up to 64k of memory causing a memory leak. This could be
7191 exploited in a Denial Of Service attack. This issue affects OpenSSL
7192 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7193 whether SRTP is used or configured. Implementations of OpenSSL that
7194 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7196 The fix was developed by the OpenSSL team.
7201 * Session Ticket Memory Leak.
7203 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7204 integrity of that ticket is first verified. In the event of a session
7205 ticket integrity check failing, OpenSSL will fail to free memory
7206 causing a memory leak. By sending a large number of invalid session
7207 tickets an attacker could exploit this issue in a Denial Of Service
7213 * Build option no-ssl3 is incomplete.
7215 When OpenSSL is configured with "no-ssl3" as a build option, servers
7216 could accept and complete a SSL 3.0 handshake, and clients could be
7217 configured to send them.
7220 *Akamai and the OpenSSL team*
7222 * Add support for TLS_FALLBACK_SCSV.
7223 Client applications doing fallback retries should call
7224 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7227 *Adam Langley, Bodo Moeller*
7229 * Add additional DigestInfo checks.
7231 Re-encode DigestInto in DER and check against the original when
7232 verifying RSA signature: this will reject any improperly encoded
7233 DigestInfo structures.
7235 Note: this is a precautionary measure and no attacks are currently known.
7239 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7241 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7242 SRP code can be overrun an internal buffer. Add sanity check that
7243 g, A, B < N to SRP code.
7245 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7246 Group for discovering this issue.
7251 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7252 TLS 1.0 instead of higher protocol versions when the ClientHello message
7253 is badly fragmented. This allows a man-in-the-middle attacker to force a
7254 downgrade to TLS 1.0 even if both the server and the client support a
7255 higher protocol version, by modifying the client's TLS records.
7257 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7258 researching this issue.
7263 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7264 to a denial of service attack. A malicious server can crash the client
7265 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7266 ciphersuite and sending carefully crafted handshake messages.
7268 Thanks to Felix Gröbert (Google) for discovering and researching this
7274 * By sending carefully crafted DTLS packets an attacker could cause openssl
7275 to leak memory. This can be exploited through a Denial of Service attack.
7276 Thanks to Adam Langley for discovering and researching this issue.
7281 * An attacker can force openssl to consume large amounts of memory whilst
7282 processing DTLS handshake messages. This can be exploited through a
7283 Denial of Service attack.
7284 Thanks to Adam Langley for discovering and researching this issue.
7289 * An attacker can force an error condition which causes openssl to crash
7290 whilst processing DTLS packets due to memory being freed twice. This
7291 can be exploited through a Denial of Service attack.
7292 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7298 * If a multithreaded client connects to a malicious server using a resumed
7299 session and the server sends an ec point format extension it could write
7300 up to 255 bytes to freed memory.
7302 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7308 * A malicious server can crash an OpenSSL client with a null pointer
7309 dereference (read) by specifying an SRP ciphersuite even though it was not
7310 properly negotiated with the client. This can be exploited through a
7311 Denial of Service attack.
7313 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7314 discovering and researching this issue.
7319 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7320 X509_name_oneline, X509_name_print_ex et al. to leak some information
7321 from the stack. Applications may be affected if they echo pretty printing
7322 output to the attacker.
7324 Thanks to Ivan Fratric (Google) for discovering this issue.
7327 *Emilia Käsper, and Steve Henson*
7329 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7330 for corner cases. (Certain input points at infinity could lead to
7331 bogus results, with non-infinity inputs mapped to infinity too.)
7335 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7337 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7338 handshake can force the use of weak keying material in OpenSSL
7339 SSL/TLS clients and servers.
7341 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7342 researching this issue. ([CVE-2014-0224])
7344 *KIKUCHI Masashi, Steve Henson*
7346 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7347 OpenSSL DTLS client the code can be made to recurse eventually crashing
7350 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7353 *Imre Rad, Steve Henson*
7355 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7356 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7357 client or server. This is potentially exploitable to run arbitrary
7358 code on a vulnerable client or server.
7360 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7362 *Jüri Aedla, Steve Henson*
7364 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7365 are subject to a denial of service attack.
7367 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7368 this issue. ([CVE-2014-3470])
7370 *Felix Gröbert, Ivan Fratric, Steve Henson*
7372 * Harmonize version and its documentation. -f flag is used to display
7375 *mancha <mancha1@zoho.com>*
7377 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7378 in i2d_ECPrivateKey.
7380 *mancha <mancha1@zoho.com>*
7382 * Fix some double frees. These are not thought to be exploitable.
7384 *mancha <mancha1@zoho.com>*
7386 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7388 * A missing bounds check in the handling of the TLS heartbeat extension
7389 can be used to reveal up to 64k of memory to a connected client or
7392 Thanks for Neel Mehta of Google Security for discovering this bug and to
7393 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7394 preparing the fix ([CVE-2014-0160])
7396 *Adam Langley, Bodo Moeller*
7398 * Fix for the attack described in the paper "Recovering OpenSSL
7399 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7400 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7401 <http://eprint.iacr.org/2014/140>
7403 Thanks to Yuval Yarom and Naomi Benger for discovering this
7404 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7406 *Yuval Yarom and Naomi Benger*
7408 * TLS pad extension: draft-agl-tls-padding-03
7410 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7411 TLS client Hello record length value would otherwise be > 255 and
7412 less that 512 pad with a dummy extension containing zeroes so it
7413 is at least 512 bytes long.
7415 *Adam Langley, Steve Henson*
7417 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7419 * Fix for TLS record tampering bug. A carefully crafted invalid
7420 handshake could crash OpenSSL with a NULL pointer exception.
7421 Thanks to Anton Johansson for reporting this issues.
7424 * Keep original DTLS digest and encryption contexts in retransmission
7425 structures so we can use the previous session parameters if they need
7426 to be resent. ([CVE-2013-6450])
7430 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7431 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7432 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7433 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7434 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7435 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7437 *Rob Stradling, Adam Langley*
7439 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7441 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7442 supporting platforms or when small records were transferred.
7444 *Andy Polyakov, Steve Henson*
7446 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7448 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7450 This addresses the flaw in CBC record processing discovered by
7451 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7452 at: <http://www.isg.rhul.ac.uk/tls/>
7454 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7455 Security Group at Royal Holloway, University of London
7456 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7457 Emilia Käsper for the initial patch.
7460 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7462 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7463 ciphersuites which can be exploited in a denial of service attack.
7464 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7465 and detecting this bug and to Wolfgang Ettlinger
7466 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7471 * Return an error when checking OCSP signatures when key is NULL.
7472 This fixes a DoS attack. ([CVE-2013-0166])
7476 * Make openssl verify return errors.
7478 *Chris Palmer <palmer@google.com> and Ben Laurie*
7480 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7481 the right response is stapled. Also change SSL_get_certificate()
7482 so it returns the certificate actually sent.
7483 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7485 *Rob Stradling <rob.stradling@comodo.com>*
7487 * Fix possible deadlock when decoding public keys.
7491 * Don't use TLS 1.0 record version number in initial client hello
7496 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7498 * Sanity check record length before skipping explicit IV in TLS
7499 1.2, 1.1 and DTLS to fix DoS attack.
7501 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7502 fuzzing as a service testing platform.
7507 * Initialise tkeylen properly when encrypting CMS messages.
7508 Thanks to Solar Designer of Openwall for reporting this issue.
7512 * In FIPS mode don't try to use composite ciphers as they are not
7517 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7519 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7520 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7521 mean any application compiled against OpenSSL 1.0.0 headers setting
7522 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7523 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7524 0x10000000L Any application which was previously compiled against
7525 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7526 will need to be recompiled as a result. Letting be results in
7527 inability to disable specifically TLS 1.1 and in client context,
7528 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7532 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7533 disable just protocol X, but all protocols above X *if* there are
7534 protocols *below* X still enabled. In more practical terms it means
7535 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7536 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7537 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7542 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7544 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7545 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7546 in CRYPTO_realloc_clean.
7548 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7549 issue and to Adam Langley <agl@chromium.org> for fixing it.
7552 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7554 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7558 * Workarounds for some broken servers that "hang" if a client hello
7559 record length exceeds 255 bytes.
7561 1. Do not use record version number > TLS 1.0 in initial client
7562 hello: some (but not all) hanging servers will now work.
7563 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7564 the number of ciphers sent in the client hello. This should be
7565 set to an even number, such as 50, for example by passing:
7566 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7567 Most broken servers should now work.
7568 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7569 TLS 1.2 client support entirely.
7573 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7577 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7579 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7580 STRING form instead of a DigestInfo.
7584 * The format used for MDC2 RSA signatures is inconsistent between EVP
7585 and the RSA_sign/RSA_verify functions. This was made more apparent when
7586 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7587 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7588 the correct format in RSA_verify so both forms transparently work.
7592 * Some servers which support TLS 1.0 can choke if we initially indicate
7593 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7594 encrypted premaster secret. As a workaround use the maximum permitted
7595 client version in client hello, this should keep such servers happy
7596 and still work with previous versions of OpenSSL.
7600 * Add support for TLS/DTLS heartbeats.
7602 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7604 * Add support for SCTP.
7606 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7608 * Improved PRNG seeding for VOS.
7610 *Paul Green <Paul.Green@stratus.com>*
7612 * Extensive assembler packs updates, most notably:
7614 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7615 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7616 - x86_64: bit-sliced AES implementation;
7617 - ARM: NEON support, contemporary platforms optimizations;
7618 - s390x: z196 support;
7619 - `*`: GHASH and GF(2^m) multiplication implementations;
7623 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7624 (removal of unnecessary code)
7626 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7628 * Add TLS key material exporter from RFC 5705.
7632 * Add DTLS-SRTP negotiation from RFC 5764.
7636 * Add Next Protocol Negotiation,
7637 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7638 disabled with a no-npn flag to config or Configure. Code donated
7641 *Adam Langley <agl@google.com> and Ben Laurie*
7643 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7644 NIST-P256, NIST-P521, with constant-time single point multiplication on
7645 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7646 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7647 Code made available under Apache License version 2.0.
7649 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7650 line to include this in your build of OpenSSL, and run "make depend" (or
7651 "make update"). This enables the following EC_METHODs:
7653 EC_GFp_nistp224_method()
7654 EC_GFp_nistp256_method()
7655 EC_GFp_nistp521_method()
7657 EC_GROUP_new_by_curve_name() will automatically use these (while
7658 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7661 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7663 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7664 all platforms. Move ssize_t definition from e_os.h to the public
7665 header file e_os2.h as it now appears in public header file cms.h
7669 * New -sigopt option to the ca, req and x509 utilities. Additional
7670 signature parameters can be passed using this option and in
7675 * Add RSA PSS signing function. This will generate and set the
7676 appropriate AlgorithmIdentifiers for PSS based on those in the
7677 corresponding EVP_MD_CTX structure. No application support yet.
7681 * Support for companion algorithm specific ASN1 signing routines.
7682 New function ASN1_item_sign_ctx() signs a pre-initialised
7683 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7684 the appropriate parameters.
7688 * Add new algorithm specific ASN1 verification initialisation function
7689 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7690 handling will be the same no matter what EVP_PKEY_METHOD is used.
7691 Add a PSS handler to support verification of PSS signatures: checked
7692 against a number of sample certificates.
7696 * Add signature printing for PSS. Add PSS OIDs.
7698 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7700 * Add algorithm specific signature printing. An individual ASN1 method
7701 can now print out signatures instead of the standard hex dump.
7703 More complex signatures (e.g. PSS) can print out more meaningful
7704 information. Include DSA version that prints out the signature
7709 * Password based recipient info support for CMS library: implementing
7714 * Split password based encryption into PBES2 and PBKDF2 functions. This
7715 neatly separates the code into cipher and PBE sections and is required
7716 for some algorithms that split PBES2 into separate pieces (such as
7717 password based CMS).
7721 * Session-handling fixes:
7722 - Fix handling of connections that are resuming with a session ID,
7723 but also support Session Tickets.
7724 - Fix a bug that suppressed issuing of a new ticket if the client
7725 presented a ticket with an expired session.
7726 - Try to set the ticket lifetime hint to something reasonable.
7727 - Make tickets shorter by excluding irrelevant information.
7728 - On the client side, don't ignore renewed tickets.
7730 *Adam Langley, Bodo Moeller (Google)*
7732 * Fix PSK session representation.
7736 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7738 This work was sponsored by Intel.
7742 * Add GCM support to TLS library. Some custom code is needed to split
7743 the IV between the fixed (from PRF) and explicit (from TLS record)
7744 portions. This adds all GCM ciphersuites supported by RFC5288 and
7745 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7746 add a special AESGCM string for GCM only.
7750 * Expand range of ctrls for AES GCM. Permit setting invocation
7751 field on decrypt and retrieval of invocation field only on encrypt.
7755 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7756 As required by RFC5289 these ciphersuites cannot be used if for
7757 versions of TLS earlier than 1.2.
7761 * For FIPS capable OpenSSL interpret a NULL default public key method
7762 as unset and return the appropriate default but do *not* set the default.
7763 This means we can return the appropriate method in applications that
7764 switch between FIPS and non-FIPS modes.
7768 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7769 ENGINE is used then we cannot handle that in the FIPS module so we
7770 keep original code iff non-FIPS operations are allowed.
7774 * Add -attime option to openssl utilities.
7776 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7778 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7782 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7783 FIPS EC methods unconditionally for now.
7787 * New build option no-ec2m to disable characteristic 2 code.
7791 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7792 all cases can be covered as some introduce binary incompatibilities.
7796 * Redirect RSA operations to FIPS module including keygen,
7797 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7801 * Add similar low-level API blocking to ciphers.
7805 * low-level digest APIs are not approved in FIPS mode: any attempt
7806 to use these will cause a fatal error. Applications that *really* want
7807 to use them can use the `private_*` version instead.
7811 * Redirect cipher operations to FIPS module for FIPS builds.
7815 * Redirect digest operations to FIPS module for FIPS builds.
7819 * Update build system to add "fips" flag which will link in fipscanister.o
7820 for static and shared library builds embedding a signature if needed.
7824 * Output TLS supported curves in preference order instead of numerical
7825 order. This is currently hardcoded for the highest order curves first.
7826 This should be configurable so applications can judge speed vs strength.
7830 * Add TLS v1.2 server support for client authentication.
7834 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7839 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7840 FIPS modules versions.
7844 * Add TLS v1.2 client side support for client authentication. Keep cache
7845 of handshake records longer as we don't know the hash algorithm to use
7846 until after the certificate request message is received.
7850 * Initial TLS v1.2 client support. Add a default signature algorithms
7851 extension including all the algorithms we support. Parse new signature
7852 format in client key exchange. Relax some ECC signing restrictions for
7853 TLS v1.2 as indicated in RFC5246.
7857 * Add server support for TLS v1.2 signature algorithms extension. Switch
7858 to new signature format when needed using client digest preference.
7859 All server ciphersuites should now work correctly in TLS v1.2. No client
7860 support yet and no support for client certificates.
7864 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7865 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7866 ciphersuites. At present only RSA key exchange ciphersuites work with
7867 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7868 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7869 and version checking.
7873 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7874 with this defined it will not be affected by any changes to ssl internal
7875 structures. Add several utility functions to allow openssl application
7876 to work with OPENSSL_NO_SSL_INTERN defined.
7880 * A long standing patch to add support for SRP from EdelWeb (Peter
7881 Sylvester and Christophe Renou) was integrated.
7882 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7883 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7886 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7890 * Permit abbreviated handshakes when renegotiating using the function
7891 SSL_renegotiate_abbreviated().
7893 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7895 * Add call to ENGINE_register_all_complete() to
7896 ENGINE_load_builtin_engines(), so some implementations get used
7897 automatically instead of needing explicit application support.
7901 * Add support for TLS key exporter as described in RFC5705.
7903 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7905 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7906 a few changes are required:
7908 Add SSL_OP_NO_TLSv1_1 flag.
7909 Add TLSv1_1 methods.
7910 Update version checking logic to handle version 1.1.
7911 Add explicit IV handling (ported from DTLS code).
7912 Add command line options to s_client/s_server.
7919 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7921 * X509_ATTRIBUTE memory leak
7923 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7924 memory. This structure is used by the PKCS#7 and CMS routines so any
7925 application which reads PKCS#7 or CMS data from untrusted sources is
7926 affected. SSL/TLS is not affected.
7928 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7934 * Race condition handling PSK identify hint
7936 If PSK identity hints are received by a multi-threaded client then
7937 the values are wrongly updated in the parent SSL_CTX structure. This can
7938 result in a race condition potentially leading to a double free of the
7944 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7946 * Malformed ECParameters causes infinite loop
7948 When processing an ECParameters structure OpenSSL enters an infinite loop
7949 if the curve specified is over a specially malformed binary polynomial
7952 This can be used to perform denial of service against any
7953 system which processes public keys, certificate requests or
7954 certificates. This includes TLS clients and TLS servers with
7955 client authentication enabled.
7957 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7962 * Exploitable out-of-bounds read in X509_cmp_time
7964 X509_cmp_time does not properly check the length of the ASN1_TIME
7965 string and can read a few bytes out of bounds. In addition,
7966 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7969 An attacker can use this to craft malformed certificates and CRLs of
7970 various sizes and potentially cause a segmentation fault, resulting in
7971 a DoS on applications that verify certificates or CRLs. TLS clients
7972 that verify CRLs are affected. TLS clients and servers with client
7973 authentication enabled may be affected if they use custom verification
7976 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7977 independently by Hanno Böck.
7982 * PKCS7 crash with missing EnvelopedContent
7984 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7985 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7986 with missing content and trigger a NULL pointer dereference on parsing.
7988 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7989 structures from untrusted sources are affected. OpenSSL clients and
7990 servers are not affected.
7992 This issue was reported to OpenSSL by Michal Zalewski (Google).
7997 * CMS verify infinite loop with unknown hash function
7999 When verifying a signedData message the CMS code can enter an infinite loop
8000 if presented with an unknown hash function OID. This can be used to perform
8001 denial of service against any system which verifies signedData messages using
8003 This issue was reported to OpenSSL by Johannes Bauer.
8008 * Race condition handling NewSessionTicket
8010 If a NewSessionTicket is received by a multi-threaded client when attempting to
8011 reuse a previous ticket then a race condition can occur potentially leading to
8012 a double free of the ticket data.
8017 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8019 * Segmentation fault in ASN1_TYPE_cmp fix
8021 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8022 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8023 certificate signature algorithm consistency this can be used to crash any
8024 certificate verification operation and exploited in a DoS attack. Any
8025 application which performs certificate verification is vulnerable including
8026 OpenSSL clients and servers which enable client authentication.
8031 * ASN.1 structure reuse memory corruption fix
8033 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8034 memory corruption via an invalid write. Such reuse is and has been
8035 strongly discouraged and is believed to be rare.
8037 Applications that parse structures containing CHOICE or ANY DEFINED BY
8038 components may be affected. Certificate parsing (d2i_X509 and related
8039 functions) are however not affected. OpenSSL clients and servers are
8045 * PKCS7 NULL pointer dereferences fix
8047 The PKCS#7 parsing code does not handle missing outer ContentInfo
8048 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8049 missing content and trigger a NULL pointer dereference on parsing.
8051 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8052 otherwise parse PKCS#7 structures from untrusted sources are
8053 affected. OpenSSL clients and servers are not affected.
8055 This issue was reported to OpenSSL by Michal Zalewski (Google).
8060 * DoS via reachable assert in SSLv2 servers fix
8062 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8063 servers that both support SSLv2 and enable export cipher suites by sending
8064 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8066 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8067 (OpenSSL development team).
8072 * Use After Free following d2i_ECPrivatekey error fix
8074 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8075 could cause a use after free condition. This, in turn, could cause a double
8076 free in several private key parsing functions (such as d2i_PrivateKey
8077 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8078 for applications that receive EC private keys from untrusted
8079 sources. This scenario is considered rare.
8081 This issue was discovered by the BoringSSL project and fixed in their
8087 * X509_to_X509_REQ NULL pointer deref fix
8089 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8090 the certificate key is invalid. This function is rarely used in practice.
8092 This issue was discovered by Brian Carpenter.
8097 * Removed the export ciphers from the DEFAULT ciphers
8101 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8103 * Build fixes for the Windows and OpenVMS platforms
8105 *Matt Caswell and Richard Levitte*
8107 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8109 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8110 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8111 dereference. This could lead to a Denial Of Service attack. Thanks to
8112 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8117 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8118 dtls1_buffer_record function under certain conditions. In particular this
8119 could occur if an attacker sent repeated DTLS records with the same
8120 sequence number but for the next epoch. The memory leak could be exploited
8121 by an attacker in a Denial of Service attack through memory exhaustion.
8122 Thanks to Chris Mueller for reporting this issue.
8127 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8128 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8129 method would be set to NULL which could later result in a NULL pointer
8130 dereference. Thanks to Frank Schmirler for reporting this issue.
8135 * Abort handshake if server key exchange message is omitted for ephemeral
8138 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8139 reporting this issue.
8144 * Remove non-export ephemeral RSA code on client and server. This code
8145 violated the TLS standard by allowing the use of temporary RSA keys in
8146 non-export ciphersuites and could be used by a server to effectively
8147 downgrade the RSA key length used to a value smaller than the server
8148 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8149 INRIA or reporting this issue.
8154 * Fixed issue where DH client certificates are accepted without verification.
8155 An OpenSSL server will accept a DH certificate for client authentication
8156 without the certificate verify message. This effectively allows a client to
8157 authenticate without the use of a private key. This only affects servers
8158 which trust a client certificate authority which issues certificates
8159 containing DH keys: these are extremely rare and hardly ever encountered.
8160 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8166 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8167 results on some platforms, including x86_64. This bug occurs at random
8168 with a very low probability, and is not known to be exploitable in any
8169 way, though its exact impact is difficult to determine. Thanks to Pieter
8170 Wuille (Blockstream) who reported this issue and also suggested an initial
8171 fix. Further analysis was conducted by the OpenSSL development team and
8172 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8173 the OpenSSL core team.
8178 * Fix various certificate fingerprint issues.
8180 By using non-DER or invalid encodings outside the signed portion of a
8181 certificate the fingerprint can be changed without breaking the signature.
8182 Although no details of the signed portion of the certificate can be changed
8183 this can cause problems with some applications: e.g. those using the
8184 certificate fingerprint for blacklists.
8186 1. Reject signatures with non zero unused bits.
8188 If the BIT STRING containing the signature has non zero unused bits reject
8189 the signature. All current signature algorithms require zero unused bits.
8191 2. Check certificate algorithm consistency.
8193 Check the AlgorithmIdentifier inside TBS matches the one in the
8194 certificate signature. NB: this will result in signature failure
8195 errors for some broken certificates.
8197 Thanks to Konrad Kraszewski from Google for reporting this issue.
8199 3. Check DSA/ECDSA signatures use DER.
8201 Reencode DSA/ECDSA signatures and compare with the original received
8202 signature. Return an error if there is a mismatch.
8204 This will reject various cases including garbage after signature
8205 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8206 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8207 (negative or with leading zeroes).
8209 Further analysis was conducted and fixes were developed by Stephen Henson
8210 of the OpenSSL core team.
8216 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8218 * Session Ticket Memory Leak.
8220 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8221 integrity of that ticket is first verified. In the event of a session
8222 ticket integrity check failing, OpenSSL will fail to free memory
8223 causing a memory leak. By sending a large number of invalid session
8224 tickets an attacker could exploit this issue in a Denial Of Service
8230 * Build option no-ssl3 is incomplete.
8232 When OpenSSL is configured with "no-ssl3" as a build option, servers
8233 could accept and complete a SSL 3.0 handshake, and clients could be
8234 configured to send them.
8237 *Akamai and the OpenSSL team*
8239 * Add support for TLS_FALLBACK_SCSV.
8240 Client applications doing fallback retries should call
8241 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8244 *Adam Langley, Bodo Moeller*
8246 * Add additional DigestInfo checks.
8248 Reencode DigestInto in DER and check against the original when
8249 verifying RSA signature: this will reject any improperly encoded
8250 DigestInfo structures.
8252 Note: this is a precautionary measure and no attacks are currently known.
8256 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8258 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8259 to a denial of service attack. A malicious server can crash the client
8260 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8261 ciphersuite and sending carefully crafted handshake messages.
8263 Thanks to Felix Gröbert (Google) for discovering and researching this
8269 * By sending carefully crafted DTLS packets an attacker could cause openssl
8270 to leak memory. This can be exploited through a Denial of Service attack.
8271 Thanks to Adam Langley for discovering and researching this issue.
8276 * An attacker can force openssl to consume large amounts of memory whilst
8277 processing DTLS handshake messages. This can be exploited through a
8278 Denial of Service attack.
8279 Thanks to Adam Langley for discovering and researching this issue.
8284 * An attacker can force an error condition which causes openssl to crash
8285 whilst processing DTLS packets due to memory being freed twice. This
8286 can be exploited through a Denial of Service attack.
8287 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8293 * If a multithreaded client connects to a malicious server using a resumed
8294 session and the server sends an ec point format extension it could write
8295 up to 255 bytes to freed memory.
8297 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8303 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8304 X509_name_oneline, X509_name_print_ex et al. to leak some information
8305 from the stack. Applications may be affected if they echo pretty printing
8306 output to the attacker.
8308 Thanks to Ivan Fratric (Google) for discovering this issue.
8311 *Emilia Käsper, and Steve Henson*
8313 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8314 for corner cases. (Certain input points at infinity could lead to
8315 bogus results, with non-infinity inputs mapped to infinity too.)
8319 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8321 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8322 handshake can force the use of weak keying material in OpenSSL
8323 SSL/TLS clients and servers.
8325 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8326 researching this issue. ([CVE-2014-0224])
8328 *KIKUCHI Masashi, Steve Henson*
8330 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8331 OpenSSL DTLS client the code can be made to recurse eventually crashing
8334 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8337 *Imre Rad, Steve Henson*
8339 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8340 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8341 client or server. This is potentially exploitable to run arbitrary
8342 code on a vulnerable client or server.
8344 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8346 *Jüri Aedla, Steve Henson*
8348 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8349 are subject to a denial of service attack.
8351 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8352 this issue. ([CVE-2014-3470])
8354 *Felix Gröbert, Ivan Fratric, Steve Henson*
8356 * Harmonize version and its documentation. -f flag is used to display
8359 *mancha <mancha1@zoho.com>*
8361 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8362 in i2d_ECPrivateKey.
8364 *mancha <mancha1@zoho.com>*
8366 * Fix some double frees. These are not thought to be exploitable.
8368 *mancha <mancha1@zoho.com>*
8370 * Fix for the attack described in the paper "Recovering OpenSSL
8371 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8372 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8373 <http://eprint.iacr.org/2014/140>
8375 Thanks to Yuval Yarom and Naomi Benger for discovering this
8376 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8378 *Yuval Yarom and Naomi Benger*
8380 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8382 * Keep original DTLS digest and encryption contexts in retransmission
8383 structures so we can use the previous session parameters if they need
8384 to be resent. ([CVE-2013-6450])
8388 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8389 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8390 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8391 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8392 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8393 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8395 *Rob Stradling, Adam Langley*
8397 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8399 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8401 This addresses the flaw in CBC record processing discovered by
8402 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8403 at: <http://www.isg.rhul.ac.uk/tls/>
8405 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8406 Security Group at Royal Holloway, University of London
8407 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8408 Emilia Käsper for the initial patch.
8411 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8413 * Return an error when checking OCSP signatures when key is NULL.
8414 This fixes a DoS attack. ([CVE-2013-0166])
8418 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8419 the right response is stapled. Also change SSL_get_certificate()
8420 so it returns the certificate actually sent.
8421 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8422 (This is a backport)
8424 *Rob Stradling <rob.stradling@comodo.com>*
8426 * Fix possible deadlock when decoding public keys.
8430 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8432 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8435 * Sanity check record length before skipping explicit IV in DTLS
8438 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8439 fuzzing as a service testing platform.
8444 * Initialise tkeylen properly when encrypting CMS messages.
8445 Thanks to Solar Designer of Openwall for reporting this issue.
8449 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8451 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8452 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8453 in CRYPTO_realloc_clean.
8455 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8456 issue and to Adam Langley <agl@chromium.org> for fixing it.
8459 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8461 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8463 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8464 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8465 content decryption and always return the same error. Note: this attack
8466 needs on average 2^20 messages so it only affects automated senders. The
8467 old behaviour can be re-enabled in the CMS code by setting the
8468 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8469 an MMA defence is not necessary.
8470 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8471 this issue. ([CVE-2012-0884])
8475 * Fix CVE-2011-4619: make sure we really are receiving a
8476 client hello before rejecting multiple SGC restarts. Thanks to
8477 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8481 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8483 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8484 Thanks to Antonio Martin, Enterprise Secure Access Research and
8485 Development, Cisco Systems, Inc. for discovering this bug and
8486 preparing a fix. ([CVE-2012-0050])
8490 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8492 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8493 of the Vaudenay padding oracle attack on CBC mode encryption
8494 which enables an efficient plaintext recovery attack against
8495 the OpenSSL implementation of DTLS. Their attack exploits timing
8496 differences arising during decryption processing. A research
8497 paper describing this attack can be found at:
8498 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8499 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8500 Security Group at Royal Holloway, University of London
8501 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8502 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8503 for preparing the fix. ([CVE-2011-4108])
8505 *Robin Seggelmann, Michael Tuexen*
8507 * Clear bytes used for block padding of SSL 3.0 records.
8510 *Adam Langley (Google)*
8512 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8513 Kadianakis <desnacked@gmail.com> for discovering this issue and
8514 Adam Langley for preparing the fix. ([CVE-2011-4619])
8516 *Adam Langley (Google)*
8518 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8520 *Andrey Kulikov <amdeich@gmail.com>*
8522 * Prevent malformed RFC3779 data triggering an assertion failure.
8523 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8524 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8526 *Rob Austein <sra@hactrn.net>*
8528 * Improved PRNG seeding for VOS.
8530 *Paul Green <Paul.Green@stratus.com>*
8532 * Fix ssl_ciph.c set-up race.
8534 *Adam Langley (Google)*
8536 * Fix spurious failures in ecdsatest.c.
8538 *Emilia Käsper (Google)*
8540 * Fix the BIO_f_buffer() implementation (which was mixing different
8541 interpretations of the `..._len` fields).
8543 *Adam Langley (Google)*
8545 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8546 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8547 threads won't reuse the same blinding coefficients.
8549 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8550 lock to call BN_BLINDING_invert_ex, and avoids one use of
8551 BN_BLINDING_update for each BN_BLINDING structure (previously,
8552 the last update always remained unused).
8554 *Emilia Käsper (Google)*
8556 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8558 *Bob Buckholz (Google)*
8560 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8562 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8563 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8565 *Kaspar Brand <ossl@velox.ch>*
8567 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8568 for multi-threaded use of ECDH. ([CVE-2011-3210])
8570 *Adam Langley (Google)*
8572 * Fix x509_name_ex_d2i memory leak on bad inputs.
8576 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8577 signature public key algorithm by using OID xref utilities instead.
8578 Before this you could only use some ECC ciphersuites with SHA1 only.
8582 * Add protection against ECDSA timing attacks as mentioned in the paper
8583 by Billy Bob Brumley and Nicola Tuveri, see:
8584 <http://eprint.iacr.org/2011/232.pdf>
8586 *Billy Bob Brumley and Nicola Tuveri*
8588 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8590 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8592 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8594 * Fix bug in string printing code: if *any* escaping is enabled we must
8595 escape the escape character (backslash) or the resulting string is
8600 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8602 * Disable code workaround for ancient and obsolete Netscape browsers
8603 and servers: an attacker can use it in a ciphersuite downgrade attack.
8604 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8608 * Fixed J-PAKE implementation error, originally discovered by
8609 Sebastien Martini, further info and confirmation from Stefan
8610 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8614 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8616 * Fix extension code to avoid race conditions which can result in a buffer
8617 overrun vulnerability: resumed sessions must not be modified as they can
8618 be shared by multiple threads. CVE-2010-3864
8622 * Fix WIN32 build system to correctly link an ENGINE directory into
8627 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8629 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8632 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8634 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8636 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8637 context. The operation can be customised via the ctrl mechanism in
8638 case ENGINEs want to include additional functionality.
8642 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8646 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8647 output hashes compatible with older versions of OpenSSL.
8649 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8651 * Fix compression algorithm handling: if resuming a session use the
8652 compression algorithm of the resumed session instead of determining
8653 it from client hello again. Don't allow server to change algorithm.
8657 * Add load_crls() function to commands tidying load_certs() too. Add option
8658 to verify utility to allow additional CRLs to be included.
8662 * Update OCSP request code to permit adding custom headers to the request:
8663 some responders need this.
8667 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8670 *Julia Lawall <julia@diku.dk>*
8672 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8673 needlessly dereferenced structures, used obsolete functions and
8674 didn't handle all updated verify codes correctly.
8678 * Disable MD2 in the default configuration.
8682 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8683 indicate the initial BIO being pushed or popped. This makes it possible
8684 to determine whether the BIO is the one explicitly called or as a result
8685 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8686 it handles reference counts correctly and doesn't zero out the I/O bio
8687 when it is not being explicitly popped. WARNING: applications which
8688 included workarounds for the old buggy behaviour will need to be modified
8689 or they could free up already freed BIOs.
8693 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8694 renaming to all platforms (within the 0.9.8 branch, this was
8695 done conditionally on Netware platforms to avoid a name clash).
8697 *Guenter <lists@gknw.net>*
8699 * Add ECDHE and PSK support to DTLS.
8701 *Michael Tuexen <tuexen@fh-muenster.de>*
8703 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8708 * Add "missing" function EVP_MD_flags() (without this the only way to
8709 retrieve a digest flags is by accessing the structure directly. Update
8710 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8711 or cipher is registered as in the "from" argument. Print out all
8712 registered digests in the dgst usage message instead of manually
8713 attempting to work them out.
8717 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8718 this allows the use of compression and extensions. Change default cipher
8719 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8720 by default unless an application cipher string requests it.
8724 * Alter match criteria in PKCS12_parse(). It used to try to use local
8725 key ids to find matching certificates and keys but some PKCS#12 files
8726 don't follow the (somewhat unwritten) rules and this strategy fails.
8727 Now just gather all certificates together and the first private key
8728 then look for the first certificate that matches the key.
8732 * Support use of registered digest and cipher names for dgst and cipher
8733 commands instead of having to add each one as a special case. So now
8740 openssl dgst -sha256 foo
8742 and this works for ENGINE based algorithms too.
8746 * Update Gost ENGINE to support parameter files.
8748 *Victor B. Wagner <vitus@cryptocom.ru>*
8750 * Support GeneralizedTime in ca utility.
8752 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8754 * Enhance the hash format used for certificate directory links. The new
8755 form uses the canonical encoding (meaning equivalent names will work
8756 even if they aren't identical) and uses SHA1 instead of MD5. This form
8757 is incompatible with the older format and as a result c_rehash should
8758 be used to rebuild symbolic links.
8762 * Make PKCS#8 the default write format for private keys, replacing the
8763 traditional format. This form is standardised, more secure and doesn't
8764 include an implicit MD5 dependency.
8768 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8769 committed to OpenSSL should pass this lot as a minimum.
8773 * Add session ticket override functionality for use by EAP-FAST.
8775 *Jouni Malinen <j@w1.fi>*
8777 * Modify HMAC functions to return a value. Since these can be implemented
8778 in an ENGINE errors can occur.
8782 * Type-checked OBJ_bsearch_ex.
8786 * Type-checked OBJ_bsearch. Also some constification necessitated
8787 by type-checking. Still to come: TXT_DB, bsearch(?),
8788 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8793 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8794 seconds to a tm structure directly, instead of going through OS
8795 specific date routines. This avoids any issues with OS routines such
8796 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8797 and X509_time_adj_ex() to cover the extended range. The existing
8798 X509_time_adj() is still usable and will no longer have any date issues.
8802 * Delta CRL support. New use deltas option which will attempt to locate
8803 and search any appropriate delta CRLs available.
8805 This work was sponsored by Google.
8809 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8810 code and add additional score elements. Validate alternate CRL paths
8811 as part of the CRL checking and indicate a new error "CRL path validation
8812 error" in this case. Applications wanting additional details can use
8813 the verify callback and check the new "parent" field. If this is not
8814 NULL CRL path validation is taking place. Existing applications won't
8815 see this because it requires extended CRL support which is off by
8818 This work was sponsored by Google.
8822 * Support for freshest CRL extension.
8824 This work was sponsored by Google.
8828 * Initial indirect CRL support. Currently only supported in the CRLs
8829 passed directly and not via lookup. Process certificate issuer
8830 CRL entry extension and lookup CRL entries by bother issuer name
8831 and serial number. Check and process CRL issuer entry in IDP extension.
8833 This work was sponsored by Google.
8837 * Add support for distinct certificate and CRL paths. The CRL issuer
8838 certificate is validated separately in this case. Only enabled if
8839 an extended CRL support flag is set: this flag will enable additional
8840 CRL functionality in future.
8842 This work was sponsored by Google.
8846 * Add support for policy mappings extension.
8848 This work was sponsored by Google.
8852 * Fixes to pathlength constraint, self issued certificate handling,
8853 policy processing to align with RFC3280 and PKITS tests.
8855 This work was sponsored by Google.
8859 * Support for name constraints certificate extension. DN, email, DNS
8860 and URI types are currently supported.
