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 * Add more SRTP protection profiles from RFC8723 and RFC8269.
31 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
33 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
35 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
36 supported and enabled.
40 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
41 to the list of ciphersuites providing Perfect Forward Secrecy as
42 required by SECLEVEL >= 3.
44 *Dmitry Belyavskiy, Nicola Tuveri*
46 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting. The
47 SSL_CTRL_GET_IANA_GROUPS control code, exposed as the SSL_get0_iana_groups()
48 function-like macro, retrieves the list of supported groups sent by the peer,
49 and the function SSL_client_hello_get_extension_order() populates a caller-supplied
50 array with the list of extension types present in the ClientHello, in order of
55 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
56 to make it possible to use empty passphrase strings.
60 * RNDR and RNDRRS support in provider functions to provide
61 random number generation for Arm CPUs (aarch64).
65 * s_client and s_server apps now explicitly say when the TLS version
66 does not include the renegotiation mechanism. This avoids confusion
67 between that scenario versus when the TLS version includes secure
68 renegotiation but the peer lacks support for it.
72 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
74 *Tomasz Kantecki, Andrey Matyukov*
76 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
77 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
78 of 160 bits and above and less than 224 bits were previously accepted by
79 default but are now no longer allowed. By default TLS compression was
80 already disabled in previous OpenSSL versions. At security level 2 it cannot
85 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
90 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
91 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
92 will need to load the legacy crypto provider.
96 * The various OBJ_* functions have been made thread safe.
100 * CCM8 cipher suites in TLS have been downgraded to security level zero
101 because they use a short authentication tag which lowers their strength.
105 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
110 * Parallel dual-prime 1536/2048-bit modular exponentiation for
111 AVX512_IFMA capable processors.
113 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
118 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
119 listed here are only a brief description.
120 The migration guide contains more detailed information related to new features,
121 breaking changes, and mappings for the large list of deprecated functions.
123 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
125 ### Changes between 3.0.2 and 3.0.3
127 * Case insensitive string comparison is reimplemented via new locale-agnostic
128 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
129 comparison. The previous implementation had problems when the Turkish locale
134 * Fixed a bug in the c_rehash script which was not properly sanitising shell
135 metacharacters to prevent command injection. This script is distributed by
136 some operating systems in a manner where it is automatically executed. On
137 such operating systems, an attacker could execute arbitrary commands with the
138 privileges of the script.
140 Use of the c_rehash script is considered obsolete and should be replaced
141 by the OpenSSL rehash command line tool.
146 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
147 certificate on an OCSP response. The bug caused the function in the case
148 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
149 response (meaning a successful verification) even in the case where the
150 response signing certificate fails to verify.
152 It is anticipated that most users of `OCSP_basic_verify` will not use the
153 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
154 a negative value (indicating a fatal error) in the case of a certificate
155 verification failure. The normal expected return value in this case would be
158 This issue also impacts the command line OpenSSL "ocsp" application. When
159 verifying an ocsp response with the "-no_cert_checks" option the command line
160 application will report that the verification is successful even though it
161 has in fact failed. In this case the incorrect successful response will also
162 be accompanied by error messages showing the failure and contradicting the
163 apparently successful result.
168 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
169 AAD data as the MAC key. This made the MAC key trivially predictable.
171 An attacker could exploit this issue by performing a man-in-the-middle attack
172 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
173 that the modified data would still pass the MAC integrity check.
175 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
176 endpoint will always be rejected by the recipient and the connection will
177 fail at that point. Many application protocols require data to be sent from
178 the client to the server first. Therefore, in such a case, only an OpenSSL
179 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
181 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
182 sent in both directions. In this case both clients and servers could be
183 affected, regardless of the application protocol.
185 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
186 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
187 the handshake when using this ciphersuite.
189 The confidentiality of data is not impacted by this issue, i.e. an attacker
190 cannot decrypt data that has been encrypted using this ciphersuite - they can
193 In order for this attack to work both endpoints must legitimately negotiate
194 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
195 OpenSSL 3.0, and is not available within the default provider or the default
196 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
197 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
198 following must have occurred:
200 1) OpenSSL must have been compiled with the (non-default) compile time option
201 enable-weak-ssl-ciphers
203 2) OpenSSL must have had the legacy provider explicitly loaded (either
204 through application code or via configuration)
206 3) The ciphersuite must have been explicitly added to the ciphersuite list
208 4) The libssl security level must have been set to 0 (default is 1)
210 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
212 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
213 others that both endpoints have in common
218 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
219 occuppied by the removed hash table entries.
221 This function is used when decoding certificates or keys. If a long lived
222 process periodically decodes certificates or keys its memory usage will
223 expand without bounds and the process might be terminated by the operating
224 system causing a denial of service. Also traversing the empty hash table
225 entries will take increasingly more time.
227 Typically such long lived processes might be TLS clients or TLS servers
228 configured to accept client certificate authentication.
231 *Hugo Landau, Aliaksei Levin*
233 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
234 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
235 statistics are no longer supported. For compatibility, these statistics are
236 still listed in the output but are now always reported as zero.
240 ### Changes between 3.0.1 and 3.0.2 [15 mar 2022]
242 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
243 for non-prime moduli.
245 Internally this function is used when parsing certificates that contain
246 elliptic curve public keys in compressed form or explicit elliptic curve
247 parameters with a base point encoded in compressed form.
249 It is possible to trigger the infinite loop by crafting a certificate that
250 has invalid explicit curve parameters.
252 Since certificate parsing happens prior to verification of the certificate
253 signature, any process that parses an externally supplied certificate may thus
254 be subject to a denial of service attack. The infinite loop can also be
255 reached when parsing crafted private keys as they can contain explicit
256 elliptic curve parameters.
258 Thus vulnerable situations include:
260 - TLS clients consuming server certificates
261 - TLS servers consuming client certificates
262 - Hosting providers taking certificates or private keys from customers
263 - Certificate authorities parsing certification requests from subscribers
264 - Anything else which parses ASN.1 elliptic curve parameters
266 Also any other applications that use the BN_mod_sqrt() where the attacker
267 can control the parameter values are vulnerable to this DoS issue.
272 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
273 to the list of ciphersuites providing Perfect Forward Secrecy as
274 required by SECLEVEL >= 3.
276 *Dmitry Belyavskiy, Nicola Tuveri*
278 * Made the AES constant time code for no-asm configurations
279 optional due to the resulting 95% performance degradation.
280 The AES constant time code can be enabled, for no assembly
281 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
285 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
290 * The negative return value handling of the certificate verification callback
291 was reverted. The replacement is to set the verification retry state with
292 the SSL_set_retry_verify() function.
296 ### Changes between 3.0.0 and 3.0.1 [14 dec 2021]
298 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
299 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
300 verify a certificate supplied by a server. That function may return a
301 negative return value to indicate an internal error (for example out of
302 memory). Such a negative return value is mishandled by OpenSSL and will cause
303 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
304 success and a subsequent call to SSL_get_error() to return the value
305 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
306 returned by OpenSSL if the application has previously called
307 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
308 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
309 totally unexpected and applications may not behave correctly as a result. The
310 exact behaviour will depend on the application but it could result in
311 crashes, infinite loops or other similar incorrect responses.
313 This issue is made more serious in combination with a separate bug in OpenSSL
314 3.0 that will cause X509_verify_cert() to indicate an internal error when
315 processing a certificate chain. This will occur where a certificate does not
316 include the Subject Alternative Name extension but where a Certificate
317 Authority has enforced name constraints. This issue can occur even with valid
323 * Corrected a few file name and file reference bugs in the build,
324 installation and setup scripts, which lead to installation verification
325 failures. Slightly enhanced the installation verification script.
329 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
334 * Fixed PVK encoder to properly query for the passphrase.
338 * Multiple fixes in the OSSL_HTTP API functions.
342 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
343 OSSL_PARAM_INTEGER data type and return error on negative numbers
344 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
345 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
349 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
353 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
357 * Multiple threading fixes.
361 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
365 * Allow fetching an operation from the provider that owns an unexportable key
366 as a fallback if that is still allowed by the property query.
370 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
372 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
377 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
378 S390X capability vector to zero. This simplifies testing of different code
379 paths on S390X architecture.
383 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
384 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
385 SP 800-38D". The communication will fail at this point.
389 * The EC_GROUP_clear_free() function is deprecated as there is nothing
390 confidential in EC_GROUP data.
394 * The byte order mark (BOM) character is ignored if encountered at the
395 beginning of a PEM-formatted file.
399 * Added CMS support for the Russian GOST algorithms.
403 * Due to move of the implementation of cryptographic operations
404 to the providers, validation of various operation parameters can
405 be postponed until the actual operation is executed where previously
406 it happened immediately when an operation parameter was set.
408 For example when setting an unsupported curve with
409 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
410 fail but later keygen operations with the EVP_PKEY_CTX will fail.
412 *OpenSSL team members and many third party contributors*
414 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
415 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
416 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
417 instead to retrieve these algorithms from a provider.
421 * On build targets where the multilib postfix is set in the build
422 configuration the libdir directory was changing based on whether
423 the lib directory with the multilib postfix exists on the system
424 or not. This unpredictable behavior was removed and eventual
425 multilib postfix is now always added to the default libdir. Use
426 `--libdir=lib` to override the libdir if adding the postfix is
431 * The triple DES key wrap functionality now conforms to RFC 3217 but is
432 no longer interoperable with OpenSSL 1.1.1.
436 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
437 function codes, this function can only cause problems for calling
442 * Add a configurable flag to output date formats as ISO 8601. Does not
443 change the default date format.
447 * Version of MSVC earlier than 1300 could get link warnings, which could
448 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
449 Support for this flag has been removed.
453 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
454 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
455 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
456 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
457 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
461 * The signatures of the functions to get and set options on SSL and
462 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
463 Some source code changes may be required.
467 * The public definitions of conf_method_st and conf_st have been
468 deprecated. They will be made opaque in a future release.
470 *Rich Salz and Tomáš Mráz*
472 * Client-initiated renegotiation is disabled by default. To allow it, use
473 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
474 flag, or the "ClientRenegotiation" config parameter as appropriate.
478 * Add "abspath" and "includedir" pragma's to config files, to prevent,
479 or modify relative pathname inclusion.
483 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
484 validated. Please consult the README-FIPS and
485 README-PROVIDERS files, as well as the migration guide.
487 *OpenSSL team members and many third party contributors*
489 * For the key types DH and DHX the allowed settable parameters are now different.
493 * The openssl commands that read keys, certificates, and CRLs now
494 automatically detect the PEM or DER format of the input files.
496 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
498 * Added enhanced PKCS#12 APIs which accept a library context.
502 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
506 * Added support for Kernel TLS (KTLS).
508 *Boris Pismenny, John Baldwin and Andrew Gallatin*
510 * Support for RFC 5746 secure renegotiation is now required by default for
511 SSL or TLS connections to succeed.
515 * The signature of the `copy` functional parameter of the
516 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
517 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
518 the signature of the `pub_decode` functional parameter of the
519 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
520 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
524 * The error return values from some control calls (ctrl) have changed.
528 * A public key check is now performed during EVP_PKEY_derive_set_peer().
532 * Many functions in the EVP_ namespace that are getters of values from
533 implementations or contexts were renamed to include get or get0 in their
534 names. Old names are provided as macro aliases for compatibility and
539 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
540 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
541 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
546 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
549 * The output from the command line applications may have minor
554 * The output from numerous "printing" may have minor changes.
558 * Windows thread synchronization uses read/write primitives (SRWLock) when
559 supported by the OS, otherwise CriticalSection continues to be used.
563 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
564 work on read only BIO source/sinks that do not support these functions.
565 This allows piping or redirection of a file BIO using stdin to be buffered
566 into memory. This is used internally in OSSL_DECODER_from_bio().
570 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
571 this function would return one of the values OSSL_STORE_INFO_NAME,
572 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
573 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
574 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
575 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
576 using this function should be amended to handle the changed return value.
580 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
581 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
582 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
583 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
584 contains more than one certificate identifier: This means that all
585 certificates referenced there MUST be part of the validation chain.
589 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
590 RC5, DESX and DES have been moved to the legacy provider.
594 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
595 RIPEMD-160 have been moved to the legacy provider.
599 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
604 * The deprecated functions EVP_PKEY_get0_RSA(),
605 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
606 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
607 well as the similarly named "get1" functions behave differently in
612 * A number of functions handling low-level keys or engines were deprecated
613 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
614 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
615 EVP_PKEY_get0_siphash().
619 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
620 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
621 will need to load the legacy crypto provider. This includes these PBE
622 algorithms which use this KDF:
623 - NID_pbeWithMD2AndDES_CBC
624 - NID_pbeWithMD5AndDES_CBC
625 - NID_pbeWithSHA1AndRC2_CBC
626 - NID_pbeWithMD2AndRC2_CBC
627 - NID_pbeWithMD5AndRC2_CBC
628 - NID_pbeWithSHA1AndDES_CBC
632 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
633 BIO_debug_callback() functions.
637 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
638 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
642 * The RAND_METHOD APIs have been deprecated.
646 * The SRP APIs have been deprecated.
650 * Add a compile time option to prevent the caching of provider fetched
651 algorithms. This is enabled by including the no-cached-fetch option
652 at configuration time.
656 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
657 count of PKCS12_DEFAULT_ITER.
659 *Tomáš Mráz and Sahana Prasad*
661 * The openssl speed command does not use low-level API calls anymore.
665 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
668 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
670 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
674 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
675 providers may supply their own group implementations (using either the "key
676 exchange" or the "key encapsulation" methods) which will automatically be
677 detected and used by libssl.
679 *Matt Caswell, Nicola Tuveri*
681 * The undocumented function X509_certificate_type() has been deprecated;
685 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
689 * Removed RSA padding mode for SSLv23 (which was only used for
690 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
691 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
696 * Deprecated the obsolete X9.31 RSA key generation related functions.
698 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
699 is not allowed to return a value > 1, this is no more taken as failure.
701 *Viktor Dukhovni and David von Oheimb*
703 * Deprecated the obsolete X9.31 RSA key generation related functions
704 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
705 BN_X931_generate_prime_ex().
709 * The default key generation method for the regular 2-prime RSA keys was
710 changed to the FIPS 186-4 B.3.6 method.
714 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
718 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
722 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
723 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
725 *Rich Salz, Richard Levitte, and David von Oheimb*
727 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
731 * Deprecated `OCSP_parse_url()`.
735 * Validation of SM2 keys has been separated from the validation of regular EC
740 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
741 switches: a validation failure triggers an early exit, returning a failure
742 exit status to the parent process.
746 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
747 to ignore unknown ciphers.
751 * The `-cipher-commands` and `-digest-commands` options
752 of the command line utility `list` have been deprecated.
753 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
757 * Added convenience functions for generating asymmetric key pairs:
758 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
759 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
763 * All of the low level EC_KEY functions have been deprecated.
765 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
767 * Deprecated all the libcrypto and libssl error string loading
772 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
773 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
778 * The `-crypt` option to the `passwd` command line tool has been removed.
782 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
787 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
791 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
792 EVP_PKEY_get1_tls_encodedpoint().
796 * The security callback, which can be customised by application code, supports
797 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
798 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
802 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
803 interface. Their functionality remains unchanged.
807 * Added new option for 'openssl list', '-providers', which will display the
808 list of loaded providers, their names, version and status. It optionally
809 displays their gettable parameters.
813 * Removed EVP_PKEY_set_alias_type().
817 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
818 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
822 * Changed all "STACK" functions to be macros instead of inline functions. Macro
823 parameters are still checked for type safety at compile time via helper
828 * Remove the RAND_DRBG API
830 *Paul Dale and Matthias St. Pierre*
832 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
833 as well as actual hostnames.
837 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
838 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
839 conversely, silently ignore DTLS protocol version bounds when configuring
840 TLS-based contexts. The commands can be repeated to set bounds of both
841 types. The same applies with the corresponding "min_protocol" and
842 "max_protocol" command-line switches, in case some application uses both TLS
845 SSL_CTX instances that are created for a fixed protocol version (e.g.
846 `TLSv1_server_method()`) also silently ignore version bounds. Previously
847 attempts to apply bounds to these protocol versions would result in an
848 error. Now only the "version-flexible" SSL_CTX instances are subject to
849 limits in configuration files in command-line options.
853 * Deprecated the `ENGINE` API. Engines should be replaced with providers
858 * Reworked the recorded ERR codes to make better space for system errors.
859 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
860 given code is a system error (true) or an OpenSSL error (false).
864 * Reworked the test perl framework to better allow parallel testing.
866 *Nicola Tuveri and David von Oheimb*
868 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
869 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
873 * 'Configure' has been changed to figure out the configuration target if
874 none is given on the command line. Consequently, the 'config' script is
875 now only a mere wrapper. All documentation is changed to only mention
878 *Rich Salz and Richard Levitte*
880 * Added a library context `OSSL_LIB_CTX` that applications as well as
881 other libraries can use to form a separate context within which
882 libcrypto operations are performed.
886 * Added various `_ex` functions to the OpenSSL API that support using
887 a non-default `OSSL_LIB_CTX`.