8862 This work was sponsored by Google.
8866 * To cater for systems that provide a pointer-based thread ID rather
8867 than numeric, deprecate the current numeric thread ID mechanism and
8868 replace it with a structure and associated callback type. This
8869 mechanism allows a numeric "hash" to be extracted from a thread ID in
8870 either case, and on platforms where pointers are larger than 'long',
8871 mixing is done to help ensure the numeric 'hash' is usable even if it
8872 can't be guaranteed unique. The default mechanism is to use "&errno"
8873 as a pointer-based thread ID to distinguish between threads.
8875 Applications that want to provide their own thread IDs should now use
8876 CRYPTO_THREADID_set_callback() to register a callback that will call
8877 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8879 Note that ERR_remove_state() is now deprecated, because it is tied
8880 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8881 to free the current thread's error state should be replaced by
8882 ERR_remove_thread_state(NULL).
8884 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8885 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8886 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8887 application was previously providing a numeric thread callback that
8888 was inappropriate for distinguishing threads, then uniqueness might
8889 have been obtained with &errno that happened immediately in the
8890 intermediate development versions of OpenSSL; this is no longer the
8891 case, the numeric thread callback will now override the automatic use
8894 *Geoff Thorpe, with help from Bodo Moeller*
8896 * Initial support for different CRL issuing certificates. This covers a
8897 simple case where the self issued certificates in the chain exist and
8898 the real CRL issuer is higher in the existing chain.
8900 This work was sponsored by Google.
8904 * Removed effectively defunct crypto/store from the build.
8908 * Revamp of STACK to provide stronger type-checking. Still to come:
8909 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8910 ASN1_STRING, CONF_VALUE.
8914 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8915 RAM on SSL connections. This option can save about 34k per idle SSL.
8919 * Revamp of LHASH to provide stronger type-checking. Still to come:
8920 STACK, TXT_DB, bsearch, qsort.
8924 * Initial support for Cryptographic Message Syntax (aka CMS) based
8925 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8926 support for data, signedData, compressedData, digestedData and
8927 encryptedData, envelopedData types included. Scripts to check against
8928 RFC4134 examples draft and interop and consistency checks of many
8929 content types and variants.
8933 * Add options to enc utility to support use of zlib compression BIO.
8937 * Extend mk1mf to support importing of options and assembly language
8938 files from Configure script, currently only included in VC-WIN32.
8939 The assembly language rules can now optionally generate the source
8940 files from the associated perl scripts.
8944 * Implement remaining functionality needed to support GOST ciphersuites.
8945 Interop testing has been performed using CryptoPro implementations.
8947 *Victor B. Wagner <vitus@cryptocom.ru>*
8949 * s390x assembler pack.
8953 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8958 * Implement Opaque PRF Input TLS extension as specified in
8959 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8960 official specification yet and no extension type assignment by
8961 IANA exists, this extension (for now) will have to be explicitly
8962 enabled when building OpenSSL by providing the extension number
8963 to use. For example, specify an option
8965 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8967 to the "config" or "Configure" script to enable the extension,
8968 assuming extension number 0x9527 (which is a completely arbitrary
8969 and unofficial assignment based on the MD5 hash of the Internet
8970 Draft). Note that by doing so, you potentially lose
8971 interoperability with other TLS implementations since these might
8972 be using the same extension number for other purposes.
8974 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8975 opaque PRF input value to use in the handshake. This will create
8976 an internal copy of the length-'len' string at 'src', and will
8977 return non-zero for success.
8979 To get more control and flexibility, provide a callback function
8982 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8983 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8987 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8990 Callback function 'cb' will be called in handshakes, and is
8991 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8992 Argument 'arg' is for application purposes (the value as given to
8993 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8994 be provided to the callback function). The callback function
8995 has to return non-zero to report success: usually 1 to use opaque
8996 PRF input just if possible, or 2 to enforce use of the opaque PRF
8997 input. In the latter case, the library will abort the handshake
8998 if opaque PRF input is not successfully negotiated.
9000 Arguments 'peerinput' and 'len' given to the callback function
9001 will always be NULL and 0 in the case of a client. A server will
9002 see the client's opaque PRF input through these variables if
9003 available (NULL and 0 otherwise). Note that if the server
9004 provides an opaque PRF input, the length must be the same as the
9005 length of the client's opaque PRF input.
9007 Note that the callback function will only be called when creating
9008 a new session (session resumption can resume whatever was
9009 previously negotiated), and will not be called in SSL 2.0
9010 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9011 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9012 for applications that need to enforce opaque PRF input.
9016 * Update ssl code to support digests other than SHA1+MD5 for handshake
9019 *Victor B. Wagner <vitus@cryptocom.ru>*
9021 * Add RFC4507 support to OpenSSL. This includes the corrections in
9022 RFC4507bis. The encrypted ticket format is an encrypted encoded
9023 SSL_SESSION structure, that way new session features are automatically
9026 If a client application caches session in an SSL_SESSION structure
9027 support is transparent because tickets are now stored in the encoded
9030 The SSL_CTX structure automatically generates keys for ticket
9031 protection in servers so again support should be possible
9032 with no application modification.
9034 If a client or server wishes to disable RFC4507 support then the option
9035 SSL_OP_NO_TICKET can be set.
9037 Add a TLS extension debugging callback to allow the contents of any client
9038 or server extensions to be examined.
9040 This work was sponsored by Google.
9044 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9045 OpenSSL should now compile cleanly on gcc 4.2
9047 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9049 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9050 support including streaming MAC support: this is required for GOST
9051 ciphersuite support.
9053 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9055 * Add option -stream to use PKCS#7 streaming in smime utility. New
9056 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9057 to output in BER and PEM format.
9061 * Experimental support for use of HMAC via EVP_PKEY interface. This
9062 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9063 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9064 ENGINE support for HMAC keys which are unextractable. New -mac and
9065 -macopt options to dgst utility.
9069 * New option -sigopt to dgst utility. Update dgst to use
9070 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9071 alternative signing parameters such as X9.31 or PSS in the dgst
9076 * Change ssl_cipher_apply_rule(), the internal function that does
9077 the work each time a ciphersuite string requests enabling
9078 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9079 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9080 the order of disabled ciphersuites such that those ciphersuites
9081 that most recently went from enabled to disabled not only stay
9082 in order with respect to each other, but also have higher priority
9083 than other disabled ciphersuites the next time ciphersuites are
9086 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9087 the same ciphersuites as with "HIGH" alone, but in a specific
9088 order where the PSK ciphersuites come first (since they are the
9089 most recently disabled ciphersuites when "HIGH" is parsed).
9091 Also, change ssl_create_cipher_list() (using this new
9092 functionality) such that between otherwise identical
9093 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9098 * Change ssl_create_cipher_list() so that it automatically
9099 arranges the ciphersuites in reasonable order before starting
9100 to process the rule string. Thus, the definition for "DEFAULT"
9101 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9102 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9103 This makes it much easier to arrive at a reasonable default order
9104 in applications for which anonymous ciphers are OK (meaning
9105 that you can't actually use DEFAULT).
9107 *Bodo Moeller; suggested by Victor Duchovni*
9109 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9110 processing) into multiple integers instead of setting
9111 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9112 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9113 (These masks as well as the individual bit definitions are hidden
9114 away into the non-exported interface ssl/ssl_locl.h, so this
9115 change to the definition of the SSL_CIPHER structure shouldn't
9116 affect applications.) This give us more bits for each of these
9117 categories, so there is no longer a need to coagulate AES128 and
9118 AES256 into a single algorithm bit, and to coagulate Camellia128
9119 and Camellia256 into a single algorithm bit, which has led to all
9122 Thus, among other things, the kludge introduced in 0.9.7m and
9123 0.9.8e for masking out AES256 independently of AES128 or masking
9124 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9126 With the change, we also introduce new ciphersuite aliases that
9127 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9132 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9133 Use the leftmost N bytes of the signature input if the input is
9134 larger than the prime q (with N being the size in bytes of q).
9138 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9139 it yet and it is largely untested.
9143 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9147 * Initial incomplete changes to avoid need for function casts in OpenSSL
9148 some compilers (gcc 4.2 and later) reject their use. Safestack is
9149 reimplemented. Update ASN1 to avoid use of legacy functions.
9153 * Win32/64 targets are linked with Winsock2.
9157 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9158 to external functions. This can be used to increase CRL handling
9159 efficiency especially when CRLs are very large by (for example) storing
9160 the CRL revoked certificates in a database.
9164 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9165 new CRLs added to a directory can be used. New command line option
9166 -verify_return_error to s_client and s_server. This causes real errors
9167 to be returned by the verify callback instead of carrying on no matter
9168 what. This reflects the way a "real world" verify callback would behave.
9172 * GOST engine, supporting several GOST algorithms and public key formats.
9173 Kindly donated by Cryptocom.
9177 * Partial support for Issuing Distribution Point CRL extension. CRLs
9178 partitioned by DP are handled but no indirect CRL or reason partitioning
9179 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9180 selected via a scoring technique which handles IDP and AKID in CRLs.
9184 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9185 will ultimately be used for all verify operations: this will remove the
9186 X509_STORE dependency on certificate verification and allow alternative
9187 lookup methods. X509_STORE based implementations of these two callbacks.
9191 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9192 Modify get_crl() to find a valid (unexpired) CRL if possible.
9196 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9197 this would be called X509_CRL_cmp() but that name is already used by
9198 a function that just compares CRL issuer names. Cache several CRL
9199 extensions in X509_CRL structure and cache CRLDP in X509.
9203 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9204 this maps equivalent X509_NAME structures into a consistent structure.
9205 Name comparison can then be performed rapidly using memcmp().
9209 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9214 * Allow digests to supply their own micalg string for S/MIME type using
9215 the ctrl EVP_MD_CTRL_MICALG.
9219 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9220 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9221 ctrl. It can then customise the structure before and/or after signing
9226 * New function OBJ_add_sigid() to allow application defined signature OIDs
9227 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9228 to free up any added signature OIDs.
9232 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9233 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9234 digest and cipher tables. New options added to openssl utility:
9235 list-message-digest-algorithms and list-cipher-algorithms.
9239 * Change the array representation of binary polynomials: the list
9240 of degrees of non-zero coefficients is now terminated with -1.
9241 Previously it was terminated with 0, which was also part of the
9242 value; thus, the array representation was not applicable to
9243 polynomials where t^0 has coefficient zero. This change makes
9244 the array representation useful in a more general context.
9248 * Various modifications and fixes to SSL/TLS cipher string
9249 handling. For ECC, the code now distinguishes between fixed ECDH
9250 with RSA certificates on the one hand and with ECDSA certificates
9251 on the other hand, since these are separate ciphersuites. The
9252 unused code for Fortezza ciphersuites has been removed.
9254 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9255 (not "ECDHE"). For consistency with the code for DH
9256 certificates, use of ECDH certificates is now considered ECDH
9257 authentication, not RSA or ECDSA authentication (the latter is
9258 merely the CA's signing algorithm and not actively used in the
9261 The temporary ciphersuite alias "ECCdraft" is no longer
9262 available, and ECC ciphersuites are no longer excluded from "ALL"
9263 and "DEFAULT". The following aliases now exist for RFC 4492
9264 ciphersuites, most of these by analogy with the DH case:
9266 kECDHr - ECDH cert, signed with RSA
9267 kECDHe - ECDH cert, signed with ECDSA
9268 kECDH - ECDH cert (signed with either RSA or ECDSA)
9269 kEECDH - ephemeral ECDH
9270 ECDH - ECDH cert or ephemeral ECDH
9276 AECDH - anonymous ECDH
9277 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9281 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9282 Use correct micalg parameters depending on digest(s) in signed message.
9286 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9287 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9291 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9292 an engine to register a method. Add ENGINE lookups for methods and
9293 functional reference processing.
9297 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9298 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9303 * New -resign option to smime utility. This adds one or more signers
9304 to an existing PKCS#7 signedData structure. Also -md option to use an
9305 alternative message digest algorithm for signing.
9309 * Tidy up PKCS#7 routines and add new functions to make it easier to
9310 create PKCS7 structures containing multiple signers. Update smime
9311 application to support multiple signers.
9315 * New -macalg option to pkcs12 utility to allow setting of an alternative
9320 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9321 Reorganize PBE internals to lookup from a static table using NIDs,
9322 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9323 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9324 PRF which will be automatically used with PBES2.
9328 * Replace the algorithm specific calls to generate keys in "req" with the
9333 * Update PKCS#7 enveloped data routines to use new API. This is now
9334 supported by any public key method supporting the encrypt operation. A
9335 ctrl is added to allow the public key algorithm to examine or modify
9336 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9341 * Add a ctrl to asn1 method to allow a public key algorithm to express
9342 a default digest type to use. In most cases this will be SHA1 but some
9343 algorithms (such as GOST) need to specify an alternative digest. The
9344 return value indicates how strong the preference is 1 means optional and
9345 2 is mandatory (that is it is the only supported type). Modify
9346 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9347 use the default md. Update openssl utilities to use the default digest
9348 type for signing if it is not explicitly indicated.
9352 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9353 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9354 signing method from the key type. This effectively removes the link
9355 between digests and public key types.
9359 * Add an OID cross reference table and utility functions. Its purpose is to
9360 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9361 rsaEncryption. This will allow some of the algorithm specific hackery
9362 needed to use the correct OID to be removed.
9366 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9367 structures for PKCS7_sign(). They are now set up by the relevant public
9372 * Add provisional EC pkey method with support for ECDSA and ECDH.
9376 * Add support for key derivation (agreement) in the API, DH method and
9381 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9382 public and private key formats. As a side effect these add additional
9383 command line functionality not previously available: DSA signatures can be
9384 generated and verified using pkeyutl and DH key support and generation in
9391 *Oliver Tappe <zooey@hirschkaefer.de>*
9393 * New make target "install_html_docs" installs HTML renditions of the
9396 *Oliver Tappe <zooey@hirschkaefer.de>*
9398 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9399 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9400 support key and parameter generation and add initial key generation
9401 functionality for RSA.
9405 * Add functions for main EVP_PKEY_method operations. The undocumented
9406 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9407 `EVP_PKEY_{encrypt,decrypt}_old`.
9411 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9412 key API, doesn't do much yet.
9416 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9417 public key algorithms. New option to openssl utility:
9418 "list-public-key-algorithms" to print out info.
9422 * Implement the Supported Elliptic Curves Extension for
9423 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9427 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9428 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9432 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9433 utilities such as rsa, dsa, dsaparam etc except they process any key
9438 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9439 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9440 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9445 * Initial support for pluggable public key ASN1.
9446 De-spaghettify the public key ASN1 handling. Move public and private
9447 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9448 algorithm specific handling to a single module within the relevant
9449 algorithm directory. Add functions to allow (near) opaque processing
9450 of public and private key structures.
9454 * Implement the Supported Point Formats Extension for
9455 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9459 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9460 for the psk identity [hint] and the psk callback functions to the
9461 SSL_SESSION, SSL and SSL_CTX structure.
9464 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9468 SSL_CTX_use_psk_identity_hint
9469 SSL_get_psk_identity_hint
9470 SSL_get_psk_identity
9471 SSL_use_psk_identity_hint
9473 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9475 * Add RFC 3161 compliant time stamp request creation, response generation
9476 and response verification functionality.
9478 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9480 * Add initial support for TLS extensions, specifically for the server_name
9481 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9482 have new members for a host name. The SSL data structure has an
9483 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9484 stored in that context to allow for session resumption, even after the
9485 SSL has been switched to a new SSL_CTX in reaction to a client's
9486 server_name extension.
9488 New functions (subject to change):
9490 SSL_get_servername()
9491 SSL_get_servername_type()
9494 New CTRL codes and macros (subject to change):
9496 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9497 - SSL_CTX_set_tlsext_servername_callback()
9498 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9499 - SSL_CTX_set_tlsext_servername_arg()
9500 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9502 openssl s_client has a new '-servername ...' option.
9504 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9505 '-key2 ...', '-servername_fatal' (subject to change). This allows
9506 testing the HostName extension for a specific single host name ('-cert'
9507 and '-key' remain fallbacks for handshakes without HostName
9508 negotiation). If the unrecognized_name alert has to be sent, this by
9509 default is a warning; it becomes fatal with the '-servername_fatal'
9512 *Peter Sylvester, Remy Allais, Christophe Renou*
9514 * Whirlpool hash implementation is added.
9518 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9519 bn(64,32). Because of instruction set limitations it doesn't have
9520 any negative impact on performance. This was done mostly in order
9521 to make it possible to share assembler modules, such as bn_mul_mont
9522 implementations, between 32- and 64-bit builds without hassle.
9526 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9527 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9532 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9533 dedicated Montgomery multiplication procedure, is introduced.
9534 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9535 "64-bit" performance on certain 32-bit targets.
9539 * New option SSL_OP_NO_COMP to disable use of compression selectively
9540 in SSL structures. New SSL ctrl to set maximum send fragment size.
9541 Save memory by setting the I/O buffer sizes dynamically instead of
9542 using the maximum available value.
9546 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9547 in addition to the text details.
9551 * Very, very preliminary EXPERIMENTAL support for printing of general
9552 ASN1 structures. This currently produces rather ugly output and doesn't
9553 handle several customised structures at all.
9557 * Integrated support for PVK file format and some related formats such
9558 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9559 these in the 'rsa' and 'dsa' utilities.
9563 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9567 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9568 place for the (very old) "NETSCAPE" format certificates which are now
9569 handled using new ASN1 code equivalents.
9573 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9574 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9575 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9579 * Modify CRL distribution points extension code to print out previously
9580 unsupported fields. Enhance extension setting code to allow setting of
9585 * Add print and set support for Issuing Distribution Point CRL extension.
9589 * Change 'Configure' script to enable Camellia by default.
9596 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9598 * When rejecting SSL/TLS records due to an incorrect version number, never
9599 update s->server with a new major version number. As of
9600 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9601 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9602 the previous behavior could result in a read attempt at NULL when
9603 receiving specific incorrect SSL/TLS records once record payload
9604 protection is active. ([CVE-2010-0740])
9606 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9608 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9609 could be crashed if the relevant tables were not present (e.g. chrooted).
9611 *Tomas Hoger <thoger@redhat.com>*
9613 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9615 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9617 *Martin Olsson, Neel Mehta*
9619 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9620 accommodate for stack sorting, always a write lock!).
9624 * On some versions of WIN32 Heap32Next is very slow. This can cause
9625 excessive delays in the RAND_poll(): over a minute. As a workaround
9626 include a time check in the inner Heap32Next loop too.
9630 * The code that handled flushing of data in SSL/TLS originally used the
9631 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9632 the problem outlined in PR#1949. The fix suggested there however can
9633 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9634 of Apache). So instead simplify the code to flush unconditionally.
9635 This should be fine since flushing with no data to flush is a no op.
9639 * Handle TLS versions 2.0 and later properly and correctly use the
9640 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9641 off ancient servers have a habit of sticking around for a while...
9645 * Modify compression code so it frees up structures without using the
9646 ex_data callbacks. This works around a problem where some applications
9647 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9648 restarting) then use compression (e.g. SSL with compression) later.
9649 This results in significant per-connection memory leaks and
9650 has caused some security issues including CVE-2008-1678 and
9655 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9656 change when encrypting or decrypting.
9660 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9661 connect and renegotiate with servers which do not support RI.
9662 Until RI is more widely deployed this option is enabled by default.
9666 * Add "missing" ssl ctrls to clear options and mode.
9670 * If client attempts to renegotiate and doesn't support RI respond with
9671 a no_renegotiation alert as required by RFC5746. Some renegotiating
9672 TLS clients will continue a connection gracefully when they receive
9673 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9674 waiting for a server hello which it will never receive. Now we treat a
9675 received no_renegotiation alert as a fatal error. This is because
9676 applications requesting a renegotiation might well expect it to succeed
9677 and would have no code in place to handle the server denying it so the
9678 only safe thing to do is to terminate the connection.
9682 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9683 peer supports secure renegotiation and 0 otherwise. Print out peer
9684 renegotiation support in s_client/s_server.
9688 * Replace the highly broken and deprecated SPKAC certification method with
9689 the updated NID creation version. This should correctly handle UTF8.
9693 * Implement RFC5746. Re-enable renegotiation but require the extension
9694 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9695 turns out to be a bad idea. It has been replaced by
9696 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9697 SSL_CTX_set_options(). This is really not recommended unless you
9698 know what you are doing.
9700 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9702 * Fixes to stateless session resumption handling. Use initial_ctx when
9703 issuing and attempting to decrypt tickets in case it has changed during
9704 servername handling. Use a non-zero length session ID when attempting
9705 stateless session resumption: this makes it possible to determine if
9706 a resumption has occurred immediately after receiving server hello
9707 (several places in OpenSSL subtly assume this) instead of later in
9712 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9713 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9714 fixes for a few places where the return code is not checked
9717 *Julia Lawall <julia@diku.dk>*
9719 * Add --strict-warnings option to Configure script to include devteam
9720 warnings in other configurations.
9724 * Add support for --libdir option and LIBDIR variable in makefiles. This
9725 makes it possible to install openssl libraries in locations which
9726 have names other than "lib", for example "/usr/lib64" which some
9729 *Steve Henson, based on patch from Jeremy Utley*
9731 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9732 X690 8.9.12 and can produce some misleading textual output of OIDs.
9734 *Steve Henson, reported by Dan Kaminsky*
9736 * Delete MD2 from algorithm tables. This follows the recommendation in
9737 several standards that it is not used in new applications due to
9738 several cryptographic weaknesses. For binary compatibility reasons
9739 the MD2 API is still compiled in by default.
9743 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9748 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9749 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9752 *Guenter <lists@gknw.net>*
9754 * Fix the server certificate chain building code to use X509_verify_cert(),
9755 it used to have an ad-hoc builder which was unable to cope with anything
9756 other than a simple chain.
9758 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9760 * Don't check self signed certificate signatures in X509_verify_cert()
9761 by default (a flag can override this): it just wastes time without
9762 adding any security. As a useful side effect self signed root CAs
9763 with non-FIPS digests are now usable in FIPS mode.
9767 * In dtls1_process_out_of_seq_message() the check if the current message
9768 is already buffered was missing. For every new message was memory
9769 allocated, allowing an attacker to perform an denial of service attack
9770 with sending out of seq handshake messages until there is no memory
9771 left. Additionally every future message was buffered, even if the
9772 sequence number made no sense and would be part of another handshake.
9773 So only messages with sequence numbers less than 10 in advance will be
9774 buffered. ([CVE-2009-1378])
9776 *Robin Seggelmann, discovered by Daniel Mentz*
9778 * Records are buffered if they arrive with a future epoch to be
9779 processed after finishing the corresponding handshake. There is
9780 currently no limitation to this buffer allowing an attacker to perform
9781 a DOS attack with sending records with future epochs until there is no
9782 memory left. This patch adds the pqueue_size() function to determine
9783 the size of a buffer and limits the record buffer to 100 entries.
9786 *Robin Seggelmann, discovered by Daniel Mentz*
9788 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9789 parent structure is freed. ([CVE-2009-1379])
9793 * Handle non-blocking I/O properly in SSL_shutdown() call.
9795 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9797 * Add `2.5.4.*` OIDs
9799 *Ilya O. <vrghost@gmail.com>*
9801 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9803 * Disable renegotiation completely - this fixes a severe security
9804 problem ([CVE-2009-3555]) at the cost of breaking all
9805 renegotiation. Renegotiation can be re-enabled by setting
9806 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9807 run-time. This is really not recommended unless you know what
9812 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9814 * Don't set val to NULL when freeing up structures, it is freed up by
9815 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9816 zeroing past the valid field. ([CVE-2009-0789])
9818 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9820 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9821 checked correctly. This would allow some invalid signed attributes to
9822 appear to verify correctly. ([CVE-2009-0591])
9824 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9826 * Reject UniversalString and BMPString types with invalid lengths. This
9827 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9828 a legal length. ([CVE-2009-0590])
9832 * Set S/MIME signing as the default purpose rather than setting it
9833 unconditionally. This allows applications to override it at the store
9838 * Permit restricted recursion of ASN1 strings. This is needed in practice
9839 to handle some structures.
9843 * Improve efficiency of mem_gets: don't search whole buffer each time
9846 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9848 * New -hex option for openssl rand.
9852 * Print out UTF8String and NumericString when parsing ASN1.
9856 * Support NumericString type for name components.
9860 * Allow CC in the environment to override the automatically chosen
9861 compiler. Note that nothing is done to ensure flags work with the
9866 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9868 * Properly check EVP_VerifyFinal() and similar return values
9871 *Ben Laurie, Bodo Moeller, Google Security Team*
9873 * Enable TLS extensions by default.
9877 * Allow the CHIL engine to be loaded, whether the application is
9878 multithreaded or not. (This does not release the developer from the
9879 obligation to set up the dynamic locking callbacks.)
9881 *Sander Temme <sander@temme.net>*
9883 * Use correct exit code if there is an error in dgst command.
9885 *Steve Henson; problem pointed out by Roland Dirlewanger*
9887 * Tweak Configure so that you need to say "experimental-jpake" to enable
9888 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9892 * Add experimental JPAKE support, including demo authentication in
9893 s_client and s_server.
9897 * Set the comparison function in v3_addr_canonize().
9899 *Rob Austein <sra@hactrn.net>*
9901 * Add support for XMPP STARTTLS in s_client.
9903 *Philip Paeps <philip@freebsd.org>*
9905 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9906 to ensure that even with this option, only ciphersuites in the
9907 server's preference list will be accepted. (Note that the option
9908 applies only when resuming a session, so the earlier behavior was
9909 just about the algorithm choice for symmetric cryptography.)
9913 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9915 * Fix NULL pointer dereference if a DTLS server received
9916 ChangeCipherSpec as first record ([CVE-2009-1386]).
9920 * Fix a state transition in s3_srvr.c and d1_srvr.c
9921 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9925 * The fix in 0.9.8c that supposedly got rid of unsafe
9926 double-checked locking was incomplete for RSA blinding,
9927 addressing just one layer of what turns out to have been
9928 doubly unsafe triple-checked locking.
9930 So now fix this for real by retiring the MONT_HELPER macro
9931 in crypto/rsa/rsa_eay.c.
9933 *Bodo Moeller; problem pointed out by Marius Schilder*
9935 * Various precautionary measures:
9937 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9939 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9940 (NB: This would require knowledge of the secret session ticket key
9941 to exploit, in which case you'd be SOL either way.)
9943 - Change bn_nist.c so that it will properly handle input BIGNUMs
9944 outside the expected range.
9946 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9949 *Neel Mehta, Bodo Moeller*
9951 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9952 the load fails. Useful for distros.
9954 *Ben Laurie and the FreeBSD team*
9956 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9960 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9964 * Expand ENGINE to support engine supplied SSL client certificate functions.
9966 This work was sponsored by Logica.
9970 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9971 keystores. Support for SSL/TLS client authentication too.
9972 Not compiled unless enable-capieng specified to Configure.
9974 This work was sponsored by Logica.
9978 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9979 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9980 attribute creation routines such as certificate requests and PKCS#12
9985 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9987 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9988 handshake which could lead to a client crash as found using the
9989 Codenomicon TLS test suite ([CVE-2008-1672])
9991 *Steve Henson, Mark Cox*
9993 * Fix double free in TLS server name extensions which could lead to
9994 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9998 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10000 Clear the error queue to ensure that error entries left from
10001 older function calls do not interfere with the correct operation.
10003 *Lutz Jaenicke, Erik de Castro Lopo*
10005 * Remove root CA certificates of commercial CAs:
10007 The OpenSSL project does not recommend any specific CA and does not
10008 have any policy with respect to including or excluding any CA.
10009 Therefore it does not make any sense to ship an arbitrary selection
10010 of root CA certificates with the OpenSSL software.
10014 * RSA OAEP patches to fix two separate invalid memory reads.
10015 The first one involves inputs when 'lzero' is greater than
10016 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10017 before the beginning of from). The second one involves inputs where
10018 the 'db' section contains nothing but zeroes (there is a one-byte
10019 invalid read after the end of 'db').
10021 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10023 * Partial backport from 0.9.9-dev:
10025 Introduce bn_mul_mont (dedicated Montgomery multiplication
10026 procedure) as a candidate for BIGNUM assembler implementation.
10027 While 0.9.9-dev uses assembler for various architectures, only
10028 x86_64 is available by default here in the 0.9.8 branch, and
10029 32-bit x86 is available through a compile-time setting.
10031 To try the 32-bit x86 assembler implementation, use Configure
10032 option "enable-montasm" (which exists only for this backport).
10034 As "enable-montasm" for 32-bit x86 disclaims code stability
10035 anyway, in this constellation we activate additional code
10036 backported from 0.9.9-dev for further performance improvements,
10037 namely BN_from_montgomery_word. (To enable this otherwise,
10038 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10040 *Andy Polyakov (backport partially by Bodo Moeller)*
10042 * Add TLS session ticket callback. This allows an application to set
10043 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10044 values. This is useful for key rollover for example where several key
10045 sets may exist with different names.
10049 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10050 This was broken until now in 0.9.8 releases, such that the only way
10051 a registered ENGINE could be used (assuming it initialises
10052 successfully on the host) was to explicitly set it as the default
10053 for the relevant algorithms. This is in contradiction with 0.9.7
10054 behaviour and the documentation. With this fix, when an ENGINE is
10055 registered into a given algorithm's table of implementations, the
10056 'uptodate' flag is reset so that auto-discovery will be used next
10057 time a new context for that algorithm attempts to select an
10060 *Ian Lister (tweaked by Geoff Thorpe)*
10062 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10063 implementation in the following ways:
10065 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10068 Lack of BER streaming support means one pass streaming processing is
10069 only supported if data is detached: setting the streaming flag is
10070 ignored for embedded content.
10072 CMS support is disabled by default and must be explicitly enabled
10073 with the enable-cms configuration option.
10077 * Update the GMP engine glue to do direct copies between BIGNUM and
10078 mpz_t when openssl and GMP use the same limb size. Otherwise the
10079 existing "conversion via a text string export" trick is still used.
10081 *Paul Sheer <paulsheer@gmail.com>*
10083 * Zlib compression BIO. This is a filter BIO which compressed and
10084 uncompresses any data passed through it.
10088 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10089 RFC3394 compatible AES key wrapping.
10093 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10094 sets string data without copying. X509_ALGOR_set0() and
10095 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10096 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10097 from an X509_ATTRIBUTE structure optionally checking it occurs only
10098 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10103 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10104 to get the expected BN_FLG_CONSTTIME behavior.
10106 *Bodo Moeller (Google)*
10110 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10111 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10112 - added some more tests to do_tests.pl
10113 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10114 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10115 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10116 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10117 - various changes to netware.pl to enable gcc-cross builds on Win32
10119 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10120 - various changes to fix missing prototype warnings
10121 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10122 - added AES, WHIRLPOOL and CPUID assembler code to build files
10123 - added missing AES assembler make rules to mk1mf.pl
10124 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10126 *Guenter Knauf <eflash@gmx.net>*
10128 * Implement certificate status request TLS extension defined in RFC3546.
10129 A client can set the appropriate parameters and receive the encoded
10130 OCSP response via a callback. A server can query the supplied parameters
10131 and set the encoded OCSP response in the callback. Add simplified examples
10132 to s_client and s_server.
10136 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10138 * Fix various bugs:
10139 + Binary incompatibility of ssl_ctx_st structure
10140 + DTLS interoperation with non-compliant servers
10141 + Don't call get_session_cb() without proposed session
10142 + Fix ia64 assembler code
10144 *Andy Polyakov, Steve Henson*
10146 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10148 * DTLS Handshake overhaul. There were longstanding issues with
10149 OpenSSL DTLS implementation, which were making it impossible for
10150 RFC 4347 compliant client to communicate with OpenSSL server.
10151 Unfortunately just fixing these incompatibilities would "cut off"
10152 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10153 server keeps tolerating non RFC compliant syntax. The opposite is
10154 not true, 0.9.8f client can not communicate with earlier server.
10155 This update even addresses CVE-2007-4995.
10159 * Changes to avoid need for function casts in OpenSSL: some compilers
10160 (gcc 4.2 and later) reject their use.
10161 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10164 * Add RFC4507 support to OpenSSL. This includes the corrections in
10165 RFC4507bis. The encrypted ticket format is an encrypted encoded
10166 SSL_SESSION structure, that way new session features are automatically
10169 If a client application caches session in an SSL_SESSION structure
10170 support is transparent because tickets are now stored in the encoded
10173 The SSL_CTX structure automatically generates keys for ticket
10174 protection in servers so again support should be possible
10175 with no application modification.
10177 If a client or server wishes to disable RFC4507 support then the option
10178 SSL_OP_NO_TICKET can be set.
10180 Add a TLS extension debugging callback to allow the contents of any client
10181 or server extensions to be examined.
10183 This work was sponsored by Google.
10187 * Add initial support for TLS extensions, specifically for the server_name
10188 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10189 have new members for a host name. The SSL data structure has an
10190 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10191 stored in that context to allow for session resumption, even after the
10192 SSL has been switched to a new SSL_CTX in reaction to a client's
10193 server_name extension.
10195 New functions (subject to change):
10197 SSL_get_servername()
10198 SSL_get_servername_type()
10201 New CTRL codes and macros (subject to change):
10203 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10204 - SSL_CTX_set_tlsext_servername_callback()
10205 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10206 - SSL_CTX_set_tlsext_servername_arg()
10207 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10209 openssl s_client has a new '-servername ...' option.
10211 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10212 '-key2 ...', '-servername_fatal' (subject to change). This allows
10213 testing the HostName extension for a specific single host name ('-cert'
10214 and '-key' remain fallbacks for handshakes without HostName
10215 negotiation). If the unrecognized_name alert has to be sent, this by
10216 default is a warning; it becomes fatal with the '-servername_fatal'
10219 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10221 * Add AES and SSE2 assembly language support to VC++ build.
10225 * Mitigate attack on final subtraction in Montgomery reduction.
10229 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10230 (which previously caused an internal error).
10234 * Squeeze another 10% out of IGE mode when in != out.
10238 * AES IGE mode speedup.
10240 *Dean Gaudet (Google)*
10242 * Add the Korean symmetric 128-bit cipher SEED (see
10243 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10244 add SEED ciphersuites from RFC 4162:
10246 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10247 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10248 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10249 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10251 To minimize changes between patchlevels in the OpenSSL 0.9.8
10252 series, SEED remains excluded from compilation unless OpenSSL
10253 is configured with 'enable-seed'.
10255 *KISA, Bodo Moeller*
10257 * Mitigate branch prediction attacks, which can be practical if a
10258 single processor is shared, allowing a spy process to extract
10259 information. For detailed background information, see
10260 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10261 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10262 and Necessary Software Countermeasures"). The core of the change
10263 are new versions BN_div_no_branch() and
10264 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10265 respectively, which are slower, but avoid the security-relevant
10266 conditional branches. These are automatically called by BN_div()
10267 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10268 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10269 remove a conditional branch.
10271 BN_FLG_CONSTTIME is the new name for the previous
10272 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10273 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10274 in the exponent causes BN_mod_exp_mont() to use the alternative
10275 implementation in BN_mod_exp_mont_consttime().) The old name
10276 remains as a deprecated alias.
10278 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10279 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10280 constant-time implementations for more than just exponentiation.
10281 Here too the old name is kept as a deprecated alias.
10283 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10284 the BN_BLINDING structure gets an independent copy of the
10285 modulus. This means that the previous `BIGNUM *m` argument to
10286 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10287 essentially becomes `const BIGNUM *m`, although we can't actually
10288 change this in the header file before 0.9.9. It allows
10289 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10290 enable BN_FLG_CONSTTIME.
10292 *Matthew D Wood (Intel Corp)*
10294 * In the SSL/TLS server implementation, be strict about session ID
10295 context matching (which matters if an application uses a single
10296 external cache for different purposes). Previously,
10297 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10298 set. This did ensure strict client verification, but meant that,
10299 with applications using a single external cache for quite
10300 different requirements, clients could circumvent ciphersuite
10301 restrictions for a given session ID context by starting a session
10302 in a different context.
10306 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10307 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10308 authentication-only ciphersuites.
10312 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10313 not complete and could lead to a possible single byte overflow
10314 ([CVE-2007-5135]) [Ben Laurie]
10316 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10318 * Since AES128 and AES256 (and similarly Camellia128 and
10319 Camellia256) share a single mask bit in the logic of
10320 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10321 kludge to work properly if AES128 is available and AES256 isn't
10322 (or if Camellia128 is available and Camellia256 isn't).
10326 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10327 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10328 When a point or a seed is encoded in a BIT STRING, we need to
10329 prevent the removal of trailing zero bits to get the proper DER
10330 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10331 of a NamedBitList, for which trailing 0 bits need to be removed.)
10335 * Have SSL/TLS server implementation tolerate "mismatched" record
10336 protocol version while receiving ClientHello even if the
10337 ClientHello is fragmented. (The server can't insist on the
10338 particular protocol version it has chosen before the ServerHello
10339 message has informed the client about his choice.)
10343 * Add RFC 3779 support.
10345 *Rob Austein for ARIN, Ben Laurie*
10347 * Load error codes if they are not already present instead of using a
10348 static variable. This allows them to be cleanly unloaded and reloaded.
10349 Improve header file function name parsing.
10353 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10354 or CAPABILITY handshake as required by RFCs.
10356 *Goetz Babin-Ebell*
10358 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10360 * Introduce limits to prevent malicious keys being able to
10361 cause a denial of service. ([CVE-2006-2940])
10363 *Steve Henson, Bodo Moeller*
10365 * Fix ASN.1 parsing of certain invalid structures that can result
10366 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10368 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10369 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10371 * Fix SSL client code which could crash if connecting to a
10372 malicious SSLv2 server. ([CVE-2006-4343])
10374 *Tavis Ormandy and Will Drewry, Google Security Team*
10376 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10377 match only those. Before that, "AES256-SHA" would be interpreted
10378 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10379 the same strength classification in 0.9.7h) as we currently only
10380 have a single AES bit in the ciphersuite description bitmap.
10381 That change, however, also applied to ciphersuite strings such as
10382 "RC4-MD5" that intentionally matched multiple ciphersuites --
10383 namely, SSL 2.0 ciphersuites in addition to the more common ones
10384 from SSL 3.0/TLS 1.0.
10386 So we change the selection algorithm again: Naming an explicit
10387 ciphersuite selects this one ciphersuite, and any other similar
10388 ciphersuite (same bitmap) from *other* protocol versions.
10389 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10390 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10392 Since SSL 2.0 does not have any ciphersuites for which the
10393 128/256 bit distinction would be relevant, this works for now.
10394 The proper fix will be to use different bits for AES128 and
10395 AES256, which would have avoided the problems from the beginning;
10396 however, bits are scarce, so we can only do this in a new release
10397 (not just a patchlevel) when we can change the SSL_CIPHER
10398 definition to split the single 'unsigned long mask' bitmap into
10399 multiple values to extend the available space.
10403 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10405 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10406 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10408 * Add AES IGE and biIGE modes.
10412 * Change the Unix randomness entropy gathering to use poll() when
10413 possible instead of select(), since the latter has some
10414 undesirable limitations.
10416 *Darryl Miles via Richard Levitte and Bodo Moeller*
10418 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10419 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10420 cannot be implicitly activated as part of, e.g., the "AES" alias.
10421 However, please upgrade to OpenSSL 0.9.9[-dev] for
10422 non-experimental use of the ECC ciphersuites to get TLS extension
10423 support, which is required for curve and point format negotiation
10424 to avoid potential handshake problems.
10428 * Disable rogue ciphersuites:
10430 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10431 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10432 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10434 The latter two were purportedly from
10435 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10438 Also deactivate the remaining ciphersuites from
10439 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10440 unofficial, and the ID has long expired.
10444 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10445 dual-core machines) and other potential thread-safety issues.
10449 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10450 versions), which is now available for royalty-free use
10451 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10452 Also, add Camellia TLS ciphersuites from RFC 4132.
10454 To minimize changes between patchlevels in the OpenSSL 0.9.8
10455 series, Camellia remains excluded from compilation unless OpenSSL
10456 is configured with 'enable-camellia'.
10460 * Disable the padding bug check when compression is in use. The padding
10461 bug check assumes the first packet is of even length, this is not
10462 necessarily true if compression is enabled and can result in false
10463 positives causing handshake failure. The actual bug test is ancient
10464 code so it is hoped that implementations will either have fixed it by
10465 now or any which still have the bug do not support compression.
10469 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10471 * When applying a cipher rule check to see if string match is an explicit
10472 cipher suite and only match that one cipher suite if it is.
10476 * Link in manifests for VC++ if needed.
10478 *Austin Ziegler <halostatue@gmail.com>*
10480 * Update support for ECC-based TLS ciphersuites according to
10481 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10482 TLS extensions, which are supported starting with the 0.9.9
10483 branch, not in the OpenSSL 0.9.8 branch).
10487 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10488 opaque EVP_CIPHER_CTX handling.
10492 * Fixes and enhancements to zlib compression code. We now only use
10493 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10494 to conform with the standards mentioned here:
10495 <http://www.zlib.net/DLL_FAQ.txt>
10496 Static zlib linking now works on Windows and the new --with-zlib-include
10497 --with-zlib-lib options to Configure can be used to supply the location
10498 of the headers and library. Gracefully handle case where zlib library
10503 * Several fixes and enhancements to the OID generation code. The old code
10504 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10505 handle numbers larger than ULONG_MAX, truncated printing and had a
10506 non standard OBJ_obj2txt() behaviour.
10510 * Add support for building of engines under engine/ as shared libraries
10511 under VC++ build system.
10515 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10516 Hopefully, we will not see any false combination of paths any more.
10520 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10522 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10523 (part of SSL_OP_ALL). This option used to disable the
10524 countermeasure against man-in-the-middle protocol-version
10525 rollback in the SSL 2.0 server implementation, which is a bad
10526 idea. ([CVE-2005-2969])
10528 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10529 for Information Security, National Institute of Advanced Industrial
10530 Science and Technology [AIST], Japan)*
10532 * Add two function to clear and return the verify parameter flags.
10536 * Keep cipherlists sorted in the source instead of sorting them at
10537 runtime, thus removing the need for a lock.
10541 * Avoid some small subgroup attacks in Diffie-Hellman.
10543 *Nick Mathewson and Ben Laurie*
10545 * Add functions for well-known primes.
10549 * Extended Windows CE support.
10551 *Satoshi Nakamura and Andy Polyakov*
10553 * Initialize SSL_METHOD structures at compile time instead of during
10554 runtime, thus removing the need for a lock.
10558 * Make PKCS7_decrypt() work even if no certificate is supplied by
10559 attempting to decrypt each encrypted key in turn. Add support to
10564 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10566 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10569 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10573 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10574 key into the same file any more.
10578 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10582 * Add -utf8 command line and config file option to 'ca'.
10584 *Stefan <stf@udoma.org*
10586 * Removed the macro des_crypt(), as it seems to conflict with some
10587 libraries. Use DES_crypt().
10591 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10592 involves renaming the source and generated shared-libs for
10593 both. The engines will accept the corrected or legacy ids
10594 ('ncipher' and '4758_cca' respectively) when binding. NB,
10595 this only applies when building 'shared'.
10597 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10599 * Add attribute functions to EVP_PKEY structure. Modify
10600 PKCS12_create() to recognize a CSP name attribute and
10601 use it. Make -CSP option work again in pkcs12 utility.
10605 * Add new functionality to the bn blinding code:
10606 - automatic re-creation of the BN_BLINDING parameters after
10607 a fixed number of uses (currently 32)
10608 - add new function for parameter creation
10609 - introduce flags to control the update behaviour of the
10610 BN_BLINDING parameters
10611 - hide BN_BLINDING structure
10612 Add a second BN_BLINDING slot to the RSA structure to improve
10613 performance when a single RSA object is shared among several
10618 * Add support for DTLS.
10620 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10622 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10623 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10627 * Remove buggy and incomplete DH cert support from
10628 ssl/ssl_rsa.c and ssl/s3_both.c
10632 * Use SHA-1 instead of MD5 as the default digest algorithm for
10633 the `apps/openssl` commands.
10637 * Compile clean with "-Wall -Wmissing-prototypes
10638 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10639 DEBUG_SAFESTACK must also be set.
10643 * Change ./Configure so that certain algorithms can be disabled by default.
10644 The new counterpiece to "no-xxx" is "enable-xxx".
10646 The patented RC5 and MDC2 algorithms will now be disabled unless
10647 "enable-rc5" and "enable-mdc2", respectively, are specified.
10649 (IDEA remains enabled despite being patented. This is because IDEA
10650 is frequently required for interoperability, and there is no license
10651 fee for non-commercial use. As before, "no-idea" can be used to
10652 avoid this algorithm.)
10656 * Add processing of proxy certificates (see RFC 3820). This work was
10657 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10658 EGEE (Enabling Grids for E-science in Europe).
10662 * RC4 performance overhaul on modern architectures/implementations, such
10663 as Intel P4, IA-64 and AMD64.
10667 * New utility extract-section.pl. This can be used specify an alternative
10668 section number in a pod file instead of having to treat each file as
10669 a separate case in Makefile. This can be done by adding two lines to the
10672 =for comment openssl_section:XXX
10674 The blank line is mandatory.
10678 * New arguments -certform, -keyform and -pass for s_client and s_server
10679 to allow alternative format key and certificate files and passphrase
10684 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10685 update associated structures and add various utility functions.
10687 Add new policy related verify parameters, include policy checking in
10688 standard verify code. Enhance 'smime' application with extra parameters
10689 to support policy checking and print out.
10693 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10694 Nehemiah processors. These extensions support AES encryption in hardware
10695 as well as RNG (though RNG support is currently disabled).
10697 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10699 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10703 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10705 *Andy Polyakov and a number of other people*
10707 * Improved PowerPC platform support. Most notably BIGNUM assembler
10708 implementation contributed by IBM.
10710 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10712 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10713 exponent rather than 'unsigned long'. There is a corresponding change to
10714 the new 'rsa_keygen' element of the RSA_METHOD structure.
10716 *Jelte Jansen, Geoff Thorpe*
10718 * Functionality for creating the initial serial number file is now
10719 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10721 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10722 number file to 1, which is bound to cause problems. To avoid
10723 the problems while respecting compatibility between different 0.9.7
10724 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10725 CA.pl for serial number initialization. With the new release 0.9.8,
10726 we can fix the problem directly in the 'ca' utility.)
10730 * Reduced header interdependencies by declaring more opaque objects in
10731 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10732 give fewer recursive includes, which could break lazy source code - so
10733 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10734 developers should define this symbol when building and using openssl to
10735 ensure they track the recommended behaviour, interfaces, [etc], but
10736 backwards-compatible behaviour prevails when this isn't defined.
10740 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10744 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10745 This will generate a random key of the appropriate length based on the
10746 cipher context. The EVP_CIPHER can provide its own random key generation
10747 routine to support keys of a specific form. This is used in the des and
10748 3des routines to generate a key of the correct parity. Update S/MIME
10749 code to use new functions and hence generate correct parity DES keys.
10750 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10751 valid (weak or incorrect parity).
10755 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10756 as looking them up. This is useful when the verified structure may contain
10757 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10758 present unless the new PKCS7_NO_CRL flag is asserted.
10762 * Extend ASN1 oid configuration module. It now additionally accepts the
10765 shortName = some long name, 1.2.3.4
10769 * Reimplemented the BN_CTX implementation. There is now no more static
10770 limitation on the number of variables it can handle nor the depth of the
10771 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10772 information can now expand as required, and rather than having a single
10773 static array of bignums, BN_CTX now uses a linked-list of such arrays
10774 allowing it to expand on demand whilst maintaining the usefulness of
10775 BN_CTX's "bundling".
10779 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10780 to allow all RSA operations to function using a single BN_CTX.
10784 * Preliminary support for certificate policy evaluation and checking. This
10785 is initially intended to pass the tests outlined in "Conformance Testing
10786 of Relying Party Client Certificate Path Processing Logic" v1.07.
10790 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10791 remained unused and not that useful. A variety of other little bignum
10792 tweaks and fixes have also been made continuing on from the audit (see
10797 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10798 associated ASN1, EVP and SSL functions and old ASN1 macros.
10802 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10803 and this should never fail. So the return value from the use of
10804 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10805 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10809 * BN_CTX_get() should return zero-valued bignums, providing the same
10810 initialised value as BN_new().
10812 *Geoff Thorpe, suggested by Ulf Möller*
10814 * Support for inhibitAnyPolicy certificate extension.
10818 * An audit of the BIGNUM code is underway, for which debugging code is
10819 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10820 is considered valid when processing BIGNUMs, and causes execution to
10821 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10822 further steps are taken to deliberately pollute unused data in BIGNUM
10823 structures to try and expose faulty code further on. For now, openssl will
10824 (in its default mode of operation) continue to tolerate the inconsistent
10825 forms that it has tolerated in the past, but authors and packagers should
10826 consider trying openssl and their own applications when compiled with
10827 these debugging symbols defined. It will help highlight potential bugs in
10828 their own code, and will improve the test coverage for OpenSSL itself. At
10829 some point, these tighter rules will become openssl's default to improve
10830 maintainability, though the assert()s and other overheads will remain only
10831 in debugging configurations. See bn.h for more details.
10833 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10835 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10836 that can only be obtained through BN_CTX_new() (which implicitly
10837 initialises it). The presence of this function only made it possible
10838 to overwrite an existing structure (and cause memory leaks).
10842 * Because of the callback-based approach for implementing LHASH as a
10843 template type, lh_insert() adds opaque objects to hash-tables and
10844 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10845 to clean up those corresponding objects before destroying the hash table
10846 (and losing the object pointers). So some over-zealous constifications in
10847 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10848 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10849 prototyped to have "const" restrictions on the object pointers they are
10850 given (and so aren't required to cast them away any more).
10854 * The tmdiff.h API was so ugly and minimal that our own timing utility
10855 (speed) prefers to use its own implementation. The two implementations
10856 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10857 its object type properly exposed (MS_TM) instead of casting to/from
10858 `char *`. This may still change yet if someone realises MS_TM and
10860 aren't necessarily the greatest nomenclatures - but this is what was used
10861 internally to the implementation so I've used that for now.
10865 * Ensure that deprecated functions do not get compiled when
10866 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10867 the self-tests were still using deprecated key-generation functions so
10868 these have been updated also.
10872 * Reorganise PKCS#7 code to separate the digest location functionality
10873 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10874 New function PKCS7_set_digest() to set the digest type for PKCS#7
10875 digestedData type. Add additional code to correctly generate the
10876 digestedData type and add support for this type in PKCS7 initialization
10881 * New function PKCS7_set0_type_other() this initializes a PKCS7
10882 structure of type "other".
10886 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10887 sure the loop does correctly stop and breaking ("division by zero")
10888 modulus operations are not performed. The (pre-generated) prime
10889 table crypto/bn/bn_prime.h was already correct, but it could not be
10890 re-generated on some platforms because of the "division by zero"
10891 situation in the script.
10893 *Ralf S. Engelschall*
10895 * Update support for ECC-based TLS ciphersuites according to
10896 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10897 SHA-1 now is only used for "small" curves (where the
10898 representation of a field element takes up to 24 bytes); for
10899 larger curves, the field element resulting from ECDH is directly
10900 used as premaster secret.
10902 *Douglas Stebila (Sun Microsystems Laboratories)*
10904 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10905 curve secp160r1 to the tests.
10907 *Douglas Stebila (Sun Microsystems Laboratories)*
10909 * Add the possibility to load symbols globally with DSO.
10911 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10913 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10914 control of the error stack.
10918 * Add support for STORE in ENGINE.
10922 * Add the STORE type. The intention is to provide a common interface
10923 to certificate and key stores, be they simple file-based stores, or
10924 HSM-type store, or LDAP stores, or...
10925 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10929 * Add a generic structure called OPENSSL_ITEM. This can be used to
10930 pass a list of arguments to any function as well as provide a way
10931 for a function to pass data back to the caller.
10935 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10936 works like BUF_strdup() but can be used to duplicate a portion of
10937 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10942 * Add the function sk_find_ex() which works like sk_find(), but will
10943 return an index to an element even if an exact match couldn't be
10944 found. The index is guaranteed to point at the element where the
10945 searched-for key would be inserted to preserve sorting order.
10949 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10950 takes an extra flags argument for optional functionality. Currently,
10951 the following flags are defined:
10953 OBJ_BSEARCH_VALUE_ON_NOMATCH
10954 This one gets OBJ_bsearch_ex() to return a pointer to the first
10955 element where the comparing function returns a negative or zero
10958 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10959 This one gets OBJ_bsearch_ex() to return a pointer to the first
10960 element where the comparing function returns zero. This is useful
10961 if there are more than one element where the comparing function
10966 * Make it possible to create self-signed certificates with 'openssl ca'
10967 in such a way that the self-signed certificate becomes part of the
10968 CA database and uses the same mechanisms for serial number generation
10969 as all other certificate signing. The new flag '-selfsign' enables
10970 this functionality. Adapt CA.sh and CA.pl.in.
10974 * Add functionality to check the public key of a certificate request
10975 against a given private. This is useful to check that a certificate
10976 request can be signed by that key (self-signing).
10980 * Make it possible to have multiple active certificates with the same
10981 subject in the CA index file. This is done only if the keyword
10982 'unique_subject' is set to 'no' in the main CA section (default
10983 if 'CA_default') of the configuration file. The value is saved
10984 with the database itself in a separate index attribute file,
10985 named like the index file with '.attr' appended to the name.
10989 * Generate multi-valued AVAs using '+' notation in config files for
10994 * Support for nameConstraints certificate extension.
10998 * Support for policyConstraints certificate extension.
11002 * Support for policyMappings certificate extension.
11006 * Make sure the default DSA_METHOD implementation only uses its
11007 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11008 and change its own handlers to be NULL so as to remove unnecessary
11009 indirection. This lets alternative implementations fallback to the
11010 default implementation more easily.
11014 * Support for directoryName in GeneralName related extensions
11019 * Make it possible to link applications using Makefile.shared.
11020 Make that possible even when linking against static libraries!
11024 * Support for single pass processing for S/MIME signing. This now
11025 means that S/MIME signing can be done from a pipe, in addition
11026 cleartext signing (multipart/signed type) is effectively streaming
11027 and the signed data does not need to be all held in memory.
11029 This is done with a new flag PKCS7_STREAM. When this flag is set
11030 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11031 is done after the data is output (and digests calculated) in
11032 SMIME_write_PKCS7().
11036 * Add full support for -rpath/-R, both in shared libraries and
11037 applications, at least on the platforms where it's known how
11042 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11043 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11044 will now compute a table of multiples of the generator that
11045 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11046 faster (notably in the case of a single point multiplication,
11047 scalar * generator).
11049 *Nils Larsch, Bodo Moeller*
11051 * IPv6 support for certificate extensions. The various extensions
11052 which use the IP:a.b.c.d can now take IPv6 addresses using the
11053 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11058 * Added an ENGINE that implements RSA by performing private key
11059 exponentiations with the GMP library. The conversions to and from
11060 GMP's mpz_t format aren't optimised nor are any montgomery forms
11061 cached, and on x86 it appears OpenSSL's own performance has caught up.
11062 However there are likely to be other architectures where GMP could
11063 provide a boost. This ENGINE is not built in by default, but it can be
11064 specified at Configure time and should be accompanied by the necessary
11065 linker additions, eg;
11066 ./config -DOPENSSL_USE_GMP -lgmp
11070 * "openssl engine" will not display ENGINE/DSO load failure errors when
11071 testing availability of engines with "-t" - the old behaviour is
11072 produced by increasing the feature's verbosity with "-tt".
11076 * ECDSA routines: under certain error conditions uninitialized BN objects
11077 could be freed. Solution: make sure initialization is performed early
11078 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11083 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11084 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11085 software implementations. For DSA and DH, parameter generation can
11086 also be overridden by providing the appropriate method callbacks.
11090 * Change the "progress" mechanism used in key-generation and
11091 primality testing to functions that take a new BN_GENCB pointer in
11092 place of callback/argument pairs. The new API functions have `_ex`
11093 postfixes and the older functions are reimplemented as wrappers for
11094 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11095 declarations of the old functions to help (graceful) attempts to
11096 migrate to the new functions. Also, the new key-generation API
11097 functions operate on a caller-supplied key-structure and return
11098 success/failure rather than returning a key or NULL - this is to
11099 help make "keygen" another member function of RSA_METHOD etc.
11101 Example for using the new callback interface:
11103 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11104 void *my_arg = ...;
11107 BN_GENCB_set(&my_cb, my_callback, my_arg);
11109 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11110 /* For the meaning of a, b in calls to my_callback(), see the
11111 * documentation of the function that calls the callback.
11112 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11113 * my_callback should return 1 if it wants BN_is_prime_ex()
11114 * to continue, or 0 to stop.
11119 * Change the ZLIB compression method to be stateful, and make it
11120 available to TLS with the number defined in
11121 draft-ietf-tls-compression-04.txt.
11125 * Add the ASN.1 structures and functions for CertificatePair, which
11126 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11128 CertificatePair ::= SEQUENCE {
11129 forward [0] Certificate OPTIONAL,
11130 reverse [1] Certificate OPTIONAL,
11131 -- at least one of the pair shall be present -- }
11133 Also implement the PEM functions to read and write certificate
11134 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11136 This needed to be defined, mostly for the sake of the LDAP
11137 attribute crossCertificatePair, but may prove useful elsewhere as
11142 * Make it possible to inhibit symlinking of shared libraries in
11143 Makefile.shared, for Cygwin's sake.
11147 * Extend the BIGNUM API by creating a function
11148 void BN_set_negative(BIGNUM *a, int neg);
11149 and a macro that behave like
11150 int BN_is_negative(const BIGNUM *a);
11152 to avoid the need to access 'a->neg' directly in applications.
11156 * Implement fast modular reduction for pseudo-Mersenne primes
11157 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11158 EC_GROUP_new_curve_GFp() will now automatically use this
11161 *Nils Larsch <nla@trustcenter.de>*
11163 * Add new lock type (CRYPTO_LOCK_BN).
11167 * Change the ENGINE framework to automatically load engines
11168 dynamically from specific directories unless they could be
11169 found to already be built in or loaded. Move all the
11170 current engines except for the cryptodev one to a new
11171 directory engines/.
11172 The engines in engines/ are built as shared libraries if
11173 the "shared" options was given to ./Configure or ./config.
11174 Otherwise, they are inserted in libcrypto.a.
11175 /usr/local/ssl/engines is the default directory for dynamic
11176 engines, but that can be overridden at configure time through
11177 the usual use of --prefix and/or --openssldir, and at run
11178 time with the environment variable OPENSSL_ENGINES.
11180 *Geoff Thorpe and Richard Levitte*
11182 * Add Makefile.shared, a helper makefile to build shared
11183 libraries. Adapt Makefile.org.
11187 * Add version info to Win32 DLLs.
11189 *Peter 'Luna' Runestig" <peter@runestig.com>*
11191 * Add new 'medium level' PKCS#12 API. Certificates and keys
11192 can be added using this API to created arbitrary PKCS#12
11193 files while avoiding the low-level API.
11195 New options to PKCS12_create(), key or cert can be NULL and
11196 will then be omitted from the output file. The encryption
11197 algorithm NIDs can be set to -1 for no encryption, the mac
11198 iteration count can be set to 0 to omit the mac.
11200 Enhance pkcs12 utility by making the -nokeys and -nocerts
11201 options work when creating a PKCS#12 file. New option -nomac
11202 to omit the mac, NONE can be set for an encryption algorithm.
11203 New code is modified to use the enhanced PKCS12_create()
11204 instead of the low-level API.
11208 * Extend ASN1 encoder to support indefinite length constructed
11209 encoding. This can output sequences tags and octet strings in
11210 this form. Modify pk7_asn1.c to support indefinite length
11211 encoding. This is experimental and needs additional code to
11212 be useful, such as an ASN1 bio and some enhanced streaming
11215 Extend template encode functionality so that tagging is passed
11216 down to the template encoder.
11220 * Let 'openssl req' fail if an argument to '-newkey' is not
11221 recognized instead of using RSA as a default.
11225 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11226 As these are not official, they are not included in "ALL";
11227 the "ECCdraft" ciphersuite group alias can be used to select them.
11229 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11231 * Add ECDH engine support.
11233 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11235 * Add ECDH in new directory crypto/ecdh/.
11237 *Douglas Stebila (Sun Microsystems Laboratories)*
11239 * Let BN_rand_range() abort with an error after 100 iterations
11240 without success (which indicates a broken PRNG).
11244 * Change BN_mod_sqrt() so that it verifies that the input value
11245 is really the square of the return value. (Previously,
11246 BN_mod_sqrt would show GIGO behaviour.)
11250 * Add named elliptic curves over binary fields from X9.62, SECG,
11251 and WAP/WTLS; add OIDs that were still missing.
11253 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11255 * Extend the EC library for elliptic curves over binary fields
11256 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11259 EC_GF2m_simple_method
11263 EC_GROUP_new_curve_GF2m
11264 EC_GROUP_set_curve_GF2m
11265 EC_GROUP_get_curve_GF2m
11266 EC_POINT_set_affine_coordinates_GF2m
11267 EC_POINT_get_affine_coordinates_GF2m
11268 EC_POINT_set_compressed_coordinates_GF2m
11270 Point compression for binary fields is disabled by default for
11271 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11274 As binary polynomials are represented as BIGNUMs, various members
11275 of the EC_GROUP and EC_POINT data structures can be shared
11276 between the implementations for prime fields and binary fields;
11277 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11278 are essentially identical to their `..._GFp` counterparts.
11279 (For simplicity, the `..._GFp` prefix has been dropped from
11280 various internal method names.)
11282 An internal 'field_div' method (similar to 'field_mul' and
11283 'field_sqr') has been added; this is used only for binary fields.
11285 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11287 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11288 through methods ('mul', 'precompute_mult').
11290 The generic implementations (now internally called 'ec_wNAF_mul'
11291 and 'ec_wNAF_precomputed_mult') remain the default if these
11292 methods are undefined.
11294 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11296 * New function EC_GROUP_get_degree, which is defined through
11297 EC_METHOD. For curves over prime fields, this returns the bit
11298 length of the modulus.
11300 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11302 * New functions EC_GROUP_dup, EC_POINT_dup.
11303 (These simply call ..._new and ..._copy).
11305 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11307 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11308 Polynomials are represented as BIGNUMs (where the sign bit is not
11309 used) in the following functions [macros]:
11312 BN_GF2m_sub [= BN_GF2m_add]
11313 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11314 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11315 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11317 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11318 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11319 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11320 BN_GF2m_cmp [= BN_ucmp]
11322 (Note that only the 'mod' functions are actually for fields GF(2^m).
11323 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11325 For some functions, an the irreducible polynomial defining a
11326 field can be given as an 'unsigned int[]' with strictly
11327 decreasing elements giving the indices of those bits that are set;
11328 i.e., p[] represents the polynomial
11329 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11331 p[0] > p[1] > ... > p[k] = 0.
11332 This applies to the following functions:
11335 BN_GF2m_mod_mul_arr
11336 BN_GF2m_mod_sqr_arr
11337 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11338 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11339 BN_GF2m_mod_exp_arr
11340 BN_GF2m_mod_sqrt_arr
11341 BN_GF2m_mod_solve_quad_arr
11345 Conversion can be performed by the following functions:
11350 bntest.c has additional tests for binary polynomial arithmetic.
11352 Two implementations for BN_GF2m_mod_div() are available.
11353 The default algorithm simply uses BN_GF2m_mod_inv() and
11354 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11355 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11356 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11358 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11360 * Add new error code 'ERR_R_DISABLED' that can be used when some
11361 functionality is disabled at compile-time.
11363 *Douglas Stebila <douglas.stebila@sun.com>*
11365 * Change default behaviour of 'openssl asn1parse' so that more
11366 information is visible when viewing, e.g., a certificate:
11368 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11369 mode the content of non-printable OCTET STRINGs is output in a
11370 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11371 avoid the appearance of a printable string.
11373 *Nils Larsch <nla@trustcenter.de>*
11375 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11377 EC_GROUP_set_asn1_flag()
11378 EC_GROUP_get_asn1_flag()
11379 EC_GROUP_set_point_conversion_form()
11380 EC_GROUP_get_point_conversion_form()
11381 These control ASN1 encoding details:
11382 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11383 has been set to OPENSSL_EC_NAMED_CURVE.
11384 - Points are encoded in uncompressed form by default; options for
11385 asn1_for are as for point2oct, namely
11386 POINT_CONVERSION_COMPRESSED
11387 POINT_CONVERSION_UNCOMPRESSED
11388 POINT_CONVERSION_HYBRID
11390 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11392 EC_GROUP_set_seed()
11393 EC_GROUP_get0_seed()
11394 EC_GROUP_get_seed_len()
11395 This is used only for ASN1 purposes (so far).
11397 *Nils Larsch <nla@trustcenter.de>*
11399 * Add 'field_type' member to EC_METHOD, which holds the NID
11400 of the appropriate field type OID. The new function
11401 EC_METHOD_get_field_type() returns this value.
11403 *Nils Larsch <nla@trustcenter.de>*
11406 EC_POINT_point2bn()
11407 EC_POINT_bn2point()
11408 EC_POINT_point2hex()
11409 EC_POINT_hex2point()
11410 providing useful interfaces to EC_POINT_point2oct() and
11411 EC_POINT_oct2point().
11413 *Nils Larsch <nla@trustcenter.de>*
11415 * Change internals of the EC library so that the functions
11416 EC_GROUP_set_generator()
11417 EC_GROUP_get_generator()
11418 EC_GROUP_get_order()
11419 EC_GROUP_get_cofactor()
11420 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11421 to methods, which would lead to unnecessary code duplication when
11422 adding different types of curves.
11424 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11426 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11427 arithmetic, and such that modified wNAFs are generated
11428 (which avoid length expansion in many cases).
11432 * Add a function EC_GROUP_check_discriminant() (defined via
11433 EC_METHOD) that verifies that the curve discriminant is non-zero.
11435 Add a function EC_GROUP_check() that makes some sanity tests
11436 on a EC_GROUP, its generator and order. This includes
11437 EC_GROUP_check_discriminant().
11439 *Nils Larsch <nla@trustcenter.de>*
11441 * Add ECDSA in new directory crypto/ecdsa/.
11443 Add applications 'openssl ecparam' and 'openssl ecdsa'
11444 (these are based on 'openssl dsaparam' and 'openssl dsa').
11446 ECDSA support is also included in various other files across the
11447 library. Most notably,
11448 - 'openssl req' now has a '-newkey ecdsa:file' option;
11449 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11450 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11451 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11452 them suitable for ECDSA where domain parameters must be
11453 extracted before the specific public key;
11454 - ECDSA engine support has been added.
11456 *Nils Larsch <nla@trustcenter.de>*
11458 * Include some named elliptic curves, and add OIDs from X9.62,
11459 SECG, and WAP/WTLS. Each curve can be obtained from the new
11461 EC_GROUP_new_by_curve_name(),
11462 and the list of available named curves can be obtained with
11463 EC_get_builtin_curves().
11464 Also add a 'curve_name' member to EC_GROUP objects, which can be
11466 EC_GROUP_set_curve_name()
11467 EC_GROUP_get_curve_name()
11469 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11471 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11472 was actually never needed) and in BN_mul(). The removal in BN_mul()
11473 required a small change in bn_mul_part_recursive() and the addition
11474 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11475 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11476 bn_sub_words() and bn_add_words() except they take arrays with
11481 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11483 * Cleanse PEM buffers before freeing them since they may contain
11486 *Benjamin Bennett <ben@psc.edu>*
11488 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11489 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11490 authentication-only ciphersuites.
11494 * Since AES128 and AES256 share a single mask bit in the logic of
11495 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11496 kludge to work properly if AES128 is available and AES256 isn't.
11500 * Expand security boundary to match 1.1.1 module.
11504 * Remove redundant features: hash file source, editing of test vectors
11505 modify fipsld to use external fips_premain.c signature.
11509 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11510 run algorithm test programs.
11514 * Make algorithm test programs more tolerant of whitespace.
11518 * Have SSL/TLS server implementation tolerate "mismatched" record
11519 protocol version while receiving ClientHello even if the
11520 ClientHello is fragmented. (The server can't insist on the
11521 particular protocol version it has chosen before the ServerHello
11522 message has informed the client about his choice.)
11526 * Load error codes if they are not already present instead of using a
11527 static variable. This allows them to be cleanly unloaded and reloaded.
11531 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11533 * Introduce limits to prevent malicious keys being able to
11534 cause a denial of service. ([CVE-2006-2940])
11536 *Steve Henson, Bodo Moeller*
11538 * Fix ASN.1 parsing of certain invalid structures that can result
11539 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11541 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11542 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11544 * Fix SSL client code which could crash if connecting to a
11545 malicious SSLv2 server. ([CVE-2006-4343])
11547 *Tavis Ormandy and Will Drewry, Google Security Team*
11549 * Change ciphersuite string processing so that an explicit
11550 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11551 will no longer include "AES128-SHA"), and any other similar
11552 ciphersuite (same bitmap) from *other* protocol versions (so that
11553 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11554 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11555 changes from 0.9.8b and 0.9.8d.
11559 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11561 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11562 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11564 * Change the Unix randomness entropy gathering to use poll() when
11565 possible instead of select(), since the latter has some
11566 undesirable limitations.
11568 *Darryl Miles via Richard Levitte and Bodo Moeller*
11570 * Disable rogue ciphersuites:
11572 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11573 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11574 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11576 The latter two were purportedly from
11577 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11580 Also deactivate the remaining ciphersuites from
11581 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11582 unofficial, and the ID has long expired.
11586 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11587 dual-core machines) and other potential thread-safety issues.
11591 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11593 * Adapt fipsld and the build system to link against the validated FIPS
11594 module in FIPS mode.
11598 * Fixes for VC++ 2005 build under Windows.
11602 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11603 from a Windows bash shell such as MSYS. It is autodetected from the
11604 "config" script when run from a VC++ environment. Modify standard VC++
11605 build to use fipscanister.o from the GNU make build.
11609 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11611 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11612 The value now differs depending on if you build for FIPS or not.
11613 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11614 safely run with a non-FIPSed libcrypto, as it may crash because of
11615 the difference induced by this change.
11619 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11621 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11622 (part of SSL_OP_ALL). This option used to disable the
11623 countermeasure against man-in-the-middle protocol-version
11624 rollback in the SSL 2.0 server implementation, which is a bad
11625 idea. ([CVE-2005-2969])
11627 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11628 for Information Security, National Institute of Advanced Industrial
11629 Science and Technology [AIST, Japan)]*
11631 * Minimal support for X9.31 signatures and PSS padding modes. This is
11632 mainly for FIPS compliance and not fully integrated at this stage.
11636 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11637 the exponentiation using a fixed-length exponent. (Otherwise,
11638 the information leaked through timing could expose the secret key
11639 after many signatures; cf. Bleichenbacher's attack on DSA with
11644 * Make a new fixed-window mod_exp implementation the default for
11645 RSA, DSA, and DH private-key operations so that the sequence of
11646 squares and multiplies and the memory access pattern are
11647 independent of the particular secret key. This will mitigate
11648 cache-timing and potential related attacks.
11650 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11651 and this is automatically used by BN_mod_exp_mont() if the new flag
11652 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11653 will use this BN flag for private exponents unless the flag
11654 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11655 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11657 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11659 * Change the client implementation for SSLv23_method() and
11660 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11661 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11662 (Previously, the SSL 2.0 backwards compatible Client Hello
11663 message format would be used even with SSL_OP_NO_SSLv2.)
11667 * Add support for smime-type MIME parameter in S/MIME messages which some
11672 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11673 a threadsafe manner. Modify rsa code to use new function and add calls
11674 to dsa and dh code (which had race conditions before).
11678 * Include the fixed error library code in the C error file definitions
11679 instead of fixing them up at runtime. This keeps the error code
11680 structures constant.
11684 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11686 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11689 * Fixes for newer kerberos headers. NB: the casts are needed because
11690 the 'length' field is signed on one version and unsigned on another
11691 with no (?) obvious way to tell the difference, without these VC++
11692 complains. Also the "definition" of FAR (blank) is no longer included
11693 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11694 some needed definitions.
11698 * Undo Cygwin change.
11702 * Added support for proxy certificates according to RFC 3820.
11703 Because they may be a security thread to unaware applications,
11704 they must be explicitly allowed in run-time. See
11705 docs/HOWTO/proxy_certificates.txt for further information.
11709 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11711 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11712 server and client random values. Previously
11713 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11714 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11716 This change has negligible security impact because:
11718 1. Server and client random values still have 24 bytes of pseudo random
11721 2. Server and client random values are sent in the clear in the initial
11724 3. The master secret is derived using the premaster secret (48 bytes in
11725 size for static RSA ciphersuites) as well as client server and random
11728 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11731 *Stephen Henson, reported by UK NISCC*
11733 * Use Windows randomness collection on Cygwin.
11737 * Fix hang in EGD/PRNGD query when communication socket is closed
11738 prematurely by EGD/PRNGD.
11740 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11742 * Prompt for pass phrases when appropriate for PKCS12 input format.
11746 * Back-port of selected performance improvements from development
11747 branch, as well as improved support for PowerPC platforms.
11751 * Add lots of checks for memory allocation failure, error codes to indicate
11752 failure and freeing up memory if a failure occurs.
11754 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11756 * Add new -passin argument to dgst.
11760 * Perform some character comparisons of different types in X509_NAME_cmp:
11761 this is needed for some certificates that re-encode DNs into UTF8Strings
11762 (in violation of RFC3280) and can't or won't issue name rollover
11767 * Make an explicit check during certificate validation to see that
11768 the CA setting in each certificate on the chain is correct. As a
11769 side effect always do the following basic checks on extensions,
11770 not just when there's an associated purpose to the check:
11772 - if there is an unhandled critical extension (unless the user
11773 has chosen to ignore this fault)
11774 - if the path length has been exceeded (if one is set at all)
11775 - that certain extensions fit the associated purpose (if one has
11780 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11782 * Avoid a race condition when CRLs are checked in a multi threaded
11783 environment. This would happen due to the reordering of the revoked
11784 entries during signature checking and serial number lookup. Now the
11785 encoding is cached and the serial number sort performed under a lock.
11786 Add new STACK function sk_is_sorted().
11790 * Add Delta CRL to the extension code.
11794 * Various fixes to s3_pkt.c so alerts are sent properly.
11796 *David Holmes <d.holmes@f5.com>*
11798 * Reduce the chances of duplicate issuer name and serial numbers (in
11799 violation of RFC3280) using the OpenSSL certificate creation utilities.
11800 This is done by creating a random 64 bit value for the initial serial
11801 number when a serial number file is created or when a self signed
11802 certificate is created using 'openssl req -x509'. The initial serial
11803 number file is created using 'openssl x509 -next_serial' in CA.pl
11804 rather than being initialized to 1.