891 * Handshake now fails if Extended Master Secret extension is dropped
896 * Dropped interactive mode from the `openssl` program.
900 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
902 *David von Oheimb and Shane Lontis*
904 * Deprecated `EC_METHOD_get_field_type()`.
908 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
909 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
910 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
914 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
918 * Add CAdES-BES signature verification support, mostly derived
919 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
921 *Filipe Raimundo da Silva*
923 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
927 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
928 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
932 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
936 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
937 EC_KEY_precompute_mult().
941 * Deprecated EC_POINTs_mul().
945 * Removed FIPS_mode() and FIPS_mode_set().
949 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
953 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
954 EC_POINT_get_Jprojective_coordinates_GFp().
958 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
959 arrays to be more easily constructed via a series of utility functions.
960 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
961 the various push functions and finally convert to a passable OSSL_PARAM
962 array using OSSL_PARAM_BLD_to_param().
966 * The security strength of SHA1 and MD5 based signatures in TLS has been
971 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
972 contain a provider side internal key.
976 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
980 * Project text documents not yet having a proper file name extension
981 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
982 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
986 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
987 have been converted to Markdown with the goal to produce documents
988 which not only look pretty when viewed online in the browser, but
989 remain well readable inside a plain text editor.
991 To achieve this goal, a 'minimalistic' Markdown style has been applied
992 which avoids formatting elements that interfere too much with the
993 reading flow in the text file. For example, it
995 * avoids [ATX headings][] and uses [setext headings][] instead
996 (which works for `<h1>` and `<h2>` headings only).
997 * avoids [inline links][] and uses [reference links][] instead.
998 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1000 [ATX headings]: https://github.github.com/gfm/#atx-headings
1001 [setext headings]: https://github.github.com/gfm/#setext-headings
1002 [inline links]: https://github.github.com/gfm/#inline-link
1003 [reference links]: https://github.github.com/gfm/#reference-link
1004 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1005 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1007 *Matthias St. Pierre*
1009 * The test suite is changed to preserve results of each test recipe.
1010 A new directory test-runs/ with subdirectories named like the
1011 test recipes are created in the build tree for this purpose.
1015 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1016 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1017 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1019 *David von Oheimb, Martin Peylo*
1021 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1022 It supports arbitrary request and response content types, GET redirection,
1023 TLS, connections via HTTP(S) proxies, connections and exchange via
1024 user-defined BIOs (allowing implicit connections), persistent connections,
1025 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1026 The legacy OCSP-focused (and only partly documented) API
1027 is retained for backward compatibility, while most of it is deprecated.
1031 * Added `util/check-format.pl`, a tool for checking adherence to the
1032 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1033 The checks performed are incomplete and yield some false positives.
1034 Still the tool should be useful for detecting most typical glitches.
1038 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1039 If domain name resolution yields multiple IP addresses all of them are tried
1040 after `connect()` failures.
1044 * All of the low level RSA functions have been deprecated.
1048 * X509 certificates signed using SHA1 are no longer allowed at security
1053 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1054 modified to use PKEY APIs. These commands are now in maintenance mode
1055 and no new features will be added to them.
1059 * The command line utility rsautl has been deprecated.
1063 * The command line utilities genrsa and rsa have been modified to use PKEY
1064 APIs. They now write PKCS#8 keys by default. These commands are now in
1065 maintenance mode and no new features will be added to them.
1069 * All of the low level DH functions have been deprecated.
1071 *Paul Dale and Matt Caswell*
1073 * All of the low level DSA functions have been deprecated.
1077 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1078 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1082 * Deprecated low level ECDH and ECDSA functions.
1086 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1090 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1091 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1092 a new formulation to include all the things it can be used for,
1093 as well as words of caution.
1097 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1101 * All of the low level HMAC functions have been deprecated.
1103 *Paul Dale and David von Oheimb*
1105 * Over two thousand fixes were made to the documentation, including:
1106 - Common options (such as -rand/-writerand, TLS version control, etc)
1107 were refactored and point to newly-enhanced descriptions in openssl.pod.
1108 - Added style conformance for all options (with help from Richard Levitte),
1109 documented all reported missing options, added a CI build to check
1110 that all options are documented and that no unimplemented options
1112 - Documented some internals, such as all use of environment variables.
1113 - Addressed all internal broken L<> references.
1117 * All of the low level CMAC functions have been deprecated.
1121 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1122 functions have been deprecated.
1124 *Paul Dale and David von Oheimb*
1126 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1127 set of functions. The documentation mentioned negative values for some
1128 errors, but this was never the case, so the mention of negative values
1131 Code that followed the documentation and thereby check with something
1132 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1136 * All of the low level cipher functions have been deprecated.
1138 *Matt Caswell and Paul Dale*
1140 * Removed include/openssl/opensslconf.h.in and replaced it with
1141 include/openssl/configuration.h.in, which differs in not including
1142 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1143 was added to include both.
1145 This allows internal hacks where one might need to modify the set
1146 of configured macros, for example this if deprecated symbols are
1147 still supposed to be available internally:
1149 #include <openssl/configuration.h>
1151 #undef OPENSSL_NO_DEPRECATED
1152 #define OPENSSL_SUPPRESS_DEPRECATED
1154 #include <openssl/macros.h>
1156 This should not be used by applications that use the exported
1157 symbols, as that will lead to linking errors.
1161 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1162 used in exponentiation with 512-bit moduli. No EC algorithms are
1163 affected. Analysis suggests that attacks against 2-prime RSA1024,
1164 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1165 difficult to perform and are not believed likely. Attacks against DH512
1166 are considered just feasible. However, for an attack the target would
1167 have to re-use the DH512 private key, which is not recommended anyway.
1168 Also applications directly using the low-level API BN_mod_exp may be
1169 affected if they use BN_FLG_CONSTTIME.
1174 * Most memory-debug features have been deprecated, and the functionality
1175 replaced with no-ops.
1179 * Added documentation for the STACK API.
1183 * Introduced a new method type and API, OSSL_ENCODER, to represent
1184 generic encoders. These do the same sort of job that PEM writers
1185 and d2i functions do, but with support for methods supplied by
1186 providers, and the possibility for providers to support other
1191 * Introduced a new method type and API, OSSL_DECODER, to represent
1192 generic decoders. These do the same sort of job that PEM readers
1193 and i2d functions do, but with support for methods supplied by
1194 providers, and the possibility for providers to support other
1199 * Added a .pragma directive to the syntax of configuration files, to
1200 allow varying behavior in a supported and predictable manner.
1201 Currently added pragma:
1205 This allows dollar signs to be a keyword character unless it's
1206 followed by a opening brace or parenthesis. This is useful for
1207 platforms where dollar signs are commonly used in names, such as
1208 volume names and system directory names on VMS.
1212 * Added functionality to create an EVP_PKEY from user data.
1216 * Change the interpretation of the '--api' configuration option to
1217 mean that this is a desired API compatibility level with no
1218 further meaning. The previous interpretation, that this would
1219 also mean to remove all deprecated symbols up to and including
1220 the given version, no requires that 'no-deprecated' is also used
1221 in the configuration.
1223 When building applications, the desired API compatibility level
1224 can be set with the OPENSSL_API_COMPAT macro like before. For
1225 API compatibility version below 3.0, the old style numerical
1226 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1227 For version 3.0 and on, the value is expected to be the decimal
1228 value calculated from the major and minor version like this:
1230 MAJOR * 10000 + MINOR * 100
1234 -DOPENSSL_API_COMPAT=30000 For 3.0
1235 -DOPENSSL_API_COMPAT=30200 For 3.2
1237 To hide declarations that are deprecated up to and including the
1238 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1239 given when building the application as well.
1243 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1244 access to certificate and CRL stores via URIs and OSSL_STORE
1247 This adds the following functions:
1249 - X509_LOOKUP_store()
1250 - X509_STORE_load_file()
1251 - X509_STORE_load_path()
1252 - X509_STORE_load_store()
1253 - SSL_add_store_cert_subjects_to_stack()
1254 - SSL_CTX_set_default_verify_store()
1255 - SSL_CTX_load_verify_file()
1256 - SSL_CTX_load_verify_dir()
1257 - SSL_CTX_load_verify_store()
1261 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1262 The presence of this system service is determined at run-time.
1266 * Added functionality to create an EVP_PKEY context based on data
1267 for methods from providers. This takes an algorithm name and a
1268 property query string and simply stores them, with the intent
1269 that any operation that uses this context will use those strings
1270 to fetch the needed methods implicitly, thereby making the port
1271 of application written for pre-3.0 OpenSSL easier.
1275 * The undocumented function NCONF_WIN32() has been deprecated; for
1276 conversion details see the HISTORY section of doc/man5/config.pod
1280 * Introduced the new functions EVP_DigestSignInit_ex() and
1281 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1282 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1283 pages for further details.
1287 * Over two thousand fixes were made to the documentation, including:
1288 adding missing command flags, better style conformance, documentation
1291 *Rich Salz, Richard Levitte*
1293 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1294 X25519, X448, Ed25519 and Ed448.
1298 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1303 * Deprecated the public definition of `ERR_STATE` as well as the function
1304 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1309 * Added ERR functionality to give callers access to the stored function
1310 names that have replaced the older function code based functions.
1312 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1313 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1314 ERR_peek_error_all() and ERR_peek_last_error_all().
1316 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1317 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1318 ERR_func_error_string().
1322 * Extended testing to be verbose for failing tests only. The make variables
1323 VERBOSE_FAILURE or VF can be used to enable this:
1325 $ make VF=1 test # Unix
1326 $ mms /macro=(VF=1) test ! OpenVMS
1327 $ nmake VF=1 test # Windows
1331 * Added the `-copy_extensions` option to the `x509` command for use with
1332 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1333 all extensions in the request are copied to the certificate or vice versa.
1335 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1337 * Added the `-copy_extensions` option to the `req` command for use with
1338 `-x509`. When given with the `copy` or `copyall` argument,
1339 all extensions in the certification request are copied to the certificate.
1343 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1344 they generate are by default RFC 5280 compliant in the following sense:
1345 There is a subjectKeyIdentifier extension with a hash value of the public key
1346 and for not self-signed certs there is an authorityKeyIdentifier extension
1347 with a keyIdentifier field or issuer information identifying the signing key.
1348 This is done unless some configuration overrides the new default behavior,
1349 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1353 * Added several checks to `X509_verify_cert()` according to requirements in
1354 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1355 (which may be done by using the CLI option `-x509_strict`):
1356 * The basicConstraints of CA certificates must be marked critical.
1357 * CA certificates must explicitly include the keyUsage extension.
1358 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1359 * The issuer name of any certificate must not be empty.
1360 * The subject name of CA certs, certs with keyUsage crlSign,
1361 and certs without subjectAlternativeName must not be empty.
1362 * If a subjectAlternativeName extension is given it must not be empty.
1363 * The signatureAlgorithm field and the cert signature must be consistent.
1364 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1365 must not be marked critical.
1366 * The authorityKeyIdentifier must be given for X.509v3 certs
1367 unless they are self-signed.
1368 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1372 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1373 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1377 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1378 used even when parsing explicit parameters, when loading a encoded key
1379 or calling `EC_GROUP_new_from_ecpkparameters()`/
1380 `EC_GROUP_new_from_ecparameters()`.
1381 This prevents bypass of security hardening and performance gains,
1382 especially for curves with specialized EC_METHODs.
1383 By default, if a key encoded with explicit parameters is loaded and later
1384 encoded, the output is still encoded with explicit parameters, even if
1385 internally a "named" EC_GROUP is used for computation.
1389 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1390 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1391 NULL. After this change, only the cofactor parameter can be NULL. It also
1392 does some minimal sanity checks on the passed order.
1397 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1398 An attack is simple, if the first CMS_recipientInfo is valid but the
1399 second CMS_recipientInfo is chosen ciphertext. If the second
1400 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1401 encryption key will be replaced by garbage, and the message cannot be
1402 decoded, but if the RSA decryption fails, the correct encryption key is
1403 used and the recipient will not notice the attack.
1404 As a work around for this potential attack the length of the decrypted
1405 key must be equal to the cipher default key length, in case the
1406 certifiate is not given and all recipientInfo are tried out.
1407 The old behaviour can be re-enabled in the CMS code by setting the
1408 CMS_DEBUG_DECRYPT flag.
1412 * Early start up entropy quality from the DEVRANDOM seed source has been
1413 improved for older Linux systems. The RAND subsystem will wait for
1414 /dev/random to be producing output before seeding from /dev/urandom.
1415 The seeded state is stored for future library initialisations using
1416 a system global shared memory segment. The shared memory identifier
1417 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1418 the desired value. The default identifier is 114.
1422 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1423 when primes for RSA keys are computed.
1424 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1425 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1426 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1427 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1428 This avoids possible fingerprinting of newly generated RSA modules.
1432 * Correct the extended master secret constant on EBCDIC systems. Without this
1433 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1434 negotiate EMS will fail. Unfortunately this also means that TLS connections
1435 between EBCDIC systems with this fix, and EBCDIC systems without this
1436 fix will fail if they negotiate EMS.
1440 * Changed the library initialisation so that the config file is now loaded
1441 by default. This was already the case for libssl. It now occurs for both
1442 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1443 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1447 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1448 where the former acts as a replacement for `ERR_put_error()`, and the
1449 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1450 `ERR_raise_data()` adds more flexibility by taking a format string and
1451 an arbitrary number of arguments following it, to be processed with
1456 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1457 to check if a named provider is loaded and available. When called, it
1458 will also activate all fallback providers if such are still present.
1462 * Enforce a minimum DH modulus size of 512 bits.
1466 * Changed DH parameters to generate the order q subgroup instead of 2q.
1467 Previously generated DH parameters are still accepted by DH_check
1468 but DH_generate_key works around that by clearing bit 0 of the
1469 private key for those. This avoids leaking bit 0 of the private key.
1473 * Significantly reduce secure memory usage by the randomness pools.
1477 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1482 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1483 algorithms. An implementation of a key exchange algorithm can be obtained
1484 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1485 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1486 the older EVP_PKEY_derive_init() function. See the man pages for the new
1487 functions for further details.
1491 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1495 * Removed the function names from error messages and deprecated the
1500 * Removed NextStep support and the macro OPENSSL_UNISTD
1504 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1505 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1506 Also removed "export var as function" capability; we do not export
1507 variables, only functions.
1511 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1512 an error and 1 indicating success. In previous versions of OpenSSL this
1513 was a void type. If a key was set longer than the maximum possible this
1518 * Support SM2 signing and verification schemes with X509 certificate.
1522 * Use SHA256 as the default digest for TS query in the `ts` app.
1526 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1530 * Default cipher lists/suites are now available via a function, the
1531 #defines are deprecated.
1535 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1536 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1537 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1541 * Join the directories crypto/x509 and crypto/x509v3
1545 * Added command 'openssl kdf' that uses the EVP_KDF API.
1549 * Added command 'openssl mac' that uses the EVP_MAC API.
1553 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1554 as default directories. Also added the command 'openssl info'
1555 for scripting purposes.
1559 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1564 * Add prediction resistance to the DRBG reseeding process.
1568 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1569 mandated by IEEE Std 1619-2018.
1573 * Added newline escaping functionality to a filename when using openssl dgst.
1574 This output format is to replicate the output format found in the `*sum`
1575 checksum programs. This aims to preserve backward compatibility.
1577 *Matt Eaton, Richard Levitte, and Paul Dale*
1579 * Removed the heartbeat message in DTLS feature, as it has very
1580 little usage and doesn't seem to fulfill a valuable purpose.
1581 The configuration option is now deprecated.
1585 * Changed the output of 'openssl {digestname} < file' to display the
1586 digest name in its output.
1590 * Added a new generic trace API which provides support for enabling
1591 instrumentation through trace output.
1593 *Richard Levitte & Matthias St. Pierre*
1595 * Added build tests for C++. These are generated files that only do one
1596 thing, to include one public OpenSSL head file each. This tests that
1597 the public header files can be usefully included in a C++ application.
1599 This test isn't enabled by default. It can be enabled with the option
1600 'enable-buildtest-c++'.
1604 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1608 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1612 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1616 * Added KMAC to EVP_MAC.
1620 * Added property based algorithm implementation selection framework to
1625 * Added SCA hardening for modular field inversion in EC_GROUP through
1626 a new dedicated field_inv() pointer in EC_METHOD.
1627 This also addresses a leakage affecting conversions from projective
1628 to affine coordinates.
1630 *Billy Bob Brumley, Nicola Tuveri*
1632 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1633 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1634 those algorithms that were already supported through the EVP_PKEY API
1635 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1636 and scrypt are now wrappers that call EVP_KDF.
1640 * Build devcrypto engine as a dynamic engine.
1642 *Eneas U de Queiroz*
1644 * Add keyed BLAKE2 to EVP_MAC.
1648 * Fix a bug in the computation of the endpoint-pair shared secret used
1649 by DTLS over SCTP. This breaks interoperability with older versions
1650 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1651 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1652 interoperability with such broken implementations. However, enabling
1653 this switch breaks interoperability with correct implementations.