11808 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11810 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11811 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11813 *Joe Orton, Steve Henson*
11815 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11818 *Joe Orton, Steve Henson*
11820 * Make it possible to have multiple active certificates with the same
11821 subject in the CA index file. This is done only if the keyword
11822 'unique_subject' is set to 'no' in the main CA section (default
11823 if 'CA_default') of the configuration file. The value is saved
11824 with the database itself in a separate index attribute file,
11825 named like the index file with '.attr' appended to the name.
11829 * X509 verify fixes. Disable broken certificate workarounds when
11830 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11831 keyUsage extension present. Don't accept CRLs with unhandled critical
11832 extensions: since verify currently doesn't process CRL extensions this
11833 rejects a CRL with *any* critical extensions. Add new verify error codes
11838 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11839 A clarification of RFC2560 will require the use of OCTET STRINGs and
11840 some implementations cannot handle the current raw format. Since OpenSSL
11841 copies and compares OCSP nonces as opaque blobs without any attempt at
11842 parsing them this should not create any compatibility issues.
11846 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11847 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11848 this HMAC (and other) operations are several times slower than OpenSSL
11853 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11855 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11857 * Use the correct content when signing type "other".
11861 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11863 * Fix various bugs revealed by running the NISCC test suite:
11865 Stop out of bounds reads in the ASN1 code when presented with
11866 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11868 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11870 If verify callback ignores invalid public key errors don't try to check
11871 certificate signature with the NULL public key.
11875 * New -ignore_err option in ocsp application to stop the server
11876 exiting on the first error in a request.
11880 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11881 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11886 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11887 extra data after the compression methods not only for TLS 1.0
11888 but also for SSL 3.0 (as required by the specification).
11890 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11892 * Change X509_certificate_type() to mark the key as exported/exportable
11893 when it's 512 *bits* long, not 512 bytes.
11897 * Change AES_cbc_encrypt() so it outputs exact multiple of
11898 blocks during encryption.
11902 * Various fixes to base64 BIO and non blocking I/O. On write
11903 flushes were not handled properly if the BIO retried. On read
11904 data was not being buffered properly and had various logic bugs.
11905 This also affects blocking I/O when the data being decoded is a
11910 * Various S/MIME bugfixes and compatibility changes:
11911 output correct application/pkcs7 MIME type if
11912 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11913 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11914 of files as .eml work). Correctly handle very long lines in MIME
11919 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11921 * Countermeasure against the Klima-Pokorny-Rosa extension of
11922 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11923 a protocol version number mismatch like a decryption error
11924 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11928 * Turn on RSA blinding by default in the default implementation
11929 to avoid a timing attack. Applications that don't want it can call
11930 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11931 They would be ill-advised to do so in most cases.
11933 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11935 * Change RSA blinding code so that it works when the PRNG is not
11936 seeded (in this case, the secret RSA exponent is abused as
11937 an unpredictable seed -- if it is not unpredictable, there
11938 is no point in blinding anyway). Make RSA blinding thread-safe
11939 by remembering the creator's thread ID in rsa->blinding and
11940 having all other threads use local one-time blinding factors
11941 (this requires more computation than sharing rsa->blinding, but
11942 avoids excessive locking; and if an RSA object is not shared
11943 between threads, blinding will still be very fast).
11947 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11948 ENGINE as defaults for all supported algorithms irrespective of
11949 the 'flags' parameter. 'flags' is now honoured, so applications
11950 should make sure they are passing it correctly.
11954 * Target "mingw" now allows native Windows code to be generated in
11955 the Cygwin environment as well as with the MinGW compiler.
11959 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11961 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11962 via timing by performing a MAC computation even if incorrect
11963 block cipher padding has been found. This is a countermeasure
11964 against active attacks where the attacker has to distinguish
11965 between bad padding and a MAC verification error. ([CVE-2003-0078])
11967 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11968 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11969 Martin Vuagnoux (EPFL, Ilion)*
11971 * Make the no-err option work as intended. The intention with no-err
11972 is not to have the whole error stack handling routines removed from
11973 libcrypto, it's only intended to remove all the function name and
11974 reason texts, thereby removing some of the footprint that may not
11975 be interesting if those errors aren't displayed anyway.
11977 NOTE: it's still possible for any application or module to have its
11978 own set of error texts inserted. The routines are there, just not
11979 used by default when no-err is given.
11983 * Add support for FreeBSD on IA64.
11985 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11987 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11988 Kerberos function mit_des_cbc_cksum(). Before this change,
11989 the value returned by DES_cbc_cksum() was like the one from
11990 mit_des_cbc_cksum(), except the bytes were swapped.
11992 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11994 * Allow an application to disable the automatic SSL chain building.
11995 Before this a rather primitive chain build was always performed in
11996 ssl3_output_cert_chain(): an application had no way to send the
11997 correct chain if the automatic operation produced an incorrect result.
11999 Now the chain builder is disabled if either:
12001 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12003 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12005 The reasoning behind this is that an application would not want the
12006 auto chain building to take place if extra chain certificates are
12007 present and it might also want a means of sending no additional
12008 certificates (for example the chain has two certificates and the
12013 * Add the possibility to build without the ENGINE framework.
12015 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12017 * Under Win32 gmtime() can return NULL: check return value in
12018 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12022 * DSA routines: under certain error conditions uninitialized BN objects
12023 could be freed. Solution: make sure initialization is performed early
12024 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12025 Nils Larsch <nla@trustcenter.de> via PR#459)
12029 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12030 checked on reconnect on the client side, therefore session resumption
12031 could still fail with a "ssl session id is different" error. This
12032 behaviour is masked when SSL_OP_ALL is used due to
12033 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12034 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12035 followup to PR #377.
12039 * IA-32 assembler support enhancements: unified ELF targets, support
12040 for SCO/Caldera platforms, fix for Cygwin shared build.
12044 * Add support for FreeBSD on sparc64. As a consequence, support for
12045 FreeBSD on non-x86 processors is separate from x86 processors on
12046 the config script, much like the NetBSD support.
12048 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12050 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12052 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12055 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12056 code (06) was taken as the first octet of the session ID and the last
12057 octet was ignored consequently. As a result SSLv2 client side session
12058 caching could not have worked due to the session ID mismatch between
12060 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12065 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12066 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12071 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12072 seems that in spite of existing for more than a year, many application
12073 author have done nothing to provide the necessary callbacks, which
12074 means that this particular engine will not work properly anywhere.
12075 This is a very unfortunate situation which forces us, in the name
12076 of usability, to give the hw_ncipher.c a static lock, which is part
12078 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12079 appear in 0.9.8 or later. We EXPECT application authors to have
12080 dealt properly with this when 0.9.8 is released (unless we actually
12081 make such changes in the libcrypto locking code that changes will
12082 have to be made anyway).
12086 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12087 octets have been read, EOF or an error occurs. Without this change
12088 some truncated ASN1 structures will not produce an error.
12092 * Disable Heimdal support, since it hasn't been fully implemented.
12093 Still give the possibility to force the use of Heimdal, but with
12094 warnings and a request that patches get sent to openssl-dev.
12098 * Add the VC-CE target, introduce the WINCE sysname, and add
12099 INSTALL.WCE and appropriate conditionals to make it build.
12101 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12103 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12104 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12105 edit numbers of the version.
12107 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12109 * Introduce safe string copy and catenation functions
12110 (BUF_strlcpy() and BUF_strlcat()).
12112 *Ben Laurie (CHATS) and Richard Levitte*
12114 * Avoid using fixed-size buffers for one-line DNs.
12116 *Ben Laurie (CHATS)*
12118 * Add BUF_MEM_grow_clean() to avoid information leakage when
12119 resizing buffers containing secrets, and use where appropriate.
12121 *Ben Laurie (CHATS)*
12123 * Avoid using fixed size buffers for configuration file location.
12125 *Ben Laurie (CHATS)*
12127 * Avoid filename truncation for various CA files.
12129 *Ben Laurie (CHATS)*
12131 * Use sizeof in preference to magic numbers.
12133 *Ben Laurie (CHATS)*
12135 * Avoid filename truncation in cert requests.
12137 *Ben Laurie (CHATS)*
12139 * Add assertions to check for (supposedly impossible) buffer
12142 *Ben Laurie (CHATS)*
12144 * Don't cache truncated DNS entries in the local cache (this could
12145 potentially lead to a spoofing attack).
12147 *Ben Laurie (CHATS)*
12149 * Fix various buffers to be large enough for hex/decimal
12150 representations in a platform independent manner.
12152 *Ben Laurie (CHATS)*
12154 * Add CRYPTO_realloc_clean() to avoid information leakage when
12155 resizing buffers containing secrets, and use where appropriate.
12157 *Ben Laurie (CHATS)*
12159 * Add BIO_indent() to avoid much slightly worrying code to do
12162 *Ben Laurie (CHATS)*
12164 * Convert sprintf()/BIO_puts() to BIO_printf().
12166 *Ben Laurie (CHATS)*
12168 * buffer_gets() could terminate with the buffer only half
12171 *Ben Laurie (CHATS)*
12173 * Add assertions to prevent user-supplied crypto functions from
12174 overflowing internal buffers by having large block sizes, etc.
12176 *Ben Laurie (CHATS)*
12178 * New OPENSSL_assert() macro (similar to assert(), but enabled
12181 *Ben Laurie (CHATS)*
12183 * Eliminate unused copy of key in RC4.
12185 *Ben Laurie (CHATS)*
12187 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12189 *Ben Laurie (CHATS)*
12191 * Fix off-by-one error in EGD path.
12193 *Ben Laurie (CHATS)*
12195 * If RANDFILE path is too long, ignore instead of truncating.
12197 *Ben Laurie (CHATS)*
12199 * Eliminate unused and incorrectly sized X.509 structure
12202 *Ben Laurie (CHATS)*
12204 * Eliminate unused and dangerous function knumber().
12206 *Ben Laurie (CHATS)*
12208 * Eliminate unused and dangerous structure, KSSL_ERR.
12210 *Ben Laurie (CHATS)*
12212 * Protect against overlong session ID context length in an encoded
12213 session object. Since these are local, this does not appear to be
12216 *Ben Laurie (CHATS)*
12218 * Change from security patch (see 0.9.6e below) that did not affect
12219 the 0.9.6 release series:
12221 Remote buffer overflow in SSL3 protocol - an attacker could
12222 supply an oversized master key in Kerberos-enabled versions.
12225 *Ben Laurie (CHATS)*
12227 * Change the SSL kerb5 codes to match RFC 2712.
12231 * Make -nameopt work fully for req and add -reqopt switch.
12233 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12235 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12237 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12239 * Make sure tests can be performed even if the corresponding algorithms
12240 have been removed entirely. This was also the last step to make
12241 OpenSSL compilable with DJGPP under all reasonable conditions.
12243 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12245 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12246 to allow version independent disabling of normally unselected ciphers,
12247 which may be activated as a side-effect of selecting a single cipher.
12249 (E.g., cipher list string "RSA" enables ciphersuites that are left
12250 out of "ALL" because they do not provide symmetric encryption.
12251 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12253 *Lutz Jaenicke, Bodo Moeller*
12255 * Add appropriate support for separate platform-dependent build
12256 directories. The recommended way to make a platform-dependent
12257 build directory is the following (tested on Linux), maybe with
12260 # Place yourself outside of the OpenSSL source tree. In
12261 # this example, the environment variable OPENSSL_SOURCE
12262 # is assumed to contain the absolute OpenSSL source directory.
12263 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12264 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12265 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12266 mkdir -p `dirname $F`
12267 ln -s $OPENSSL_SOURCE/$F $F
12270 To be absolutely sure not to disturb the source tree, a "make clean"
12271 is a good thing. If it isn't successful, don't worry about it,
12272 it probably means the source directory is very clean.
12276 * Make sure any ENGINE control commands make local copies of string
12277 pointers passed to them whenever necessary. Otherwise it is possible
12278 the caller may have overwritten (or deallocated) the original string
12279 data when a later ENGINE operation tries to use the stored values.
12281 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12283 * Improve diagnostics in file reading and command-line digests.
12285 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12287 * Add AES modes CFB and OFB to the object database. Correct an
12288 error in AES-CFB decryption.
12292 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12293 allows existing EVP_CIPHER_CTX structures to be reused after
12294 calling `EVP_*Final()`. This behaviour is used by encryption
12295 BIOs and some applications. This has the side effect that
12296 applications must explicitly clean up cipher contexts with
12297 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12301 * Check the values of dna and dnb in bn_mul_recursive before calling
12302 bn_mul_comba (a non zero value means the a or b arrays do not contain
12303 n2 elements) and fallback to bn_mul_normal if either is not zero.
12307 * Fix escaping of non-ASCII characters when using the -subj option
12308 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12312 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12313 form for "surname", serialNumber has no short form.
12314 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12315 therefore remove "mail" short name for "internet 7".
12316 The OID for unique identifiers in X509 certificates is
12317 x500UniqueIdentifier, not uniqueIdentifier.
12318 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12322 * Add an "init" command to the ENGINE config module and auto initialize
12323 ENGINEs. Without any "init" command the ENGINE will be initialized
12324 after all ctrl commands have been executed on it. If init=1 the
12325 ENGINE is initialized at that point (ctrls before that point are run
12326 on the uninitialized ENGINE and after on the initialized one). If
12327 init=0 then the ENGINE will not be initialized at all.
12331 * Fix the 'app_verify_callback' interface so that the user-defined
12332 argument is actually passed to the callback: In the
12333 SSL_CTX_set_cert_verify_callback() prototype, the callback
12334 declaration has been changed from
12337 int (*cb)(X509_STORE_CTX *,void *);
12338 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12339 i=s->ctx->app_verify_callback(&ctx)
12340 has been changed into
12341 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12343 To update applications using SSL_CTX_set_cert_verify_callback(),
12344 a dummy argument can be added to their callback functions.
12346 *D. K. Smetters <smetters@parc.xerox.com>*
12348 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12350 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12352 * Add and OPENSSL_LOAD_CONF define which will cause
12353 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12354 This allows older applications to transparently support certain
12355 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12356 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12357 load the config file and OPENSSL_add_all_algorithms_conf() which will
12358 always load it have also been added.
12362 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12363 Adjust NIDs and EVP layer.
12365 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12367 * Config modules support in openssl utility.
12369 Most commands now load modules from the config file,
12370 though in a few (such as version) this isn't done
12371 because it couldn't be used for anything.
12373 In the case of ca and req the config file used is
12374 the same as the utility itself: that is the -config
12375 command line option can be used to specify an
12380 * Move default behaviour from OPENSSL_config(). If appname is NULL
12381 use "openssl_conf" if filename is NULL use default openssl config file.
12385 * Add an argument to OPENSSL_config() to allow the use of an alternative
12386 config section name. Add a new flag to tolerate a missing config file
12387 and move code to CONF_modules_load_file().
12391 * Support for crypto accelerator cards from Accelerated Encryption
12392 Processing, www.aep.ie. (Use engine 'aep')
12393 The support was copied from 0.9.6c [engine] and adapted/corrected
12394 to work with the new engine framework.
12396 *AEP Inc. and Richard Levitte*
12398 * Support for SureWare crypto accelerator cards from Baltimore
12399 Technologies. (Use engine 'sureware')
12400 The support was copied from 0.9.6c [engine] and adapted
12401 to work with the new engine framework.
12405 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12406 make the newer ENGINE framework commands for the CHIL engine work.
12408 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12410 * Make it possible to produce shared libraries on ReliantUNIX.
12412 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12414 * Add the configuration target debug-linux-ppro.
12415 Make 'openssl rsa' use the general key loading routines
12416 implemented in `apps.c`, and make those routines able to
12417 handle the key format FORMAT_NETSCAPE and the variant
12420 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12422 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12424 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12426 * Add -keyform to rsautl, and document -engine.
12428 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12430 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12431 BIO_R_NO_SUCH_FILE error code rather than the generic
12432 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12436 * Add new functions
12437 ERR_peek_last_error
12438 ERR_peek_last_error_line
12439 ERR_peek_last_error_line_data.
12440 These are similar to
12442 ERR_peek_error_line
12443 ERR_peek_error_line_data,
12444 but report on the latest error recorded rather than the first one
12445 still in the error queue.
12447 *Ben Laurie, Bodo Moeller*
12449 * default_algorithms option in ENGINE config module. This allows things
12451 default_algorithms = ALL
12452 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12456 * Preliminary ENGINE config module.
12460 * New experimental application configuration code.
12464 * Change the AES code to follow the same name structure as all other
12465 symmetric ciphers, and behave the same way. Move everything to
12466 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12468 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12470 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12472 *Ben Laurie and Theo de Raadt*
12474 * Add option to output public keys in req command.
12476 *Massimiliano Pala madwolf@openca.org*
12478 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12479 (up to about 10% better than before for P-192 and P-224).
12483 * New functions/macros
12485 SSL_CTX_set_msg_callback(ctx, cb)
12486 SSL_CTX_set_msg_callback_arg(ctx, arg)
12487 SSL_set_msg_callback(ssl, cb)
12488 SSL_set_msg_callback_arg(ssl, arg)
12490 to request calling a callback function
12492 void cb(int write_p, int version, int content_type,
12493 const void *buf, size_t len, SSL *ssl, void *arg)
12495 whenever a protocol message has been completely received
12496 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12497 protocol version according to which the SSL library interprets
12498 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12499 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12500 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12501 specification (change_cipher_spec(20), alert(21), handshake(22)).
12502 'buf' and 'len' point to the actual message, 'ssl' to the
12503 SSL object, and 'arg' is the application-defined value set by
12504 SSL[_CTX]_set_msg_callback_arg().
12506 'openssl s_client' and 'openssl s_server' have new '-msg' options
12507 to enable a callback that displays all protocol messages.
12511 * Change the shared library support so shared libraries are built as
12512 soon as the corresponding static library is finished, and thereby get
12513 openssl and the test programs linked against the shared library.
12514 This still only happens when the keyword "shard" has been given to
12515 the configuration scripts.
12517 NOTE: shared library support is still an experimental thing, and
12518 backward binary compatibility is still not guaranteed.
12520 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12522 * Add support for Subject Information Access extension.
12524 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12526 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12527 additional bytes when new memory had to be allocated, not just
12528 when reusing an existing buffer.
12532 * New command line and configuration option 'utf8' for the req command.
12533 This allows field values to be specified as UTF8 strings.
12537 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12538 runs for the former and machine-readable output for the latter.
12542 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12543 of the e-mail address in the DN (i.e., it will go into a certificate
12544 extension only). The new configuration file option 'email_in_dn = no'
12545 has the same effect.
12547 *Massimiliano Pala madwolf@openca.org*
12549 * Change all functions with names starting with `des_` to be starting
12550 with `DES_` instead. Add wrappers that are compatible with libdes,
12551 but are named `_ossl_old_des_*`. Finally, add macros that map the
12552 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12553 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12554 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12557 Since we provide two compatibility mappings, the user needs to
12558 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12559 compatibility is desired. The default (i.e., when that macro
12560 isn't defined) is OpenSSL 0.9.6c compatibility.
12562 There are also macros that enable and disable the support of old
12563 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12564 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12565 are defined, the default will apply: to support the old des routines.
12567 In either case, one must include openssl/des.h to get the correct
12568 definitions. Do not try to just include openssl/des_old.h, that
12571 NOTE: This is a major break of an old API into a new one. Software
12572 authors are encouraged to switch to the `DES_` style functions. Some
12573 time in the future, des_old.h and the libdes compatibility functions
12574 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12575 default), and then completely removed.
12579 * Test for certificates which contain unsupported critical extensions.
12580 If such a certificate is found during a verify operation it is
12581 rejected by default: this behaviour can be overridden by either
12582 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12583 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12584 X509_supported_extension() has also been added which returns 1 if a
12585 particular extension is supported.
12589 * Modify the behaviour of EVP cipher functions in similar way to digests
12590 to retain compatibility with existing code.
12594 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12595 compatibility with existing code. In particular the 'ctx' parameter does
12596 not have to be to be initialized before the call to EVP_DigestInit() and
12597 it is tidied up after a call to EVP_DigestFinal(). New function
12598 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12599 EVP_MD_CTX_copy() changed to not require the destination to be
12600 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12601 requires the destination to be valid.
12603 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12604 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12608 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12609 so that complete 'Handshake' protocol structures are kept in memory
12610 instead of overwriting 'msg_type' and 'length' with 'body' data.
12614 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12616 *Massimo Santin via Richard Levitte*
12618 * Major restructuring to the underlying ENGINE code. This includes
12619 reduction of linker bloat, separation of pure "ENGINE" manipulation
12620 (initialisation, etc) from functionality dealing with implementations
12621 of specific crypto interfaces. This change also introduces integrated
12622 support for symmetric ciphers and digest implementations - so ENGINEs
12623 can now accelerate these by providing EVP_CIPHER and EVP_MD
12624 implementations of their own. This is detailed in
12625 [crypto/engine/README.md](crypto/engine/README.md)
12626 as it couldn't be adequately described here. However, there are a few
12627 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12628 were changed in the original introduction of ENGINE code have now
12629 reverted back - the hooking from this code to ENGINE is now a good
12630 deal more passive and at run-time, operations deal directly with
12631 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12632 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12633 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12634 they were not being used by the framework as there is no concept of a
12635 BIGNUM_METHOD and they could not be generalised to the new
12636 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12637 ENGINE_cpy() has been removed as it cannot be consistently defined in
12642 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12646 * Change mkdef.pl to sort symbols that get the same entry number,
12647 and make sure the automatically generated functions `ERR_load_*`
12648 become part of libeay.num as well.
12652 * New function SSL_renegotiate_pending(). This returns true once
12653 renegotiation has been requested (either SSL_renegotiate() call
12654 or HelloRequest/ClientHello received from the peer) and becomes
12655 false once a handshake has been completed.
12656 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12657 sends a HelloRequest, but does not ensure that a handshake takes
12658 place. SSL_renegotiate_pending() is useful for checking if the
12659 client has followed the request.)
12663 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12664 By default, clients may request session resumption even during
12665 renegotiation (if session ID contexts permit); with this option,
12666 session resumption is possible only in the first handshake.
12668 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12669 more bits available for options that should not be part of
12670 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12674 * Add some demos for certificate and certificate request creation.
12678 * Make maximum certificate chain size accepted from the peer application
12679 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12680 "Douglas E. Engert" <deengert@anl.gov>.
12684 * Add support for shared libraries for Unixware-7
12685 (Boyd Lynn Gerber <gerberb@zenez.com>).
12689 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12690 be done prior to destruction. Use this to unload error strings from
12691 ENGINEs that load their own error strings. NB: This adds two new API
12692 functions to "get" and "set" this destroy handler in an ENGINE.
12696 * Alter all existing ENGINE implementations (except "openssl" and
12697 "openbsd") to dynamically instantiate their own error strings. This
12698 makes them more flexible to be built both as statically-linked ENGINEs
12699 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12700 Also, add stub code to each that makes building them as self-contained
12701 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12705 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12706 implementations into applications that are completely implemented in
12707 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12708 commands that can be used to configure what shared-library to load and
12709 to control aspects of the way it is handled. Also, made an update to
12710 the [README-Engine.md](README-Engine.md) file
12711 that brings its information up-to-date and
12712 provides some information and instructions on the "dynamic" ENGINE
12713 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12717 * Make it possible to unload ranges of ERR strings with a new
12718 "ERR_unload_strings" function.
12722 * Add a copy() function to EVP_MD.
12726 * Make EVP_MD routines take a context pointer instead of just the
12727 md_data void pointer.
12731 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12732 that the digest can only process a single chunk of data
12733 (typically because it is provided by a piece of
12734 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12735 is only going to provide a single chunk of data, and hence the
12736 framework needn't accumulate the data for oneshot drivers.
12740 * As with "ERR", make it possible to replace the underlying "ex_data"
12741 functions. This change also alters the storage and management of global
12742 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12743 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12744 index counters. The API functions that use this state have been changed
12745 to take a "class_index" rather than pointers to the class's local STACK
12746 and counter, and there is now an API function to dynamically create new
12747 classes. This centralisation allows us to (a) plug a lot of the
12748 thread-safety problems that existed, and (b) makes it possible to clean
12749 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12750 such data would previously have always leaked in application code and
12751 workarounds were in place to make the memory debugging turn a blind eye
12752 to it. Application code that doesn't use this new function will still
12753 leak as before, but their memory debugging output will announce it now
12754 rather than letting it slide.
12756 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12757 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12758 has a return value to indicate success or failure.
12762 * Make it possible to replace the underlying "ERR" functions such that the
12763 global state (2 LHASH tables and 2 locks) is only used by the "default"
12764 implementation. This change also adds two functions to "get" and "set"
12765 the implementation prior to it being automatically set the first time
12766 any other ERR function takes place. Ie. an application can call "get",
12767 pass the return value to a module it has just loaded, and that module
12768 can call its own "set" function using that value. This means the
12769 module's "ERR" operations will use (and modify) the error state in the
12770 application and not in its own statically linked copy of OpenSSL code.
12774 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12775 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12776 the operation, and provides a more encapsulated way for external code
12777 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12778 to use these functions rather than manually incrementing the counts.
12780 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12784 * Add EVP test program.
12788 * Add symmetric cipher support to ENGINE. Expect the API to change!
12792 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12793 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12794 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12795 These allow a CRL to be built without having to access X509_CRL fields
12796 directly. Modify 'ca' application to use new functions.
12800 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12801 bug workarounds. Rollback attack detection is a security feature.
12802 The problem will only arise on OpenSSL servers when TLSv1 is not
12803 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12804 Software authors not wanting to support TLSv1 will have special reasons
12805 for their choice and can explicitly enable this option.
12807 *Bodo Moeller, Lutz Jaenicke*
12809 * Rationalise EVP so it can be extended: don't include a union of
12810 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12811 (similar to those existing for EVP_CIPHER_CTX).
12816 EVP_MD_CTX_init(&md); /* new function call */
12817 EVP_DigestInit(&md, EVP_sha1());
12818 EVP_DigestUpdate(&md, in, len);
12819 EVP_DigestFinal(&md, out, NULL);
12820 EVP_MD_CTX_cleanup(&md); /* new function call */
12824 * Make DES key schedule conform to the usual scheme, as well as
12825 correcting its structure. This means that calls to DES functions
12826 now have to pass a pointer to a des_key_schedule instead of a
12827 plain des_key_schedule (which was actually always a pointer
12830 des_key_schedule ks;
12832 des_set_key_checked(..., &ks);
12833 des_ncbc_encrypt(..., &ks, ...);
12835 (Note that a later change renames 'des_...' into 'DES_...'.)
12839 * Initial reduction of linker bloat: the use of some functions, such as
12840 PEM causes large amounts of unused functions to be linked in due to
12841 poor organisation. For example pem_all.c contains every PEM function
12842 which has a knock on effect of linking in large amounts of (unused)
12843 ASN1 code. Grouping together similar functions and splitting unrelated
12844 functions prevents this.
12848 * Cleanup of EVP macros.
12852 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12853 correct `_ecb suffix`.
12857 * Add initial OCSP responder support to ocsp application. The
12858 revocation information is handled using the text based index
12859 use by the ca application. The responder can either handle
12860 requests generated internally, supplied in files (for example
12861 via a CGI script) or using an internal minimal server.
12865 * Add configuration choices to get zlib compression for TLS.
12869 * Changes to Kerberos SSL for RFC 2712 compliance:
12870 1. Implemented real KerberosWrapper, instead of just using
12871 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12872 2. Implemented optional authenticator field of KerberosWrapper.
12874 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12875 and authenticator structs; see crypto/krb5/.
12877 Generalized Kerberos calls to support multiple Kerberos libraries.
12878 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12879 via Richard Levitte*
12881 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12882 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12883 values for each of the key sizes rather than having just
12884 parameters (and 'speed' generating keys each time).
12888 * Speed up EVP routines.
12891 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12892 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12893 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12894 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12896 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12897 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12898 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12901 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12903 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12907 * Added the OS2-EMX target.
12909 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12911 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12912 New functions to support `NCONF` routines in extension code.
12913 New function `CONF_set_nconf()`
12914 to allow functions which take an `NCONF` to also handle the old `LHASH`
12915 structure: this means that the old `CONF` compatible routines can be
12916 retained (in particular w.rt. extensions) without having to duplicate the
12917 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12921 * Enhance the general user interface with mechanisms for inner control
12922 and with possibilities to have yes/no kind of prompts.
12926 * Change all calls to low-level digest routines in the library and
12927 applications to use EVP. Add missing calls to HMAC_cleanup() and
12928 don't assume HMAC_CTX can be copied using memcpy().
12930 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12932 * Add the possibility to control engines through control names but with
12933 arbitrary arguments instead of just a string.
12934 Change the key loaders to take a UI_METHOD instead of a callback
12935 function pointer. NOTE: this breaks binary compatibility with earlier
12936 versions of OpenSSL [engine].
12937 Adapt the nCipher code for these new conditions and add a card insertion
12942 * Enhance the general user interface with mechanisms to better support
12943 dialog box interfaces, application-defined prompts, the possibility
12944 to use defaults (for example default passwords from somewhere else)
12945 and interrupts/cancellations.
12949 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12950 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12954 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12955 tidy up some unnecessarily weird code in 'sk_new()').
12957 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12959 * Change the key loading routines for ENGINEs to use the same kind
12960 callback (pem_password_cb) as all other routines that need this
12965 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12966 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12967 than this minimum value is recommended.
12971 * New random seeder for OpenVMS, using the system process statistics
12972 that are easily reachable.
12976 * Windows apparently can't transparently handle global
12977 variables defined in DLLs. Initialisations such as:
12979 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12981 won't compile. This is used by the any applications that need to
12982 declare their own ASN1 modules. This was fixed by adding the option
12983 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12984 needed for static libraries under Win32.
12988 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12989 setting of purpose and trust fields. New X509_STORE trust and
12990 purpose functions and tidy up setting in other SSL functions.
12994 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12995 structure. These are inherited by X509_STORE_CTX when it is
12996 initialised. This allows various defaults to be set in the
12997 X509_STORE structure (such as flags for CRL checking and custom
12998 purpose or trust settings) for functions which only use X509_STORE_CTX
12999 internally such as S/MIME.
13001 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13002 trust settings if they are not set in X509_STORE. This allows X509_STORE
13003 purposes and trust (in S/MIME for example) to override any set by default.
13005 Add command line options for CRL checking to smime, s_client and s_server
13010 * Initial CRL based revocation checking. If the CRL checking flag(s)
13011 are set then the CRL is looked up in the X509_STORE structure and
13012 its validity and signature checked, then if the certificate is found
13013 in the CRL the verify fails with a revoked error.
13015 Various new CRL related callbacks added to X509_STORE_CTX structure.
13017 Command line options added to 'verify' application to support this.
13019 This needs some additional work, such as being able to handle multiple
13020 CRLs with different times, extension based lookup (rather than just
13021 by subject name) and ultimately more complete V2 CRL extension
13026 * Add a general user interface API (crypto/ui/). This is designed
13027 to replace things like des_read_password and friends (backward
13028 compatibility functions using this new API are provided).
13029 The purpose is to remove prompting functions from the DES code
13030 section as well as provide for prompting through dialog boxes in
13031 a window system and the like.
13035 * Add "ex_data" support to ENGINE so implementations can add state at a
13036 per-structure level rather than having to store it globally.
13040 * Make it possible for ENGINE structures to be copied when retrieved by
13041 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13042 This causes the "original" ENGINE structure to act like a template,
13043 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13044 operational state can be localised to each ENGINE structure, despite the
13045 fact they all share the same "methods". New ENGINE structures returned in
13046 this case have no functional references and the return value is the single
13047 structural reference. This matches the single structural reference returned
13048 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13053 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13054 needs to match any other type at all we need to manually clear the
13059 * Changes to the "openssl engine" utility to include;
13060 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13061 about an ENGINE's available control commands.
13062 - executing control commands from command line arguments using the
13063 '-pre' and '-post' switches. '-post' is only used if '-t' is
13064 specified and the ENGINE is successfully initialised. The syntax for
13065 the individual commands are colon-separated, for example;
13066 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13070 * New dynamic control command support for ENGINEs. ENGINEs can now
13071 declare their own commands (numbers), names (strings), descriptions,
13072 and input types for run-time discovery by calling applications. A
13073 subset of these commands are implicitly classed as "executable"
13074 depending on their input type, and only these can be invoked through
13075 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13076 can be based on user input, config files, etc). The distinction is
13077 that "executable" commands cannot return anything other than a boolean
13078 result and can only support numeric or string input, whereas some
13079 discoverable commands may only be for direct use through
13080 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13081 pointers, or other custom uses. The "executable" commands are to
13082 support parameterisations of ENGINE behaviour that can be
13083 unambiguously defined by ENGINEs and used consistently across any
13084 OpenSSL-based application. Commands have been added to all the
13085 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13086 control over shared-library paths without source code alterations.
13090 * Changed all ENGINE implementations to dynamically allocate their
13091 ENGINEs rather than declaring them statically. Apart from this being
13092 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13093 this also allows the implementations to compile without using the
13094 internal engine_int.h header.
13098 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13099 'const' value. Any code that should be able to modify a RAND_METHOD
13100 should already have non-const pointers to it (ie. they should only
13101 modify their own ones).
13105 * Made a variety of little tweaks to the ENGINE code.
13106 - "atalla" and "ubsec" string definitions were moved from header files
13107 to C code. "nuron" string definitions were placed in variables
13108 rather than hard-coded - allowing parameterisation of these values
13109 later on via ctrl() commands.
13110 - Removed unused "#if 0"'d code.
13111 - Fixed engine list iteration code so it uses ENGINE_free() to release
13112 structural references.
13113 - Constified the RAND_METHOD element of ENGINE structures.
13114 - Constified various get/set functions as appropriate and added
13115 missing functions (including a catch-all ENGINE_cpy that duplicates
13116 all ENGINE values onto a new ENGINE except reference counts/state).
13117 - Removed NULL parameter checks in get/set functions. Setting a method
13118 or function to NULL is a way of cancelling out a previously set
13119 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13120 and doesn't justify the extra error symbols and code.
13121 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13122 flags from engine_int.h to engine.h.
13123 - Changed prototypes for ENGINE handler functions (init(), finish(),
13124 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13128 * Implement binary inversion algorithm for BN_mod_inverse in addition
13129 to the algorithm using long division. The binary algorithm can be
13130 used only if the modulus is odd. On 32-bit systems, it is faster
13131 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13132 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13133 up to 450 bits. In 64-bit environments, the binary algorithm
13134 appears to be advantageous for much longer moduli; here we use it
13135 for moduli up to 2048 bits.
13139 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13140 could not support the combine flag in choice fields.
13144 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13145 extensions from a certificate request to the certificate.
13149 * Allow multiple 'certopt' and 'nameopt' options to be separated
13150 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13151 file: this allows the display of the certificate about to be
13152 signed to be customised, to allow certain fields to be included
13153 or excluded and extension details. The old system didn't display
13154 multicharacter strings properly, omitted fields not in the policy
13155 and couldn't display additional details such as extensions.
13159 * Function EC_POINTs_mul for multiple scalar multiplication
13160 of an arbitrary number of elliptic curve points
13161 \sum scalars[i]*points[i],
13162 optionally including the generator defined for the EC_GROUP:
13163 scalar*generator + \sum scalars[i]*points[i].
13165 EC_POINT_mul is a simple wrapper function for the typical case
13166 that the point list has just one item (besides the optional
13171 * First EC_METHODs for curves over GF(p):
13173 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13174 operations and provides various method functions that can also
13175 operate with faster implementations of modular arithmetic.
13177 EC_GFp_mont_method() reuses most functions that are part of
13178 EC_GFp_simple_method, but uses Montgomery arithmetic.
13180 *Bodo Moeller; point addition and point doubling
13181 implementation directly derived from source code provided by
13182 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13184 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13185 crypto/ec/ec_lib.c):
13187 Curves are EC_GROUP objects (with an optional group generator)
13188 based on EC_METHODs that are built into the library.
13190 Points are EC_POINT objects based on EC_GROUP objects.
13192 Most of the framework would be able to handle curves over arbitrary
13193 finite fields, but as there are no obvious types for fields other
13194 than GF(p), some functions are limited to that for now.
13198 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13199 that the file contains a complete HTTP response.
13203 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13204 change the def and num file printf format specifier from "%-40sXXX"
13205 to "%-39s XXX". The latter will always guarantee a space after the
13206 field while the former will cause them to run together if the field
13207 is 40 of more characters long.
13211 * Constify the cipher and digest 'method' functions and structures
13212 and modify related functions to take constant EVP_MD and EVP_CIPHER
13217 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13218 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13222 * Modify `EVP_Digest*()` routines so they now return values. Although the
13223 internal software routines can never fail additional hardware versions
13228 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13230 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13231 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13234 ERR_R_NESTED_ASN1_ERROR
13236 ERR_R_MISSING_ASN1_EOS
13237 were 4 .. 9, conflicting with
13238 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13240 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13241 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13243 Add new error code 'ERR_R_INTERNAL_ERROR'.
13247 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13252 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13253 sets the subject name for a new request or supersedes the
13254 subject name in a given request. Formats that can be parsed are
13255 'CN=Some Name, OU=myOU, C=IT'
13257 'CN=Some Name/OU=myOU/C=IT'.
13259 Add options '-batch' and '-verbose' to 'openssl req'.
13261 *Massimiliano Pala <madwolf@hackmasters.net>*
13263 * Introduce the possibility to access global variables through
13264 functions on platform were that's the best way to handle exporting
13265 global variables in shared libraries. To enable this functionality,
13266 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13267 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13268 is normally done by Configure or something similar).