1655 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1656 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1660 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1664 * Changed the license to the Apache License v2.0.
1668 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1670 - Major releases (indicated by incrementing the MAJOR release number)
1671 may introduce incompatible API/ABI changes.
1672 - Minor releases (indicated by incrementing the MINOR release number)
1673 may introduce new features but retain API/ABI compatibility.
1674 - Patch releases (indicated by incrementing the PATCH number)
1675 are intended for bug fixes and other improvements of existing
1676 features only (like improving performance or adding documentation)
1677 and retain API/ABI compatibility.
1681 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1685 * Remove the 'dist' target and add a tarball building script. The
1686 'dist' target has fallen out of use, and it shouldn't be
1687 necessary to configure just to create a source distribution.
1691 * Recreate the OS390-Unix config target. It no longer relies on a
1692 special script like it did for OpenSSL pre-1.1.0.
1696 * Instead of having the source directories listed in Configure, add
1697 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1702 * Add GMAC to EVP_MAC.
1706 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1710 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1711 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1712 to facilitate the continued use of MACs through raw private keys in
1713 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1717 * Deprecate ECDH_KDF_X9_62().
1721 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1722 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1723 are retained for backwards compatibility.
1727 * AES-XTS mode now enforces that its two keys are different to mitigate
1728 the attacked described in "Efficient Instantiations of Tweakable
1729 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1730 Details of this attack can be obtained from:
1731 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1735 * Rename the object files, i.e. give them other names than in previous
1736 versions. Their names now include the name of the final product, as
1737 well as its type mnemonic (bin, lib, shlib).
1741 * Added new option for 'openssl list', '-objects', which will display the
1742 list of built in objects, i.e. OIDs with names.
1746 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1747 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1752 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1753 improves application performance by removing data copies and providing
1754 applications with zero-copy system calls such as sendfile and splice.
1758 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1762 * `PKCS12_parse` now maintains the order of the parsed certificates
1763 when outputting them via `*ca` (rather than reversing it).
1767 * Deprecated pthread fork support methods.
1771 * Added support for FFDHE key exchange in TLS 1.3.
1775 * Added a new concept for OpenSSL plugability: providers. This
1776 functionality is designed to replace the ENGINE API and ENGINE
1777 implementations, and to be much more dynamic, allowing provider
1778 authors to introduce new algorithms among other things, as long as
1779 there's an API that supports the algorithm type.
1781 With this concept comes a new core API for interaction between
1782 libcrypto and provider implementations. Public libcrypto functions
1783 that want to use providers do so through this core API.
1785 The main documentation for this core API is found in
1786 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1787 refer to other manuals describing the API specific for supported
1788 algorithm types (also called operations).
1795 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
1797 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
1799 * Avoid loading of a dynamic engine twice.
1803 * Prioritise DANE TLSA issuer certs over peer certs
1807 * Fixed random API for MacOS prior to 10.12
1809 These MacOS versions don't support the CommonCrypto APIs
1813 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1815 * Fixed an SM2 Decryption Buffer Overflow.
1817 In order to decrypt SM2 encrypted data an application is expected to
1818 call the API function EVP_PKEY_decrypt(). Typically an application will
1819 call this function twice. The first time, on entry, the "out" parameter
1820 can be NULL and, on exit, the "outlen" parameter is populated with the
1821 buffer size required to hold the decrypted plaintext. The application
1822 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1823 again, but this time passing a non-NULL value for the "out" parameter.
1825 A bug in the implementation of the SM2 decryption code means that the
1826 calculation of the buffer size required to hold the plaintext returned
1827 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1828 size required by the second call. This can lead to a buffer overflow
1829 when EVP_PKEY_decrypt() is called by the application a second time with
1830 a buffer that is too small.
1832 A malicious attacker who is able present SM2 content for decryption to
1833 an application could cause attacker chosen data to overflow the buffer
1834 by up to a maximum of 62 bytes altering the contents of other data held
1835 after the buffer, possibly changing application behaviour or causing
1836 the application to crash. The location of the buffer is application
1837 dependent but is typically heap allocated.
1842 * Fixed various read buffer overruns processing ASN.1 strings
1844 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1845 structure which contains a buffer holding the string data and a field
1846 holding the buffer length. This contrasts with normal C strings which
1847 are repesented as a buffer for the string data which is terminated
1848 with a NUL (0) byte.
1850 Although not a strict requirement, ASN.1 strings that are parsed using
1851 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1852 well as any string whose value has been set with the ASN1_STRING_set()
1853 function will additionally NUL terminate the byte array in the
1854 ASN1_STRING structure.
1856 However, it is possible for applications to directly construct valid
1857 ASN1_STRING structures which do not NUL terminate the byte array by
1858 directly setting the "data" and "length" fields in the ASN1_STRING
1859 array. This can also happen by using the ASN1_STRING_set0() function.
1861 Numerous OpenSSL functions that print ASN.1 data have been found to
1862 assume that the ASN1_STRING byte array will be NUL terminated, even
1863 though this is not guaranteed for strings that have been directly
1864 constructed. Where an application requests an ASN.1 structure to be
1865 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1866 been directly constructed by the application without NUL terminating
1867 the "data" field, then a read buffer overrun can occur.
1869 The same thing can also occur during name constraints processing
1870 of certificates (for example if a certificate has been directly
1871 constructed by the application instead of loading it via the OpenSSL
1872 parsing functions, and the certificate contains non NUL terminated
1873 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1874 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1876 If a malicious actor can cause an application to directly construct an
1877 ASN1_STRING and then process it through one of the affected OpenSSL
1878 functions then this issue could be hit. This might result in a crash
1879 (causing a Denial of Service attack). It could also result in the
1880 disclosure of private memory contents (such as private keys, or
1881 sensitive plaintext).
1886 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1888 * Fixed a problem with verifying a certificate chain when using the
1889 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1890 the certificates present in a certificate chain. It is not set by default.
1892 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1893 the chain that have explicitly encoded elliptic curve parameters was added
1894 as an additional strict check.
1896 An error in the implementation of this check meant that the result of a
1897 previous check to confirm that certificates in the chain are valid CA
1898 certificates was overwritten. This effectively bypasses the check
1899 that non-CA certificates must not be able to issue other certificates.
1901 If a "purpose" has been configured then there is a subsequent opportunity
1902 for checks that the certificate is a valid CA. All of the named "purpose"
1903 values implemented in libcrypto perform this check. Therefore, where
1904 a purpose is set the certificate chain will still be rejected even when the
1905 strict flag has been used. A purpose is set by default in libssl client and
1906 server certificate verification routines, but it can be overridden or
1907 removed by an application.
1909 In order to be affected, an application must explicitly set the
1910 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1911 for the certificate verification or, in the case of TLS client or server
1912 applications, override the default purpose.
1917 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1918 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1919 renegotiation ClientHello omits the signature_algorithms extension (where it
1920 was present in the initial ClientHello), but includes a
1921 signature_algorithms_cert extension then a NULL pointer dereference will
1922 result, leading to a crash and a denial of service attack.
1924 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1925 (which is the default configuration). OpenSSL TLS clients are not impacted by
1929 *Peter Kästle and Samuel Sapalski*
1931 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1933 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1934 create a unique hash value based on the issuer and serial number data
1935 contained within an X509 certificate. However it was failing to correctly
1936 handle any errors that may occur while parsing the issuer field (which might
1937 occur if the issuer field is maliciously constructed). This may subsequently
1938 result in a NULL pointer deref and a crash leading to a potential denial of
1944 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1945 padding mode to correctly check for rollback attacks. This is considered a
1946 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1951 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1952 functions. Previously they could overflow the output length argument in some
1953 cases where the input length is close to the maximum permissable length for
1954 an integer on the platform. In such cases the return value from the function
1955 call would be 1 (indicating success), but the output length value would be
1956 negative. This could cause applications to behave incorrectly or crash.
1961 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1962 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1963 could be exploited in a side channel attack to recover the password. Since
1964 the attack is local host only this is outside of the current OpenSSL
1965 threat model and therefore no CVE is assigned.
1967 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1972 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1974 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1975 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1976 If an attacker can control both items being compared then this could lead
1977 to a possible denial of service attack. OpenSSL itself uses the
1978 GENERAL_NAME_cmp function for two purposes:
1979 1) Comparing CRL distribution point names between an available CRL and a
1980 CRL distribution point embedded in an X509 certificate
1981 2) When verifying that a timestamp response token signer matches the
1982 timestamp authority name (exposed via the API functions
1983 TS_RESP_verify_response and TS_RESP_verify_token)
1988 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1990 * Certificates with explicit curve parameters are now disallowed in
1991 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1995 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1996 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1997 conversely, silently ignore DTLS protocol version bounds when configuring
1998 TLS-based contexts. The commands can be repeated to set bounds of both
1999 types. The same applies with the corresponding "min_protocol" and
2000 "max_protocol" command-line switches, in case some application uses both TLS
2003 SSL_CTX instances that are created for a fixed protocol version (e.g.
2004 TLSv1_server_method()) also silently ignore version bounds. Previously
2005 attempts to apply bounds to these protocol versions would result in an
2006 error. Now only the "version-flexible" SSL_CTX instances are subject to
2007 limits in configuration files in command-line options.
2011 * Handshake now fails if Extended Master Secret extension is dropped
2016 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2018 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2020 * Fixed segmentation fault in SSL_check_chain()
2021 Server or client applications that call the SSL_check_chain() function
2022 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2023 dereference as a result of incorrect handling of the
2024 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2025 or unrecognised signature algorithm is received from the peer. This could
2026 be exploited by a malicious peer in a Denial of Service attack.
2031 * Added AES consttime code for no-asm configurations
2032 an optional constant time support for AES was added
2033 when building openssl for no-asm.
2034 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2035 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2036 At this time this feature is by default disabled.
2037 It will be enabled by default in 3.0.
2041 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2043 * Revert the change of EOF detection while reading in libssl to avoid
2044 regressions in applications depending on the current way of reporting
2045 the EOF. As the existing method is not fully accurate the change to
2046 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2047 branch and will be present in the 3.0 release.
2051 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2052 when primes for RSA keys are computed.
2053 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2054 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2055 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
2056 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2057 This avoids possible fingerprinting of newly generated RSA modules.
2061 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2063 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2064 while reading in libssl then we would report an error back to the
2065 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2066 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2067 therefore give a hint as to what went wrong.
2071 * Check that ed25519 and ed448 are allowed by the security level. Previously
2072 signature algorithms not using an MD were not being checked that they were
2073 allowed by the security level.
2077 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2078 was not quite right. The behaviour was not consistent between resumption
2079 and normal handshakes, and also not quite consistent with historical
2080 behaviour. The behaviour in various scenarios has been clarified and
2081 it has been updated to make it match historical behaviour as closely as
2086 * *[VMS only]* The header files that the VMS compilers include automatically,
2087 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2088 that the C++ compiler doesn't understand. This is a shortcoming in the
2089 compiler, but can be worked around with `__cplusplus` guards.
2091 C++ applications that use OpenSSL libraries must be compiled using the
2092 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2093 functions. Otherwise, only functions with symbols of less than 31
2094 characters can be used, as the linker will not be able to successfully
2095 resolve symbols with longer names.
2099 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2100 The presence of this system service is determined at run-time.
2104 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2109 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2111 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2112 number generator (RNG). This was intended to include protection in the
2113 event of a fork() system call in order to ensure that the parent and child
2114 processes did not share the same RNG state. However this protection was not
2115 being used in the default case.
2117 A partial mitigation for this issue is that the output from a high
2118 precision timer is mixed into the RNG state so the likelihood of a parent
2119 and child process sharing state is significantly reduced.
2121 If an application already calls OPENSSL_init_crypto() explicitly using
2122 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2125 *Matthias St. Pierre*
2127 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2128 used even when parsing explicit parameters, when loading a encoded key
2129 or calling `EC_GROUP_new_from_ecpkparameters()`/
2130 `EC_GROUP_new_from_ecparameters()`.
2131 This prevents bypass of security hardening and performance gains,
2132 especially for curves with specialized EC_METHODs.
2133 By default, if a key encoded with explicit parameters is loaded and later
2134 encoded, the output is still encoded with explicit parameters, even if
2135 internally a "named" EC_GROUP is used for computation.
2139 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2140 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2141 NULL. After this change, only the cofactor parameter can be NULL. It also
2142 does some minimal sanity checks on the passed order.
2147 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2148 An attack is simple, if the first CMS_recipientInfo is valid but the
2149 second CMS_recipientInfo is chosen ciphertext. If the second
2150 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2151 encryption key will be replaced by garbage, and the message cannot be
2152 decoded, but if the RSA decryption fails, the correct encryption key is
2153 used and the recipient will not notice the attack.
2154 As a work around for this potential attack the length of the decrypted
2155 key must be equal to the cipher default key length, in case the
2156 certifiate is not given and all recipientInfo are tried out.
2157 The old behaviour can be re-enabled in the CMS code by setting the
2158 CMS_DEBUG_DECRYPT flag.
2163 * Early start up entropy quality from the DEVRANDOM seed source has been
2164 improved for older Linux systems. The RAND subsystem will wait for
2165 /dev/random to be producing output before seeding from /dev/urandom.
2166 The seeded state is stored for future library initialisations using
2167 a system global shared memory segment. The shared memory identifier
2168 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2169 the desired value. The default identifier is 114.
2173 * Correct the extended master secret constant on EBCDIC systems. Without this
2174 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2175 negotiate EMS will fail. Unfortunately this also means that TLS connections
2176 between EBCDIC systems with this fix, and EBCDIC systems without this
2177 fix will fail if they negotiate EMS.
2181 * Use Windows installation paths in the mingw builds
2183 Mingw isn't a POSIX environment per se, which means that Windows
2184 paths should be used for installation.
2189 * Changed DH_check to accept parameters with order q and 2q subgroups.
2190 With order 2q subgroups the bit 0 of the private key is not secret
2191 but DH_generate_key works around that by clearing bit 0 of the
2192 private key for those. This avoids leaking bit 0 of the private key.
2196 * Significantly reduce secure memory usage by the randomness pools.
2200 * Revert the DEVRANDOM_WAIT feature for Linux systems
2202 The DEVRANDOM_WAIT feature added a select() call to wait for the
2203 /dev/random device to become readable before reading from the
2204 /dev/urandom device.
2206 It turned out that this change had negative side effects on
2207 performance which were not acceptable. After some discussion it
2208 was decided to revert this feature and leave it up to the OS
2209 resp. the platform maintainer to ensure a proper initialization
2210 during early boot time.
2212 *Matthias St. Pierre*
2214 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2216 * Add build tests for C++. These are generated files that only do one
2217 thing, to include one public OpenSSL head file each. This tests that
2218 the public header files can be usefully included in a C++ application.
2220 This test isn't enabled by default. It can be enabled with the option
2221 'enable-buildtest-c++'.
2225 * Enable SHA3 pre-hashing for ECDSA and DSA.
2229 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2230 This changes the size when using the `genpkey` command when no size is given.
2231 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2232 generation commands to use 2048 bits by default.
2236 * Reorganize the manual pages to consistently have RETURN VALUES,
2237 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2238 util/fix-doc-nits accordingly.
2240 *Paul Yang, Joshua Lock*
2242 * Add the missing accessor EVP_PKEY_get0_engine()
2246 * Have commands like `s_client` and `s_server` output the signature scheme
2247 along with other cipher suite parameters when debugging.
2251 * Make OPENSSL_config() error agnostic again.
2255 * Do the error handling in RSA decryption constant time.
2259 * Prevent over long nonces in ChaCha20-Poly1305.
2261 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2262 for every encryption operation. RFC 7539 specifies that the nonce value
2263 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2264 and front pads the nonce with 0 bytes if it is less than 12
2265 bytes. However it also incorrectly allows a nonce to be set of up to 16
2266 bytes. In this case only the last 12 bytes are significant and any
2267 additional leading bytes are ignored.
2269 It is a requirement of using this cipher that nonce values are
2270 unique. Messages encrypted using a reused nonce value are susceptible to
2271 serious confidentiality and integrity attacks. If an application changes
2272 the default nonce length to be longer than 12 bytes and then makes a
2273 change to the leading bytes of the nonce expecting the new value to be a
2274 new unique nonce then such an application could inadvertently encrypt
2275 messages with a reused nonce.
2277 Additionally the ignored bytes in a long nonce are not covered by the
2278 integrity guarantee of this cipher. Any application that relies on the
2279 integrity of these ignored leading bytes of a long nonce may be further
2280 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2281 is safe because no such use sets such a long nonce value. However user
2282 applications that use this cipher directly and set a non-default nonce
2283 length to be longer than 12 bytes may be vulnerable.
2285 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2291 * Add DEVRANDOM_WAIT feature for Linux systems
2293 On older Linux systems where the getrandom() system call is not available,
2294 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2295 Contrary to getrandom(), the /dev/urandom device will not block during
2296 early boot when the kernel CSPRNG has not been seeded yet.
2298 To mitigate this known weakness, use select() to wait for /dev/random to
2299 become readable before reading from /dev/urandom.