13270 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13271 in the source file (foo.c) like this:
13273 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13274 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13276 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13277 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13279 OPENSSL_DECLARE_GLOBAL(int,foo);
13280 #define foo OPENSSL_GLOBAL_REF(foo)
13281 OPENSSL_DECLARE_GLOBAL(double,bar);
13282 #define bar OPENSSL_GLOBAL_REF(bar)
13284 The #defines are very important, and therefore so is including the
13285 header file everywhere where the defined globals are used.
13287 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13288 of ASN.1 items, but that structure is a bit different.
13290 The largest change is in util/mkdef.pl which has been enhanced with
13291 better and easier to understand logic to choose which symbols should
13292 go into the Windows .def files as well as a number of fixes and code
13293 cleanup (among others, algorithm keywords are now sorted
13294 lexicographically to avoid constant rewrites).
13298 * In BN_div() keep a copy of the sign of 'num' before writing the
13299 result to 'rm' because if rm==num the value will be overwritten
13300 and produce the wrong result if 'num' is negative: this caused
13301 problems with BN_mod() and BN_nnmod().
13305 * Function OCSP_request_verify(). This checks the signature on an
13306 OCSP request and verifies the signer certificate. The signer
13307 certificate is just checked for a generic purpose and OCSP request
13312 * Add OCSP_check_validity() function to check the validity of OCSP
13313 responses. OCSP responses are prepared in real time and may only
13314 be a few seconds old. Simply checking that the current time lies
13315 between thisUpdate and nextUpdate max reject otherwise valid responses
13316 caused by either OCSP responder or client clock inaccuracy. Instead
13317 we allow thisUpdate and nextUpdate to fall within a certain period of
13318 the current time. The age of the response can also optionally be
13319 checked. Two new options -validity_period and -status_age added to
13324 * If signature or public key algorithm is unrecognized print out its
13325 OID rather that just UNKNOWN.
13329 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13330 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13331 ID to be generated from the issuer certificate alone which can then be
13332 passed to OCSP_id_issuer_cmp().
13336 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13337 ASN1 modules to export functions returning ASN1_ITEM pointers
13338 instead of the ASN1_ITEM structures themselves. This adds several
13339 new macros which allow the underlying ASN1 function/structure to
13340 be accessed transparently. As a result code should not use ASN1_ITEM
13341 references directly (such as &X509_it) but instead use the relevant
13342 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13343 use of the new ASN1 code on platforms where exporting structures
13344 is problematical (for example in shared libraries) but exporting
13345 functions returning pointers to structures is not.
13349 * Add support for overriding the generation of SSL/TLS session IDs.
13350 These callbacks can be registered either in an SSL_CTX or per SSL.
13351 The purpose of this is to allow applications to control, if they wish,
13352 the arbitrary values chosen for use as session IDs, particularly as it
13353 can be useful for session caching in multiple-server environments. A
13354 command-line switch for testing this (and any client code that wishes
13355 to use such a feature) has been added to "s_server".
13357 *Geoff Thorpe, Lutz Jaenicke*
13359 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13360 of the form `#if defined(...) || defined(...) || ...` and
13361 `#if !defined(...) && !defined(...) && ...`. This also avoids
13362 the growing number of special cases it was previously handling.
13366 * Make all configuration macros available for application by making
13367 sure they are available in opensslconf.h, by giving them names starting
13368 with `OPENSSL_` to avoid conflicts with other packages and by making
13369 sure e_os2.h will cover all platform-specific cases together with
13371 Additionally, it is now possible to define configuration/platform-
13372 specific names (called "system identities"). In the C code, these
13373 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13374 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13375 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13380 * New option -set_serial to 'req' and 'x509' this allows the serial
13381 number to use to be specified on the command line. Previously self
13382 signed certificates were hard coded with serial number 0 and the
13383 CA options of 'x509' had to use a serial number in a file which was
13388 * New options to 'ca' utility to support V2 CRL entry extensions.
13389 Currently CRL reason, invalidity date and hold instruction are
13390 supported. Add new CRL extensions to V3 code and some new objects.
13394 * New function EVP_CIPHER_CTX_set_padding() this is used to
13395 disable standard block padding (aka PKCS#5 padding) in the EVP
13396 API, which was previously mandatory. This means that the data is
13397 not padded in any way and so the total length much be a multiple
13398 of the block size, otherwise an error occurs.
13402 * Initial (incomplete) OCSP SSL support.
13406 * New function OCSP_parse_url(). This splits up a URL into its host,
13407 port and path components: primarily to parse OCSP URLs. New -url
13408 option to ocsp utility.
13412 * New nonce behavior. The return value of OCSP_check_nonce() now
13413 reflects the various checks performed. Applications can decide
13414 whether to tolerate certain situations such as an absent nonce
13415 in a response when one was present in a request: the ocsp application
13416 just prints out a warning. New function OCSP_add1_basic_nonce()
13417 this is to allow responders to include a nonce in a response even if
13418 the request is nonce-less.
13422 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13423 skipped when using openssl x509 multiple times on a single input file,
13424 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13428 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13429 set string type: to handle setting ASN1_TIME structures. Fix ca
13430 utility to correctly initialize revocation date of CRLs.
13434 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13435 the clients preferred ciphersuites and rather use its own preferences.
13436 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13437 Internet Explorer by ensuring unchanged hash method during stepup.
13438 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13442 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13443 to aes and add a new 'exist' option to print out symbols that don't
13448 * Additional options to ocsp utility to allow flags to be set and
13449 additional certificates supplied.
13453 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13454 OCSP client a number of certificate to only verify the response
13459 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13460 handle the new API. Currently only ECB, CBC modes supported. Add new
13463 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13464 Encryption Standard (AES) Ciphersuites for Transport Layer
13465 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13466 not enabled by default and were not part of the "ALL" ciphersuite
13467 alias because they were not yet official; they could be
13468 explicitly requested by specifying the "AESdraft" ciphersuite
13469 group alias. In the final release of OpenSSL 0.9.7, the group
13470 alias is called "AES" and is part of "ALL".)
13472 *Ben Laurie, Steve Henson, Bodo Moeller*
13474 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13475 request to response.
13479 * Functions for OCSP responders. OCSP_request_onereq_count(),
13480 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13481 extract information from a certificate request. OCSP_response_create()
13482 creates a response and optionally adds a basic response structure.
13483 OCSP_basic_add1_status() adds a complete single response to a basic
13484 response and returns the OCSP_SINGLERESP structure just added (to allow
13485 extensions to be included for example). OCSP_basic_add1_cert() adds a
13486 certificate to a basic response and OCSP_basic_sign() signs a basic
13487 response with various flags. New helper functions ASN1_TIME_check()
13488 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13489 (converts ASN1_TIME to GeneralizedTime).
13493 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13494 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13495 structure from a certificate. X509_pubkey_digest() digests the public_key
13496 contents: this is used in various key identifiers.
13500 * Make sk_sort() tolerate a NULL argument.
13502 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13504 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13505 passed by the function are trusted implicitly. If any of them signed the
13506 response then it is assumed to be valid and is not verified.
13510 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13511 to data. This was previously part of the PKCS7 ASN1 code. This
13512 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13513 *Steve Henson, reported by Kenneth R. Robinette
13514 <support@securenetterm.com>*
13516 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13517 routines: without these tracing memory leaks is very painful.
13518 Fix leaks in PKCS12 and PKCS7 routines.
13522 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13523 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13524 effectively meant GeneralizedTime would never be used. Now it
13525 is initialised to -1 but X509_time_adj() now has to check the value
13526 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13527 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13528 *Steve Henson, reported by Kenneth R. Robinette
13529 <support@securenetterm.com>*
13531 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13532 result in a zero length in the ASN1_INTEGER structure which was
13533 not consistent with the structure when d2i_ASN1_INTEGER() was used
13534 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13535 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13536 where it did not print out a minus for negative ASN1_INTEGER.
13540 * Add summary printout to ocsp utility. The various functions which
13541 convert status values to strings have been renamed to:
13542 OCSP_response_status_str(), OCSP_cert_status_str() and
13543 OCSP_crl_reason_str() and are no longer static. New options
13544 to verify nonce values and to disable verification. OCSP response
13545 printout format cleaned up.
13549 * Add additional OCSP certificate checks. These are those specified
13550 in RFC2560. This consists of two separate checks: the CA of the
13551 certificate being checked must either be the OCSP signer certificate
13552 or the issuer of the OCSP signer certificate. In the latter case the
13553 OCSP signer certificate must contain the OCSP signing extended key
13554 usage. This check is performed by attempting to match the OCSP
13555 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13556 in the OCSP_CERTID structures of the response.
13560 * Initial OCSP certificate verification added to OCSP_basic_verify()
13561 and related routines. This uses the standard OpenSSL certificate
13562 verify routines to perform initial checks (just CA validity) and
13563 to obtain the certificate chain. Then additional checks will be
13564 performed on the chain. Currently the root CA is checked to see
13565 if it is explicitly trusted for OCSP signing. This is used to set
13566 a root CA as a global signing root: that is any certificate that
13567 chains to that CA is an acceptable OCSP signing certificate.
13571 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13572 extensions from a separate configuration file.
13573 As when reading extensions from the main configuration file,
13574 the '-extensions ...' option may be used for specifying the
13577 *Massimiliano Pala <madwolf@comune.modena.it>*
13579 * New OCSP utility. Allows OCSP requests to be generated or
13580 read. The request can be sent to a responder and the output
13581 parsed, outputted or printed in text form. Not complete yet:
13582 still needs to check the OCSP response validity.
13586 * New subcommands for 'openssl ca':
13587 `openssl ca -status <serial>` prints the status of the cert with
13588 the given serial number (according to the index file).
13589 `openssl ca -updatedb` updates the expiry status of certificates
13592 *Massimiliano Pala <madwolf@comune.modena.it>*
13594 * New '-newreq-nodes' command option to CA.pl. This is like
13595 '-newreq', but calls 'openssl req' with the '-nodes' option
13596 so that the resulting key is not encrypted.
13598 *Damien Miller <djm@mindrot.org>*
13600 * New configuration for the GNU Hurd.
13602 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13604 * Initial code to implement OCSP basic response verify. This
13605 is currently incomplete. Currently just finds the signer's
13606 certificate and verifies the signature on the response.
13610 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13611 value of OPENSSLDIR. This is available via the new '-d' option
13612 to 'openssl version', and is also included in 'openssl version -a'.
13616 * Allowing defining memory allocation callbacks that will be given
13617 file name and line number information in additional arguments
13618 (a `const char*` and an int). The basic functionality remains, as
13619 well as the original possibility to just replace malloc(),
13620 realloc() and free() by functions that do not know about these
13621 additional arguments. To register and find out the current
13622 settings for extended allocation functions, the following
13623 functions are provided:
13625 CRYPTO_set_mem_ex_functions
13626 CRYPTO_set_locked_mem_ex_functions
13627 CRYPTO_get_mem_ex_functions
13628 CRYPTO_get_locked_mem_ex_functions
13630 These work the same way as CRYPTO_set_mem_functions and friends.
13631 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13632 extended allocation function is enabled.
13633 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13634 a conventional allocation function is enabled.
13636 *Richard Levitte, Bodo Moeller*
13638 * Finish off removing the remaining LHASH function pointer casts.
13639 There should no longer be any prototype-casting required when using
13640 the LHASH abstraction, and any casts that remain are "bugs". See
13641 the callback types and macros at the head of lhash.h for details
13642 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13646 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13647 If /dev/[u]random devices are not available or do not return enough
13648 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13650 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13651 /etc/entropy will be queried once each in this sequence, querying stops
13652 when enough entropy was collected without querying more sockets.
13656 * Change the Unix RAND_poll() variant to be able to poll several
13657 random devices, as specified by DEVRANDOM, until a sufficient amount
13658 of data has been collected. We spend at most 10 ms on each file
13659 (select timeout) and read in non-blocking mode. DEVRANDOM now
13660 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13661 (previously it was just the string "/dev/urandom"), so on typical
13662 platforms the 10 ms delay will never occur.
13663 Also separate out the Unix variant to its own file, rand_unix.c.
13664 For VMS, there's a currently-empty rand_vms.c.
13668 * Move OCSP client related routines to ocsp_cl.c. These
13669 provide utility functions which an application needing
13670 to issue a request to an OCSP responder and analyse the
13671 response will typically need: as opposed to those which an
13672 OCSP responder itself would need which will be added later.
13674 OCSP_request_sign() signs an OCSP request with an API similar
13675 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13676 response. OCSP_response_get1_basic() extracts basic response
13677 from response. OCSP_resp_find_status(): finds and extracts status
13678 information from an OCSP_CERTID structure (which will be created
13679 when the request structure is built). These are built from lower
13680 level functions which work on OCSP_SINGLERESP structures but
13681 won't normally be used unless the application wishes to examine
13682 extensions in the OCSP response for example.
13684 Replace nonce routines with a pair of functions.
13685 OCSP_request_add1_nonce() adds a nonce value and optionally
13686 generates a random value. OCSP_check_nonce() checks the
13687 validity of the nonce in an OCSP response.
13691 * Change function OCSP_request_add() to OCSP_request_add0_id().
13692 This doesn't copy the supplied OCSP_CERTID and avoids the
13693 need to free up the newly created id. Change return type
13694 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13695 This can then be used to add extensions to the request.
13696 Deleted OCSP_request_new(), since most of its functionality
13697 is now in OCSP_REQUEST_new() (and the case insensitive name
13698 clash) apart from the ability to set the request name which
13699 will be added elsewhere.
13703 * Update OCSP API. Remove obsolete extensions argument from
13704 various functions. Extensions are now handled using the new
13705 OCSP extension code. New simple OCSP HTTP function which
13706 can be used to send requests and parse the response.
13710 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13711 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13712 uses the special reorder version of SET OF to sort the attributes
13713 and reorder them to match the encoded order. This resolves a long
13714 standing problem: a verify on a PKCS7 structure just after signing
13715 it used to fail because the attribute order did not match the
13716 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13717 it uses the received order. This is necessary to tolerate some broken
13718 software that does not order SET OF. This is handled by encoding
13719 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13720 to produce the required SET OF.
13724 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13725 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13726 files to get correct declarations of the ASN.1 item variables.
13730 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13731 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13732 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13733 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13734 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13735 ASN1_ITEM and no wrapper functions.
13739 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13740 replace the old function pointer based I/O routines. Change most of
13741 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13745 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13746 lines, recognize more "algorithms" that can be deselected, and make
13747 it complain about algorithm deselection that isn't recognised.
13751 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13752 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13753 to use new functions. Add NO_ASN1_OLD which can be set to remove
13754 some old style ASN1 functions: this can be used to determine if old
13755 code will still work when these eventually go away.
13759 * New extension functions for OCSP structures, these follow the
13760 same conventions as certificates and CRLs.
13764 * New function X509V3_add1_i2d(). This automatically encodes and
13765 adds an extension. Its behaviour can be customised with various
13766 flags to append, replace or delete. Various wrappers added for
13767 certificates and CRLs.
13771 * Fix to avoid calling the underlying ASN1 print routine when
13772 an extension cannot be parsed. Correct a typo in the
13773 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13777 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13778 entries for variables.
13782 * Add functionality to `apps/openssl.c` for detecting locking
13783 problems: As the program is single-threaded, all we have
13784 to do is register a locking callback using an array for
13785 storing which locks are currently held by the program.
13789 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13790 SSL_get_ex_data_X509_STORE_idx(), which is used in
13791 ssl_verify_cert_chain() and thus can be called at any time
13792 during TLS/SSL handshakes so that thread-safety is essential.
13793 Unfortunately, the ex_data design is not at all suited
13794 for multi-threaded use, so it probably should be abolished.
13798 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13800 *Broadcom, tweaked and integrated by Geoff Thorpe*
13802 * Move common extension printing code to new function
13803 X509V3_print_extensions(). Reorganise OCSP print routines and
13804 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13808 * New function X509_signature_print() to remove duplication in some
13813 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13814 set (this was treated exactly the same as SET OF previously). This
13815 is used to reorder the STACK representing the structure to match the
13816 encoding. This will be used to get round a problem where a PKCS7
13817 structure which was signed could not be verified because the STACK
13818 order did not reflect the encoded order.
13822 * Reimplement the OCSP ASN1 module using the new code.
13826 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13827 for its ASN1 operations. The old style function pointers still exist
13828 for now but they will eventually go away.
13832 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13833 completely replaces the old ASN1 functionality with a table driven
13834 encoder and decoder which interprets an ASN1_ITEM structure describing
13835 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13836 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13837 has also been converted to the new form.
13841 * Change BN_mod_exp_recp so that negative moduli are tolerated
13842 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13843 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13844 for negative moduli.
13848 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13849 of not touching the result's sign bit.
13853 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13858 * Changed the LHASH code to use prototypes for callbacks, and created
13859 macros to declare and implement thin (optionally static) functions
13860 that provide type-safety and avoid function pointer casting for the
13861 type-specific callbacks.
13865 * Added Kerberos Cipher Suites to be used with TLS, as written in
13867 *Veers Staats <staatsvr@asc.hpc.mil>,
13868 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13870 * Reformat the FAQ so the different questions and answers can be divided
13871 in sections depending on the subject.
13875 * Have the zlib compression code load ZLIB.DLL dynamically under
13880 * New function BN_mod_sqrt for computing square roots modulo a prime
13881 (using the probabilistic Tonelli-Shanks algorithm unless
13882 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13883 be handled deterministically).
13885 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13887 * Make BN_mod_inverse faster by explicitly handling small quotients
13888 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13889 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13893 * New function BN_kronecker.
13897 * Fix BN_gcd so that it works on negative inputs; the result is
13898 positive unless both parameters are zero.
13899 Previously something reasonably close to an infinite loop was
13900 possible because numbers could be growing instead of shrinking
13901 in the implementation of Euclid's algorithm.
13905 * Fix BN_is_word() and BN_is_one() macros to take into account the
13906 sign of the number in question.
13908 Fix BN_is_word(a,w) to work correctly for w == 0.
13910 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13911 because its test if the absolute value of 'a' equals 'w'.
13912 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13913 it exists mostly for use in the implementations of BN_is_zero(),
13914 BN_is_one(), and BN_is_word().
13918 * New function BN_swap.
13922 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13923 the exponentiation functions are more likely to produce reasonable
13924 results on negative inputs.
13928 * Change BN_mod_mul so that the result is always non-negative.
13929 Previously, it could be negative if one of the factors was negative;
13930 I don't think anyone really wanted that behaviour.
13934 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13935 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13936 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13937 and add new functions:
13946 BN_mod_lshift1_quick
13948 BN_mod_lshift_quick
13950 These functions always generate non-negative results.
13952 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13953 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13955 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13956 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13957 be reduced modulo `m`.
13959 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13962 The following entry accidentally appeared in the CHANGES file
13963 distributed with OpenSSL 0.9.7. The modifications described in
13964 it do *not* apply to OpenSSL 0.9.7.
13966 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13967 was actually never needed) and in BN_mul(). The removal in BN_mul()
13968 required a small change in bn_mul_part_recursive() and the addition
13969 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13970 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13971 bn_sub_words() and bn_add_words() except they take arrays with
13977 * In 'openssl passwd', verify passwords read from the terminal
13978 unless the '-salt' option is used (which usually means that
13979 verification would just waste user's time since the resulting
13980 hash is going to be compared with some given password hash)
13981 or the new '-noverify' option is used.
13983 This is an incompatible change, but it does not affect
13984 non-interactive use of 'openssl passwd' (passwords on the command
13985 line, '-stdin' option, '-in ...' option) and thus should not
13986 cause any problems.
13990 * Remove all references to RSAref, since there's no more need for it.
13994 * Make DSO load along a path given through an environment variable
13995 (SHLIB_PATH) with shl_load().
13999 * Constify the ENGINE code as a result of BIGNUM constification.
14000 Also constify the RSA code and most things related to it. In a
14001 few places, most notable in the depth of the ASN.1 code, ugly
14002 casts back to non-const were required (to be solved at a later
14007 * Make it so the openssl application has all engines loaded by default.
14011 * Constify the BIGNUM routines a little more.
14015 * Add the following functions:
14017 ENGINE_load_cswift()
14019 ENGINE_load_atalla()
14020 ENGINE_load_nuron()
14021 ENGINE_load_builtin_engines()
14023 That way, an application can itself choose if external engines that
14024 are built-in in OpenSSL shall ever be used or not. The benefit is
14025 that applications won't have to be linked with libdl or other dso
14026 libraries unless it's really needed.
14028 Changed 'openssl engine' to load all engines on demand.
14029 Changed the engine header files to avoid the duplication of some
14030 declarations (they differed!).
14034 * 'openssl engine' can now list capabilities.
14038 * Better error reporting in 'openssl engine'.
14042 * Never call load_dh_param(NULL) in s_server.
14046 * Add engine application. It can currently list engines by name and
14047 identity, and test if they are actually available.
14051 * Improve RPM specification file by forcing symbolic linking and making
14052 sure the installed documentation is also owned by root.root.
14054 *Damien Miller <djm@mindrot.org>*
14056 * Give the OpenSSL applications more possibilities to make use of
14057 keys (public as well as private) handled by engines.
14061 * Add OCSP code that comes from CertCo.
14065 * Add VMS support for the Rijndael code.
14069 * Added untested support for Nuron crypto accelerator.
14073 * Add support for external cryptographic devices. This code was
14074 previously distributed separately as the "engine" branch.
14076 *Geoff Thorpe, Richard Levitte*
14078 * Rework the filename-translation in the DSO code. It is now possible to
14079 have far greater control over how a "name" is turned into a filename
14080 depending on the operating environment and any oddities about the
14081 different shared library filenames on each system.
14085 * Support threads on FreeBSD-elf in Configure.
14089 * Fix for SHA1 assembly problem with MASM: it produces
14090 warnings about corrupt line number information when assembling
14091 with debugging information. This is caused by the overlapping
14094 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14097 NCONF_get_number() has no error checking at all. As a replacement,
14098 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14099 promoted strongly. The old NCONF_get_number is kept around for
14100 binary backward compatibility.
14101 Make it possible for methods to load from something other than a BIO,
14102 by providing a function pointer that is given a name instead of a BIO.
14103 For example, this could be used to load configuration data from an
14108 * Fix for non blocking accept BIOs. Added new I/O special reason
14109 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14110 with non blocking I/O was not possible because no retry code was
14111 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14116 * Added the beginnings of Rijndael support.
14120 * Fix for bug in DirectoryString mask setting. Add support for
14121 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14122 to allow certificate printing to more controllable, additional
14123 'certopt' option to 'x509' to allow new printing options to be
14128 * Clean old EAY MD5 hack from e_os.h.
14132 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14134 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14135 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14137 *Joe Orton, Steve Henson*
14139 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14141 * Fix additional bug revealed by the NISCC test suite:
14143 Stop bug triggering large recursion when presented with
14144 certain ASN.1 tags ([CVE-2003-0851])
14148 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14150 * Fix various bugs revealed by running the NISCC test suite:
14152 Stop out of bounds reads in the ASN1 code when presented with
14153 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14155 If verify callback ignores invalid public key errors don't try to check
14156 certificate signature with the NULL public key.
14160 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14161 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14166 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14167 extra data after the compression methods not only for TLS 1.0
14168 but also for SSL 3.0 (as required by the specification).
14170 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14172 * Change X509_certificate_type() to mark the key as exported/exportable
14173 when it's 512 *bits* long, not 512 bytes.
14177 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14179 * Countermeasure against the Klima-Pokorny-Rosa extension of
14180 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14181 a protocol version number mismatch like a decryption error
14182 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14186 * Turn on RSA blinding by default in the default implementation
14187 to avoid a timing attack. Applications that don't want it can call
14188 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14189 They would be ill-advised to do so in most cases.
14191 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14193 * Change RSA blinding code so that it works when the PRNG is not
14194 seeded (in this case, the secret RSA exponent is abused as
14195 an unpredictable seed -- if it is not unpredictable, there
14196 is no point in blinding anyway). Make RSA blinding thread-safe
14197 by remembering the creator's thread ID in rsa->blinding and
14198 having all other threads use local one-time blinding factors
14199 (this requires more computation than sharing rsa->blinding, but
14200 avoids excessive locking; and if an RSA object is not shared
14201 between threads, blinding will still be very fast).
14205 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14207 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14208 via timing by performing a MAC computation even if incorrect
14209 block cipher padding has been found. This is a countermeasure
14210 against active attacks where the attacker has to distinguish
14211 between bad padding and a MAC verification error. ([CVE-2003-0078])
14213 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14214 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14215 Martin Vuagnoux (EPFL, Ilion)*
14217 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14219 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14220 memory from its contents. This is done with a counter that will
14221 place alternating values in each byte. This can be used to solve
14222 two issues: 1) the removal of calls to memset() by highly optimizing
14223 compilers, and 2) cleansing with other values than 0, since those can
14224 be read through on certain media, for example a swap space on disk.
14228 * Bugfix: client side session caching did not work with external caching,
14229 because the session->cipher setting was not restored when reloading
14230 from the external cache. This problem was masked, when
14231 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14232 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14236 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14237 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14239 *Zeev Lieber <zeev-l@yahoo.com>*
14241 * Undo an undocumented change introduced in 0.9.6e which caused
14242 repeated calls to OpenSSL_add_all_ciphers() and
14243 OpenSSL_add_all_digests() to be ignored, even after calling
14248 * Change the default configuration reader to deal with last line not
14249 being properly terminated.
14253 * Change X509_NAME_cmp() so it applies the special rules on handling
14254 DN values that are of type PrintableString, as well as RDNs of type
14255 emailAddress where the value has the type ia5String.
14257 *stefank@valicert.com via Richard Levitte*
14259 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14260 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14261 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14262 the bitwise-OR of the two for use by the majority of applications
14263 wanting this behaviour, and update the docs. The documented
14264 behaviour and actual behaviour were inconsistent and had been
14265 changing anyway, so this is more a bug-fix than a behavioural
14268 *Geoff Thorpe, diagnosed by Nadav Har'El*
14270 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14271 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14275 * Fix initialization code race conditions in
14276 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14277 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14278 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14279 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14280 ssl2_get_cipher_by_char(),
14281 ssl3_get_cipher_by_char().
14283 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14285 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14286 the cached sessions are flushed, as the remove_cb() might use ex_data
14287 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14288 (see [openssl.org #212]).
14290 *Geoff Thorpe, Lutz Jaenicke*
14292 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14293 length, instead of the encoding length to d2i_ASN1_OBJECT.
14297 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14299 * [In 0.9.6g-engine release:]
14300 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14302 *Lynn Gazis <lgazis@rainbow.com>*
14304 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14306 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14307 and get fix the header length calculation.
14308 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14309 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14311 * Use proper error handling instead of 'assertions' in buffer
14312 overflow checks added in 0.9.6e. This prevents DoS (the
14313 assertions could call abort()).
14315 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14317 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14319 * Add various sanity checks to asn1_get_length() to reject
14320 the ASN1 length bytes if they exceed sizeof(long), will appear
14321 negative or the content length exceeds the length of the
14324 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14326 * Fix cipher selection routines: ciphers without encryption had no flags
14327 for the cipher strength set and where therefore not handled correctly
14328 by the selection routines (PR #130).
14332 * Fix EVP_dsa_sha macro.
14337 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14338 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14339 that was added in OpenSSL 0.9.6d.
14341 As the countermeasure turned out to be incompatible with some
14342 broken SSL implementations, the new option is part of SSL_OP_ALL.
14343 SSL_OP_ALL is usually employed when compatibility with weird SSL
14344 implementations is desired (e.g. '-bugs' option to 's_client' and
14345 's_server'), so the new option is automatically set in many
14350 * Changes in security patch:
14352 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14353 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14354 Air Force Materiel Command, USAF, under agreement number
14357 * Add various sanity checks to asn1_get_length() to reject
14358 the ASN1 length bytes if they exceed sizeof(long), will appear
14359 negative or the content length exceeds the length of the
14360 supplied buffer. ([CVE-2002-0659])
14362 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14364 * Assertions for various potential buffer overflows, not known to
14365 happen in practice.
14367 *Ben Laurie (CHATS)*
14369 * Various temporary buffers to hold ASCII versions of integers were
14370 too small for 64 bit platforms. ([CVE-2002-0655])
14371 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14373 * Remote buffer overflow in SSL3 protocol - an attacker could
14374 supply an oversized session ID to a client. ([CVE-2002-0656])
14376 *Ben Laurie (CHATS)*
14378 * Remote buffer overflow in SSL2 protocol - an attacker could
14379 supply an oversized client master key. ([CVE-2002-0656])
14381 *Ben Laurie (CHATS)*
14383 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14385 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14386 encoded as NULL) with id-dsa-with-sha1.
14388 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14390 * Check various `X509_...()` return values in `apps/req.c`.
14392 *Nils Larsch <nla@trustcenter.de>*
14394 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14395 an end-of-file condition would erroneously be flagged, when the CRLF
14396 was just at the end of a processed block. The bug was discovered when
14397 processing data through a buffering memory BIO handing the data to a
14398 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14399 <ptsekov@syntrex.com> and Nedelcho Stanev.
14403 * Implement a countermeasure against a vulnerability recently found
14404 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14405 before application data chunks to avoid the use of known IVs
14406 with data potentially chosen by the attacker.
14410 * Fix length checks in ssl3_get_client_hello().
14414 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14415 to prevent ssl3_read_internal() from incorrectly assuming that
14416 ssl3_read_bytes() found application data while handshake
14417 processing was enabled when in fact s->s3->in_read_app_data was
14418 merely automatically cleared during the initial handshake.
14420 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14422 * Fix object definitions for Private and Enterprise: they were not
14423 recognized in their shortname (=lowercase) representation. Extend
14424 obj_dat.pl to issue an error when using undefined keywords instead
14425 of silently ignoring the problem (Svenning Sorensen
14426 <sss@sss.dnsalias.net>).
14430 * Fix DH_generate_parameters() so that it works for 'non-standard'
14431 generators, i.e. generators other than 2 and 5. (Previously, the
14432 code did not properly initialise the 'add' and 'rem' values to
14433 BN_generate_prime().)
14435 In the new general case, we do not insist that 'generator' is
14436 actually a primitive root: This requirement is rather pointless;
14437 a generator of the order-q subgroup is just as good, if not
14442 * Map new X509 verification errors to alerts. Discovered and submitted by
14443 Tom Wu <tom@arcot.com>.
14447 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14448 returning non-zero before the data has been completely received
14449 when using non-blocking I/O.
14451 *Bodo Moeller; problem pointed out by John Hughes*
14453 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14455 *Ben Laurie, Lutz Jaenicke*
14457 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14458 Yoram Zahavi <YoramZ@gilian.com>).
14462 * Add information about CygWin 1.3 and on, and preserve proper
14463 configuration for the versions before that.
14465 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14467 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14468 check whether we deal with a copy of a session and do not delete from
14469 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14470 <izhar@checkpoint.com>.
14474 * Do not store session data into the internal session cache, if it
14475 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14476 flag is set). Proposed by Aslam <aslam@funk.com>.
14480 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14485 * [In 0.9.6d-engine release:]
14486 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14488 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14490 * Add the configuration target linux-s390x.
14492 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14494 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14495 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14496 variable as an indication that a ClientHello message has been
14497 received. As the flag value will be lost between multiple
14498 invocations of ssl3_accept when using non-blocking I/O, the
14499 function may not be aware that a handshake has actually taken
14500 place, thus preventing a new session from being added to the
14503 To avoid this problem, we now set s->new_session to 2 instead of
14504 using a local variable.
14506 *Lutz Jaenicke, Bodo Moeller*
14508 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14509 if the SSL_R_LENGTH_MISMATCH error is detected.
14511 *Geoff Thorpe, Bodo Moeller*
14513 * New 'shared_ldflag' column in Configure platform table.
14517 * Fix EVP_CIPHER_mode macro.
14519 *"Dan S. Camper" <dan@bti.net>*
14521 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14522 type, we must throw them away by setting rr->length to 0.
14524 *D P Chang <dpc@qualys.com>*
14526 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14528 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14529 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14530 worked incorrectly for those cases where range = `10..._2` and
14531 `3*range` is two bits longer than range.)
14535 * Only add signing time to PKCS7 structures if it is not already
14540 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14541 OBJ_ld_ce should be OBJ_id_ce.
14542 Also some ip-pda OIDs in crypto/objects/objects.txt were
14543 incorrect (cf. RFC 3039).
14545 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14547 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14548 returns early because it has nothing to do.
14550 *Andy Schneider <andy.schneider@bjss.co.uk>*
14552 * [In 0.9.6c-engine release:]
14553 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14555 *Andy Schneider <andy.schneider@bjss.co.uk>*
14557 * [In 0.9.6c-engine release:]
14558 Add support for Cryptographic Appliance's keyserver technology.
14559 (Use engine 'keyclient')
14561 *Cryptographic Appliances and Geoff Thorpe*
14563 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14564 is called via tools/c89.sh because arguments have to be
14565 rearranged (all '-L' options must appear before the first object
14568 *Richard Shapiro <rshapiro@abinitio.com>*
14570 * [In 0.9.6c-engine release:]
14571 Add support for Broadcom crypto accelerator cards, backported
14574 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14576 * [In 0.9.6c-engine release:]
14577 Add support for SureWare crypto accelerator cards from
14578 Baltimore Technologies. (Use engine 'sureware')
14580 *Baltimore Technologies and Mark Cox*
14582 * [In 0.9.6c-engine release:]
14583 Add support for crypto accelerator cards from Accelerated
14584 Encryption Processing, www.aep.ie. (Use engine 'aep')
14586 *AEP Inc. and Mark Cox*
14588 * Add a configuration entry for gcc on UnixWare.
14590 *Gary Benson <gbenson@redhat.com>*
14592 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14593 messages are stored in a single piece (fixed-length part and
14594 variable-length part combined) and fix various bugs found on the way.
14598 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14599 instead. BIO_gethostbyname() does not know what timeouts are
14600 appropriate, so entries would stay in cache even when they have
14602 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14604 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14605 faced with a pathologically small ClientHello fragment that does
14606 not contain client_version: Instead of aborting with an error,
14607 simply choose the highest available protocol version (i.e.,
14608 TLS 1.0 unless it is disabled). In practice, ClientHello
14609 messages are never sent like this, but this change gives us
14610 strictly correct behaviour at least for TLS.
14614 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14615 never resets s->method to s->ctx->method when called from within
14616 one of the SSL handshake functions.
14618 *Bodo Moeller; problem pointed out by Niko Baric*
14620 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14621 (sent using the client's version number) if client_version is
14622 smaller than the protocol version in use. Also change
14623 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14624 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14625 the client will at least see that alert.
14629 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14634 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14635 client receives HelloRequest while in a handshake.
14637 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14639 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14640 should end in 'break', not 'goto end' which circumvents various
14641 cleanups done in state SSL_ST_OK. But session related stuff
14642 must be disabled for SSL_ST_OK in the case that we just sent a
14645 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14646 before just sending a HelloRequest.
14648 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14650 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14651 reveal whether illegal block cipher padding was found or a MAC
14652 verification error occurred. (Neither SSLerr() codes nor alerts
14653 are directly visible to potential attackers, but the information
14654 may leak via logfiles.)
14656 Similar changes are not required for the SSL 2.0 implementation
14657 because the number of padding bytes is sent in clear for SSL 2.0,
14658 and the extra bytes are just ignored. However ssl/s2_pkt.c
14659 failed to verify that the purported number of padding bytes is in
14664 * Add OpenUNIX-8 support including shared libraries
14665 (Boyd Lynn Gerber <gerberb@zenez.com>).
14669 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14670 'wristwatch attack' using huge encoding parameters (cf.
14671 James H. Manger's CRYPTO 2001 paper). Note that the
14672 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14673 encoding parameters and hence was not vulnerable.
14677 * BN_sqr() bug fix.
14679 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14681 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14682 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14683 followed by modular reduction.
14685 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14687 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14688 equivalent based on BN_pseudo_rand() instead of BN_rand().
14692 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14693 This function was broken, as the check for a new client hello message
14694 to handle SGC did not allow these large messages.
14695 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14699 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14703 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14704 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14708 * Rework the configuration and shared library support for Tru64 Unix.
14709 The configuration part makes use of modern compiler features and
14710 still retains old compiler behavior for those that run older versions
14711 of the OS. The shared library support part includes a variant that
14712 uses the RPATH feature, and is available through the special
14713 configuration target "alpha-cc-rpath", which will never be selected
14716 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14718 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14719 with the same message size as in ssl3_get_certificate_request().
14720 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14721 messages might inadvertently be reject as too long.
14723 *Petr Lampa <lampa@fee.vutbr.cz>*
14725 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14729 * Modified SSL library such that the verify_callback that has been set
14730 specifically for an SSL object with SSL_set_verify() is actually being
14731 used. Before the change, a verify_callback set with this function was
14732 ignored and the verify_callback() set in the SSL_CTX at the time of
14733 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14734 to allow the necessary settings.
14738 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14739 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14740 done automatically (in contradiction to the requirements of the C
14741 standard). This made problems when used from OpenSSH.
14745 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14746 dh->length and always used
14748 BN_rand_range(priv_key, dh->p).
14750 BN_rand_range() is not necessary for Diffie-Hellman, and this
14751 specific range makes Diffie-Hellman unnecessarily inefficient if
14752 dh->length (recommended exponent length) is much smaller than the
14753 length of dh->p. We could use BN_rand_range() if the order of
14754 the subgroup was stored in the DH structure, but we only have
14759 BN_rand(priv_key, l, ...)
14761 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14768 RSA_eay_public_encrypt
14769 RSA_eay_private_decrypt
14770 RSA_eay_private_encrypt (signing)
14771 RSA_eay_public_decrypt (signature verification)
14773 (default implementations for RSA_public_encrypt,
14774 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14775 always reject numbers >= n.
14779 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14780 to synchronize access to 'locking_thread'. This is necessary on
14781 systems where access to 'locking_thread' (an 'unsigned long'
14782 variable) is not atomic.
14786 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14787 *before* setting the 'crypto_lock_rand' flag. The previous code had
14788 a race condition if 0 is a valid thread ID.
14790 *Travis Vitek <vitek@roguewave.com>*
14792 * Add support for shared libraries under Irix.
14794 *Albert Chin-A-Young <china@thewrittenword.com>*
14796 * Add configuration option to build on Linux on both big-endian and
14797 little-endian MIPS.
14799 *Ralf Baechle <ralf@uni-koblenz.de>*
14801 * Add the possibility to create shared libraries on HP-UX.
14805 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14807 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14808 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14809 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14810 PRNG state recovery was possible based on the output of
14811 one PRNG request appropriately sized to gain knowledge on
14812 'md' followed by enough consecutive 1-byte PRNG requests
14813 to traverse all of 'state'.
14815 1. When updating 'md_local' (the current thread's copy of 'md')
14816 during PRNG output generation, hash all of the previous
14817 'md_local' value, not just the half used for PRNG output.
14819 2. Make the number of bytes from 'state' included into the hash
14820 independent from the number of PRNG bytes requested.
14822 The first measure alone would be sufficient to avoid
14823 Markku-Juhani's attack. (Actually it had never occurred
14824 to me that the half of 'md_local' used for chaining was the
14825 half from which PRNG output bytes were taken -- I had always
14826 assumed that the secret half would be used.) The second
14827 measure makes sure that additional data from 'state' is never
14828 mixed into 'md_local' in small portions; this heuristically
14829 further strengthens the PRNG.
14833 * Fix crypto/bn/asm/mips3.s.
14837 * When only the key is given to "enc", the IV is undefined. Print out
14838 an error message in this case.
14842 * Handle special case when X509_NAME is empty in X509 printing routines.
14846 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14847 positive and less than q.
14851 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14852 used: it isn't thread safe and the add_lock_callback should handle
14855 *Paul Rose <Paul.Rose@bridge.com>*
14857 * Verify that incoming data obeys the block size in
14858 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14864 *Ulf Möller, Bodo Möller*
14866 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14867 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14868 when fixing the server behaviour for backwards-compatible 'client
14869 hello' messages. (Note that the attack is impractical against
14870 SSL 3.0 and TLS 1.0 anyway because length and version checking
14871 means that the probability of guessing a valid ciphertext is
14872 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14875 Before 0.9.5, the countermeasure (hide the error by generating a
14876 random 'decryption result') did not work properly because
14877 ERR_clear_error() was missing, meaning that SSL_get_error() would
14878 detect the supposedly ignored error.
14880 Both problems are now fixed.
14884 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14885 (previously it was 1024).
14889 * Fix for compatibility mode trust settings: ignore trust settings
14890 unless some valid trust or reject settings are present.
14894 * Fix for blowfish EVP: its a variable length cipher.
14898 * Fix various bugs related to DSA S/MIME verification. Handle missing
14899 parameters in DSA public key structures and return an error in the
14900 DSA routines if parameters are absent.
14904 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14905 in the current directory if neither $RANDFILE nor $HOME was set.
14906 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14907 caused some confusion to Windows users who haven't defined $HOME.
14908 Thus RAND_file_name() is changed again: e_os.h can define a
14909 DEFAULT_HOME, which will be used if $HOME is not set.
14910 For Windows, we use "C:"; on other platforms, we still require
14911 environment variables.
14913 * Move 'if (!initialized) RAND_poll()' into regions protected by
14914 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14915 having multiple threads call RAND_poll() concurrently.
14919 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14920 combination of a flag and a thread ID variable.
14921 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14922 flag), *other* threads can enter ssleay_add_bytes without obeying
14923 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14924 that they do not hold after the first thread unsets add_do_not_lock).
14928 * Change bctest again: '-x' expressions are not available in all
14929 versions of 'test'.
14933 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14935 * Fix a couple of memory leaks in PKCS7_dataDecode()
14937 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14939 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14940 the default extension for executables, if any. Also, make the perl
14941 scripts that use symlink() to test if it really exists and use "cp"
14942 if it doesn't. All this made OpenSSL compilable and installable in
14947 * Fix for asn1_GetSequence() for indefinite length constructed data.
14948 If SEQUENCE is length is indefinite just set c->slen to the total
14949 amount of data available.
14951 *Steve Henson, reported by shige@FreeBSD.org*
14953 *This change does not apply to 0.9.7.*
14955 * Change bctest to avoid here-documents inside command substitution
14956 (workaround for FreeBSD /bin/sh bug).
14957 For compatibility with Ultrix, avoid shell functions (introduced
14958 in the bctest version that searches along $PATH).
14962 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14963 with des_encrypt() defined on some operating systems, like Solaris
14968 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14969 On the Importance of Eliminating Errors in Cryptographic
14970 Computations, J. Cryptology 14 (2001) 2, 101-119,
14971 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14975 * MIPS assembler BIGNUM division bug fix.
14979 * Disabled incorrect Alpha assembler code.
14983 * Fix PKCS#7 decode routines so they correctly update the length
14984 after reading an EOC for the EXPLICIT tag.
14988 *This change does not apply to 0.9.7.*
14990 * Fix bug in PKCS#12 key generation routines. This was triggered
14991 if a 3DES key was generated with a 0 initial byte. Include
14992 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14993 (but broken) behaviour.
14997 * Enhance bctest to search for a working bc along $PATH and print
15000 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15002 * Fix memory leaks in err.c: free err_data string if necessary;
15003 don't write to the wrong index in ERR_set_error_data.
15007 * Implement ssl23_peek (analogous to ssl23_read), which previously
15012 * Replace rdtsc with `_emit` statements for VC++ version 5.
15014 *Jeremy Cooper <jeremy@baymoo.org>*
15016 * Make it possible to reuse SSLv2 sessions.
15020 * In copy_email() check for >= 0 as a return value for
15021 X509_NAME_get_index_by_NID() since 0 is a valid index.
15023 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15025 * Avoid coredump with unsupported or invalid public keys by checking if
15026 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15027 PKCS7_verify() fails with non detached data.
15031 * Don't use getenv in library functions when run as setuid/setgid.
15032 New function OPENSSL_issetugid().
15036 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15037 due to incorrect handling of multi-threading:
15039 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15041 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15043 3. Count how many times MemCheck_off() has been called so that
15044 nested use can be treated correctly. This also avoids
15045 inband-signalling in the previous code (which relied on the
15046 assumption that thread ID 0 is impossible).
15050 * Add "-rand" option also to s_client and s_server.
15054 * Fix CPU detection on Irix 6.x.
15055 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15056 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15058 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15063 *This change does not apply to 0.9.7.*
15065 * Use the cached encoding of an X509_NAME structure rather than
15066 copying it. This is apparently the reason for the libsafe "errors"
15067 but the code is actually correct.
15071 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15072 Bleichenbacher's DSA attack.
15073 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15074 to be set and top=0 forces the highest bit to be set; top=-1 is new
15075 and leaves the highest bit random.
15077 *Ulf Moeller, Bodo Moeller*
15079 * In the `NCONF_...`-based implementations for `CONF_...` queries
15080 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15081 a temporary CONF structure with the data component set to NULL
15082 (which gives segmentation faults in lh_retrieve).
15083 Instead, use NULL for the CONF pointer in CONF_get_string and
15084 CONF_get_number (which may use environment variables) and directly
15085 return NULL from CONF_get_section.
15089 * Fix potential buffer overrun for EBCDIC.
15093 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15094 keyUsage if basicConstraints absent for a CA.
15098 * Make SMIME_write_PKCS7() write mail header values with a format that
15099 is more generally accepted (no spaces before the semicolon), since
15100 some programs can't parse those values properly otherwise. Also make
15101 sure BIO's that break lines after each write do not create invalid
15106 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15107 macros previously used would not encode an empty SEQUENCE OF
15108 and break the signature.
15112 *This change does not apply to 0.9.7.*
15114 * Zero the premaster secret after deriving the master secret in
15119 * Add some EVP_add_digest_alias registrations (as found in
15120 OpenSSL_add_all_digests()) to SSL_library_init()
15121 aka OpenSSL_add_ssl_algorithms(). This provides improved
15122 compatibility with peers using X.509 certificates
15123 with unconventional AlgorithmIdentifier OIDs.
15127 * Fix for Irix with NO_ASM.
15129 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15131 * ./config script fixes.
15133 *Ulf Moeller, Richard Levitte*
15135 * Fix 'openssl passwd -1'.
15139 * Change PKCS12_key_gen_asc() so it can cope with non null
15140 terminated strings whose length is passed in the passlen
15141 parameter, for example from PEM callbacks. This was done
15142 by adding an extra length parameter to asc2uni().
15144 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15146 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15147 call failed, free the DSA structure.
15151 * Fix to uni2asc() to cope with zero length Unicode strings.
15152 These are present in some PKCS#12 files.
15156 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15157 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15158 when writing a 32767 byte record.
15160 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15162 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15163 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15165 (RSA objects have a reference count access to which is protected
15166 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15167 so they are meant to be shared between threads.)
15168 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15169 "Reddie, Steven" <Steven.Reddie@ca.com>*
15171 * Fix a deadlock in CRYPTO_mem_leaks().
15175 * Use better test patterns in bntest.
15179 * rand_win.c fix for Borland C.
15183 * BN_rshift bugfix for n == 0.
15187 * Add a 'bctest' script that checks for some known 'bc' bugs
15188 so that 'make test' does not abort just because 'bc' is broken.
15192 * Store verify_result within SSL_SESSION also for client side to
15193 avoid potential security hole. (Re-used sessions on the client side
15194 always resulted in verify_result==X509_V_OK, not using the original
15195 result of the server certificate verification.)
15199 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15200 SSL3_RT_APPLICATION_DATA, return 0.
15201 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15206 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15207 releases, have been re-implemented by renaming the previous
15208 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15209 and ssl3_read_internal, respectively, and adding 'peek' parameters
15210 to them. The new ssl[23]_{read,peek} functions are calls to
15211 ssl[23]_read_internal with the 'peek' flag set appropriately.
15212 A 'peek' parameter has also been added to ssl3_read_bytes, which
15213 does the actual work for ssl3_read_internal.
15217 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15218 the method-specific "init()" handler. Also clean up ex_data after
15219 calling the method-specific "finish()" handler. Previously, this was
15220 happening the other way round.
15224 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15225 The previous value, 12, was not always sufficient for BN_mod_exp().
15229 * Make sure that shared libraries get the internal name engine with
15230 the full version number and not just 0. This should mark the
15231 shared libraries as not backward compatible. Of course, this should
15232 be changed again when we can guarantee backward binary compatibility.
15236 * Fix typo in get_cert_by_subject() in by_dir.c
15238 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15240 * Rework the system to generate shared libraries:
15242 - Make note of the expected extension for the shared libraries and
15243 if there is a need for symbolic links from for example libcrypto.so.0
15244 to libcrypto.so.0.9.7. There is extended info in Configure for
15247 - Make as few rebuilds of the shared libraries as possible.
15249 - Still avoid linking the OpenSSL programs with the shared libraries.
15251 - When installing, install the shared libraries separately from the
15256 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15258 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15259 and not in SSL_clear because the latter is also used by the
15260 accept/connect functions; previously, the settings made by
15261 SSL_set_read_ahead would be lost during the handshake.
15263 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15265 * Correct util/mkdef.pl to be selective about disabled algorithms.
15266 Previously, it would create entries for disabled algorithms no
15271 * Added several new manual pages for SSL_* function.
15275 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15277 * In ssl23_get_client_hello, generate an error message when faced
15278 with an initial SSL 3.0/TLS record that is too small to contain the
15279 first two bytes of the ClientHello message, i.e. client_version.
15280 (Note that this is a pathologic case that probably has never happened
15281 in real life.) The previous approach was to use the version number
15282 from the record header as a substitute; but our protocol choice
15283 should not depend on that one because it is not authenticated
15284 by the Finished messages.
15288 * More robust randomness gathering functions for Windows.
15290 *Jeffrey Altman <jaltman@columbia.edu>*
15292 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15293 not set then we don't setup the error code for issuer check errors
15294 to avoid possibly overwriting other errors which the callback does
15295 handle. If an application does set the flag then we assume it knows
15296 what it is doing and can handle the new informational codes
15301 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15302 a general "ANY" type, as such it should be able to decode anything
15303 including tagged types. However it didn't check the class so it would
15304 wrongly interpret tagged types in the same way as their universal
15305 counterpart and unknown types were just rejected. Changed so that the
15306 tagged and unknown types are handled in the same way as a SEQUENCE:
15307 that is the encoding is stored intact. There is also a new type
15308 "V_ASN1_OTHER" which is used when the class is not universal, in this
15309 case we have no idea what the actual type is so we just lump them all
15314 * On VMS, stdout may very well lead to a file that is written to
15315 in a record-oriented fashion. That means that every write() will
15316 write a separate record, which will be read separately by the
15317 programs trying to read from it. This can be very confusing.
15319 The solution is to put a BIO filter in the way that will buffer
15320 text until a linefeed is reached, and then write everything a
15321 line at a time, so every record written will be an actual line,
15322 not chunks of lines and not (usually doesn't happen, but I've
15323 seen it once) several lines in one record. BIO_f_linebuffer() is
15326 Currently, it's a VMS-only method, because that's where it has
15327 been tested well enough.
15331 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15332 it can return incorrect results.
15333 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15334 but it was in 0.9.6-beta[12].)
15338 * Disable the check for content being present when verifying detached
15339 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15340 include zero length content when signing messages.
15344 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15345 BIO_ctrl (for BIO pairs).
15349 * Add DSO method for VMS.
15353 * Bug fix: Montgomery multiplication could produce results with the
15358 * Add RPM specification openssl.spec and modify it to build three
15359 packages. The default package contains applications, application
15360 documentation and run-time libraries. The devel package contains
15361 include files, static libraries and function documentation. The
15362 doc package contains the contents of the doc directory. The original
15363 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15367 * Add a large number of documentation files for many SSL routines.
15369 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15371 * Add a configuration entry for Sony News 4.
15373 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15375 * Don't set the two most significant bits to one when generating a
15376 random number < q in the DSA library.
15380 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15381 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15382 the underlying transport is blocking) if a handshake took place.
15383 (The default behaviour is needed by applications such as s_client
15384 and s_server that use select() to determine when to use SSL_read;
15385 but for applications that know in advance when to expect data, it
15386 just makes things more complicated.)
15390 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15395 * Add a few more EBCDIC conditionals that make `req` and `x509`
15396 work better on such systems.
15398 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15400 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15401 Update PKCS12_parse() so it copies the friendlyName and the
15402 keyid to the certificates aux info.
15406 * Fix bug in PKCS7_verify() which caused an infinite loop
15407 if there was more than one signature.
15409 *Sven Uszpelkat <su@celocom.de>*
15411 * Major change in util/mkdef.pl to include extra information
15412 about each symbol, as well as presenting variables as well
15413 as functions. This change means that there's n more need
15414 to rebuild the .num files when some algorithms are excluded.
15418 * Allow the verify time to be set by an application,
15419 rather than always using the current time.
15423 * Phase 2 verify code reorganisation. The certificate
15424 verify code now looks up an issuer certificate by a
15425 number of criteria: subject name, authority key id
15426 and key usage. It also verifies self signed certificates
15427 by the same criteria. The main comparison function is
15428 X509_check_issued() which performs these checks.
15430 Lot of changes were necessary in order to support this
15431 without completely rewriting the lookup code.
15433 Authority and subject key identifier are now cached.
15435 The LHASH 'certs' is X509_STORE has now been replaced
15436 by a STACK_OF(X509_OBJECT). This is mainly because an
15437 LHASH can't store or retrieve multiple objects with
15438 the same hash value.
15440 As a result various functions (which were all internal
15441 use only) have changed to handle the new X509_STORE
15442 structure. This will break anything that messed round
15443 with X509_STORE internally.
15445 The functions X509_STORE_add_cert() now checks for an
15446 exact match, rather than just subject name.
15448 The X509_STORE API doesn't directly support the retrieval
15449 of multiple certificates matching a given criteria, however
15450 this can be worked round by performing a lookup first
15451 (which will fill the cache with candidate certificates)
15452 and then examining the cache for matches. This is probably
15453 the best we can do without throwing out X509_LOOKUP
15454 entirely (maybe later...).
15456 The X509_VERIFY_CTX structure has been enhanced considerably.
15458 All certificate lookup operations now go via a get_issuer()
15459 callback. Although this currently uses an X509_STORE it
15460 can be replaced by custom lookups. This is a simple way
15461 to bypass the X509_STORE hackery necessary to make this
15462 work and makes it possible to use more efficient techniques
15463 in future. A very simple version which uses a simple
15464 STACK for its trusted certificate store is also provided
15465 using X509_STORE_CTX_trusted_stack().
15467 The verify_cb() and verify() callbacks now have equivalents
15468 in the X509_STORE_CTX structure.
15470 X509_STORE_CTX also has a 'flags' field which can be used
15471 to customise the verify behaviour.
15475 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15476 excludes S/MIME capabilities.
15480 * When a certificate request is read in keep a copy of the
15481 original encoding of the signed data and use it when outputting
15482 again. Signatures then use the original encoding rather than
15483 a decoded, encoded version which may cause problems if the
15484 request is improperly encoded.
15488 * For consistency with other BIO_puts implementations, call
15489 buffer_write(b, ...) directly in buffer_puts instead of calling
15492 In BIO_puts, increment b->num_write as in BIO_write.
15494 *Peter.Sylvester@EdelWeb.fr*
15496 * Fix BN_mul_word for the case where the word is 0. (We have to use
15497 BN_zero, we may not return a BIGNUM with an array consisting of
15498 words set to zero.)
15502 * Avoid calling abort() from within the library when problems are
15503 detected, except if preprocessor symbols have been defined
15504 (such as REF_CHECK, BN_DEBUG etc.).
15508 * New openssl application 'rsautl'. This utility can be
15509 used for low-level RSA operations. DER public key
15510 BIO/fp routines also added.
15514 * New Configure entry and patches for compiling on QNX 4.
15516 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15518 * A demo state-machine implementation was sponsored by
15519 Nuron (<http://www.nuron.com/>) and is now available in
15520 demos/state_machine.
15524 * New options added to the 'dgst' utility for signature
15525 generation and verification.
15529 * Unrecognized PKCS#7 content types are now handled via a
15530 catch all ASN1_TYPE structure. This allows unsupported
15531 types to be stored as a "blob" and an application can
15532 encode and decode it manually.
15536 * Fix various signed/unsigned issues to make a_strex.c
15537 compile under VC++.
15539 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15541 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15542 length if passed a buffer. ASN1_INTEGER_to_BN failed
15543 if passed a NULL BN and its argument was negative.
15545 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15547 * Modification to PKCS#7 encoding routines to output definite
15548 length encoding. Since currently the whole structures are in
15549 memory there's not real point in using indefinite length
15550 constructed encoding. However if OpenSSL is compiled with
15551 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15555 * Added BIO_vprintf() and BIO_vsnprintf().
15559 * Added more prefixes to parse for in the strings written
15560 through a logging bio, to cover all the levels that are available
15561 through syslog. The prefixes are now:
15563 PANIC, EMERG, EMR => LOG_EMERG
15564 ALERT, ALR => LOG_ALERT
15565 CRIT, CRI => LOG_CRIT
15566 ERROR, ERR => LOG_ERR
15567 WARNING, WARN, WAR => LOG_WARNING
15568 NOTICE, NOTE, NOT => LOG_NOTICE
15569 INFO, INF => LOG_INFO
15570 DEBUG, DBG => LOG_DEBUG
15572 and as before, if none of those prefixes are present at the
15573 beginning of the string, LOG_ERR is chosen.
15575 On Win32, the `LOG_*` levels are mapped according to this:
15577 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15578 LOG_WARNING => EVENTLOG_WARNING_TYPE
15579 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15583 * Made it possible to reconfigure with just the configuration
15584 argument "reconf" or "reconfigure". The command line arguments
15585 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15586 and are retrieved from there when reconfiguring.
15592 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15594 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15598 * The obj_dat.pl script was messing up the sorting of object
15599 names. The reason was that it compared the quoted version
15600 of strings as a result "OCSP" > "OCSP Signing" because
15601 " > SPACE. Changed script to store unquoted versions of
15602 names and add quotes on output. It was also omitting some
15603 names from the lookup table if they were given a default
15604 value (that is if SN is missing it is given the same
15605 value as LN and vice versa), these are now added on the
15606 grounds that if an object has a name we should be able to
15607 look it up. Finally added warning output when duplicate
15608 short or long names are found.
15612 * Changes needed for Tandem NSK.
15614 *Scott Uroff <scott@xypro.com>*
15616 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15617 RSA_padding_check_SSLv23(), special padding was never detected
15618 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15619 version rollback attacks was not effective.
15621 In s23_clnt.c, don't use special rollback-attack detection padding
15622 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15623 client; similarly, in s23_srvr.c, don't do the rollback check if
15624 SSL 2.0 is the only protocol enabled in the server.
15628 * Make it possible to get hexdumps of unprintable data with 'openssl
15629 asn1parse'. By implication, the functions ASN1_parse_dump() and
15630 BIO_dump_indent() are added.
15634 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15635 these print out strings and name structures based on various
15636 flags including RFC2253 support and proper handling of
15637 multibyte characters. Added options to the 'x509' utility
15638 to allow the various flags to be set.
15642 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15643 Also change the functions X509_cmp_current_time() and
15644 X509_gmtime_adj() work with an ASN1_TIME structure,
15645 this will enable certificates using GeneralizedTime in validity
15646 dates to be checked.
15650 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15651 negative public key encodings) on by default,
15652 NO_NEG_PUBKEY_BUG can be set to disable it.
15656 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15657 content octets. An i2c_ASN1_OBJECT is unnecessary because
15658 the encoding can be trivially obtained from the structure.
15662 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15663 not read locks (`CRYPTO_r_[un]lock`).
15667 * A first attempt at creating official support for shared
15668 libraries through configuration. I've kept it so the
15669 default is static libraries only, and the OpenSSL programs
15670 are always statically linked for now, but there are
15671 preparations for dynamic linking in place.
15672 This has been tested on Linux and Tru64.
15676 * Randomness polling function for Win9x, as described in:
15677 Peter Gutmann, Software Generation of Practically Strong
15682 * Fix so PRNG is seeded in req if using an already existing
15687 * New options to smime application. -inform and -outform
15688 allow alternative formats for the S/MIME message including
15689 PEM and DER. The -content option allows the content to be
15690 specified separately. This should allow things like Netscape
15691 form signing output easier to verify.
15695 * Fix the ASN1 encoding of tags using the 'long form'.
15699 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15700 STRING types. These convert content octets to and from the
15701 underlying type. The actual tag and length octets are
15702 already assumed to have been read in and checked. These
15703 are needed because all other string types have virtually
15704 identical handling apart from the tag. By having versions
15705 of the ASN1 functions that just operate on content octets
15706 IMPLICIT tagging can be handled properly. It also allows
15707 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15708 and ASN1_INTEGER are identical apart from the tag.
15712 * Change the handling of OID objects as follows:
15714 - New object identifiers are inserted in objects.txt, following
15715 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15716 - objects.pl is used to process obj_mac.num and create a new
15718 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15721 This is currently kind of a hack, and the perl code in objects.pl
15722 isn't very elegant, but it works as I intended. The simplest way
15723 to check that it worked correctly is to look in obj_dat.h and
15724 check the array nid_objs and make sure the objects haven't moved
15725 around (this is important!). Additions are OK, as well as
15726 consistent name changes.
15730 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15734 * Addition of the command line parameter '-rand file' to 'openssl req'.
15735 The given file adds to whatever has already been seeded into the
15736 random pool through the RANDFILE configuration file option or
15737 environment variable, or the default random state file.
15741 * mkstack.pl now sorts each macro group into lexical order.
15742 Previously the output order depended on the order the files
15743 appeared in the directory, resulting in needless rewriting
15748 * Patches to make OpenSSL compile under Win32 again. Mostly
15749 work arounds for the VC++ problem that it treats func() as
15750 func(void). Also stripped out the parts of mkdef.pl that
15751 added extra typesafe functions: these no longer exist.
15755 * Reorganisation of the stack code. The macros are now all
15756 collected in safestack.h . Each macro is defined in terms of
15757 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15758 DEBUG_SAFESTACK is now handled in terms of function casts,
15759 this has the advantage of retaining type safety without the
15760 use of additional functions. If DEBUG_SAFESTACK is not defined
15761 then the non typesafe macros are used instead. Also modified the
15762 mkstack.pl script to handle the new form. Needs testing to see
15763 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15764 the default if no major problems. Similar behaviour for ASN1_SET_OF
15765 and PKCS12_STACK_OF.
15769 * When some versions of IIS use the 'NET' form of private key the
15770 key derivation algorithm is different. Normally MD5(password) is
15771 used as a 128 bit RC4 key. In the modified case
15772 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15773 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15774 as the old Netscape_RSA functions except they have an additional
15775 'sgckey' parameter which uses the modified algorithm. Also added
15776 an -sgckey command line option to the rsa utility. Thanks to
15777 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15778 algorithm to openssl-dev.
15782 * The evp_local.h macros were using 'c.##kname' which resulted in
15783 invalid expansion on some systems (SCO 5.0.5 for example).
15784 Corrected to 'c.kname'.
15786 *Phillip Porch <root@theporch.com>*
15788 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15789 a STACK of email addresses from a certificate or request, these look
15790 in the subject name and the subject alternative name extensions and
15791 omit any duplicate addresses.
15795 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15796 This makes DSA verification about 2 % faster.
15800 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15801 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15802 plus overhead for 1024 bit moduli).
15803 This makes exponentiations about 0.5 % faster for 1024 bit
15804 exponents (as measured by "openssl speed rsa2048").
15808 * Rename memory handling macros to avoid conflicts with other
15810 Malloc => OPENSSL_malloc
15811 Malloc_locked => OPENSSL_malloc_locked
15812 Realloc => OPENSSL_realloc
15813 Free => OPENSSL_free
15817 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15818 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15822 * CygWin32 support.
15824 *John Jarvie <jjarvie@newsguy.com>*
15826 * The type-safe stack code has been rejigged. It is now only compiled
15827 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15828 by default all type-specific stack functions are "#define"d back to
15829 standard stack functions. This results in more streamlined output
15830 but retains the type-safety checking possibilities of the original
15835 * The STACK code has been cleaned up, and certain type declarations
15836 that didn't make a lot of sense have been brought in line. This has
15837 also involved a cleanup of sorts in safestack.h to more correctly
15838 map type-safe stack functions onto their plain stack counterparts.
15839 This work has also resulted in a variety of "const"ifications of
15840 lots of the code, especially `_cmp` operations which should normally
15841 be prototyped with "const" parameters anyway.
15845 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15846 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15847 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15848 where all of 'md' is used each time the PRNG is used, but 'state'
15849 is used only indexed by a cyclic counter. As entropy may not be
15850 well distributed from the beginning, 'md' is important as a
15851 chaining variable. However, the output function chains only half
15852 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15853 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15854 in all of 'state' being rewritten, with the new values depending
15855 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15859 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15860 the handshake is continued after ssl_verify_cert_chain();
15861 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15862 can lead to 'unexplainable' connection aborts later.
15864 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15866 * Major EVP API cipher revision.
15867 Add hooks for extra EVP features. This allows various cipher
15868 parameters to be set in the EVP interface. Support added for variable
15869 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15870 setting of RC2 and RC5 parameters.
15872 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15875 Remove lots of duplicated code from the EVP library. For example *every*
15876 cipher init() function handles the 'iv' in the same way according to the
15877 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15878 for CFB and OFB modes they zero ctx->num.
15880 New functionality allows removal of S/MIME code RC2 hack.
15882 Most of the routines have the same form and so can be declared in terms
15885 By shifting this to the top level EVP_CipherInit() it can be removed from
15886 all individual ciphers. If the cipher wants to handle IVs or keys
15887 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15890 Change lots of functions like EVP_EncryptUpdate() to now return a
15891 value: although software versions of the algorithms cannot fail
15892 any installed hardware versions can.
15896 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15897 this option is set, tolerate broken clients that send the negotiated
15898 protocol version number instead of the requested protocol version
15903 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15904 i.e. non-zero for export ciphersuites, zero otherwise.
15905 Previous versions had this flag inverted, inconsistent with
15906 rsa_tmp_cb (..._TMP_RSA_CB).
15908 *Bodo Moeller; problem reported by Amit Chopra*
15910 * Add missing DSA library text string. Work around for some IIS
15911 key files with invalid SEQUENCE encoding.
15915 * Add a document (doc/standards.txt) that list all kinds of standards
15916 and so on that are implemented in OpenSSL.
15920 * Enhance c_rehash script. Old version would mishandle certificates
15921 with the same subject name hash and wouldn't handle CRLs at all.
15922 Added -fingerprint option to crl utility, to support new c_rehash
15927 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15931 * Fix for SSL server purpose checking. Server checking was
15932 rejecting certificates which had extended key usage present
15933 but no ssl client purpose.
15935 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15937 * Make PKCS#12 code work with no password. The PKCS#12 spec
15938 is a little unclear about how a blank password is handled.
15939 Since the password in encoded as a BMPString with terminating
15940 double NULL a zero length password would end up as just the
15941 double NULL. However no password at all is different and is
15942 handled differently in the PKCS#12 key generation code. NS
15943 treats a blank password as zero length. MSIE treats it as no
15944 password on export: but it will try both on import. We now do
15945 the same: PKCS12_parse() tries zero length and no password if
15946 the password is set to "" or NULL (NULL is now a valid password:
15947 it wasn't before) as does the pkcs12 application.
15951 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15952 perror when PEM_read_bio_X509_REQ fails, the error message must
15953 be obtained from the error queue.
15957 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15958 it in ERR_remove_state if appropriate, and change ERR_get_state
15959 accordingly to avoid race conditions (this is necessary because
15960 thread_hash is no longer constant once set).
15964 * Bugfix for linux-elf makefile.one.
15968 * RSA_get_default_method() will now cause a default
15969 RSA_METHOD to be chosen if one doesn't exist already.
15970 Previously this was only set during a call to RSA_new()
15971 or RSA_new_method(NULL) meaning it was possible for
15972 RSA_get_default_method() to return NULL.
15976 * Added native name translation to the existing DSO code
15977 that will convert (if the flag to do so is set) filenames
15978 that are sufficiently small and have no path information
15979 into a canonical native form. Eg. "blah" converted to
15980 "libblah.so" or "blah.dll" etc.
15984 * New function ERR_error_string_n(e, buf, len) which is like
15985 ERR_error_string(e, buf), but writes at most 'len' bytes
15986 including the 0 terminator. For ERR_error_string_n, 'buf'
15989 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15991 * CONF library reworked to become more general. A new CONF
15992 configuration file reader "class" is implemented as well as a
15993 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15994 old `CONF_*` functions are still there, but are reimplemented to
15995 work in terms of the new functions. Also, a set of functions
15996 to handle the internal storage of the configuration data is
15997 provided to make it easier to write new configuration file
15998 reader "classes" (I can definitely see something reading a
15999 configuration file in XML format, for example), called `_CONF_*`,
16000 or "the configuration storage API"...
16002 The new configuration file reading functions are:
16004 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16005 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16007 NCONF_default, NCONF_WIN32
16009 NCONF_dump_fp, NCONF_dump_bio
16011 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16012 NCONF_new creates a new CONF object. This works in the same way
16013 as other interfaces in OpenSSL, like the BIO interface.
16014 `NCONF_dump_*` dump the internal storage of the configuration file,
16015 which is useful for debugging. All other functions take the same
16016 arguments as the old `CONF_*` functions with the exception of the
16017 first that must be a `CONF *` instead of a `LHASH *`.
16019 To make it easier to use the new classes with the old `CONF_*` functions,
16020 the function CONF_set_default_method is provided.
16024 * Add '-tls1' option to 'openssl ciphers', which was already
16025 mentioned in the documentation but had not been implemented.
16026 (This option is not yet really useful because even the additional
16027 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16031 * Initial DSO code added into libcrypto for letting OpenSSL (and
16032 OpenSSL-based applications) load shared libraries and bind to
16033 them in a portable way.
16035 *Geoff Thorpe, with contributions from Richard Levitte*
16037 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16039 * Make sure _lrotl and _lrotr are only used with MSVC.
16041 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16042 (the default implementation of RAND_status).
16044 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16045 to '-clrext' (= clear extensions), as intended and documented.
16046 *Bodo Moeller; inconsistency pointed out by Michael Attili
16047 <attili@amaxo.com>*
16049 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16050 was larger than the MD block size.
16052 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16054 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16055 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16056 using the passed key: if the passed key was a private key the result
16057 of X509_print(), for example, would be to print out all the private key
16062 * des_quad_cksum() byte order bug fix.
16063 *Ulf Möller, using the problem description in krb4-0.9.7, where
16064 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16066 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16069 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16071 * For easily testing in shell scripts whether some command
16072 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16073 returns with exit code 0 iff no command of the given name is available.
16074 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16075 the output goes to stdout and nothing is printed to stderr.
16076 Additional arguments are always ignored.
16078 Since for each cipher there is a command of the same name,
16079 the 'no-cipher' compilation switches can be tested this way.
16081 ('openssl no-XXX' is not able to detect pseudo-commands such
16082 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16086 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16090 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16091 is set; it will be thrown away anyway because each handshake creates
16093 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16094 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16095 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16096 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16100 * New s_client option -ign_eof: EOF at stdin is ignored, and
16101 'Q' and 'R' lose their special meanings (quit/renegotiate).
16102 This is part of what -quiet does; unlike -quiet, -ign_eof
16103 does not suppress any output.
16107 * Add compatibility options to the purpose and trust code. The
16108 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16109 accepts a certificate or CA, this was the previous behaviour,
16110 with all the associated security issues.
16112 X509_TRUST_COMPAT is the old trust behaviour: only and
16113 automatically trust self signed roots in certificate store. A
16114 new trust setting X509_TRUST_DEFAULT is used to specify that
16115 a purpose has no associated trust setting and it should instead
16116 use the value in the default purpose.
16120 * Fix the PKCS#8 DSA private key code so it decodes keys again
16121 and fix a memory leak.
16125 * In util/mkerr.pl (which implements 'make errors'), preserve
16126 reason strings from the previous version of the .c file, as
16127 the default to have only downcase letters (and digits) in
16128 automatically generated reasons codes is not always appropriate.
16132 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16133 using strerror. Previously, ERR_reason_error_string() returned
16134 library names as reason strings for SYSerr; but SYSerr is a special
16135 case where small numbers are errno values, not library numbers.
16139 * Add '-dsaparam' option to 'openssl dhparam' application. This
16140 converts DSA parameters into DH parameters. (When creating parameters,
16141 DSA_generate_parameters is used.)
16145 * Include 'length' (recommended exponent length) in C code generated
16146 by 'openssl dhparam -C'.
16150 * The second argument to set_label in perlasm was already being used
16151 so couldn't be used as a "file scope" flag. Moved to third argument
16156 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16157 instead of RAND_bytes for encryption IVs and salts.
16161 * Include RAND_status() into RAND_METHOD instead of implementing
16162 it only for md_rand.c Otherwise replacing the PRNG by calling
16163 RAND_set_rand_method would be impossible.
16167 * Don't let DSA_generate_key() enter an infinite loop if the random
16168 number generation fails.
16172 * New 'rand' application for creating pseudo-random output.
16176 * Added configuration support for Linux/IA64
16178 *Rolf Haberrecker <rolf@suse.de>*
16180 * Assembler module support for Mingw32.
16184 * Shared library support for HPUX (in shlib/).
16186 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16188 * Shared library support for Solaris gcc.
16190 *Lutz Behnke <behnke@trustcenter.de>*
16192 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16194 * PKCS7_encrypt() was adding text MIME headers twice because they
16195 were added manually and by SMIME_crlf_copy().
16199 * In bntest.c don't call BN_rand with zero bits argument.
16201 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16203 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16204 case was implemented. This caused BN_div_recp() to fail occasionally.
16208 * Add an optional second argument to the set_label() in the perl
16209 assembly language builder. If this argument exists and is set
16210 to 1 it signals that the assembler should use a symbol whose
16211 scope is the entire file, not just the current function. This
16212 is needed with MASM which uses the format label:: for this scope.
16214 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16216 * Change the ASN1 types so they are typedefs by default. Before
16217 almost all types were #define'd to ASN1_STRING which was causing
16218 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16223 * Change names of new functions to the new get1/get0 naming
16224 convention: After 'get1', the caller owns a reference count
16225 and has to call `..._free`; 'get0' returns a pointer to some
16226 data structure without incrementing reference counters.
16227 (Some of the existing 'get' functions increment a reference
16228 counter, some don't.)
16229 Similarly, 'set1' and 'add1' functions increase reference
16230 counters or duplicate objects.
16234 * Allow for the possibility of temp RSA key generation failure:
16235 the code used to assume it always worked and crashed on failure.
16239 * Fix potential buffer overrun problem in BIO_printf().
16240 *Ulf Möller, using public domain code by Patrick Powell; problem
16241 pointed out by David Sacerdote <das33@cornell.edu>*
16243 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16244 RAND_egd() and RAND_status(). In the command line application,
16245 the EGD socket can be specified like a seed file using RANDFILE
16250 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16251 Some CAs (e.g. Verisign) distribute certificates in this form.