2301 * Ensure that SM2 only uses SM3 as digest algorithm
2305 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2307 * Change the info callback signals for the start and end of a post-handshake
2308 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2309 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2310 confused by this and assume that a TLSv1.2 renegotiation has started. This
2311 can break KeyUpdate handling. Instead we no longer signal the start and end
2312 of a post handshake message exchange (although the messages themselves are
2313 still signalled). This could break some applications that were expecting
2314 the old signals. However without this KeyUpdate is not usable for many
2319 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2321 * Timing vulnerability in DSA signature generation
2323 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2324 timing side channel attack. An attacker could use variations in the signing
2325 algorithm to recover the private key.
2327 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2332 * Timing vulnerability in ECDSA signature generation
2334 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2335 timing side channel attack. An attacker could use variations in the signing
2336 algorithm to recover the private key.
2338 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2343 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2344 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2345 of two gigabytes and the error handling improved.
2347 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2348 categorized as a normal bug, not a security issue, because the DRBG reseeds
2349 automatically and is fully functional even without additional randomness
2350 provided by the application.
2352 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2354 * Add a new ClientHello callback. Provides a callback interface that gives
2355 the application the ability to adjust the nascent SSL object at the
2356 earliest stage of ClientHello processing, immediately after extensions have
2357 been collected but before they have been processed. In particular, this
2358 callback can adjust the supported TLS versions in response to the contents
2363 * Add SM2 base algorithm support.
2367 * s390x assembly pack: add (improved) hardware-support for the following
2368 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2369 aes-cfb/cfb8, aes-ecb.
2373 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2374 parameter is no longer accepted, as it leads to a corrupt table. NULL
2375 pem_str is reserved for alias entries only.
2379 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2380 step for prime curves. The new implementation is based on formulae from
2381 differential addition-and-doubling in homogeneous projective coordinates
2382 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2383 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2384 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2385 to work in projective coordinates.
2387 *Billy Bob Brumley, Nicola Tuveri*
2389 * Change generating and checking of primes so that the error rate of not
2390 being prime depends on the intended use based on the size of the input.
2391 For larger primes this will result in more rounds of Miller-Rabin.
2392 The maximal error rate for primes with more than 1080 bits is lowered
2395 *Kurt Roeckx, Annie Yousar*
2397 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2401 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2402 moving between systems, and to avoid confusion when a Windows build is
2403 done with mingw vs with MSVC. For POSIX installs, there's still a
2404 symlink or copy named 'tsget' to avoid that confusion as well.
2408 * Revert blinding in ECDSA sign and instead make problematic addition
2409 length-invariant. Switch even to fixed-length Montgomery multiplication.
2413 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2414 step for binary curves. The new implementation is based on formulae from
2415 differential addition-and-doubling in mixed Lopez-Dahab projective
2416 coordinates, modified to independently blind the operands.
2418 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2420 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2421 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2422 EC_METHODs to implement their own specialized "ladder step", to take
2423 advantage of more favorable coordinate systems or more efficient
2424 differential addition-and-doubling algorithms.
2426 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2428 * Modified the random device based seed sources to keep the relevant
2429 file descriptors open rather than reopening them on each access.
2430 This allows such sources to operate in a chroot() jail without
2431 the associated device nodes being available. This behaviour can be
2432 controlled using RAND_keep_random_devices_open().
2436 * Numerous side-channel attack mitigations have been applied. This may have
2437 performance impacts for some algorithms for the benefit of improved
2438 security. Specific changes are noted in this change log by their respective
2443 * AIX shared library support overhaul. Switch to AIX "natural" way of
2444 handling shared libraries, which means collecting shared objects of
2445 different versions and bitnesses in one common archive. This allows to
2446 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2447 doesn't affect the way 3rd party applications are linked, only how
2448 multi-version installation is managed.
2452 * Make ec_group_do_inverse_ord() more robust and available to other
2453 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2454 mitigations are applied to the fallback BN_mod_inverse().
2455 When using this function rather than BN_mod_inverse() directly, new
2456 EC cryptosystem implementations are then safer-by-default.
2460 * Add coordinate blinding for EC_POINT and implement projective
2461 coordinate blinding for generic prime curves as a countermeasure to
2462 chosen point SCA attacks.
2464 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2466 * Add blinding to ECDSA and DSA signatures to protect against side channel
2467 attacks discovered by Keegan Ryan (NCC Group).
2471 * Enforce checking in the `pkeyutl` command to ensure that the input
2472 length does not exceed the maximum supported digest length when performing
2473 a sign, verify or verifyrecover operation.
2477 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2478 I/O in combination with something like select() or poll() will hang. This
2479 can be turned off again using SSL_CTX_clear_mode().
2480 Many applications do not properly handle non-application data records, and
2481 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2482 around the problems in those applications, but can also break some.
2483 It's recommended to read the manpages about SSL_read(), SSL_write(),
2484 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2485 SSL_CTX_set_read_ahead() again.
2489 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2490 now allow empty (zero character) pass phrases.
2494 * Apply blinding to binary field modular inversion and remove patent
2495 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2499 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2500 binary and prime elliptic curves.
2504 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2505 constant time fixed point multiplication.
2509 * Revise elliptic curve scalar multiplication with timing attack
2510 defenses: ec_wNAF_mul redirects to a constant time implementation
2511 when computing fixed point and variable point multiplication (which
2512 in OpenSSL are mostly used with secret scalars in keygen, sign,
2513 ECDH derive operations).
2514 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2517 * Updated CONTRIBUTING
2521 * Updated DRBG / RAND to request nonce and additional low entropy
2522 randomness from the system.
2524 *Matthias St. Pierre*
2526 * Updated 'openssl rehash' to use OpenSSL consistent default.
2530 * Moved the load of the ssl_conf module to libcrypto, which helps
2531 loading engines that libssl uses before libssl is initialised.
2535 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2539 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2541 *Ingo Schwarze, Rich Salz*
2543 * Added output of accepting IP address and port for 'openssl s_server'
2547 * Added a new API for TLSv1.3 ciphersuites:
2548 SSL_CTX_set_ciphersuites()
2549 SSL_set_ciphersuites()
2553 * Memory allocation failures consistently add an error to the error
2558 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2559 in libcrypto when run as setuid/setgid.
2563 * Load any config file by default when libssl is used.
2567 * Added new public header file <openssl/rand_drbg.h> and documentation
2568 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2570 *Matthias St. Pierre*
2572 * QNX support removed (cannot find contributors to get their approval
2573 for the license change).
2577 * TLSv1.3 replay protection for early data has been implemented. See the
2578 SSL_read_early_data() man page for further details.
2582 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2583 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2584 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2585 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2586 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2587 configuration has been separated out. See the ciphers man page or the
2588 SSL_CTX_set_ciphersuites() man page for more information.
2592 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2593 in responder mode now supports the new "-multi" option, which
2594 spawns the specified number of child processes to handle OCSP
2595 requests. The "-timeout" option now also limits the OCSP
2596 responder's patience to wait to receive the full client request
2597 on a newly accepted connection. Child processes are respawned
2598 as needed, and the CA index file is automatically reloaded
2599 when changed. This makes it possible to run the "ocsp" responder
2600 as a long-running service, making the OpenSSL CA somewhat more
2601 feature-complete. In this mode, most diagnostic messages logged
2602 after entering the event loop are logged via syslog(3) rather than
2607 * Added support for X448 and Ed448. Heavily based on original work by
2612 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2613 objects loaded. This adds the functions OSSL_STORE_expect() and
2614 OSSL_STORE_find() as well as needed tools to construct searches and
2615 get the search data out of them.
2619 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2620 version of OpenSSL should review their configuration settings to ensure
2621 that they are still appropriate for TLSv1.3. For further information see:
2622 <https://wiki.openssl.org/index.php/TLS1.3>
2626 * Grand redesign of the OpenSSL random generator
2628 The default RAND method now utilizes an AES-CTR DRBG according to
2629 NIST standard SP 800-90Ar1. The new random generator is essentially
2630 a port of the default random generator from the OpenSSL FIPS 2.0
2631 object module. It is a hybrid deterministic random bit generator
2632 using an AES-CTR bit stream and which seeds and reseeds itself
2633 automatically using trusted system entropy sources.
2635 Some of its new features are:
2636 - Support for multiple DRBG instances with seed chaining.
2637 - The default RAND method makes use of a DRBG.
2638 - There is a public and private DRBG instance.
2639 - The DRBG instances are fork-safe.
2640 - Keep all global DRBG instances on the secure heap if it is enabled.
2641 - The public and private DRBG instance are per thread for lock free
2644 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2646 * Changed Configure so it only says what it does and doesn't dump
2647 so much data. Instead, ./configdata.pm should be used as a script
2648 to display all sorts of configuration data.
2652 * Added processing of "make variables" to Configure.
2656 * Added SHA512/224 and SHA512/256 algorithm support.
2660 * The last traces of Netware support, first removed in 1.1.0, have
2665 * Get rid of Makefile.shared, and in the process, make the processing
2666 of certain files (rc.obj, or the .def/.map/.opt files produced from
2667 the ordinal files) more visible and hopefully easier to trace and
2668 debug (or make silent).
2672 * Make it possible to have environment variable assignments as
2673 arguments to config / Configure.
2677 * Add multi-prime RSA (RFC 8017) support.
2681 * Add SM3 implemented according to GB/T 32905-2016
2682 *Jack Lloyd <jack.lloyd@ribose.com>,*
2683 *Ronald Tse <ronald.tse@ribose.com>,*
2684 *Erick Borsboom <erick.borsboom@ribose.com>*
2686 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2687 as documented in RFC6066.
2688 Based on a patch from Tomasz Moń
2690 *Filipe Raimundo da Silva*
2692 * Add SM4 implemented according to GB/T 32907-2016.
2693 *Jack Lloyd <jack.lloyd@ribose.com>,*
2694 *Ronald Tse <ronald.tse@ribose.com>,*
2695 *Erick Borsboom <erick.borsboom@ribose.com>*
2697 * Reimplement -newreq-nodes and ERR_error_string_n; the
2698 original author does not agree with the license change.
2702 * Add ARIA AEAD TLS support.
2706 * Some macro definitions to support VS6 have been removed. Visual
2707 Studio 6 has not worked since 1.1.0
2711 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2712 without clearing the errors.
2716 * Add "atfork" functions. If building on a system that without
2717 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2718 requirements. The RAND facility now uses/requires this.
2726 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2727 not possible to disable entirely. However, it's still possible to
2728 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2731 To disable, configure with 'no-ui-console'. 'no-ui' is still
2732 possible to use as an alias. Check at compile time with the
2733 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2734 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2738 * Add a STORE module, which implements a uniform and URI based reader of
2739 stores that can contain keys, certificates, CRLs and numerous other
2740 objects. The main API is loosely based on a few stdio functions,
2741 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2742 OSSL_STORE_error and OSSL_STORE_close.
2743 The implementation uses backends called "loaders" to implement arbitrary
2744 URI schemes. There is one built in "loader" for the 'file' scheme.
2748 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2749 then adjusted to work on FreeBSD 8.4 as well.
2750 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2751 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2755 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2756 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2757 error code calls like this:
2759 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2761 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2762 that can be encoded in C. For the foreseeable future, this will only
2765 *Richard Levitte and Tim Hudson*
2767 * Removed BSD cryptodev engine.
2771 * Add a build target 'build_all_generated', to build all generated files
2772 and only that. This can be used to prepare everything that requires
2773 things like perl for a system that lacks perl and then move everything
2774 to that system and do the rest of the build there.
2778 * In the UI interface, make it possible to duplicate the user data. This
2779 can be used by engines that need to retain the data for a longer time
2780 than just the call where this user data is passed.
2784 * Ignore the '-named_curve auto' value for compatibility of applications
2787 *Tomáš Mráz <tmraz@fedoraproject.org>*
2789 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2790 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2791 alerts across multiple records (some of which could be empty). In practice
2792 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2793 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2794 support this at all. Supporting it adds significant complexity to the
2795 record layer, and its removal is unlikely to cause interoperability
2800 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2801 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2802 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2807 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2808 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2810 *Richard Levitte, Andy Polyakov*
2812 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2817 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2818 platform rather than 'mingw'.
2822 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2823 success if they are asked to add an object which already exists
2824 in the store. This change cascades to other functions which load
2825 certificates and CRLs.
2829 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2830 facilitate stack unwinding even from assembly subroutines.
2834 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2835 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2839 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2840 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2841 which is the minimum version we support.
2845 * Certificate time validation (X509_cmp_time) enforces stricter
2846 compliance with RFC 5280. Fractional seconds and timezone offsets
2847 are no longer allowed.
2851 * Add support for ARIA
2855 * s_client will now send the Server Name Indication (SNI) extension by
2856 default unless the new "-noservername" option is used. The server name is
2857 based on the host provided to the "-connect" option unless overridden by
2858 using "-servername".
2862 * Add support for SipHash
2866 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2867 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2868 prevent issues where no progress is being made and the peer continually
2869 sends unrecognised record types, using up resources processing them.
2873 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2874 using the algorithm defined in
2875 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2879 * Heartbeat support has been removed; the ABI is changed for now.
2881 *Richard Levitte, Rich Salz*
2883 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2887 * The RSA "null" method, which was partially supported to avoid patent
2888 issues, has been replaced to always returns NULL.
2895 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2897 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2898 used even when parsing explicit parameters, when loading a encoded key
2899 or calling `EC_GROUP_new_from_ecpkparameters()`/
2900 `EC_GROUP_new_from_ecparameters()`.
2901 This prevents bypass of security hardening and performance gains,
2902 especially for curves with specialized EC_METHODs.
2903 By default, if a key encoded with explicit parameters is loaded and later
2904 encoded, the output is still encoded with explicit parameters, even if
2905 internally a "named" EC_GROUP is used for computation.
2909 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2910 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2911 NULL. After this change, only the cofactor parameter can be NULL. It also
2912 does some minimal sanity checks on the passed order.
2917 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2918 An attack is simple, if the first CMS_recipientInfo is valid but the
2919 second CMS_recipientInfo is chosen ciphertext. If the second
2920 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2921 encryption key will be replaced by garbage, and the message cannot be
2922 decoded, but if the RSA decryption fails, the correct encryption key is
2923 used and the recipient will not notice the attack.
2924 As a work around for this potential attack the length of the decrypted
2925 key must be equal to the cipher default key length, in case the
2926 certifiate is not given and all recipientInfo are tried out.
2927 The old behaviour can be re-enabled in the CMS code by setting the
2928 CMS_DEBUG_DECRYPT flag.
2933 * Use Windows installation paths in the mingw builds
2935 Mingw isn't a POSIX environment per se, which means that Windows
2936 paths should be used for installation.
2941 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2943 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2944 This changes the size when using the `genpkey` command when no size is given.
2945 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2946 generation commands to use 2048 bits by default.
2950 * Prevent over long nonces in ChaCha20-Poly1305.
2952 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2953 for every encryption operation. RFC 7539 specifies that the nonce value
2954 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2955 and front pads the nonce with 0 bytes if it is less than 12
2956 bytes. However it also incorrectly allows a nonce to be set of up to 16
2957 bytes. In this case only the last 12 bytes are significant and any
2958 additional leading bytes are ignored.
2960 It is a requirement of using this cipher that nonce values are
2961 unique. Messages encrypted using a reused nonce value are susceptible to
2962 serious confidentiality and integrity attacks. If an application changes
2963 the default nonce length to be longer than 12 bytes and then makes a
2964 change to the leading bytes of the nonce expecting the new value to be a
2965 new unique nonce then such an application could inadvertently encrypt
2966 messages with a reused nonce.
2968 Additionally the ignored bytes in a long nonce are not covered by the
2969 integrity guarantee of this cipher. Any application that relies on the
2970 integrity of these ignored leading bytes of a long nonce may be further
2971 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2972 is safe because no such use sets such a long nonce value. However user
2973 applications that use this cipher directly and set a non-default nonce
2974 length to be longer than 12 bytes may be vulnerable.
2976 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2982 * Added SCA hardening for modular field inversion in EC_GROUP through
2983 a new dedicated field_inv() pointer in EC_METHOD.
2984 This also addresses a leakage affecting conversions from projective
2985 to affine coordinates.
2987 *Billy Bob Brumley, Nicola Tuveri*
2989 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2990 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2994 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2998 * Remove the 'dist' target and add a tarball building script. The
2999 'dist' target has fallen out of use, and it shouldn't be
3000 necessary to configure just to create a source distribution.
3004 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3006 * Timing vulnerability in DSA signature generation
3008 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3009 timing side channel attack. An attacker could use variations in the signing
3010 algorithm to recover the private key.
3012 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3017 * Timing vulnerability in ECDSA signature generation
3019 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3020 timing side channel attack. An attacker could use variations in the signing
3021 algorithm to recover the private key.
3023 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3028 * Add coordinate blinding for EC_POINT and implement projective
3029 coordinate blinding for generic prime curves as a countermeasure to
3030 chosen point SCA attacks.
3032 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3034 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3036 * Client DoS due to large DH parameter
3038 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3039 malicious server can send a very large prime value to the client. This will
3040 cause the client to spend an unreasonably long period of time generating a
3041 key for this prime resulting in a hang until the client has finished. This
3042 could be exploited in a Denial Of Service attack.