16255 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16256 list to exclude them. This means that no special compilation option
16257 is needed to use anonymous DH: it just needs to be included in the
16262 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16263 EVP_MD_type. The old functionality is available in a new macro called
16264 EVP_MD_md(). Change code that uses it and update docs.
16268 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16269 where the `void *` argument is replaced by a function pointer argument.
16270 Previously `void *` was abused to point to functions, which works on
16271 many platforms, but is not correct. As these functions are usually
16272 called by macros defined in OpenSSL header files, most source code
16273 should work without changes.
16277 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16278 sections with information on -D... compiler switches used for
16279 compiling the library so that applications can see them. To enable
16280 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16281 must be defined. E.g.,
16282 #define OPENSSL_ALGORITHM_DEFINES
16283 #include <openssl/opensslconf.h>
16284 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16286 *Richard Levitte, Ulf and Bodo Möller*
16288 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16293 * Change the 'other' type in certificate aux info to a STACK_OF
16294 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16295 the required ASN1 format: arbitrary types determined by an OID.
16299 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16300 argument to 'req'. This is not because the function is newer or
16301 better than others it just uses the work 'NEW' in the certificate
16302 request header lines. Some software needs this.
16306 * Reorganise password command line arguments: now passwords can be
16307 obtained from various sources. Delete the PEM_cb function and make
16308 it the default behaviour: i.e. if the callback is NULL and the
16309 usrdata argument is not NULL interpret it as a null terminated pass
16310 phrase. If usrdata and the callback are NULL then the pass phrase
16311 is prompted for as usual.
16315 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16316 the support is automatically enabled. The resulting binaries will
16317 autodetect the card and use it if present.
16319 *Ben Laurie and Compaq Inc.*
16321 * Work around for Netscape hang bug. This sends certificate request
16322 and server done in one record. Since this is perfectly legal in the
16323 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16324 the bugs/SSLv3 entry for more info.
16328 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16332 * Add -rand argument to smime and pkcs12 applications and read/write
16337 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16341 * Add command line password options to the remaining applications.
16345 * Bug fix for BN_div_recp() for numerators with an even number of
16350 * More tests in bntest.c, and changed test_bn output.
16354 * ./config recognizes MacOS X now.
16358 * Bug fix for BN_div() when the first words of num and divisor are
16359 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16363 * Add support for various broken PKCS#8 formats, and command line
16364 options to produce them.
16368 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16369 get temporary BIGNUMs from a BN_CTX.
16373 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16378 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16379 include a #define from the old name to the new. The original intent
16380 was that statically linked binaries could for example just call
16381 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16382 link with digests. This never worked because SSLeay_add_all_digests()
16383 and SSLeay_add_all_ciphers() were in the same source file so calling
16384 one would link with the other. They are now in separate source files.
16388 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16392 * Use a less unusual form of the Miller-Rabin primality test (it used
16393 a binary algorithm for exponentiation integrated into the Miller-Rabin
16394 loop, our standard modexp algorithms are faster).
16398 * Support for the EBCDIC character set completed.
16400 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16402 * Source code cleanups: use const where appropriate, eliminate casts,
16403 use `void *` instead of `char *` in lhash.
16407 * Bugfix: ssl3_send_server_key_exchange was not restartable
16408 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16409 this the server could overwrite ephemeral keys that the client
16414 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16415 using 50 iterations of the Rabin-Miller test.
16417 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16418 iterations of the Rabin-Miller test as required by the appendix
16419 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16420 As BN_is_prime_fasttest includes trial division, DSA parameter
16421 generation becomes much faster.
16423 This implies a change for the callback functions in DSA_is_prime
16424 and DSA_generate_parameters: The callback function is called once
16425 for each positive witness in the Rabin-Miller test, not just
16426 occasionally in the inner loop; and the parameters to the
16427 callback function now provide an iteration count for the outer
16428 loop rather than for the current invocation of the inner loop.
16429 DSA_generate_parameters additionally can call the callback
16430 function with an 'iteration count' of -1, meaning that a
16431 candidate has passed the trial division test (when q is generated
16432 from an application-provided seed, trial division is skipped).
16436 * New function BN_is_prime_fasttest that optionally does trial
16437 division before starting the Rabin-Miller test and has
16438 an additional BN_CTX * argument (whereas BN_is_prime always
16439 has to allocate at least one BN_CTX).
16440 'callback(1, -1, cb_arg)' is called when a number has passed the
16441 trial division stage.
16445 * Fix for bug in CRL encoding. The validity dates weren't being handled
16450 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16454 * New function BN_pseudo_rand().
16458 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16459 bignum version of BN_from_montgomery() with the working code from
16460 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16465 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16466 made it impossible to use the same SSL_SESSION data structure in
16467 SSL2 clients in multiple threads.
16471 * The return value of RAND_load_file() no longer counts bytes obtained
16472 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16473 to seed the PRNG (previously an explicit byte count was required).
16475 *Ulf Möller, Bodo Möller*
16477 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16478 used `char *` instead of `void *` and had casts all over the place.
16482 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16486 * Retain source code compatibility for BN_prime_checks macro:
16487 BN_is_prime(..., BN_prime_checks, ...) now uses
16488 BN_prime_checks_for_size to determine the appropriate number of
16489 Rabin-Miller iterations.
16493 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16494 DH_CHECK_P_NOT_SAFE_PRIME.
16495 (Check if this is true? OpenPGP calls them "strong".)
16499 * Merge the functionality of "dh" and "gendh" programs into a new program
16500 "dhparam". The old programs are retained for now but will handle DH keys
16501 (instead of parameters) in future.
16505 * Make the ciphers, s_server and s_client programs check the return values
16506 when a new cipher list is set.
16510 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16511 ciphers. Before when the 56bit ciphers were enabled the sorting was
16514 The syntax for the cipher sorting has been extended to support sorting by
16515 cipher-strength (using the strength_bits hard coded in the tables).
16516 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16518 Fix a bug in the cipher-command parser: when supplying a cipher command
16519 string with an "undefined" symbol (neither command nor alphanumeric
16520 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16521 an error is flagged.
16523 Due to the strength-sorting extension, the code of the
16524 ssl_create_cipher_list() function was completely rearranged. I hope that
16525 the readability was also increased :-)
16527 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16529 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16530 for the first serial number and places 2 in the serial number file. This
16531 avoids problems when the root CA is created with serial number zero and
16532 the first user certificate has the same issuer name and serial number
16537 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16538 the new code. Add documentation for this stuff.
16542 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16543 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16544 structures and behave in an analogous way to the X509v3 functions:
16545 they shouldn't be called directly but wrapper functions should be used
16548 So we also now have some wrapper functions that call the X509at functions
16549 when passed certificate requests. (TO DO: similar things can be done with
16550 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16551 things. Some of these need some d2i or i2d and print functionality
16552 because they handle more complex structures.)
16556 * Add missing #ifndefs that caused missing symbols when building libssl
16557 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16558 NO_RSA in `ssl/s2*.c`.
16560 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16562 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16563 has a return value which indicates the quality of the random data
16564 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16565 error queue. New function RAND_pseudo_bytes() generates output that is
16566 guaranteed to be unique but not unpredictable. RAND_add is like
16567 RAND_seed, but takes an extra argument for an entropy estimate
16568 (RAND_seed always assumes full entropy).
16572 * Do more iterations of Rabin-Miller probable prime test (specifically,
16573 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16574 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16575 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16576 false-positive rate of at most 2^-80 for random input.
16580 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16584 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16585 in the 0.9.5 release), this returns the chain
16586 from an X509_CTX structure with a dup of the stack and all
16587 the X509 reference counts upped: so the stack will exist
16588 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16591 Also make SSL_SESSION_print() print out the verify return
16596 * Add manpage for the pkcs12 command. Also change the default
16597 behaviour so MAC iteration counts are used unless the new
16598 -nomaciter option is used. This improves file security and
16599 only older versions of MSIE (4.0 for example) need it.
16603 * Honor the no-xxx Configure options when creating .DEF files.
16607 * Add PKCS#10 attributes to field table: challengePassword,
16608 unstructuredName and unstructuredAddress. These are taken from
16609 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16610 international characters are used.
16612 More changes to X509_ATTRIBUTE code: allow the setting of types
16613 based on strings. Remove the 'loc' parameter when adding
16614 attributes because these will be a SET OF encoding which is sorted
16619 * Initial changes to the 'req' utility to allow request generation
16620 automation. This will allow an application to just generate a template
16621 file containing all the field values and have req construct the
16624 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16625 used all over the place including certificate requests and PKCS#7
16626 structures. They are currently handled manually where necessary with
16627 some primitive wrappers for PKCS#7. The new functions behave in a
16628 manner analogous to the X509 extension functions: they allow
16629 attributes to be looked up by NID and added.
16631 Later something similar to the X509V3 code would be desirable to
16632 automatically handle the encoding, decoding and printing of the
16633 more complex types. The string types like challengePassword can
16634 be handled by the string table functions.
16636 Also modified the multi byte string table handling. Now there is
16637 a 'global mask' which masks out certain types. The table itself
16638 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16639 is useful when for example there is only one permissible type
16640 (as in countryName) and using the mask might result in no valid
16645 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16646 SSL_get_peer_finished to allow applications to obtain the latest
16647 Finished messages sent to the peer or expected from the peer,
16648 respectively. (SSL_get_peer_finished is usually the Finished message
16649 actually received from the peer, otherwise the protocol will be aborted.)
16651 As the Finished message are message digests of the complete handshake
16652 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16653 be used for external authentication procedures when the authentication
16654 provided by SSL/TLS is not desired or is not enough.
16658 * Enhanced support for Alpha Linux is added. Now ./config checks if
16659 the host supports BWX extension and if Compaq C is present on the
16660 $PATH. Just exploiting of the BWX extension results in 20-30%
16661 performance kick for some algorithms, e.g. DES and RC4 to mention
16662 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16667 * Add support for MS "fast SGC". This is arguably a violation of the
16668 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16669 weak crypto and after checking the certificate is SGC a second one
16670 with strong crypto. MS SGC stops the first handshake after receiving
16671 the server certificate message and sends a second client hello. Since
16672 a server will typically do all the time consuming operations before
16673 expecting any further messages from the client (server key exchange
16674 is the most expensive) there is little difference between the two.
16676 To get OpenSSL to support MS SGC we have to permit a second client
16677 hello message after we have sent server done. In addition we have to
16678 reset the MAC if we do get this second client hello.
16682 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16683 if a DER encoded private key is RSA or DSA traditional format. Changed
16684 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16685 format DER encoded private key. Newer code should use PKCS#8 format which
16686 has the key type encoded in the ASN1 structure. Added DER private key
16687 support to pkcs8 application.
16691 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16692 ciphersuites has been selected (as required by the SSL 3/TLS 1
16693 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16694 is set, we interpret this as a request to violate the specification
16695 (the worst that can happen is a handshake failure, and 'correct'
16696 behaviour would result in a handshake failure anyway).
16700 * In SSL_CTX_add_session, take into account that there might be multiple
16701 SSL_SESSION structures with the same session ID (e.g. when two threads
16702 concurrently obtain them from an external cache).
16703 The internal cache can handle only one SSL_SESSION with a given ID,
16704 so if there's a conflict, we now throw out the old one to achieve
16709 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16710 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16711 some routines that use cipher OIDs: some ciphers do not have OIDs
16712 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16717 * Simplify the trust setting structure and code. Now we just have
16718 two sequences of OIDs for trusted and rejected settings. These will
16719 typically have values the same as the extended key usage extension
16720 and any application specific purposes.
16722 The trust checking code now has a default behaviour: it will just
16723 check for an object with the same NID as the passed id. Functions can
16724 be provided to override either the default behaviour or the behaviour
16725 for a given id. SSL client, server and email already have functions
16726 in place for compatibility: they check the NID and also return "trusted"
16727 if the certificate is self signed.
16731 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16732 traditional format into an EVP_PKEY structure.
16736 * Add a password callback function PEM_cb() which either prompts for
16737 a password if usr_data is NULL or otherwise assumes it is a null
16738 terminated password. Allow passwords to be passed on command line
16739 environment or config files in a few more utilities.
16743 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16744 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16745 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16746 Update documentation.
16750 * Support for ASN1 "NULL" type. This could be handled before by using
16751 ASN1_TYPE but there wasn't any function that would try to read a NULL
16752 and produce an error if it couldn't. For compatibility we also have
16753 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16754 don't allocate anything because they don't need to.
16758 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16761 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16763 * Rebuild of the memory allocation routines used by OpenSSL code and
16764 possibly others as well. The purpose is to make an interface that
16765 provide hooks so anyone can build a separate set of allocation and
16766 deallocation routines to be used by OpenSSL, for example memory
16767 pool implementations, or something else, which was previously hard
16768 since Malloc(), Realloc() and Free() were defined as macros having
16769 the values malloc, realloc and free, respectively (except for Win32
16770 compilations). The same is provided for memory debugging code.
16771 OpenSSL already comes with functionality to find memory leaks, but
16772 this gives people a chance to debug other memory problems.
16774 With these changes, a new set of functions and macros have appeared:
16776 CRYPTO_set_mem_debug_functions() [F]
16777 CRYPTO_get_mem_debug_functions() [F]
16778 CRYPTO_dbg_set_options() [F]
16779 CRYPTO_dbg_get_options() [F]
16780 CRYPTO_malloc_debug_init() [M]
16782 The memory debug functions are NULL by default, unless the library
16783 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16784 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16785 gives the standard debugging functions that come with OpenSSL) or
16786 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16787 provided by the library user) must be used. When the standard
16788 debugging functions are used, CRYPTO_dbg_set_options can be used to
16789 request additional information:
16790 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16791 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16793 Also, things like CRYPTO_set_mem_functions will always give the
16794 expected result (the new set of functions is used for allocation
16795 and deallocation) at all times, regardless of platform and compiler
16798 To finish it up, some functions that were never use in any other
16799 way than through macros have a new API and new semantic:
16801 CRYPTO_dbg_malloc()
16802 CRYPTO_dbg_realloc()
16805 All macros of value have retained their old syntax.
16807 *Richard Levitte and Bodo Moeller*
16809 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16810 ordering of SMIMECapabilities wasn't in "strength order" and there
16811 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16816 * Some ASN1 types with illegal zero length encoding (INTEGER,
16817 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16819 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16821 * Merge in my S/MIME library for OpenSSL. This provides a simple
16822 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16823 functionality to handle multipart/signed properly) and a utility
16824 called 'smime' to call all this stuff. This is based on code I
16825 originally wrote for Celo who have kindly allowed it to be
16826 included in OpenSSL.
16830 * Add variants des_set_key_checked and des_set_key_unchecked of
16831 des_set_key (aka des_key_sched). Global variable des_check_key
16832 decides which of these is called by des_set_key; this way
16833 des_check_key behaves as it always did, but applications and
16834 the library itself, which was buggy for des_check_key == 1,
16835 have a cleaner way to pick the version they need.
16839 * New function PKCS12_newpass() which changes the password of a
16844 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16845 dynamic mix. In both cases the ids can be used as an index into the
16846 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16847 functions so they accept a list of the field values and the
16848 application doesn't need to directly manipulate the X509_TRUST
16853 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16858 * Modify the way the V3 extension code looks up extensions. This now
16859 works in a similar way to the object code: we have some "standard"
16860 extensions in a static table which is searched with OBJ_bsearch()
16861 and the application can add dynamic ones if needed. The file
16862 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16863 updated whenever a new extension is added to the core code and kept
16864 in ext_nid order. There is a simple program 'tabtest.c' which checks
16865 this. New extensions are not added too often so this file can readily
16866 be maintained manually.
16868 There are two big advantages in doing things this way. The extensions
16869 can be looked up immediately and no longer need to be "added" using
16870 X509V3_add_standard_extensions(): this function now does nothing.
16871 Side note: I get *lots* of email saying the extension code doesn't
16872 work because people forget to call this function.
16873 Also no dynamic allocation is done unless new extensions are added:
16874 so if we don't add custom extensions there is no need to call
16875 X509V3_EXT_cleanup().
16879 * Modify enc utility's salting as follows: make salting the default. Add a
16880 magic header, so unsalted files fail gracefully instead of just decrypting
16881 to garbage. This is because not salting is a big security hole, so people
16882 should be discouraged from doing it.
16886 * Fixes and enhancements to the 'x509' utility. It allowed a message
16887 digest to be passed on the command line but it only used this
16888 parameter when signing a certificate. Modified so all relevant
16889 operations are affected by the digest parameter including the
16890 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16891 DSA key was used because it didn't fix the digest.
16895 * Initial certificate chain verify code. Currently tests the untrusted
16896 certificates for consistency with the verify purpose (which is set
16897 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16899 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16900 this is because it will reject chains with invalid extensions whereas
16901 every previous version of OpenSSL and SSLeay made no checks at all.
16903 Trust code: checks the root CA for the relevant trust settings. Trust
16904 settings have an initial value consistent with the verify purpose: e.g.
16905 if the verify purpose is for SSL client use it expects the CA to be
16906 trusted for SSL client use. However the default value can be changed to
16907 permit custom trust settings: one example of this would be to only trust
16908 certificates from a specific "secure" set of CAs.
16910 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16911 which should be used for version portability: especially since the
16912 verify structure is likely to change more often now.
16914 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16915 to set them. If not set then assume SSL clients will verify SSL servers
16918 Two new options to the verify program: -untrusted allows a set of
16919 untrusted certificates to be passed in and -purpose which sets the
16920 intended purpose of the certificate. If a purpose is set then the
16921 new chain verify code is used to check extension consistency.
16925 * Support for the authority information access extension.
16929 * Modify RSA and DSA PEM read routines to transparently handle
16930 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16931 public keys in a format compatible with certificate
16932 SubjectPublicKeyInfo structures. Unfortunately there were already
16933 functions called *_PublicKey_* which used various odd formats so
16934 these are retained for compatibility: however the DSA variants were
16935 never in a public release so they have been deleted. Changed dsa/rsa
16936 utilities to handle the new format: note no releases ever handled public
16937 keys so we should be OK.
16939 The primary motivation for this change is to avoid the same fiasco
16940 that dogs private keys: there are several incompatible private key
16941 formats some of which are standard and some OpenSSL specific and
16942 require various evil hacks to allow partial transparent handling and
16943 even then it doesn't work with DER formats. Given the option anything
16944 other than PKCS#8 should be dumped: but the other formats have to
16945 stay in the name of compatibility.
16947 With public keys and the benefit of hindsight one standard format
16948 is used which works with EVP_PKEY, RSA or DSA structures: though
16949 it clearly returns an error if you try to read the wrong kind of key.
16951 Added a -pubkey option to the 'x509' utility to output the public key.
16952 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16953 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16954 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16955 that do the same as the `EVP_PKEY_assign_*()` except they up the
16956 reference count of the added key (they don't "swallow" the
16961 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16962 CRLs would fail if the file contained no certificates or no CRLs:
16963 added a new function to read in both types and return the number
16964 read: this means that if none are read it will be an error. The
16965 DER versions of the certificate and CRL reader would always fail
16966 because it isn't possible to mix certificates and CRLs in DER format
16967 without choking one or the other routine. Changed this to just read
16968 a certificate: this is the best we can do. Also modified the code
16969 in `apps/verify.c` to take notice of return codes: it was previously
16970 attempting to read in certificates from NULL pointers and ignoring
16971 any errors: this is one reason why the cert and CRL reader seemed
16972 to work. It doesn't check return codes from the default certificate
16973 routines: these may well fail if the certificates aren't installed.
16977 * Code to support otherName option in GeneralName.
16981 * First update to verify code. Change the verify utility
16982 so it warns if it is passed a self signed certificate:
16983 for consistency with the normal behaviour. X509_verify
16984 has been modified to it will now verify a self signed
16985 certificate if *exactly* the same certificate appears
16986 in the store: it was previously impossible to trust a
16987 single self signed certificate. This means that:
16988 openssl verify ss.pem
16989 now gives a warning about a self signed certificate but
16990 openssl verify -CAfile ss.pem ss.pem
16995 * For servers, store verify_result in SSL_SESSION data structure
16996 (and add it to external session representation).
16997 This is needed when client certificate verifications fails,
16998 but an application-provided verification callback (set by
16999 SSL_CTX_set_cert_verify_callback) allows accepting the session
17000 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17001 but returns 1): When the session is reused, we have to set
17002 ssl->verify_result to the appropriate error code to avoid
17005 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17007 * Fix a bug in the new PKCS#7 code: it didn't consider the
17008 case in PKCS7_dataInit() where the signed PKCS7 structure
17009 didn't contain any existing data because it was being created.
17011 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17013 * Add a salt to the key derivation routines in enc.c. This
17014 forms the first 8 bytes of the encrypted file. Also add a
17015 -S option to allow a salt to be input on the command line.
17019 * New function X509_cmp(). Oddly enough there wasn't a function
17020 to compare two certificates. We do this by working out the SHA1
17021 hash and comparing that. X509_cmp() will be needed by the trust
17026 * SSL_get1_session() is like SSL_get_session(), but increments
17027 the reference count in the SSL_SESSION returned.
17029 *Geoff Thorpe <geoff@eu.c2.net>*
17031 * Fix for 'req': it was adding a null to request attributes.
17032 Also change the X509_LOOKUP and X509_INFO code to handle
17033 certificate auxiliary information.
17037 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17042 * Add the possibility to add extra information to the memory leak
17043 detecting output, to form tracebacks, showing from where each
17044 allocation was originated: CRYPTO_push_info("constant string") adds
17045 the string plus current file name and line number to a per-thread
17046 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17047 is like calling CYRPTO_pop_info() until the stack is empty.
17048 Also updated memory leak detection code to be multi-thread-safe.
17052 * Add options -text and -noout to pkcs7 utility and delete the
17053 encryption options which never did anything. Update docs.
17057 * Add options to some of the utilities to allow the pass phrase
17058 to be included on either the command line (not recommended on
17059 OSes like Unix) or read from the environment. Update the
17060 manpages and fix a few bugs.
17064 * Add a few manpages for some of the openssl commands.
17068 * Fix the -revoke option in ca. It was freeing up memory twice,
17069 leaking and not finding already revoked certificates.
17073 * Extensive changes to support certificate auxiliary information.
17074 This involves the use of X509_CERT_AUX structure and X509_AUX
17075 functions. An X509_AUX function such as PEM_read_X509_AUX()
17076 can still read in a certificate file in the usual way but it
17077 will also read in any additional "auxiliary information". By
17078 doing things this way a fair degree of compatibility can be
17079 retained: existing certificates can have this information added
17080 using the new 'x509' options.
17082 Current auxiliary information includes an "alias" and some trust
17083 settings. The trust settings will ultimately be used in enhanced
17084 certificate chain verification routines: currently a certificate
17085 can only be trusted if it is self signed and then it is trusted
17090 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17091 The problem was that one of the replacement routines had not been working
17092 since SSLeay releases. For now the offending routine has been replaced
17093 with non-optimised assembler. Even so, this now gives around 95%
17094 performance improvement for 1024 bit RSA signs.
17098 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17099 handling. Most clients have the effective key size in bits equal to
17100 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17101 A few however don't do this and instead use the size of the decrypted key
17102 to determine the RC2 key length and the AlgorithmIdentifier to determine
17103 the effective key length. In this case the effective key length can still
17104 be 40 bits but the key length can be 168 bits for example. This is fixed
17105 by manually forcing an RC2 key into the EVP_PKEY structure because the
17106 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17107 the key length and effective key length are equal.
17111 * Add a bunch of functions that should simplify the creation of
17112 X509_NAME structures. Now you should be able to do:
17113 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17114 and have it automatically work out the correct field type and fill in
17115 the structures. The more adventurous can try:
17116 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17117 and it will (hopefully) work out the correct multibyte encoding.
17121 * Change the 'req' utility to use the new field handling and multibyte
17122 copy routines. Before the DN field creation was handled in an ad hoc
17123 way in req, ca, and x509 which was rather broken and didn't support
17124 BMPStrings or UTF8Strings. Since some software doesn't implement
17125 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17126 using the dirstring_type option. See the new comment in the default
17127 openssl.cnf for more info.
17131 * Make crypto/rand/md_rand.c more robust:
17132 - Assure unique random numbers after fork().
17133 - Make sure that concurrent threads access the global counter and
17134 md serializably so that we never lose entropy in them
17135 or use exactly the same state in multiple threads.
17136 Access to the large state is not always serializable because
17137 the additional locking could be a performance killer, and
17138 md should be large enough anyway.
17142 * New file `apps/app_rand.c` with commonly needed functionality
17143 for handling the random seed file.
17145 Use the random seed file in some applications that previously did not:
17147 dsaparam -genkey (which also ignored its '-rand' option),
17150 x509 (when signing).
17151 Except on systems with /dev/urandom, it is crucial to have a random
17152 seed file at least for key creation, DSA signing, and for DH exchanges;
17153 for RSA signatures we could do without one.
17155 gendh and gendsa (unlike genrsa) used to read only the first byte
17156 of each file listed in the '-rand' option. The function as previously
17157 found in genrsa is now in app_rand.c and is used by all programs
17158 that support '-rand'.
17162 * In RAND_write_file, use mode 0600 for creating files;
17163 don't just chmod when it may be too late.
17167 * Report an error from X509_STORE_load_locations
17168 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17172 * New function ASN1_mbstring_copy() this copies a string in either
17173 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17174 into an ASN1_STRING type. A mask of permissible types is passed
17175 and it chooses the "minimal" type to use or an error if not type
17180 * Add function equivalents to the various macros in asn1.h. The old
17181 macros are retained with an `M_` prefix. Code inside the library can
17182 use the `M_` macros. External code (including the openssl utility)
17183 should *NOT* in order to be "shared library friendly".
17187 * Add various functions that can check a certificate's extensions
17188 to see if it usable for various purposes such as SSL client,
17189 server or S/MIME and CAs of these types. This is currently
17190 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17191 verification. Also added a -purpose flag to x509 utility to
17192 print out all the purposes.
17196 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17201 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17202 for, obtain and decode and extension and obtain its critical flag.
17203 This allows all the necessary extension code to be handled in a
17204 single function call.
17208 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17209 platforms. See crypto/rc4/rc4_enc.c for further details.
17213 * New -noout option to asn1parse. This causes no output to be produced
17214 its main use is when combined with -strparse and -out to extract data
17215 from a file (which may not be in ASN.1 format).
17219 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17220 when producing the local key id.
17222 *Richard Levitte <levitte@stacken.kth.se>*
17224 * New option -dhparam in s_server. This allows a DH parameter file to be
17225 stated explicitly. If it is not stated then it tries the first server
17226 certificate file. The previous behaviour hard coded the filename
17231 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17232 a public key to be input or output. For example:
17233 openssl rsa -in key.pem -pubout -out pubkey.pem
17234 Also added necessary DSA public key functions to handle this.
17238 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17239 in the message. This was handled by allowing
17240 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17242 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17244 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17245 to the end of the strings whereas this didn't. This would cause problems
17246 if strings read with d2i_ASN1_bytes() were later modified.
17248 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17250 * Fix for base64 decode bug. When a base64 bio reads only one line of
17251 data and it contains EOF it will end up returning an error. This is
17252 caused by input 46 bytes long. The cause is due to the way base64
17253 BIOs find the start of base64 encoded data. They do this by trying a
17254 trial decode on each line until they find one that works. When they
17255 do a flag is set and it starts again knowing it can pass all the
17256 data directly through the decoder. Unfortunately it doesn't reset
17257 the context it uses. This means that if EOF is reached an attempt
17258 is made to pass two EOFs through the context and this causes the
17259 resulting error. This can also cause other problems as well. As is
17260 usual with these problems it takes *ages* to find and the fix is
17261 trivial: move one line.
17263 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17265 * Ugly workaround to get s_client and s_server working under Windows. The
17266 old code wouldn't work because it needed to select() on sockets and the
17267 tty (for keypresses and to see if data could be written). Win32 only
17268 supports select() on sockets so we select() with a 1s timeout on the
17269 sockets and then see if any characters are waiting to be read, if none
17270 are present then we retry, we also assume we can always write data to
17271 the tty. This isn't nice because the code then blocks until we've
17272 received a complete line of data and it is effectively polling the
17273 keyboard at 1s intervals: however it's quite a bit better than not
17274 working at all :-) A dedicated Windows application might handle this
17275 with an event loop for example.
17279 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17280 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17281 will be called when RSA_sign() and RSA_verify() are used. This is useful
17282 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17283 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17284 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17285 This necessitated the support of an extra signature type NID_md5_sha1
17286 for SSL signatures and modifications to the SSL library to use it instead
17287 of calling RSA_public_decrypt() and RSA_private_encrypt().
17291 * Add new -verify -CAfile and -CApath options to the crl program, these
17292 will lookup a CRL issuers certificate and verify the signature in a
17293 similar way to the verify program. Tidy up the crl program so it
17294 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17295 less strict. It will now permit CRL extensions even if it is not
17296 a V2 CRL: this will allow it to tolerate some broken CRLs.
17300 * Initialize all non-automatic variables each time one of the openssl
17301 sub-programs is started (this is necessary as they may be started
17302 multiple times from the "OpenSSL>" prompt).
17304 *Lennart Bang, Bodo Moeller*
17306 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17307 removing all other RSA functionality (this is what NO_RSA does). This
17308 is so (for example) those in the US can disable those operations covered
17309 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17314 * Non-copying interface to BIO pairs.
17315 (still largely untested)
17319 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17320 ASCII string. This was handled independently in various places before.
17324 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17325 UTF8 strings a character at a time.
17329 * Use client_version from client hello to select the protocol
17330 (s23_srvr.c) and for RSA client key exchange verification
17331 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17335 * Add various utility functions to handle SPKACs, these were previously
17336 handled by poking round in the structure internals. Added new function
17337 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17338 print, verify and generate SPKACs. Based on an original idea from
17339 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17343 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17347 * Allow the config file extension section to be overwritten on the
17348 command line. Based on an original idea from Massimiliano Pala
17349 <madwolf@comune.modena.it>. The new option is called -extensions
17350 and can be applied to ca, req and x509. Also -reqexts to override
17351 the request extensions in req and -crlexts to override the crl extensions
17356 * Add new feature to the SPKAC handling in ca. Now you can include
17357 the same field multiple times by preceding it by "XXXX." for example:
17360 this is the same syntax as used in the req config file.
17364 * Allow certificate extensions to be added to certificate requests. These
17365 are specified in a 'req_extensions' option of the req section of the
17366 config file. They can be printed out with the -text option to req but
17367 are otherwise ignored at present.
17371 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17372 data read consists of only the final block it would not decrypted because
17373 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17374 A misplaced 'break' also meant the decrypted final block might not be
17375 copied until the next read.
17379 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17380 a few extra parameters to the DH structure: these will be useful if
17381 for example we want the value of 'q' or implement X9.42 DH.
17385 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17386 provides hooks that allow the default DSA functions or functions on a
17387 "per key" basis to be replaced. This allows hardware acceleration and
17388 hardware key storage to be handled without major modification to the
17389 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17390 associated functions.
17394 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17395 as "read only": it can't be written to and the buffer it points to will
17396 not be freed. Reading from a read only BIO is much more efficient than
17397 a normal memory BIO. This was added because there are several times when
17398 an area of memory needs to be read from a BIO. The previous method was
17399 to create a memory BIO and write the data to it, this results in two
17400 copies of the data and an O(n^2) reading algorithm. There is a new
17401 function BIO_new_mem_buf() which creates a read only memory BIO from
17402 an area of memory. Also modified the PKCS#7 routines to use read only
17407 * Bugfix: ssl23_get_client_hello did not work properly when called in
17408 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17409 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17410 but a retry condition occurred while trying to read the rest.
17414 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17415 NID_pkcs7_encrypted by default: this was wrong since this should almost
17416 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17417 the encrypted data type: this is a more sensible place to put it and it
17418 allows the PKCS#12 code to be tidied up that duplicated this
17423 * Changed obj_dat.pl script so it takes its input and output files on
17424 the command line. This should avoid shell escape redirection problems
17429 * Initial support for certificate extension requests, these are included
17430 in things like Xenroll certificate requests. Included functions to allow
17431 extensions to be obtained and added.
17435 * -crlf option to s_client and s_server for sending newlines as
17436 CRLF (as required by many protocols).
17440 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17442 * Install libRSAglue.a when OpenSSL is built with RSAref.
17444 *Ralf S. Engelschall*
17446 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17448 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17450 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17455 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17456 DH parameters/keys (q is lost during that conversion, but the resulting
17457 DH parameters contain its length).
17459 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17460 much faster than DH_generate_parameters (which creates parameters
17461 where `p = 2*q + 1`), and also the smaller q makes DH computations
17462 much more efficient (160-bit exponentiation instead of 1024-bit
17463 exponentiation); so this provides a convenient way to support DHE
17464 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17465 utter importance to use
17466 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17468 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17469 when such DH parameters are used, because otherwise small subgroup
17470 attacks may become possible!
17474 * Avoid memory leak in i2d_DHparams.
17478 * Allow the -k option to be used more than once in the enc program:
17479 this allows the same encrypted message to be read by multiple recipients.
17483 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17484 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17485 it will always use the numerical form of the OID, even if it has a short
17490 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17491 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17492 otherwise bn_mod_exp was called. In the case of hardware keys for example
17493 no private key components need be present and it might store extra data
17494 in the RSA structure, which cannot be accessed from bn_mod_exp.
17495 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17496 private key operations.
17500 * Added support for SPARC Linux.
17504 * pem_password_cb function type incompatibly changed from
17505 typedef int pem_password_cb(char *buf, int size, int rwflag);
17507 ....(char *buf, int size, int rwflag, void *userdata);
17508 so that applications can pass data to their callbacks:
17509 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17510 additional void * argument, which is just handed through whenever
17511 the password callback is called.
17513 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17515 New function SSL_CTX_set_default_passwd_cb_userdata.
17517 Compatibility note: As many C implementations push function arguments
17518 onto the stack in reverse order, the new library version is likely to
17519 interoperate with programs that have been compiled with the old
17520 pem_password_cb definition (PEM_whatever takes some data that
17521 happens to be on the stack as its last argument, and the callback
17522 just ignores this garbage); but there is no guarantee whatsoever that
17525 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17526 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17527 problems not only on Windows, but also on some Unix platforms.
17528 To avoid problematic command lines, these definitions are now in an
17529 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17530 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17534 * MIPS III/IV assembler module is reimplemented.
17538 * More DES library cleanups: remove references to srand/rand and
17539 delete an unused file.
17543 * Add support for the free Netwide assembler (NASM) under Win32,
17544 since not many people have MASM (ml) and it can be hard to obtain.
17545 This is currently experimental but it seems to work OK and pass all
17546 the tests. Check out INSTALL.W32 for info.
17550 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17551 without temporary keys kept an extra copy of the server key,
17552 and connections with temporary keys did not free everything in case
17557 * New function RSA_check_key and new openssl rsa option -check
17558 for verifying the consistency of RSA keys.
17560 *Ulf Moeller, Bodo Moeller*
17562 * Various changes to make Win32 compile work:
17563 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17564 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17565 comparison" warnings.
17566 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17570 * Add a debugging option to PKCS#5 v2 key generation function: when
17571 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17572 derived keys are printed to stderr.
17576 * Copy the flags in ASN1_STRING_dup().
17578 *Roman E. Pavlov <pre@mo.msk.ru>*
17580 * The x509 application mishandled signing requests containing DSA
17581 keys when the signing key was also DSA and the parameters didn't match.
17583 It was supposed to omit the parameters when they matched the signing key:
17584 the verifying software was then supposed to automatically use the CA's
17585 parameters if they were absent from the end user certificate.
17587 Omitting parameters is no longer recommended. The test was also
17588 the wrong way round! This was probably due to unusual behaviour in
17589 EVP_cmp_parameters() which returns 1 if the parameters match.
17590 This meant that parameters were omitted when they *didn't* match and
17591 the certificate was useless. Certificates signed with 'ca' didn't have
17594 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17596 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17597 The interface is as follows:
17598 Applications can use
17599 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17600 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17601 "off" is now the default.
17602 The library internally uses
17603 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17604 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17605 to disable memory-checking temporarily.
17607 Some inconsistent states that previously were possible (and were
17608 even the default) are now avoided.
17610 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17611 with each memory chunk allocated; this is occasionally more helpful
17612 than just having a counter.
17614 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17616 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17621 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17622 which largely parallels "options", but is for changing API behaviour,
17623 whereas "options" are about protocol behaviour.
17624 Initial "mode" flags are:
17626 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17627 a single record has been written.
17628 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17629 retries use the same buffer location.
17630 (But all of the contents must be
17635 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17638 * Fix problems with no-hmac etc.
17640 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17642 * New functions RSA_get_default_method(), RSA_set_method() and
17643 RSA_get_method(). These allows replacement of RSA_METHODs without having
17644 to mess around with the internals of an RSA structure.
17648 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17649 Also really enable memory leak checks in openssl.c and in some
17652 *Chad C. Mulligan, Bodo Moeller*
17654 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17655 up the length of negative integers. This has now been simplified to just
17656 store the length when it is first determined and use it later, rather
17657 than trying to keep track of where data is copied and updating it to
17659 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17661 * Add a new function PKCS7_signatureVerify. This allows the verification
17662 of a PKCS#7 signature but with the signing certificate passed to the
17663 function itself. This contrasts with PKCS7_dataVerify which assumes the
17664 certificate is present in the PKCS#7 structure. This isn't always the
17665 case: certificates can be omitted from a PKCS#7 structure and be
17666 distributed by "out of band" means (such as a certificate database).
17670 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17671 function prototypes in pem.h, also change util/mkdef.pl to add the
17672 necessary function names.
17676 * mk1mf.pl (used by Windows builds) did not properly read the
17677 options set by Configure in the top level Makefile, and Configure
17678 was not even able to write more than one option correctly.
17679 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17683 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17684 file to be loaded from a BIO or FILE pointer. The BIO version will
17685 for example allow memory BIOs to contain config info.
17689 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17690 Whoever hopes to achieve shared-library compatibility across versions
17691 must use this, not the compile-time macro.
17692 (Exercise 0.9.4: Which is the minimum library version required by
17694 Note: All this applies only to multi-threaded programs, others don't
17699 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17700 through a BIO pair triggered the default case, i.e.
17701 SSLerr(...,SSL_R_UNKNOWN_STATE).
17705 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17706 can use the SSL library even if none of the specific BIOs is
17711 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17712 for the encoded length.
17714 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17716 * Add initial documentation of the X509V3 functions.
17720 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17721 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17722 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17723 secure PKCS#8 private key format with a high iteration count.
17727 * Fix determination of Perl interpreter: A perl or perl5
17728 *directory* in $PATH was also accepted as the interpreter.
17730 *Ralf S. Engelschall*
17732 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17733 wrong with it but it was very old and did things like calling
17734 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17735 unusual formatting.
17739 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17740 to use the new extension code.
17744 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17745 with macros. This should make it easier to change their form, add extra
17746 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17751 * Add to configuration table a new entry that can specify an alternative
17752 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17753 according to Mark Crispin <MRC@Panda.COM>.
17757 * DES CBC did not update the IV. Weird.
17761 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17762 Changing the behaviour of the former might break existing programs --
17763 where IV updating is needed, des_ncbc_encrypt can be used.
17766 * When bntest is run from "make test" it drives bc to check its
17767 calculations, as well as internally checking them. If an internal check
17768 fails, it needs to cause bc to give a non-zero result or make test carries
17769 on without noticing the failure. Fixed.
17773 * DES library cleanups.
17777 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17778 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17779 ciphers. NOTE: although the key derivation function has been verified
17780 against some published test vectors it has not been extensively tested
17781 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17786 * Instead of "mkdir -p", which is not fully portable, use new
17787 Perl script "util/mkdir-p.pl".
17791 * Rewrite the way password based encryption (PBE) is handled. It used to
17792 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17793 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17794 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17795 the 'parameter' field of the AlgorithmIdentifier is passed to the
17796 underlying key generation function so it must do its own ASN1 parsing.
17797 This has also changed the EVP_PBE_CipherInit() function which now has a
17798 'parameter' argument instead of literal salt and iteration count values
17799 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17803 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17804 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17805 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17806 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17807 value was just used as a "magic string" and not used directly its
17808 value doesn't matter.
17812 * Introduce some semblance of const correctness to BN. Shame C doesn't
17817 * "linux-sparc64" configuration (ultrapenguin).
17819 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17820 "linux-sparc" configuration.
17822 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17824 * config now generates no-xxx options for missing ciphers.
17828 * Support the EBCDIC character set (work in progress).
17829 File ebcdic.c not yet included because it has a different license.
17831 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17833 * Support BS2000/OSD-POSIX.
17835 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17837 * Make callbacks for key generation use `void *` instead of `char *`.
17841 * Make S/MIME samples compile (not yet tested).
17845 * Additional typesafe stacks.
17849 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17853 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17855 * New configuration variant "sco5-gcc".
17857 * Updated some demos.
17859 *Sean O Riordain, Wade Scholine*
17861 * Add missing BIO_free at exit of pkcs12 application.
17865 * Fix memory leak in conf.c.
17869 * Updates for Win32 to assembler version of MD5.
17873 * Set #! path to perl in `apps/der_chop` to where we found it
17874 instead of using a fixed path.
17878 * SHA library changes for irix64-mips4-cc.
17882 * Improvements for VMS support.
17886 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17888 * Bignum library bug fix. IRIX 6 passes "make test" now!
17889 This also avoids the problems with SC4.2 and unpatched SC5.
17891 *Andy Polyakov <appro@fy.chalmers.se>*
17893 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17894 These are required because of the typesafe stack would otherwise break
17895 existing code. If old code used a structure member which used to be STACK
17896 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17897 sk_num or sk_value it would produce an error because the num, data members
17898 are not present in STACK_OF. Now it just produces a warning. sk_set
17899 replaces the old method of assigning a value to sk_value
17900 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17901 that does this will no longer work (and should use sk_set instead) but
17902 this could be regarded as a "questionable" behaviour anyway.
17906 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17907 correctly handle encrypted S/MIME data.
17911 * Change type of various DES function arguments from des_cblock
17912 (which means, in function argument declarations, pointer to char)
17913 to des_cblock * (meaning pointer to array with 8 char elements),
17914 which allows the compiler to do more typechecking; it was like
17915 that back in SSLeay, but with lots of ugly casts.
17917 Introduce new type const_des_cblock.
17921 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17922 problems: find RecipientInfo structure that matches recipient certificate
17923 and initialise the ASN1 structures properly based on passed cipher.
17927 * Belatedly make the BN tests actually check the results.
17931 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17932 to and from BNs: it was completely broken. New compilation option
17933 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17934 key elements as negative integers.
17938 * Reorganize and speed up MD5.
17940 *Andy Polyakov <appro@fy.chalmers.se>*
17944 *Richard Levitte <richard@levitte.org>*
17946 * New option -out to asn1parse to allow the parsed structure to be
17947 output to a file. This is most useful when combined with the -strparse
17948 option to examine the output of things like OCTET STRINGS.
17952 * Make SSL library a little more fool-proof by not requiring any longer
17953 that `SSL_set_{accept,connect}_state` be called before
17954 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17955 in many applications because usually everything *appeared* to work as
17956 intended anyway -- now it really works as intended).
17960 * Move openssl.cnf out of lib/.
17964 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17965 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17966 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17968 *Ralf S. Engelschall*
17970 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17971 handle PKCS#7 enveloped data properly.
17973 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17975 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17976 copying pointers. The cert_st handling is changed by this in
17977 various ways (and thus what used to be known as ctx->default_cert
17978 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17979 any longer when s->cert does not give us what we need).
17980 ssl_cert_instantiate becomes obsolete by this change.
17981 As soon as we've got the new code right (possibly it already is?),
17982 we have solved a couple of bugs of the earlier code where s->cert
17983 was used as if it could not have been shared with other SSL structures.
17985 Note that using the SSL API in certain dirty ways now will result
17986 in different behaviour than observed with earlier library versions:
17987 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17988 does not influence s as it used to.
17990 In order to clean up things more thoroughly, inside SSL_SESSION
17991 we don't use CERT any longer, but a new structure SESS_CERT
17992 that holds per-session data (if available); currently, this is
17993 the peer's certificate chain and, for clients, the server's certificate
17994 and temporary key. CERT holds only those values that can have
17995 meaningful defaults in an SSL_CTX.
17999 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18000 from the internal representation. Various PKCS#7 fixes: remove some
18001 evil casts and set the enc_dig_alg field properly based on the signing
18006 * Allow PKCS#12 password to be set from the command line or the
18007 environment. Let 'ca' get its config file name from the environment
18008 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18013 * Allow certificate policies extension to use an IA5STRING for the
18014 organization field. This is contrary to the PKIX definition but
18015 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18020 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18021 without disallowing inline assembler and the like for non-pedantic builds.
18025 * Support Borland C++ builder.
18027 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18033 * SHA-1 cleanups and performance enhancements.
18035 *Andy Polyakov <appro@fy.chalmers.se>*
18037 * Sparc v8plus assembler for the bignum library.
18039 *Andy Polyakov <appro@fy.chalmers.se>*
18041 * Accept any -xxx and +xxx compiler options in Configure.
18045 * Update HPUX configuration.
18049 * Add missing `sk_<type>_unshift()` function to safestack.h
18051 *Ralf S. Engelschall*
18053 * New function SSL_CTX_use_certificate_chain_file that sets the
18054 "extra_cert"s in addition to the certificate. (This makes sense
18055 only for "PEM" format files, as chains as a whole are not
18060 * Support verify_depth from the SSL API.
18061 x509_vfy.c had what can be considered an off-by-one-error:
18062 Its depth (which was not part of the external interface)
18063 was actually counting the number of certificates in a chain;
18064 now it really counts the depth.
18068 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18069 instead of X509err, which often resulted in confusing error
18070 messages since the error codes are not globally unique
18071 (e.g. an alleged error in ssl3_accept when a certificate
18072 didn't match the private key).
18074 * New function SSL_CTX_set_session_id_context that allows to set a default
18075 value (so that you don't need SSL_set_session_id_context for each
18076 connection using the SSL_CTX).
18080 * OAEP decoding bug fix.
18084 * Support INSTALL_PREFIX for package builders, as proposed by
18089 * New Configure options "threads" and "no-threads". For systems
18090 where the proper compiler options are known (currently Solaris
18091 and Linux), "threads" is the default.
18095 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18099 * Install various scripts to $(OPENSSLDIR)/misc, not to
18100 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18101 such as /usr/local/bin.
18105 * "make linux-shared" to build shared libraries.
18107 *Niels Poppe <niels@netbox.org>*
18109 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18113 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18114 extension adding in x509 utility.
18118 * Remove NOPROTO sections and error code comments.
18122 * Partial rewrite of the DEF file generator to now parse the ANSI
18127 * New Configure options --prefix=DIR and --openssldir=DIR.
18131 * Complete rewrite of the error code script(s). It is all now handled
18132 by one script at the top level which handles error code gathering,
18133 header rewriting and C source file generation. It should be much better
18134 than the old method: it now uses a modified version of Ulf's parser to
18135 read the ANSI prototypes in all header files (thus the old K&R definitions
18136 aren't needed for error creation any more) and do a better job of
18137 translating function codes into names. The old 'ASN1 error code embedded
18138 in a comment' is no longer necessary and it doesn't use .err files which
18139 have now been deleted. Also the error code call doesn't have to appear all
18140 on one line (which resulted in some large lines...).
18144 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18148 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18149 0 (which usually indicates a closed connection), but continue reading.
18153 * Fix some race conditions.
18157 * Add support for CRL distribution points extension. Add Certificate
18158 Policies and CRL distribution points documentation.
18162 * Move the autogenerated header file parts to crypto/opensslconf.h.
18166 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18167 8 of keying material. Merlin has also confirmed interop with this fix
18168 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18170 *Merlin Hughes <merlin@baltimore.ie>*
18172 * Fix lots of warnings.
18174 *Richard Levitte <levitte@stacken.kth.se>*
18176 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18177 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18179 *Richard Levitte <levitte@stacken.kth.se>*
18181 * Fix problems with sizeof(long) == 8.
18183 *Andy Polyakov <appro@fy.chalmers.se>*
18185 * Change functions to ANSI C.
18189 * Fix typos in error codes.
18191 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18193 * Remove defunct assembler files from Configure.
18197 * SPARC v8 assembler BIGNUM implementation.
18199 *Andy Polyakov <appro@fy.chalmers.se>*
18201 * Support for Certificate Policies extension: both print and set.
18202 Various additions to support the r2i method this uses.
18206 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18207 return a const string when you are expecting an allocated buffer.
18211 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18212 types DirectoryString and DisplayText.
18216 * Add code to allow r2i extensions to access the configuration database,
18217 add an LHASH database driver and add several ctx helper functions.
18221 * Fix an evil bug in bn_expand2() which caused various BN functions to
18222 fail when they extended the size of a BIGNUM.
18226 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18227 support typesafe stack.
18231 * Fix typo in SSL_[gs]et_options().
18233 *Nils Frostberg <nils@medcom.se>*
18235 * Delete various functions and files that belonged to the (now obsolete)
18236 old X509V3 handling code.
18240 * New Configure option "rsaref".
18244 * Don't auto-generate pem.h.
18248 * Introduce type-safe ASN.1 SETs.
18252 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18254 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18256 * Introduce type-safe STACKs. This will almost certainly break lots of code
18257 that links with OpenSSL (well at least cause lots of warnings), but fear
18258 not: the conversion is trivial, and it eliminates loads of evil casts. A
18259 few STACKed things have been converted already. Feel free to convert more.
18260 In the fullness of time, I'll do away with the STACK type altogether.
18264 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18265 specified in `<certfile>` by updating the entry in the index.txt file.
18266 This way one no longer has to edit the index.txt file manually for
18267 revoking a certificate. The -revoke option does the gory details now.
18269 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18271 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18272 `-text` option at all and this way the `-noout -text` combination was
18273 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18275 *Ralf S. Engelschall*
18277 * Make sure a corresponding plain text error message exists for the
18278 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18279 verify callback function determined that a certificate was revoked.
18281 *Ralf S. Engelschall*
18283 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18284 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18285 all available ciphers including rc5, which was forgotten until now.
18286 In order to let the testing shell script know which algorithms
18287 are available, a new (up to now undocumented) command
18288 `openssl list-cipher-commands` is used.
18292 * Bugfix: s_client occasionally would sleep in select() when
18293 it should have checked SSL_pending() first.
18297 * New functions DSA_do_sign and DSA_do_verify to provide access to
18298 the raw DSA values prior to ASN.1 encoding.
18302 * Tweaks to Configure
18304 *Niels Poppe <niels@netbox.org>*
18306 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18311 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18315 * New config option to avoid instructions that are illegal on the 80386.
18316 The default code is faster, but requires at least a 486.
18320 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18321 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18322 same as SSL2_VERSION anyway.
18326 * New "-showcerts" option for s_client.
18330 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18331 application. Various cleanups and fixes.
18335 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18336 modify error routines to work internally. Add error codes and PBE init
18337 to library startup routines.
18341 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18342 packing functions to asn1 and evp. Changed function names and error
18343 codes along the way.
18347 * PKCS12 integration: and so it begins... First of several patches to
18348 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18349 objects to objects.h
18353 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18354 and display support for Thawte strong extranet extension.
18358 * Add LinuxPPC support.
18360 *Jeff Dubrule <igor@pobox.org>*
18362 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18363 bn_div_words in alpha.s.
18365 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18367 * Make sure the RSA OAEP test is skipped under -DRSAref because
18368 OAEP isn't supported when OpenSSL is built with RSAref.
18370 *Ulf Moeller <ulf@fitug.de>*
18372 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18373 so they no longer are missing under -DNOPROTO.
18375 *Soren S. Jorvang <soren@t.dk>*
18377 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18379 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18380 doesn't work when the session is reused. Coming soon!
18384 * Fix a security hole, that allows sessions to be reused in the wrong
18385 context thus bypassing client cert protection! All software that uses
18386 client certs and session caches in multiple contexts NEEDS PATCHING to
18387 allow session reuse! A fuller solution is in the works.
18389 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18391 * Some more source tree cleanups (removed obsolete files
18392 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18393 permission on "config" script to be executable) and a fix for the INSTALL
18396 *Ulf Moeller <ulf@fitug.de>*
18398 * Remove some legacy and erroneous uses of malloc, free instead of
18401 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18403 * Make rsa_oaep_test return non-zero on error.
18405 *Ulf Moeller <ulf@fitug.de>*
18407 * Add support for native Solaris shared libraries. Configure
18408 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18409 if someone would make that last step automatic.
18411 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18413 * ctx_size was not built with the right compiler during "make links". Fixed.
18417 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18418 except NULL ciphers". This means the default cipher list will no longer
18419 enable NULL ciphers. They need to be specifically enabled e.g. with
18420 the string "DEFAULT:eNULL".
18424 * Fix to RSA private encryption routines: if p < q then it would
18425 occasionally produce an invalid result. This will only happen with
18426 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18430 * Be less restrictive and allow also `perl util/perlpath.pl
18431 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18432 because this way one can also use an interpreter named `perl5` (which is
18433 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18434 installed as `perl`).
18436 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18438 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18440 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18442 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18443 advapi32.lib to Win32 build and change the pem test comparison
18444 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18445 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18446 and crypto/des/ede_cbcm_enc.c.
18450 * DES quad checksum was broken on big-endian architectures. Fixed.
18454 * Comment out two functions in bio.h that aren't implemented. Fix up the
18455 Win32 test batch file so it (might) work again. The Win32 test batch file
18456 is horrible: I feel ill....
18460 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18461 in e_os.h. Audit of header files to check ANSI and non ANSI
18462 sections: 10 functions were absent from non ANSI section and not exported
18463 from Windows DLLs. Fixed up libeay.num for new functions.
18467 * Make `openssl version` output lines consistent.
18469 *Ralf S. Engelschall*
18471 * Fix Win32 symbol export lists for BIO functions: Added
18472 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18473 to ms/libeay{16,32}.def.
18475 *Ralf S. Engelschall*
18477 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18478 fine under Unix and passes some trivial tests I've now added. But the
18479 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18480 added to make sure no one expects that this stuff really works in the
18481 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18482 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18485 *Ralf S. Engelschall*
18487 * Fix the generation of two part addresses in perl.
18489 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18491 * Add config entry for Linux on MIPS.
18493 *John Tobey <jtobey@channel1.com>*
18495 * Make links whenever Configure is run, unless we are on Windoze.
18499 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18500 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18505 * Add a useful kludge to allow package maintainers to specify compiler and
18506 other platforms details on the command line without having to patch the
18507 Configure script every time: One now can use
18508 `perl Configure <id>:<details>`,
18509 i.e. platform ids are allowed to have details appended
18510 to them (separated by colons). This is treated as there would be a static
18511 pre-configured entry in Configure's %table under key `<id>` with value
18512 `<details>` and `perl Configure <id>` is called. So, when you want to
18513 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18514 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18515 now, which overrides the FreeBSD-elf entry on-the-fly.
18517 *Ralf S. Engelschall*
18519 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18523 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18524 on the `perl Configure ...` command line. This way one can compile
18525 OpenSSL libraries with Position Independent Code (PIC) which is needed
18526 for linking it into DSOs.
18528 *Ralf S. Engelschall*
18530 * Remarkably, export ciphers were totally broken and no-one had noticed!
18535 * Cleaned up the LICENSE document: The official contact for any license
18536 questions now is the OpenSSL core team under openssl-core@openssl.org.
18537 And add a paragraph about the dual-license situation to make sure people
18538 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18539 to the OpenSSL toolkit.
18541 *Ralf S. Engelschall*
18543 * General source tree makefile cleanups: Made `making xxx in yyy...`
18544 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18545 Additionally cleaned up the `make links` target: Remove unnecessary
18546 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18547 to speed processing and no longer clutter the display with confusing
18548 stuff. Instead only the actually done links are displayed.
18550 *Ralf S. Engelschall*
18552 * Permit null encryption ciphersuites, used for authentication only. It used
18553 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18554 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18559 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18560 signed attributes when verifying signatures (this would break them),
18561 the detached data encoding was wrong and public keys obtained using
18562 X509_get_pubkey() weren't freed.
18566 * Add text documentation for the BUFFER functions. Also added a work around
18567 to a Win95 console bug. This was triggered by the password read stuff: the
18568 last character typed gets carried over to the next fread(). If you were
18569 generating a new cert request using 'req' for example then the last
18570 character of the passphrase would be CR which would then enter the first
18575 * Added the new 'Includes OpenSSL Cryptography Software' button as
18576 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18577 button and can be used by applications based on OpenSSL to show the
18578 relationship to the OpenSSL project.
18580 *Ralf S. Engelschall*
18582 * Remove confusing variables in function signatures in files
18583 ssl/ssl_lib.c and ssl/ssl.h.
18585 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18587 * Don't install bss_file.c under PREFIX/include/
18589 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18591 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18592 functions that return function pointers and has support for NT specific
18593 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18594 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18595 unsigned to signed types: this was killing the Win32 compile.
18599 * Add new certificate file to stack functions,
18600 SSL_add_dir_cert_subjects_to_stack() and
18601 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18602 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18603 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18604 This means that Apache-SSL and similar packages don't have to mess around
18605 to add as many CAs as they want to the preferred list.
18609 * Experiment with doxygen documentation. Currently only partially applied to
18611 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18612 openssl.doxy as the configuration file.
18616 * Get rid of remaining C++-style comments which strict C compilers hate.
18618 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18620 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18621 compiled in by default: it has problems with large keys.
18625 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18626 DH private keys and/or callback functions which directly correspond to
18627 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18628 is needed for applications which have to configure certificates on a
18629 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18631 For the RSA certificate situation is makes no difference, but
18632 for the DSA certificate situation this fixes the "no shared cipher"
18633 problem where the OpenSSL cipher selection procedure failed because the
18634 temporary keys were not overtaken from the context and the API provided
18635 no way to reconfigure them.
18636 The new functions now let applications reconfigure the stuff and they
18637 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18638 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18639 non-public-API function ssl_cert_instantiate() is used as a helper
18640 function and also to reduce code redundancy inside ssl_rsa.c.
18642 *Ralf S. Engelschall*
18644 * Move s_server -dcert and -dkey options out of the undocumented feature
18645 area because they are useful for the DSA situation and should be
18646 recognized by the users.
18648 *Ralf S. Engelschall*
18650 * Fix the cipher decision scheme for export ciphers: the export bits are
18651 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18652 SSL_EXP_MASK. So, the original variable has to be used instead of the
18653 already masked variable.
18655 *Richard Levitte <levitte@stacken.kth.se>*
18657 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18659 *Richard Levitte <levitte@stacken.kth.se>*
18661 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18662 from `int` to `unsigned int` because it is a length and initialized by
18663 EVP_DigestFinal() which expects an `unsigned int *`.
18665 *Richard Levitte <levitte@stacken.kth.se>*
18667 * Don't hard-code path to Perl interpreter on shebang line of Configure
18668 script. Instead use the usual Shell->Perl transition trick.
18670 *Ralf S. Engelschall*
18672 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18673 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18674 -noout -modulus` as it's already the case for `openssl rsa -noout
18675 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18676 currently the public key is printed (a decision which was already done by
18677 `openssl dsa -modulus` in the past) which serves a similar purpose.
18678 Additionally the NO_RSA no longer completely removes the whole -modulus
18679 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18682 *Ralf S. Engelschall*
18684 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18685 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18687 *Arne Ansper <arne@ats.cyber.ee>*
18689 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18690 to be added. Now both 'req' and 'ca' can use new objects defined in the
18695 * Add cool BIO that does syslog (or event log on NT).
18697 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18699 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18700 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18701 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18702 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18706 * Add preliminary config info for new extension code.
18710 * Make RSA_NO_PADDING really use no padding.
18712 *Ulf Moeller <ulf@fitug.de>*
18714 * Generate errors when private/public key check is done.
18718 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18719 for some CRL extensions and new objects added.
18723 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18724 key usage extension and fuller support for authority key id.
18728 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18729 padding method for RSA, which is recommended for new applications in PKCS
18730 #1 v2.0 (RFC 2437, October 1998).
18731 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18732 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18733 against Bleichbacher's attack on RSA.
18734 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18737 * Updates to the new SSL compression code
18739 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18741 * Fix so that the version number in the master secret, when passed
18742 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18743 (because the server will not accept higher), that the version number
18744 is 0x03,0x01, not 0x03,0x00
18746 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18748 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18749 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18750 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18754 * Support for RAW extensions where an arbitrary extension can be
18755 created by including its DER encoding. See `apps/openssl.cnf` for
18760 * Make sure latest Perl versions don't interpret some generated C array
18761 code as Perl array code in the crypto/err/err_genc.pl script.
18763 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18765 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18766 not many people have the assembler. Various Win32 compilation fixes and
18767 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18768 build instructions.
18772 * Modify configure script 'Configure' to automatically create crypto/date.h
18773 file under Win32 and also build pem.h from pem.org. New script
18774 util/mkfiles.pl to create the MINFO file on environments that can't do a
18775 'make files': perl util/mkfiles.pl >MINFO should work.
18779 * Major rework of DES function declarations, in the pursuit of correctness
18780 and purity. As a result, many evil casts evaporated, and some weirdness,
18781 too. You may find this causes warnings in your code. Zapping your evil
18782 casts will probably fix them. Mostly.
18786 * Fix for a typo in asn1.h. Bug fix to object creation script
18787 obj_dat.pl. It considered a zero in an object definition to mean
18788 "end of object": none of the objects in objects.h have any zeros
18789 so it wasn't spotted.
18791 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18793 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18794 Masking (CBCM). In the absence of test vectors, the best I have been able
18795 to do is check that the decrypt undoes the encrypt, so far. Send me test
18796 vectors if you have them.
18800 * Correct calculation of key length for export ciphers (too much space was
18801 allocated for null ciphers). This has not been tested!
18805 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18806 message is now correct (it understands "crypto" and "ssl" on its
18807 command line). There is also now an "update" option. This will update
18808 the util/ssleay.num and util/libeay.num files with any new functions.
18810 perl util/mkdef.pl crypto ssl update
18811 it will update them.
18815 * Overhauled the Perl interface:
18816 - ported BN stuff to OpenSSL's different BN library
18817 - made the perl/ source tree CVS-aware
18818 - renamed the package from SSLeay to OpenSSL (the files still contain
18819 their history because I've copied them in the repository)
18820 - removed obsolete files (the test scripts will be replaced
18821 by better Test::Harness variants in the future)
18823 *Ralf S. Engelschall*
18825 * First cut for a very conservative source tree cleanup:
18826 1. merge various obsolete readme texts into doc/ssleay.txt
18827 where we collect the old documents and readme texts.
18828 2. remove the first part of files where I'm already sure that we no
18829 longer need them because of three reasons: either they are just temporary
18830 files which were left by Eric or they are preserved original files where
18831 I've verified that the diff is also available in the CVS via "cvs diff
18832 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18833 the crypto/md/ stuff).
18835 *Ralf S. Engelschall*
18837 * More extension code. Incomplete support for subject and issuer alt
18838 name, issuer and authority key id. Change the i2v function parameters
18839 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18840 what that's for :-) Fix to ASN1 macro which messed up
18841 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18845 * Preliminary support for ENUMERATED type. This is largely copied from the
18850 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18852 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18854 * Make sure `make rehash` target really finds the `openssl` program.
18856 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18858 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18859 like to hear about it if this slows down other processors.
18863 * Add CygWin32 platform information to Configure script.
18865 *Alan Batie <batie@aahz.jf.intel.com>*
18867 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18869 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18871 * New program nseq to manipulate netscape certificate sequences
18875 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18880 * Fixes to BN code. Previously the default was to define BN_RECURSION
18881 but the BN code had some problems that would cause failures when
18882 doing certificate verification and some other functions.
18884 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18886 * Add ASN1 and PEM code to support netscape certificate sequences.
18890 * Add ASN1 and PEM code to support netscape certificate sequences.
18894 * Add several PKIX and private extended key usage OIDs.
18898 * Modify the 'ca' program to handle the new extension code. Modify
18899 openssl.cnf for new extension format, add comments.
18903 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18904 and add a sample to openssl.cnf so req -x509 now adds appropriate
18909 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18910 error code, add initial support to X509_print() and x509 application.
18914 * Takes a deep breath and start adding X509 V3 extension support code. Add
18915 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18916 stuff is currently isolated and isn't even compiled yet.
18920 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18921 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18922 Removed the versions check from X509 routines when loading extensions:
18923 this allows certain broken certificates that don't set the version
18924 properly to be processed.
18928 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18929 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18930 can still be regenerated with "make depend".
18934 * Spelling mistake in C version of CAST-128.
18936 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18938 * Changes to the error generation code. The perl script err-code.pl
18939 now reads in the old error codes and retains the old numbers, only
18940 adding new ones if necessary. It also only changes the .err files if new
18941 codes are added. The makefiles have been modified to only insert errors
18942 when needed (to avoid needlessly modifying header files). This is done
18943 by only inserting errors if the .err file is newer than the auto generated
18944 C file. To rebuild all the error codes from scratch (the old behaviour)
18945 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18946 or delete all the .err files.
18950 * CAST-128 was incorrectly implemented for short keys. The C version has
18951 been fixed, but is untested. The assembler versions are also fixed, but
18952 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18953 to regenerate it if needed.
18954 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18955 Hagino <itojun@kame.net>*
18957 * File was opened incorrectly in randfile.c.
18959 *Ulf Möller <ulf@fitug.de>*
18961 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18962 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18963 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18964 al: it's just almost always a UTCTime. Note this patch adds new error
18965 codes so do a "make errors" if there are problems.
18969 * Correct Linux 1 recognition in config.
18971 *Ulf Möller <ulf@fitug.de>*
18973 * Remove pointless MD5 hash when using DSA keys in ca.
18975 *Anonymous <nobody@replay.com>*
18977 * Generate an error if given an empty string as a cert directory. Also
18978 generate an error if handed NULL (previously returned 0 to indicate an
18979 error, but didn't set one).
18981 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18983 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18987 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18988 parameters. This was causing a warning which killed off the Win32 compile.
18992 * Remove C++ style comments from crypto/bn/bn_local.h.
18994 *Neil Costigan <neil.costigan@celocom.com>*
18996 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18997 based on a text string, looking up short and long names and finally
18998 "dot" format. The "dot" format stuff didn't work. Added new function
18999 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19000 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19001 OID is not part of the table.
19005 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19006 X509_LOOKUP_by_alias().
19010 * Sort openssl functions by name.
19014 * Get the `gendsa` command working and add it to the `list` command. Remove
19015 encryption from sample DSA keys (in case anyone is interested the password
19020 * Make *all* `*_free` functions accept a NULL pointer.
19022 *Frans Heymans <fheymans@isaserver.be>*
19024 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19027 *Anonymous <nobody@replay.com>*
19029 * s_server should send the CAfile as acceptable CAs, not its own cert.
19031 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19033 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19035 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19037 * Temp key "for export" tests were wrong in s3_srvr.c.
19039 *Anonymous <nobody@replay.com>*
19041 * Add prototype for temp key callback functions
19042 SSL_CTX_set_tmp_{rsa,dh}_callback().
19046 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19047 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19051 * X509_name_add_entry() freed the wrong thing after an error.
19053 *Arne Ansper <arne@ats.cyber.ee>*
19055 * rsa_eay.c would attempt to free a NULL context.
19057 *Arne Ansper <arne@ats.cyber.ee>*
19059 * BIO_s_socket() had a broken should_retry() on Windoze.
19061 *Arne Ansper <arne@ats.cyber.ee>*
19063 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19065 *Arne Ansper <arne@ats.cyber.ee>*
19067 * Make sure the already existing X509_STORE->depth variable is initialized
19068 in X509_STORE_new(), but document the fact that this variable is still
19069 unused in the certificate verification process.
19071 *Ralf S. Engelschall*
19073 * Fix the various library and `apps/` files to free up pkeys obtained from
19074 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19078 * Fix reference counting in X509_PUBKEY_get(). This makes
19079 demos/maurice/example2.c work, amongst others, probably.
19081 *Steve Henson and Ben Laurie*
19083 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19084 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19085 are no longer created. This way we have a single and consistent command
19086 line interface `openssl <command>`, similar to `cvs <command>`.
19088 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19090 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19091 BIT STRING wrapper always have zero unused bits.
19095 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19099 * Make the top-level INSTALL documentation easier to understand.
19103 * Makefiles updated to exit if an error occurs in a sub-directory
19104 make (including if user presses ^C) [Paul Sutton]
19106 * Make Montgomery context stuff explicit in RSA data structure.
19110 * Fix build order of pem and err to allow for generated pem.h.
19114 * Fix renumbering bug in X509_NAME_delete_entry().
19118 * Enhanced the err-ins.pl script so it makes the error library number
19119 global and can add a library name. This is needed for external ASN1 and
19120 other error libraries.
19124 * Fixed sk_insert which never worked properly.
19128 * Fix ASN1 macros so they can handle indefinite length constructed
19129 EXPLICIT tags. Some non standard certificates use these: they can now
19134 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19135 into a single doc/ssleay.txt bundle. This way the information is still
19136 preserved but no longer messes up this directory. Now it's new room for
19137 the new set of documentation files.
19139 *Ralf S. Engelschall*
19141 * SETs were incorrectly DER encoded. This was a major pain, because they
19142 shared code with SEQUENCEs, which aren't coded the same. This means that
19143 almost everything to do with SETs or SEQUENCEs has either changed name or
19144 number of arguments.
19146 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19148 * Fix test data to work with the above.
19152 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19153 was already fixed by Eric for 0.9.1 it seems.
19155 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19157 * Autodetect FreeBSD3.
19161 * Fix various bugs in Configure. This affects the following platforms:
19165 unixware-2.0-pentium
19170 * Eliminate generated files from CVS. Reorder tests to regenerate files
19171 before they are needed.
19175 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19179 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19181 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19182 changed SSLeay to OpenSSL in version strings.
19184 *Ralf S. Engelschall*
19186 * Some fixups to the top-level documents.
19190 * Fixed the nasty bug where rsaref.h was not found under compile-time
19191 because the symlink to include/ was missing.
19193 *Ralf S. Engelschall*
19195 * Incorporated the popular no-RSA/DSA-only patches
19196 which allow to compile a RSA-free SSLeay.
19198 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19200 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19201 when "ssleay" is still not found.
19203 *Ralf S. Engelschall*
19205 * Added more platforms to Configure: Cray T3E, HPUX 11,
19207 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19209 * Updated the README file.
19211 *Ralf S. Engelschall*
19213 * Added various .cvsignore files in the CVS repository subdirs
19214 to make a "cvs update" really silent.
19216 *Ralf S. Engelschall*
19218 * Recompiled the error-definition header files and added
19219 missing symbols to the Win32 linker tables.
19221 *Ralf S. Engelschall*
19223 * Cleaned up the top-level documents;
19224 o new files: CHANGES and LICENSE
19225 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19226 o merged COPYRIGHT into LICENSE
19227 o removed obsolete TODO file
19228 o renamed MICROSOFT to INSTALL.W32
19230 *Ralf S. Engelschall*
19232 * Removed dummy files from the 0.9.1b source tree:
19233 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19234 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19235 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19236 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19237 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19239 *Ralf S. Engelschall*
19241 * Added various platform portability fixes.
19245 * The Genesis of the OpenSSL rpject:
19246 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19247 Young and Tim J. Hudson created while they were working for C2Net until
19250 *The OpenSSL Project*
19252 ### Changes between 0.9.0b and 0.9.1b [not released]
19254 * Updated a few CA certificates under certs/
19258 * Changed some BIGNUM api stuff.
19262 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19263 DGUX x86, Linux Alpha, etc.
19267 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19268 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19273 * Add -strparse option to asn1pars program which parses nested
19276 *Dr Stephen Henson <shenson@bigfoot.com>*
19278 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19282 * DSA fix for "ca" program.
19286 * Added "-genkey" option to "dsaparam" program.
19290 * Added RIPE MD160 (rmd160) message digest.
19294 * Added -a (all) option to "ssleay version" command.
19298 * Added PLATFORM define which is the id given to Configure.
19302 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19306 * Extended the ASN.1 parser routines.
19310 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19314 * Added a BN_CTX to the BN library.
19318 * Fixed the weak key values in DES library
19322 * Changed API in EVP library for cipher aliases.
19326 * Added support for RC2/64bit cipher.
19330 * Converted the lhash library to the crypto/mem.c functions.
19334 * Added more recognized ASN.1 object ids.
19338 * Added more RSA padding checks for SSL/TLS.
19342 * Added BIO proxy/filter functionality.
19346 * Added extra_certs to SSL_CTX which can be used
19347 send extra CA certificates to the client in the CA cert chain sending
19348 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19352 * Now Fortezza is denied in the authentication phase because
19353 this is key exchange mechanism is not supported by SSLeay at all.
19357 * Additional PKCS1 checks.
19361 * Support the string "TLSv1" for all TLS v1 ciphers.
19365 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19366 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19370 * Fixed a few memory leaks.
19374 * Fixed various code and comment typos.
19378 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19379 bytes sent in the client random.
19381 *Edward Bishop <ebishop@spyglass.com>*
19385 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19386 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19387 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19388 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19389 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19390 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19391 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19392 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19393 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19394 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19395 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19396 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19397 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19398 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19399 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19400 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19401 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19402 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19403 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19404 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19405 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19406 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19407 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19408 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19409 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19410 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19411 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19412 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19413 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19414 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19415 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19416 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19417 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19418 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19419 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19420 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19421 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19422 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19423 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19424 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19425 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19426 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19427 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19428 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19429 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19430 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19431 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19432 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19433 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19434 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19435 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19436 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19437 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19438 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19439 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19440 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19441 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19442 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19443 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19444 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19445 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19446 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19447 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19448 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19449 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19450 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19451 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19452 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19453 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19454 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19455 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19456 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19457 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19458 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19459 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19460 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19461 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19462 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19463 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19464 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19465 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19466 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19467 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19468 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19469 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19470 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19471 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19472 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19473 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19474 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19475 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19476 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19477 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19478 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19479 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19480 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19481 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19482 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19483 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19484 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19485 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19486 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19487 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19488 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19489 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19490 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19491 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19492 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19493 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19494 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19495 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19496 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19497 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19498 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19499 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19500 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19501 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19502 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19503 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19504 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19505 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19506 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19507 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19508 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19509 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19510 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19511 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19512 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19513 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19514 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19515 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19516 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19517 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19518 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19519 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19520 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19521 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19522 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19523 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19524 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19525 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19526 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19527 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19528 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19529 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19530 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19531 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19532 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19533 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19534 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19535 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19536 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19537 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19538 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19539 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19540 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19541 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19542 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19543 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19544 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19545 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19546 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655