3044 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3049 * Cache timing vulnerability in RSA Key Generation
3051 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3052 a cache timing side channel attack. An attacker with sufficient access to
3053 mount cache timing attacks during the RSA key generation process could
3054 recover the private key.
3056 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3057 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3062 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3063 parameter is no longer accepted, as it leads to a corrupt table. NULL
3064 pem_str is reserved for alias entries only.
3068 * Revert blinding in ECDSA sign and instead make problematic addition
3069 length-invariant. Switch even to fixed-length Montgomery multiplication.
3073 * Change generating and checking of primes so that the error rate of not
3074 being prime depends on the intended use based on the size of the input.
3075 For larger primes this will result in more rounds of Miller-Rabin.
3076 The maximal error rate for primes with more than 1080 bits is lowered
3079 *Kurt Roeckx, Annie Yousar*
3081 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3085 * Add blinding to ECDSA and DSA signatures to protect against side channel
3086 attacks discovered by Keegan Ryan (NCC Group).
3090 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3091 now allow empty (zero character) pass phrases.
3095 * Certificate time validation (X509_cmp_time) enforces stricter
3096 compliance with RFC 5280. Fractional seconds and timezone offsets
3097 are no longer allowed.
3101 * Fixed a text canonicalisation bug in CMS
3103 Where a CMS detached signature is used with text content the text goes
3104 through a canonicalisation process first prior to signing or verifying a
3105 signature. This process strips trailing space at the end of lines, converts
3106 line terminators to CRLF and removes additional trailing line terminators
3107 at the end of a file. A bug in the canonicalisation process meant that
3108 some characters, such as form-feed, were incorrectly treated as whitespace
3109 and removed. This is contrary to the specification (RFC5485). This fix
3110 could mean that detached text data signed with an earlier version of
3111 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3112 signed with a fixed OpenSSL may fail to verify with an earlier version of
3113 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3114 and use the "-binary" flag (for the "cms" command line application) or set
3115 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3119 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3121 * Constructed ASN.1 types with a recursive definition could exceed the stack
3123 Constructed ASN.1 types with a recursive definition (such as can be found
3124 in PKCS7) could eventually exceed the stack given malicious input with
3125 excessive recursion. This could result in a Denial Of Service attack. There
3126 are no such structures used within SSL/TLS that come from untrusted sources
3127 so this is considered safe.
3129 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3135 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3137 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3138 effectively reduced to only comparing the least significant bit of each
3139 byte. This allows an attacker to forge messages that would be considered as
3140 authenticated in an amount of tries lower than that guaranteed by the
3141 security claims of the scheme. The module can only be compiled by the
3142 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3144 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3150 * Add a build target 'build_all_generated', to build all generated files
3151 and only that. This can be used to prepare everything that requires
3152 things like perl for a system that lacks perl and then move everything
3153 to that system and do the rest of the build there.
3157 * Backport SSL_OP_NO_RENGOTIATION
3159 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3160 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3161 changes this is no longer possible in 1.1.0. Therefore the new
3162 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3163 1.1.0 to provide equivalent functionality.
3165 Note that if an application built against 1.1.0h headers (or above) is run
3166 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3167 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3171 * Removed the OS390-Unix config target. It relied on a script that doesn't
3176 * rsaz_1024_mul_avx2 overflow bug on x86_64
3178 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3179 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3180 Analysis suggests that attacks against RSA and DSA as a result of this
3181 defect would be very difficult to perform and are not believed likely.
3182 Attacks against DH1024 are considered just feasible, because most of the
3183 work necessary to deduce information about a private key may be performed
3184 offline. The amount of resources required for such an attack would be
3185 significant. However, for an attack on TLS to be meaningful, the server
3186 would have to share the DH1024 private key among multiple clients, which is
3187 no longer an option since CVE-2016-0701.
3189 This only affects processors that support the AVX2 but not ADX extensions
3190 like Intel Haswell (4th generation).
3192 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3193 was originally found via the OSS-Fuzz project.
3198 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3200 * bn_sqrx8x_internal carry bug on x86_64
3202 There is a carry propagating bug in the x86_64 Montgomery squaring
3203 procedure. No EC algorithms are affected. Analysis suggests that attacks
3204 against RSA and DSA as a result of this defect would be very difficult to
3205 perform and are not believed likely. Attacks against DH are considered just
3206 feasible (although very difficult) because most of the work necessary to
3207 deduce information about a private key may be performed offline. The amount
3208 of resources required for such an attack would be very significant and
3209 likely only accessible to a limited number of attackers. An attacker would
3210 additionally need online access to an unpatched system using the target
3211 private key in a scenario with persistent DH parameters and a private
3212 key that is shared between multiple clients.
3214 This only affects processors that support the BMI1, BMI2 and ADX extensions
3215 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3217 This issue was reported to OpenSSL by the OSS-Fuzz project.
3222 * Malformed X.509 IPAddressFamily could cause OOB read
3224 If an X.509 certificate has a malformed IPAddressFamily extension,
3225 OpenSSL could do a one-byte buffer overread. The most likely result
3226 would be an erroneous display of the certificate in text format.
3228 This issue was reported to OpenSSL by the OSS-Fuzz project.
3233 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3235 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3236 platform rather than 'mingw'.
3240 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3241 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3242 which is the minimum version we support.
3246 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3248 * Encrypt-Then-Mac renegotiation crash
3250 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3251 negotiated where it was not in the original handshake (or vice-versa) then
3252 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3253 and servers are affected.
3255 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3260 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3262 * Truncated packet could crash via OOB read
3264 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3265 cipher is being used, then a truncated packet can cause that host to
3266 perform an out-of-bounds read, usually resulting in a crash.
3268 This issue was reported to OpenSSL by Robert Święcki of Google.
3273 * Bad (EC)DHE parameters cause a client crash
3275 If a malicious server supplies bad parameters for a DHE or ECDHE key
3276 exchange then this can result in the client attempting to dereference a
3277 NULL pointer leading to a client crash. This could be exploited in a Denial
3280 This issue was reported to OpenSSL by Guido Vranken.
3285 * BN_mod_exp may produce incorrect results on x86_64
3287 There is a carry propagating bug in the x86_64 Montgomery squaring
3288 procedure. No EC algorithms are affected. Analysis suggests that attacks
3289 against RSA and DSA as a result of this defect would be very difficult to
3290 perform and are not believed likely. Attacks against DH are considered just
3291 feasible (although very difficult) because most of the work necessary to
3292 deduce information about a private key may be performed offline. The amount
3293 of resources required for such an attack would be very significant and
3294 likely only accessible to a limited number of attackers. An attacker would
3295 additionally need online access to an unpatched system using the target
3296 private key in a scenario with persistent DH parameters and a private
3297 key that is shared between multiple clients. For example this can occur by
3298 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3299 similar to CVE-2015-3193 but must be treated as a separate problem.
3301 This issue was reported to OpenSSL by the OSS-Fuzz project.
3306 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3308 * ChaCha20/Poly1305 heap-buffer-overflow
3310 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3311 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3312 crash. This issue is not considered to be exploitable beyond a DoS.
3314 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3319 * CMS Null dereference
3321 Applications parsing invalid CMS structures can crash with a NULL pointer
3322 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3323 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3324 structure callback if an attempt is made to free certain invalid encodings.
3325 Only CHOICE structures using a callback which do not handle NULL value are
3328 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3333 * Montgomery multiplication may produce incorrect results
3335 There is a carry propagating bug in the Broadwell-specific Montgomery
3336 multiplication procedure that handles input lengths divisible by, but
3337 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3338 and DH private keys are impossible. This is because the subroutine in
3339 question is not used in operations with the private key itself and an input
3340 of the attacker's direct choice. Otherwise the bug can manifest itself as
3341 transient authentication and key negotiation failures or reproducible
3342 erroneous outcome of public-key operations with specially crafted input.
3343 Among EC algorithms only Brainpool P-512 curves are affected and one
3344 presumably can attack ECDH key negotiation. Impact was not analyzed in
3345 detail, because pre-requisites for attack are considered unlikely. Namely
3346 multiple clients have to choose the curve in question and the server has to
3347 share the private key among them, neither of which is default behaviour.
3348 Even then only clients that chose the curve will be affected.
3350 This issue was publicly reported as transient failures and was not
3351 initially recognized as a security issue. Thanks to Richard Morgan for
3352 providing reproducible case.
3357 * Removed automatic addition of RPATH in shared libraries and executables,
3358 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3362 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3364 * Fix Use After Free for large message sizes
3366 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3367 message larger than approx 16k is received then the underlying buffer to
3368 store the incoming message is reallocated and moved. Unfortunately a
3369 dangling pointer to the old location is left which results in an attempt to
3370 write to the previously freed location. This is likely to result in a
3371 crash, however it could potentially lead to execution of arbitrary code.
3373 This issue only affects OpenSSL 1.1.0a.
3375 This issue was reported to OpenSSL by Robert Święcki.
3380 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3382 * OCSP Status Request extension unbounded memory growth
3384 A malicious client can send an excessively large OCSP Status Request
3385 extension. If that client continually requests renegotiation, sending a
3386 large OCSP Status Request extension each time, then there will be unbounded
3387 memory growth on the server. This will eventually lead to a Denial Of
3388 Service attack through memory exhaustion. Servers with a default
3389 configuration are vulnerable even if they do not support OCSP. Builds using
3390 the "no-ocsp" build time option are not affected.
3392 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3397 * SSL_peek() hang on empty record
3399 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3400 sends an empty record. This could be exploited by a malicious peer in a
3401 Denial Of Service attack.
3403 This issue was reported to OpenSSL by Alex Gaynor.
3408 * Excessive allocation of memory in tls_get_message_header() and
3409 dtls1_preprocess_fragment()
3411 A (D)TLS message includes 3 bytes for its length in the header for the
3412 message. This would allow for messages up to 16Mb in length. Messages of
3413 this length are excessive and OpenSSL includes a check to ensure that a
3414 peer is sending reasonably sized messages in order to avoid too much memory
3415 being consumed to service a connection. A flaw in the logic of version
3416 1.1.0 means that memory for the message is allocated too early, prior to
3417 the excessive message length check. Due to way memory is allocated in
3418 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3419 to service a connection. This could lead to a Denial of Service through
3420 memory exhaustion. However, the excessive message length check still takes
3421 place, and this would cause the connection to immediately fail. Assuming
3422 that the application calls SSL_free() on the failed connection in a timely
3423 manner then the 21Mb of allocated memory will then be immediately freed
3424 again. Therefore the excessive memory allocation will be transitory in
3425 nature. This then means that there is only a security impact if:
3427 1) The application does not call SSL_free() in a timely manner in the event
3428 that the connection fails
3430 2) The application is working in a constrained environment where there is
3431 very little free memory
3433 3) The attacker initiates multiple connection attempts such that there are
3434 multiple connections in a state where memory has been allocated for the
3435 connection; SSL_free() has not yet been called; and there is insufficient
3436 memory to service the multiple requests.
3438 Except in the instance of (1) above any Denial Of Service is likely to be
3439 transitory because as soon as the connection fails the memory is
3440 subsequently freed again in the SSL_free() call. However there is an
3441 increased risk during this period of application crashes due to the lack of
3442 memory - which would then mean a more serious Denial of Service.
3444 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3445 (CVE-2016-6307 and CVE-2016-6308)
3449 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3450 had to be removed. Primary reason is that vendor assembler can't
3451 assemble our modules with -KPIC flag. As result it, assembly
3452 support, was not even available as option. But its lack means
3453 lack of side-channel resistant code, which is incompatible with
3454 security by todays standards. Fortunately gcc is readily available
3455 prepackaged option, which we firmly point at...
3459 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3461 * Windows command-line tool supports UTF-8 opt-in option for arguments
3462 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3463 (to any value) allows Windows user to access PKCS#12 file generated
3464 with Windows CryptoAPI and protected with non-ASCII password, as well
3465 as files generated under UTF-8 locale on Linux also protected with
3470 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3471 have been disabled by default and removed from DEFAULT, just like RC4.
3472 See the RC4 item below to re-enable both.
3476 * The method for finding the storage location for the Windows RAND seed file
3477 has changed. First we check %RANDFILE%. If that is not set then we check
3478 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3479 all else fails we fall back to C:\.
3483 * The EVP_EncryptUpdate() function has had its return type changed from void
3484 to int. A return of 0 indicates and error while a return of 1 indicates
3489 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3490 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3491 off the constant time implementation for RSA, DSA and DH have been made
3492 no-ops and deprecated.
3496 * Windows RAND implementation was simplified to only get entropy by
3497 calling CryptGenRandom(). Various other RAND-related tickets
3500 *Joseph Wylie Yandle, Rich Salz*
3502 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3503 and `OPENSSL_LH_`, respectively. The old names are available
3504 with API compatibility. They new names are now completely documented.
3508 * Unify TYPE_up_ref(obj) methods signature.
3509 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3510 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3511 int (instead of void) like all others TYPE_up_ref() methods.
3512 So now these methods also check the return value of CRYPTO_atomic_add(),
3513 and the validity of object reference counter.
3515 *fdasilvayy@gmail.com*
3517 * With Windows Visual Studio builds, the .pdb files are installed
3518 alongside the installed libraries and executables. For a static
3519 library installation, ossl_static.pdb is the associate compiler
3520 generated .pdb file to be used when linking programs.
3524 * Remove openssl.spec. Packaging files belong with the packagers.
3528 * Automatic Darwin/OSX configuration has had a refresh, it will now
3529 recognise x86_64 architectures automatically. You can still decide
3530 to build for a different bitness with the environment variable
3531 KERNEL_BITS (can be 32 or 64), for example:
3533 KERNEL_BITS=32 ./config
3537 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3538 256 bit AES and HMAC with SHA256.
3542 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3546 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3550 * To enable users to have their own config files and build file templates,
3551 Configure looks in the directory indicated by the environment variable
3552 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3553 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3554 name and is used as is.
3558 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3559 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3560 X509_CERT_FILE_CTX was removed.
3564 * "shared" builds are now the default. To create only static libraries use
3565 the "no-shared" Configure option.
3569 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3570 All of these option have not worked for some while and are fundamental
3575 * Make various cleanup routines no-ops and mark them as deprecated. Most
3576 global cleanup functions are no longer required because they are handled
3577 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3578 Explicitly de-initing can cause problems (e.g. where a library that uses
3579 OpenSSL de-inits, but an application is still using it). The affected
3580 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3581 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3582 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3583 COMP_zlib_cleanup().
3587 * --strict-warnings no longer enables runtime debugging options
3588 such as REF_DEBUG. Instead, debug options are automatically
3589 enabled with '--debug' builds.
3591 *Andy Polyakov, Emilia Käsper*
3593 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3594 have been moved out of the public header files. New functions for managing
3595 these have been added.
3599 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3600 objects have been moved out of the public header files. New
3601 functions for managing these have been added.
3605 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3606 have been moved out of the public header files. New functions for managing
3607 these have been added.
3611 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3612 moved out of the public header files. New functions for managing these
3617 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3621 * Removed the mk1mf build scripts.
3625 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3626 it is always safe to #include a header now.
3630 * Removed the aged BC-32 config and all its supporting scripts
3634 * Removed support for Ultrix, Netware, and OS/2.
3638 * Add support for HKDF.
3640 *Alessandro Ghedini*
3642 * Add support for blake2b and blake2s
3646 * Added support for "pipelining". Ciphers that have the
3647 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3648 encryptions/decryptions simultaneously. There are currently no built-in
3649 ciphers with this property but the expectation is that engines will be able
3650 to offer it to significantly improve throughput. Support has been extended
3651 into libssl so that multiple records for a single connection can be
3652 processed in one go (for >=TLS 1.1).
3656 * Added the AFALG engine. This is an async capable engine which is able to
3657 offload work to the Linux kernel. In this initial version it only supports
3658 AES128-CBC. The kernel must be version 4.1.0 or greater.
3662 * OpenSSL now uses a new threading API. It is no longer necessary to
3663 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3664 are two supported threading models: pthreads and windows threads. It is
3665 also possible to configure OpenSSL at compile time for "no-threads". The
3666 old threading API should no longer be used. The functions have been
3667 replaced with "no-op" compatibility macros.
3669 *Alessandro Ghedini, Matt Caswell*
3671 * Modify behavior of ALPN to invoke callback after SNI/servername
3672 callback, such that updates to the SSL_CTX affect ALPN.
3676 * Add SSL_CIPHER queries for authentication and key-exchange.
3680 * Changes to the DEFAULT cipherlist:
3681 - Prefer (EC)DHE handshakes over plain RSA.
3682 - Prefer AEAD ciphers over legacy ciphers.
3683 - Prefer ECDSA over RSA when both certificates are available.
3684 - Prefer TLSv1.2 ciphers/PRF.
3685 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3690 * Change the ECC default curve list to be this, in order: x25519,
3691 secp256r1, secp521r1, secp384r1.
3695 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3696 disabled by default. They can be re-enabled using the
3697 enable-weak-ssl-ciphers option to Configure.
3701 * If the server has ALPN configured, but supports no protocols that the
3702 client advertises, send a fatal "no_application_protocol" alert.
3703 This behaviour is SHALL in RFC 7301, though it isn't universally
3704 implemented by other servers.
3708 * Add X25519 support.
3709 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3710 for public and private key encoding using the format documented in
3711 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3712 key generation and key derivation.
3714 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3719 * Deprecate SRP_VBASE_get_by_user.
3720 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3721 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3722 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3723 seed, even if the seed is configured.
3725 Users should use SRP_VBASE_get1_by_user instead. Note that in
3726 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3727 also that even though configuring the SRP seed attempts to hide
3728 invalid usernames by continuing the handshake with fake
3729 credentials, this behaviour is not constant time and no strong
3730 guarantees are made that the handshake is indistinguishable from
3731 that of a valid user.
3735 * Configuration change; it's now possible to build dynamic engines
3736 without having to build shared libraries and vice versa. This
3737 only applies to the engines in `engines/`, those in `crypto/engine/`
3738 will always be built into libcrypto (i.e. "static").
3740 Building dynamic engines is enabled by default; to disable, use
3741 the configuration option "disable-dynamic-engine".
3743 The only requirements for building dynamic engines are the
3744 presence of the DSO module and building with position independent
3745 code, so they will also automatically be disabled if configuring
3746 with "disable-dso" or "disable-pic".
3748 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3749 are also taken away from openssl/opensslconf.h, as they are
3754 * Configuration change; if there is a known flag to compile
3755 position independent code, it will always be applied on the
3756 libcrypto and libssl object files, and never on the application
3757 object files. This means other libraries that use routines from
3758 libcrypto / libssl can be made into shared libraries regardless
3759 of how OpenSSL was configured.
3761 If this isn't desirable, the configuration options "disable-pic"
3762 or "no-pic" can be used to disable the use of PIC. This will
3763 also disable building shared libraries and dynamic engines.
3767 * Removed JPAKE code. It was experimental and has no wide use.
3771 * The INSTALL_PREFIX Makefile variable has been renamed to
3772 DESTDIR. That makes for less confusion on what this variable
3773 is for. Also, the configuration option --install_prefix is
3778 * Heartbeat for TLS has been removed and is disabled by default
3779 for DTLS; configure with enable-heartbeats. Code that uses the
3780 old #define's might need to be updated.
3782 *Emilia Käsper, Rich Salz*
3784 * Rename REF_CHECK to REF_DEBUG.
3788 * New "unified" build system
3790 The "unified" build system is aimed to be a common system for all
3791 platforms we support. With it comes new support for VMS.
3793 This system builds supports building in a different directory tree
3794 than the source tree. It produces one Makefile (for unix family
3795 or lookalikes), or one descrip.mms (for VMS).
3797 The source of information to make the Makefile / descrip.mms is
3798 small files called 'build.info', holding the necessary
3799 information for each directory with source to compile, and a
3800 template in Configurations, like unix-Makefile.tmpl or
3803 With this change, the library names were also renamed on Windows
3804 and on VMS. They now have names that are closer to the standard
3805 on Unix, and include the major version number, and in certain
3806 cases, the architecture they are built for. See "Notes on shared
3807 libraries" in INSTALL.
3809 We rely heavily on the perl module Text::Template.
3813 * Added support for auto-initialisation and de-initialisation of the library.
3814 OpenSSL no longer requires explicit init or deinit routines to be called,
3815 except in certain circumstances. See the OPENSSL_init_crypto() and
3816 OPENSSL_init_ssl() man pages for further information.
3820 * The arguments to the DTLSv1_listen function have changed. Specifically the
3821 "peer" argument is now expected to be a BIO_ADDR object.
3823 * Rewrite of BIO networking library. The BIO library lacked consistent
3824 support of IPv6, and adding it required some more extensive
3825 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3826 which hold all types of addresses and chains of address information.
3827 It also introduces a new API, with functions like BIO_socket,
3828 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3829 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3830 have been adapted accordingly.
3834 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3839 * CRIME protection: disable compression by default, even if OpenSSL is
3840 compiled with zlib enabled. Applications can still enable compression
3841 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3842 using the SSL_CONF library to configure compression.
3846 * The signature of the session callback configured with
3847 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3848 was explicitly marked as `const unsigned char*` instead of
3853 * Always DPURIFY. Remove the use of uninitialized memory in the
3854 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3858 * Removed many obsolete configuration items, including
3859 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3860 MD2_CHAR, MD2_INT, MD2_LONG
3862 IDEA_SHORT, IDEA_LONG
3863 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3865 *Rich Salz, with advice from Andy Polyakov*
3867 * Many BN internals have been moved to an internal header file.
3869 *Rich Salz with help from Andy Polyakov*
3871 * Configuration and writing out the results from it has changed.
3872 Files such as Makefile include/openssl/opensslconf.h and are now
3873 produced through general templates, such as Makefile.in and
3874 crypto/opensslconf.h.in and some help from the perl module
3877 Also, the center of configuration information is no longer
3878 Makefile. Instead, Configure produces a perl module in
3879 configdata.pm which holds most of the config data (in the hash
3880 table %config), the target data that comes from the target
3881 configuration in one of the `Configurations/*.conf` files (in
3886 * To clarify their intended purposes, the Configure options
3887 --prefix and --openssldir change their semantics, and become more
3888 straightforward and less interdependent.
3890 --prefix shall be used exclusively to give the location INSTALLTOP
3891 where programs, scripts, libraries, include files and manuals are
3892 going to be installed. The default is now /usr/local.
3894 --openssldir shall be used exclusively to give the default
3895 location OPENSSLDIR where certificates, private keys, CRLs are
3896 managed. This is also where the default openssl.cnf gets
3898 If the directory given with this option is a relative path, the
3899 values of both the --prefix value and the --openssldir value will
3900 be combined to become OPENSSLDIR.
3901 The default for --openssldir is INSTALLTOP/ssl.
3903 Anyone who uses --openssldir to specify where OpenSSL is to be
3904 installed MUST change to use --prefix instead.
3908 * The GOST engine was out of date and therefore it has been removed. An up
3909 to date GOST engine is now being maintained in an external repository.
3910 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3911 support for GOST ciphersuites (these are only activated if a GOST engine
3916 * EGD is no longer supported by default; use enable-egd when
3919 *Ben Kaduk and Rich Salz*
3921 * The distribution now has Makefile.in files, which are used to
3922 create Makefile's when Configure is run. *Configure must be run
3923 before trying to build now.*
3927 * The return value for SSL_CIPHER_description() for error conditions
3932 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3934 Obtaining and performing DNSSEC validation of TLSA records is
3935 the application's responsibility. The application provides
3936 the TLSA records of its choice to OpenSSL, and these are then
3937 used to authenticate the peer.
3939 The TLSA records need not even come from DNS. They can, for
3940 example, be used to implement local end-entity certificate or
3941 trust-anchor "pinning", where the "pin" data takes the form
3942 of TLSA records, which can augment or replace verification
3943 based on the usual WebPKI public certification authorities.
3947 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3948 continues to support deprecated interfaces in default builds.
3949 However, applications are strongly advised to compile their
3950 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3951 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3952 or the 1.1.0 releases.
3954 In environments in which all applications have been ported to
3955 not use any deprecated interfaces OpenSSL's Configure script
3956 should be used with the --api=1.1.0 option to entirely remove
3957 support for the deprecated features from the library and
3958 unconditionally disable them in the installed headers.
3959 Essentially the same effect can be achieved with the "no-deprecated"
3960 argument to Configure, except that this will always restrict
3961 the build to just the latest API, rather than a fixed API
3964 As applications are ported to future revisions of the API,
3965 they should update their compile-time OPENSSL_API_COMPAT define
3966 accordingly, but in most cases should be able to continue to
3967 compile with later releases.
3969 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3970 0x10000000L and 0x00908000L, respectively. However those
3971 versions did not support the OPENSSL_API_COMPAT feature, and
3972 so applications are not typically tested for explicit support
3973 of just the undeprecated features of either release.
3977 * Add support for setting the minimum and maximum supported protocol.
3978 It can bet set via the SSL_set_min_proto_version() and
3979 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3980 MaxProtocol. It's recommended to use the new APIs to disable
3981 protocols instead of disabling individual protocols using
3982 SSL_set_options() or SSL_CONF's Protocol. This change also
3983 removes support for disabling TLS 1.2 in the OpenSSL TLS
3984 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3988 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3992 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3993 and integrates ECDSA and ECDH functionality into EC. Implementations can
3994 now redirect key generation and no longer need to convert to or from
3997 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3998 include the ec.h header file instead.
4002 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4003 ciphers who are no longer supported and drops support the ephemeral RSA key
4004 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4008 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4009 opaque. For HMAC_CTX, the following constructors and destructors
4012 HMAC_CTX *HMAC_CTX_new(void);
4013 void HMAC_CTX_free(HMAC_CTX *ctx);
4015 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4016 destroy such methods has been added. See EVP_MD_meth_new(3) and
4017 EVP_CIPHER_meth_new(3) for documentation.
4020 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4021 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4022 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4023 an already created structure.
4024 2) For consistency with the majority of our object creators and
4025 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4026 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4027 for deprecated builds.
4031 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4032 cryptographic operations to be performed asynchronously as long as an
4033 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4034 further details. Libssl has also had this capability integrated with the
4035 introduction of the new mode SSL_MODE_ASYNC and associated error
4036 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4037 pages. This work was developed in partnership with Intel Corp.
4041 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4042 always enabled now. If you want to disable the support you should
4043 exclude it using the list of supported ciphers. This also means that the
4044 "-no_ecdhe" option has been removed from s_server.
4048 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4049 SSL_{CTX_}set1_curves() which can set a list.
4053 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4054 curve you want to support using SSL_{CTX_}set1_curves().
4058 * State machine rewrite. The state machine code has been significantly
4059 refactored in order to remove much duplication of code and solve issues
4060 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4061 further details). This change does have some associated API changes.
4062 Notably the SSL_state() function has been removed and replaced by
4063 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4064 SSL_set_state() has been removed altogether. The previous handshake states
4065 defined in ssl.h and ssl3.h have also been removed.
4069 * All instances of the string "ssleay" in the public API were replaced
4070 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4071 Some error codes related to internal RSA_eay API's were renamed.
4075 * The demo files in crypto/threads were moved to demo/threads.
4079 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4082 *Matt Caswell, Rich Salz*
4084 * New ASN.1 embed macro.
4086 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4087 structure is not allocated: it is part of the parent. That is instead of
4095 This reduces memory fragmentation and make it impossible to accidentally
4096 set a mandatory field to NULL.
4098 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4099 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4100 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4105 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4109 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4110 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4111 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4112 DES and RC4 ciphersuites.
4116 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4117 This changes the decoding behaviour for some invalid messages,
4118 though the change is mostly in the more lenient direction, and
4119 legacy behaviour is preserved as much as possible.
4123 * Fix no-stdio build.
4124 *David Woodhouse <David.Woodhouse@intel.com> and also*
4125 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4127 * New testing framework
4128 The testing framework has been largely rewritten and is now using
4129 perl and the perl modules Test::Harness and an extended variant of
4130 Test::More called OpenSSL::Test to do its work. All test scripts in
4131 test/ have been rewritten into test recipes, and all direct calls to
4132 executables in test/Makefile have become individual recipes using the
4133 simplified testing OpenSSL::Test::Simple.
4135 For documentation on our testing modules, do:
4137 perldoc test/testlib/OpenSSL/Test/Simple.pm
4138 perldoc test/testlib/OpenSSL/Test.pm
4142 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4143 are used; the latter aborts on memory leaks (usually checked on exit).
4144 Some undocumented "set malloc, etc., hooks" functions were removed
4145 and others were changed. All are now documented.
4149 * In DSA_generate_parameters_ex, if the provided seed is too short,
4152 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4154 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4155 from RFC4279, RFC4785, RFC5487, RFC5489.
4157 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4158 original RSA_PSK patch.
4162 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4163 era flag was never set throughout the codebase (only read). Also removed
4164 SSL3_FLAGS_POP_BUFFER which was only used if
4165 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4169 * Changed the default name options in the "ca", "crl", "req" and "x509"
4170 to be "oneline" instead of "compat".
4174 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4175 not aware of clients that still exhibit this bug, and the workaround
4176 hasn't been working properly for a while.
4180 * The return type of BIO_number_read() and BIO_number_written() as well as
4181 the corresponding num_read and num_write members in the BIO structure has
4182 changed from unsigned long to uint64_t. On platforms where an unsigned
4183 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4188 * Given the pervasive nature of TLS extensions it is inadvisable to run
4189 OpenSSL without support for them. It also means that maintaining
4190 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4191 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4195 * Removed support for the two export grade static DH ciphersuites
4196 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4197 were newly added (along with a number of other static DH ciphersuites) to
4198 1.0.2. However the two export ones have *never* worked since they were
4199 introduced. It seems strange in any case to be adding new export
4200 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4204 * Version negotiation has been rewritten. In particular SSLv23_method(),
4205 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4206 and turned into macros which simply call the new preferred function names
4207 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4208 should use the new names instead. Also as part of this change the ssl23.h
4209 header file has been removed.
4213 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4214 code and the associated standard is no longer considered fit-for-purpose.
4218 * RT2547 was closed. When generating a private key, try to make the
4219 output file readable only by the owner. This behavior change might
4220 be noticeable when interacting with other software.
4222 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4227 * Added HTTP GET support to the ocsp command.
4231 * Changed default digest for the dgst and enc commands from MD5 to
4236 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4240 * Added support for TLS extended master secret from
4241 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4242 initial patch which was a great help during development.
4246 * All libssl internal structures have been removed from the public header
4247 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4248 now redundant). Users should not attempt to access internal structures
4249 directly. Instead they should use the provided API functions.
4253 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4254 Access to deprecated functions can be re-enabled by running config with
4255 "enable-deprecated". In addition applications wishing to use deprecated
4256 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4257 will, by default, disable some transitive includes that previously existed
4258 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4262 * Added support for OCB mode. OpenSSL has been granted a patent license
4263 compatible with the OpenSSL license for use of OCB. Details are available
4264 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4265 for OCB can be removed by calling config with no-ocb.
4269 * SSLv2 support has been removed. It still supports receiving a SSLv2
4270 compatible client hello.
4274 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4275 done while fixing the error code for the key-too-small case.
4277 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4279 * CA.sh has been removed; use CA.pl instead.
4283 * Removed old DES API.
4287 * Remove various unsupported platforms:
4293 Sinix/ReliantUNIX RM400
4298 16-bit platforms such as WIN16
4302 * Clean up OPENSSL_NO_xxx #define's
4303 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4304 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4305 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4306 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4307 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4308 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4309 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4310 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4311 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4312 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4316 * Cleaned up dead code
4317 Remove all but one '#ifdef undef' which is to be looked at.
4321 * Clean up calling of xxx_free routines.
4322 Just like free(), fix most of the xxx_free routines to accept
4323 NULL. Remove the non-null checks from callers. Save much code.
4327 * Add secure heap for storage of private keys (when possible).
4328 Add BIO_s_secmem(), CBIGNUM, etc.
4329 Contributed by Akamai Technologies under our Corporate CLA.
4333 * Experimental support for a new, fast, unbiased prime candidate generator,
4334 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4336 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4338 * New output format NSS in the sess_id command line tool. This allows
4339 exporting the session id and the master key in NSS keylog format.
4341 *Martin Kaiser <martin@kaiser.cx>*
4343 * Harmonize version and its documentation. -f flag is used to display
4346 *mancha <mancha1@zoho.com>*
4348 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4349 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4351 *mancha <mancha1@zoho.com>*
4353 * Fix some double frees. These are not thought to be exploitable.
4355 *mancha <mancha1@zoho.com>*
4357 * A missing bounds check in the handling of the TLS heartbeat extension
4358 can be used to reveal up to 64k of memory to a connected client or
4361 Thanks for Neel Mehta of Google Security for discovering this bug and to
4362 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4363 preparing the fix ([CVE-2014-0160])
4365 *Adam Langley, Bodo Moeller*
4367 * Fix for the attack described in the paper "Recovering OpenSSL
4368 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4369 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4370 <http://eprint.iacr.org/2014/140>
4372 Thanks to Yuval Yarom and Naomi Benger for discovering this
4373 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4375 *Yuval Yarom and Naomi Benger*
4377 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4378 this fixes a limitation in previous versions of OpenSSL.
4382 * Experimental encrypt-then-mac support.
4384 Experimental support for encrypt then mac from
4385 draft-gutmann-tls-encrypt-then-mac-02.txt
4387 To enable it set the appropriate extension number (0x42 for the test
4388 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4390 For non-compliant peers (i.e. just about everything) this should have no
4393 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4397 * Add EVP support for key wrapping algorithms, to avoid problems with
4398 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4399 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4400 algorithms and include tests cases.
4404 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4409 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4410 MGF1 digest and OAEP label.
4414 * Make openssl verify return errors.
4416 *Chris Palmer <palmer@google.com> and Ben Laurie*
4418 * New function ASN1_TIME_diff to calculate the difference between two
4419 ASN1_TIME structures or one structure and the current time.
4423 * Update fips_test_suite to support multiple command line options. New
4424 test to induce all self test errors in sequence and check expected
4429 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4430 sign or verify all in one operation.
4434 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4435 test programs and fips_test_suite. Includes functionality to parse
4436 the minimal script output of fipsalgest.pl directly.
4440 * Add authorisation parameter to FIPS_module_mode_set().
4444 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4448 * Use separate DRBG fields for internal and external flags. New function
4449 FIPS_drbg_health_check() to perform on demand health checking. Add
4450 generation tests to fips_test_suite with reduced health check interval to
4451 demonstrate periodic health checking. Add "nodh" option to
4452 fips_test_suite to skip very slow DH test.
4456 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4461 * More extensive health check for DRBG checking many more failure modes.
4462 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4463 combination: call this in fips_test_suite.
4467 * Add support for canonical generation of DSA parameter 'g'. See
4470 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4471 POST to handle HMAC cases.
4475 * Add functions FIPS_module_version() and FIPS_module_version_text()
4476 to return numerical and string versions of the FIPS module number.
4480 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4481 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4482 outside the validated module in the FIPS capable OpenSSL.
4486 * Minor change to DRBG entropy callback semantics. In some cases
4487 there is no multiple of the block length between min_len and
4488 max_len. Allow the callback to return more than max_len bytes
4489 of entropy but discard any extra: it is the callback's responsibility
4490 to ensure that the extra data discarded does not impact the
4491 requested amount of entropy.
4495 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4496 information in FIPS186-3, SP800-57 and SP800-131A.
4500 * CCM support via EVP. Interface is very similar to GCM case except we
4501 must supply all data in one chunk (i.e. no update, final) and the
4502 message length must be supplied if AAD is used. Add algorithm test
4507 * Initial version of POST overhaul. Add POST callback to allow the status
4508 of POST to be monitored and/or failures induced. Modify fips_test_suite
4509 to use callback. Always run all selftests even if one fails.
4513 * XTS support including algorithm test driver in the fips_gcmtest program.
4514 Note: this does increase the maximum key length from 32 to 64 bytes but
4515 there should be no binary compatibility issues as existing applications
4516 will never use XTS mode.
4520 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4521 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4522 performs algorithm blocking for unapproved PRNG types. Also do not
4523 set PRNG type in FIPS_mode_set(): leave this to the application.
4524 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4525 the standard OpenSSL PRNG: set additional data to a date time vector.
4529 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4530 This shouldn't present any incompatibility problems because applications
4531 shouldn't be using these directly and any that are will need to rethink
4532 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4536 * Extensive self tests and health checking required by SP800-90 DRBG.
4537 Remove strength parameter from FIPS_drbg_instantiate and always
4538 instantiate at maximum supported strength.
4542 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4546 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4550 * New function DH_compute_key_padded() to compute a DH key and pad with
4551 leading zeroes if needed: this complies with SP800-56A et al.
4555 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4556 anything, incomplete, subject to change and largely untested at present.
4560 * Modify fipscanisteronly build option to only build the necessary object
4561 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4565 * Add experimental option FIPSSYMS to give all symbols in
4566 fipscanister.o and FIPS or fips prefix. This will avoid
4567 conflicts with future versions of OpenSSL. Add perl script
4568 util/fipsas.pl to preprocess assembly language source files
4569 and rename any affected symbols.
4573 * Add selftest checks and algorithm block of non-fips algorithms in
4574 FIPS mode. Remove DES2 from selftests.
4578 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4579 return internal method without any ENGINE dependencies. Add new
4580 tiny fips sign and verify functions.
4584 * New build option no-ec2m to disable characteristic 2 code.
4588 * New build option "fipscanisteronly". This only builds fipscanister.o
4589 and (currently) associated fips utilities. Uses the file Makefile.fips
4590 instead of Makefile.org as the prototype.
4594 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4595 Update fips_gcmtest to use IV generator.
4599 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4600 setting output buffer to NULL. The `*Final` function must be
4601 called although it will not retrieve any additional data. The tag
4602 can be set or retrieved with a ctrl. The IV length is by default 12
4603 bytes (96 bits) but can be set to an alternative value. If the IV
4604 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4609 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4610 underlying do_cipher function handles all cipher semantics itself
4611 including padding and finalisation. This is useful if (for example)
4612 an ENGINE cipher handles block padding itself. The behaviour of
4613 do_cipher is subtly changed if this flag is set: the return value
4614 is the number of characters written to the output buffer (zero is
4615 no longer an error code) or a negative error code. Also if the
4616 input buffer is NULL and length 0 finalisation should be performed.
4620 * If a candidate issuer certificate is already part of the constructed
4621 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4625 * Improve forward-security support: add functions
4627 void SSL_CTX_set_not_resumable_session_callback(
4628 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4629 void SSL_set_not_resumable_session_callback(
4630 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4632 for use by SSL/TLS servers; the callback function will be called whenever a
4633 new session is created, and gets to decide whether the session may be
4634 cached to make it resumable (return 0) or not (return 1). (As by the
4635 SSL/TLS protocol specifications, the session_id sent by the server will be
4636 empty to indicate that the session is not resumable; also, the server will
4637 not generate RFC 4507 (RFC 5077) session tickets.)
4639 A simple reasonable callback implementation is to return is_forward_secure.
4640 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4641 by the SSL/TLS server library, indicating whether it can provide forward
4644 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4646 * New -verify_name option in command line utilities to set verification
4651 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4652 Add CMAC pkey methods.
4656 * Experimental renegotiation in s_server -www mode. If the client
4657 browses /reneg connection is renegotiated. If /renegcert it is
4658 renegotiated requesting a certificate.
4662 * Add an "external" session cache for debugging purposes to s_server. This
4663 should help trace issues which normally are only apparent in deployed
4664 multi-process servers.
4668 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4669 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4670 BIO_set_cipher() and some obscure PEM functions were changed so they
4671 can now return an error. The RAND changes required a change to the
4672 RAND_METHOD structure.
4676 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4677 a gcc attribute to warn if the result of a function is ignored. This
4678 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4679 whose return value is often ignored.
4683 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4684 These allow SCTs (signed certificate timestamps) to be requested and
4685 validated when establishing a connection.
4687 *Rob Percival <robpercival@google.com>*
4692 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4694 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4695 used even when parsing explicit parameters, when loading a encoded key
4696 or calling `EC_GROUP_new_from_ecpkparameters()`/
4697 `EC_GROUP_new_from_ecparameters()`.
4698 This prevents bypass of security hardening and performance gains,
4699 especially for curves with specialized EC_METHODs.
4700 By default, if a key encoded with explicit parameters is loaded and later
4701 encoded, the output is still encoded with explicit parameters, even if
4702 internally a "named" EC_GROUP is used for computation.
4706 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4707 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4708 NULL. After this change, only the cofactor parameter can be NULL. It also
4709 does some minimal sanity checks on the passed order.
4714 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4715 An attack is simple, if the first CMS_recipientInfo is valid but the
4716 second CMS_recipientInfo is chosen ciphertext. If the second
4717 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4718 encryption key will be replaced by garbage, and the message cannot be
4719 decoded, but if the RSA decryption fails, the correct encryption key is
4720 used and the recipient will not notice the attack.
4721 As a work around for this potential attack the length of the decrypted
4722 key must be equal to the cipher default key length, in case the
4723 certifiate is not given and all recipientInfo are tried out.
4724 The old behaviour can be re-enabled in the CMS code by setting the
4725 CMS_DEBUG_DECRYPT flag.
4730 * Document issue with installation paths in diverse Windows builds
4732 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4733 binaries and run-time config file.
4738 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4740 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4741 This changes the size when using the `genpkey` command when no size is given.
4742 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4743 generation commands to use 2048 bits by default.
4747 * Add FIPS support for Android Arm 64-bit
4749 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4750 Module in Version 2.0.10. For some reason, the corresponding target
4751 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4752 built with FIPS support on Android Arm 64-bit. This omission has been
4755 *Matthias St. Pierre*
4757 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4759 * 0-byte record padding oracle
4761 If an application encounters a fatal protocol error and then calls
4762 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4763 then OpenSSL can respond differently to the calling application if a 0 byte
4764 record is received with invalid padding compared to if a 0 byte record is
4765 received with an invalid MAC. If the application then behaves differently
4766 based on that in a way that is detectable to the remote peer, then this
4767 amounts to a padding oracle that could be used to decrypt data.
4769 In order for this to be exploitable "non-stitched" ciphersuites must be in
4770 use. Stitched ciphersuites are optimised implementations of certain
4771 commonly used ciphersuites. Also the application must call SSL_shutdown()
4772 twice even if a protocol error has occurred (applications should not do
4773 this but some do anyway).
4775 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4776 Aviram, with additional investigation by Steven Collison and Andrew
4777 Hourselt. It was reported to OpenSSL on 10th December 2018.
4782 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4786 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4788 * Microarchitecture timing vulnerability in ECC scalar multiplication
4790 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4791 shown to be vulnerable to a microarchitecture timing side channel attack.
4792 An attacker with sufficient access to mount local timing attacks during
4793 ECDSA signature generation could recover the private key.
4795 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4796 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4802 * Timing vulnerability in DSA signature generation
4804 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4805 timing side channel attack. An attacker could use variations in the signing
4806 algorithm to recover the private key.
4808 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4813 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4814 Module, accidentally introduced while backporting security fixes from the
4815 development branch and hindering the use of ECC in FIPS mode.
4819 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4821 * Client DoS due to large DH parameter
4823 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4824 malicious server can send a very large prime value to the client. This will
4825 cause the client to spend an unreasonably long period of time generating a
4826 key for this prime resulting in a hang until the client has finished. This
4827 could be exploited in a Denial Of Service attack.
4829 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4834 * Cache timing vulnerability in RSA Key Generation
4836 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4837 a cache timing side channel attack. An attacker with sufficient access to
4838 mount cache timing attacks during the RSA key generation process could
4839 recover the private key.
4841 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4842 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4847 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4848 parameter is no longer accepted, as it leads to a corrupt table. NULL
4849 pem_str is reserved for alias entries only.
4853 * Revert blinding in ECDSA sign and instead make problematic addition
4854 length-invariant. Switch even to fixed-length Montgomery multiplication.
4858 * Change generating and checking of primes so that the error rate of not
4859 being prime depends on the intended use based on the size of the input.
4860 For larger primes this will result in more rounds of Miller-Rabin.
4861 The maximal error rate for primes with more than 1080 bits is lowered
4864 *Kurt Roeckx, Annie Yousar*
4866 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4870 * Add blinding to ECDSA and DSA signatures to protect against side channel
4871 attacks discovered by Keegan Ryan (NCC Group).
4875 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4876 now allow empty (zero character) pass phrases.
4880 * Certificate time validation (X509_cmp_time) enforces stricter
4881 compliance with RFC 5280. Fractional seconds and timezone offsets
4882 are no longer allowed.
4886 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4888 * Constructed ASN.1 types with a recursive definition could exceed the stack
4890 Constructed ASN.1 types with a recursive definition (such as can be found
4891 in PKCS7) could eventually exceed the stack given malicious input with
4892 excessive recursion. This could result in a Denial Of Service attack. There
4893 are no such structures used within SSL/TLS that come from untrusted sources
4894 so this is considered safe.
4896 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4902 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4904 * Read/write after SSL object in error state
4906 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4907 mechanism. The intent was that if a fatal error occurred during a handshake
4908 then OpenSSL would move into the error state and would immediately fail if
4909 you attempted to continue the handshake. This works as designed for the
4910 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4911 SSL_connect()), however due to a bug it does not work correctly if
4912 SSL_read() or SSL_write() is called directly. In that scenario, if the
4913 handshake fails then a fatal error will be returned in the initial function
4914 call. If SSL_read()/SSL_write() is subsequently called by the application
4915 for the same SSL object then it will succeed and the data is passed without
4916 being decrypted/encrypted directly from the SSL/TLS record layer.
4918 In order to exploit this issue an application bug would have to be present
4919 that resulted in a call to SSL_read()/SSL_write() being issued after having
4920 already received a fatal error.
4922 This issue was reported to OpenSSL by David Benjamin (Google).
4927 * rsaz_1024_mul_avx2 overflow bug on x86_64
4929 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4930 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4931 Analysis suggests that attacks against RSA and DSA as a result of this
4932 defect would be very difficult to perform and are not believed likely.
4933 Attacks against DH1024 are considered just feasible, because most of the
4934 work necessary to deduce information about a private key may be performed
4935 offline. The amount of resources required for such an attack would be
4936 significant. However, for an attack on TLS to be meaningful, the server
4937 would have to share the DH1024 private key among multiple clients, which is
4938 no longer an option since CVE-2016-0701.
4940 This only affects processors that support the AVX2 but not ADX extensions
4941 like Intel Haswell (4th generation).
4943 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4944 was originally found via the OSS-Fuzz project.
4949 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4951 * bn_sqrx8x_internal carry bug on x86_64
4953 There is a carry propagating bug in the x86_64 Montgomery squaring
4954 procedure. No EC algorithms are affected. Analysis suggests that attacks
4955 against RSA and DSA as a result of this defect would be very difficult to
4956 perform and are not believed likely. Attacks against DH are considered just
4957 feasible (although very difficult) because most of the work necessary to
4958 deduce information about a private key may be performed offline. The amount
4959 of resources required for such an attack would be very significant and
4960 likely only accessible to a limited number of attackers. An attacker would
4961 additionally need online access to an unpatched system using the target
4962 private key in a scenario with persistent DH parameters and a private
4963 key that is shared between multiple clients.
4965 This only affects processors that support the BMI1, BMI2 and ADX extensions
4966 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4968 This issue was reported to OpenSSL by the OSS-Fuzz project.
4973 * Malformed X.509 IPAddressFamily could cause OOB read
4975 If an X.509 certificate has a malformed IPAddressFamily extension,
4976 OpenSSL could do a one-byte buffer overread. The most likely result
4977 would be an erroneous display of the certificate in text format.
4979 This issue was reported to OpenSSL by the OSS-Fuzz project.
4983 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4985 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4986 platform rather than 'mingw'.
4990 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4992 * Truncated packet could crash via OOB read
4994 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4995 cipher is being used, then a truncated packet can cause that host to
4996 perform an out-of-bounds read, usually resulting in a crash.
4998 This issue was reported to OpenSSL by Robert Święcki of Google.
5003 * BN_mod_exp may produce incorrect results on x86_64
5005 There is a carry propagating bug in the x86_64 Montgomery squaring
5006 procedure. No EC algorithms are affected. Analysis suggests that attacks
5007 against RSA and DSA as a result of this defect would be very difficult to
5008 perform and are not believed likely. Attacks against DH are considered just
5009 feasible (although very difficult) because most of the work necessary to
5010 deduce information about a private key may be performed offline. The amount
5011 of resources required for such an attack would be very significant and
5012 likely only accessible to a limited number of attackers. An attacker would
5013 additionally need online access to an unpatched system using the target
5014 private key in a scenario with persistent DH parameters and a private
5015 key that is shared between multiple clients. For example this can occur by
5016 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5017 similar to CVE-2015-3193 but must be treated as a separate problem.
5019 This issue was reported to OpenSSL by the OSS-Fuzz project.
5024 * Montgomery multiplication may produce incorrect results
5026 There is a carry propagating bug in the Broadwell-specific Montgomery
5027 multiplication procedure that handles input lengths divisible by, but
5028 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5029 and DH private keys are impossible. This is because the subroutine in
5030 question is not used in operations with the private key itself and an input
5031 of the attacker's direct choice. Otherwise the bug can manifest itself as
5032 transient authentication and key negotiation failures or reproducible
5033 erroneous outcome of public-key operations with specially crafted input.
5034 Among EC algorithms only Brainpool P-512 curves are affected and one
5035 presumably can attack ECDH key negotiation. Impact was not analyzed in
5036 detail, because pre-requisites for attack are considered unlikely. Namely
5037 multiple clients have to choose the curve in question and the server has to
5038 share the private key among them, neither of which is default behaviour.
5039 Even then only clients that chose the curve will be affected.
5041 This issue was publicly reported as transient failures and was not
5042 initially recognized as a security issue. Thanks to Richard Morgan for
5043 providing reproducible case.
5048 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5049 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5050 prevent issues where no progress is being made and the peer continually
5051 sends unrecognised record types, using up resources processing them.
5055 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5057 * Missing CRL sanity check
5059 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5060 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5061 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5063 This issue only affects the OpenSSL 1.0.2i
5068 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5070 * OCSP Status Request extension unbounded memory growth
5072 A malicious client can send an excessively large OCSP Status Request
5073 extension. If that client continually requests renegotiation, sending a
5074 large OCSP Status Request extension each time, then there will be unbounded
5075 memory growth on the server. This will eventually lead to a Denial Of
5076 Service attack through memory exhaustion. Servers with a default
5077 configuration are vulnerable even if they do not support OCSP. Builds using
5078 the "no-ocsp" build time option are not affected.
5080 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5085 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5088 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5094 * OOB write in MDC2_Update()
5096 An overflow can occur in MDC2_Update() either if called directly or
5097 through the EVP_DigestUpdate() function using MDC2. If an attacker
5098 is able to supply very large amounts of input data after a previous
5099 call to EVP_EncryptUpdate() with a partial block then a length check
5100 can overflow resulting in a heap corruption.
5102 The amount of data needed is comparable to SIZE_MAX which is impractical
5105 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5110 * Malformed SHA512 ticket DoS
5112 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5113 DoS attack where a malformed ticket will result in an OOB read which will
5116 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5117 a custom server callback and ticket lookup mechanism.
5119 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5124 * OOB write in BN_bn2dec()
5126 The function BN_bn2dec() does not check the return value of BN_div_word().
5127 This can cause an OOB write if an application uses this function with an
5128 overly large BIGNUM. This could be a problem if an overly large certificate
5129 or CRL is printed out from an untrusted source. TLS is not affected because
5130 record limits will reject an oversized certificate before it is parsed.
5132 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5137 * OOB read in TS_OBJ_print_bio()
5139 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5140 the total length the OID text representation would use and not the amount
5141 of data written. This will result in OOB reads when large OIDs are
5144 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5149 * Pointer arithmetic undefined behaviour
5151 Avoid some undefined pointer arithmetic
5153 A common idiom in the codebase is to check limits in the following manner:
5156 Where "p" points to some malloc'd data of SIZE bytes and
5159 "len" here could be from some externally supplied data (e.g. from a TLS
5162 The rules of C pointer arithmetic are such that "p + len" is only well
5163 defined where len <= SIZE. Therefore the above idiom is actually
5164 undefined behaviour.
5166 For example this could cause problems if some malloc implementation
5167 provides an address for "p" such that "p + len" actually overflows for
5168 values of len that are too big and therefore p + len < limit.
5170 This issue was reported to OpenSSL by Guido Vranken
5175 * Constant time flag not preserved in DSA signing
5177 Operations in the DSA signing algorithm should run in constant time in
5178 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5179 implementation means that a non-constant time codepath is followed for
5180 certain operations. This has been demonstrated through a cache-timing
5181 attack to be sufficient for an attacker to recover the private DSA key.
5183 This issue was reported by César Pereida (Aalto University), Billy Brumley
5184 (Tampere University of Technology), and Yuval Yarom (The University of
5185 Adelaide and NICTA).
5190 * DTLS buffered message DoS
5192 In a DTLS connection where handshake messages are delivered out-of-order
5193 those messages that OpenSSL is not yet ready to process will be buffered
5194 for later use. Under certain circumstances, a flaw in the logic means that
5195 those messages do not get removed from the buffer even though the handshake
5196 has been completed. An attacker could force up to approx. 15 messages to
5197 remain in the buffer when they are no longer required. These messages will
5198 be cleared when the DTLS connection is closed. The default maximum size for
5199 a message is 100k. Therefore the attacker could force an additional 1500k
5200 to be consumed per connection. By opening many simulataneous connections an
5201 attacker could cause a DoS attack through memory exhaustion.
5203 This issue was reported to OpenSSL by Quan Luo.
5208 * DTLS replay protection DoS
5210 A flaw in the DTLS replay attack protection mechanism means that records
5211 that arrive for future epochs update the replay protection "window" before
5212 the MAC for the record has been validated. This could be exploited by an
5213 attacker by sending a record for the next epoch (which does not have to
5214 decrypt or have a valid MAC), with a very large sequence number. This means
5215 that all subsequent legitimate packets are dropped causing a denial of
5216 service for a specific DTLS connection.
5218 This issue was reported to OpenSSL by the OCAP audit team.
5223 * Certificate message OOB reads
5225 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5226 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5227 theoretical DoS risk but this has not been observed in practice on common
5230 The messages affected are client certificate, client certificate request
5231 and server certificate. As a result the attack can only be performed
5232 against a client or a server which enables client authentication.
5234 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5239 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5241 * Prevent padding oracle in AES-NI CBC MAC check
5243 A MITM attacker can use a padding oracle attack to decrypt traffic
5244 when the connection uses an AES CBC cipher and the server support
5247 This issue was introduced as part of the fix for Lucky 13 padding
5248 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5249 constant time by making sure that always the same bytes are read and
5250 compared against either the MAC or padding bytes. But it no longer
5251 checked that there was enough data to have both the MAC and padding
5254 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5258 * Fix EVP_EncodeUpdate overflow
5260 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5261 Base64 encoding of binary data. If an attacker is able to supply very large
5262 amounts of input data then a length check can overflow resulting in a heap
5265 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5266 the `PEM_write_bio*` family of functions. These are mainly used within the
5267 OpenSSL command line applications, so any application which processes data
5268 from an untrusted source and outputs it as a PEM file should be considered
5269 vulnerable to this issue. User applications that call these APIs directly
5270 with large amounts of untrusted data may also be vulnerable.
5272 This issue was reported by Guido Vranken.
5277 * Fix EVP_EncryptUpdate overflow
5279 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5280 is able to supply very large amounts of input data after a previous call to
5281 EVP_EncryptUpdate() with a partial block then a length check can overflow
5282 resulting in a heap corruption. Following an analysis of all OpenSSL
5283 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5284 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5285 the first called function after an EVP_EncryptInit(), and therefore that
5286 specific call must be safe. The second form is where the length passed to
5287 EVP_EncryptUpdate() can be seen from the code to be some small value and
5288 therefore there is no possibility of an overflow. Since all instances are
5289 one of these two forms, it is believed that there can be no overflows in
5290 internal code due to this problem. It should be noted that
5291 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5292 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5293 of these calls have also been analysed too and it is believed there are no
5294 instances in internal usage where an overflow could occur.
5296 This issue was reported by Guido Vranken.
5301 * Prevent ASN.1 BIO excessive memory allocation
5303 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5304 a short invalid encoding can cause allocation of large amounts of memory
5305 potentially consuming excessive resources or exhausting memory.
5307 Any application parsing untrusted data through d2i BIO functions is
5308 affected. The memory based functions such as d2i_X509() are *not* affected.
5309 Since the memory based functions are used by the TLS library, TLS
5310 applications are not affected.
5312 This issue was reported by Brian Carpenter.
5319 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5320 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5321 in arbitrary stack data being returned in the buffer.
5323 This issue was reported by Guido Vranken.
5328 * Modify behavior of ALPN to invoke callback after SNI/servername
5329 callback, such that updates to the SSL_CTX affect ALPN.
5333 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5338 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5339 methods are enabled and ssl2 is disabled the methods return NULL.
5343 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5345 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5346 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5347 provide any "EXPORT" or "LOW" strength ciphers.
5351 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5352 is by default disabled at build-time. Builds that are not configured with
5353 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5354 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5355 will need to explicitly call either of:
5357 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5359 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5361 as appropriate. Even if either of those is used, or the application
5362 explicitly uses the version-specific SSLv2_method() or its client and
5363 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5364 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5365 ciphers, and SSLv2 56-bit DES are no longer available.
5370 * Fix a double-free in DSA code
5372 A double free bug was discovered when OpenSSL parses malformed DSA private
5373 keys and could lead to a DoS attack or memory corruption for applications
5374 that receive DSA private keys from untrusted sources. This scenario is
5377 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5383 * Disable SRP fake user seed to address a server memory leak.
5385 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5387 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5388 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5389 was changed to ignore the "fake user" SRP seed, even if the seed
5392 Users should use SRP_VBASE_get1_by_user instead. Note that in
5393 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5394 also that even though configuring the SRP seed attempts to hide
5395 invalid usernames by continuing the handshake with fake
5396 credentials, this behaviour is not constant time and no strong
5397 guarantees are made that the handshake is indistinguishable from
5398 that of a valid user.
5403 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5405 In the BN_hex2bn function the number of hex digits is calculated using an
5406 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5407 large values of `i` this can result in `bn_expand` not allocating any
5408 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5409 field as NULL leading to a subsequent NULL ptr deref. For very large values
5410 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5411 In this case memory is allocated to the internal BIGNUM data field, but it
5412 is insufficiently sized leading to heap corruption. A similar issue exists
5413 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5414 is ever called by user applications with very large untrusted hex/dec data.
5415 This is anticipated to be a rare occurrence.
5417 All OpenSSL internal usage of these functions use data that is not expected
5418 to be untrusted, e.g. config file data or application command line
5419 arguments. If user developed applications generate config file data based
5420 on untrusted data then it is possible that this could also lead to security
5421 consequences. This is also anticipated to be rare.
5423 This issue was reported to OpenSSL by Guido Vranken.
5428 * Fix memory issues in `BIO_*printf` functions
5430 The internal `fmtstr` function used in processing a "%s" format string in
5431 the `BIO_*printf` functions could overflow while calculating the length of a
5432 string and cause an OOB read when printing very long strings.
5434 Additionally the internal `doapr_outch` function can attempt to write to an
5435 OOB memory location (at an offset from the NULL pointer) in the event of a
5436 memory allocation failure. In 1.0.2 and below this could be caused where
5437 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5438 could be in processing a very long "%s" format string. Memory leaks can
5441 The first issue may mask the second issue dependent on compiler behaviour.
5442 These problems could enable attacks where large amounts of untrusted data
5443 is passed to the `BIO_*printf` functions. If applications use these functions
5444 in this way then they could be vulnerable. OpenSSL itself uses these
5445 functions when printing out human-readable dumps of ASN.1 data. Therefore
5446 applications that print this data could be vulnerable if the data is from
5447 untrusted sources. OpenSSL command line applications could also be
5448 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5449 as command line arguments.
5451 Libssl is not considered directly vulnerable. Additionally certificates etc
5452 received via remote connections via libssl are also unlikely to be able to
5453 trigger these issues because of message size limits enforced within libssl.
5455 This issue was reported to OpenSSL Guido Vranken.
5460 * Side channel attack on modular exponentiation
5462 A side-channel attack was found which makes use of cache-bank conflicts on
5463 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5464 of RSA keys. The ability to exploit this issue is limited as it relies on
5465 an attacker who has control of code in a thread running on the same
5466 hyper-threaded core as the victim thread which is performing decryptions.
5468 This issue was reported to OpenSSL by Yuval Yarom, The University of
5469 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5470 Nadia Heninger, University of Pennsylvania with more information at
5471 <http://cachebleed.info>.
5476 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5477 if no keysize is specified with default_bits. This fixes an
5478 omission in an earlier change that changed all RSA/DSA key generation
5479 commands to use 2048 bits by default.
5483 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5485 * DH small subgroups
5487 Historically OpenSSL only ever generated DH parameters based on "safe"
5488 primes. More recently (in version 1.0.2) support was provided for
5489 generating X9.42 style parameter files such as those required for RFC 5114
5490 support. The primes used in such files may not be "safe". Where an
5491 application is using DH configured with parameters based on primes that are
5492 not "safe" then an attacker could use this fact to find a peer's private
5493 DH exponent. This attack requires that the attacker complete multiple
5494 handshakes in which the peer uses the same private DH exponent. For example
5495 this could be used to discover a TLS server's private DH exponent if it's
5496 reusing the private DH exponent or it's using a static DH ciphersuite.
5498 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5499 TLS. It is not on by default. If the option is not set then the server
5500 reuses the same private DH exponent for the life of the server process and
5501 would be vulnerable to this attack. It is believed that many popular
5502 applications do set this option and would therefore not be at risk.
5504 The fix for this issue adds an additional check where a "q" parameter is
5505 available (as is the case in X9.42 based parameters). This detects the
5506 only known attack, and is the only possible defense for static DH
5507 ciphersuites. This could have some performance impact.
5509 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5510 default and cannot be disabled. This could have some performance impact.
5512 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5517 * SSLv2 doesn't block disabled ciphers
5519 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5520 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5521 been disabled, provided that the SSLv2 protocol was not also disabled via
5524 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5525 and Sebastian Schinzel.
5530 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5532 * BN_mod_exp may produce incorrect results on x86_64
5534 There is a carry propagating bug in the x86_64 Montgomery squaring
5535 procedure. No EC algorithms are affected. Analysis suggests that attacks
5536 against RSA and DSA as a result of this defect would be very difficult to
5537 perform and are not believed likely. Attacks against DH are considered just
5538 feasible (although very difficult) because most of the work necessary to
5539 deduce information about a private key may be performed offline. The amount
5540 of resources required for such an attack would be very significant and
5541 likely only accessible to a limited number of attackers. An attacker would
5542 additionally need online access to an unpatched system using the target
5543 private key in a scenario with persistent DH parameters and a private
5544 key that is shared between multiple clients. For example this can occur by
5545 default in OpenSSL DHE based SSL/TLS ciphersuites.
5547 This issue was reported to OpenSSL by Hanno Böck.
5552 * Certificate verify crash with missing PSS parameter
5554 The signature verification routines will crash with a NULL pointer
5555 dereference if presented with an ASN.1 signature using the RSA PSS
5556 algorithm and absent mask generation function parameter. Since these
5557 routines are used to verify certificate signature algorithms this can be
5558 used to crash any certificate verification operation and exploited in a
5559 DoS attack. Any application which performs certificate verification is
5560 vulnerable including OpenSSL clients and servers which enable client
5563 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5568 * X509_ATTRIBUTE memory leak
5570 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5571 memory. This structure is used by the PKCS#7 and CMS routines so any
5572 application which reads PKCS#7 or CMS data from untrusted sources is
5573 affected. SSL/TLS is not affected.
5575 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5581 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5582 This changes the decoding behaviour for some invalid messages,
5583 though the change is mostly in the more lenient direction, and
5584 legacy behaviour is preserved as much as possible.
5588 * In DSA_generate_parameters_ex, if the provided seed is too short,
5591 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5593 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5595 * Alternate chains certificate forgery
5597 During certificate verification, OpenSSL will attempt to find an
5598 alternative certificate chain if the first attempt to build such a chain
5599 fails. An error in the implementation of this logic can mean that an
5600 attacker could cause certain checks on untrusted certificates to be
5601 bypassed, such as the CA flag, enabling them to use a valid leaf
5602 certificate to act as a CA and "issue" an invalid certificate.
5604 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5609 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5611 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5612 incompatibility in the handling of HMAC. The previous ABI has now been
5617 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5619 * Malformed ECParameters causes infinite loop
5621 When processing an ECParameters structure OpenSSL enters an infinite loop
5622 if the curve specified is over a specially malformed binary polynomial
5625 This can be used to perform denial of service against any
5626 system which processes public keys, certificate requests or
5627 certificates. This includes TLS clients and TLS servers with
5628 client authentication enabled.
5630 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5635 * Exploitable out-of-bounds read in X509_cmp_time
5637 X509_cmp_time does not properly check the length of the ASN1_TIME
5638 string and can read a few bytes out of bounds. In addition,
5639 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5642 An attacker can use this to craft malformed certificates and CRLs of
5643 various sizes and potentially cause a segmentation fault, resulting in
5644 a DoS on applications that verify certificates or CRLs. TLS clients
5645 that verify CRLs are affected. TLS clients and servers with client
5646 authentication enabled may be affected if they use custom verification
5649 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5650 independently by Hanno Böck.
5655 * PKCS7 crash with missing EnvelopedContent
5657 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5658 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5659 with missing content and trigger a NULL pointer dereference on parsing.
5661 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5662 structures from untrusted sources are affected. OpenSSL clients and
5663 servers are not affected.
5665 This issue was reported to OpenSSL by Michal Zalewski (Google).
5670 * CMS verify infinite loop with unknown hash function
5672 When verifying a signedData message the CMS code can enter an infinite loop
5673 if presented with an unknown hash function OID. This can be used to perform
5674 denial of service against any system which verifies signedData messages using
5676 This issue was reported to OpenSSL by Johannes Bauer.
5681 * Race condition handling NewSessionTicket
5683 If a NewSessionTicket is received by a multi-threaded client when attempting to
5684 reuse a previous ticket then a race condition can occur potentially leading to
5685 a double free of the ticket data.
5690 * Only support 256-bit or stronger elliptic curves with the
5691 'ecdh_auto' setting (server) or by default (client). Of supported
5692 curves, prefer P-256 (both).
5696 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5698 * ClientHello sigalgs DoS fix
5700 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5701 invalid signature algorithms extension a NULL pointer dereference will
5702 occur. This can be exploited in a DoS attack against the server.
5704 This issue was was reported to OpenSSL by David Ramos of Stanford
5708 *Stephen Henson and Matt Caswell*
5710 * Multiblock corrupted pointer fix
5712 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5713 feature only applies on 64 bit x86 architecture platforms that support AES
5714 NI instructions. A defect in the implementation of "multiblock" can cause
5715 OpenSSL's internal write buffer to become incorrectly set to NULL when
5716 using non-blocking IO. Typically, when the user application is using a
5717 socket BIO for writing, this will only result in a failed connection.
5718 However if some other BIO is used then it is likely that a segmentation
5719 fault will be triggered, thus enabling a potential DoS attack.
5721 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5726 * Segmentation fault in DTLSv1_listen fix
5728 The DTLSv1_listen function is intended to be stateless and processes the
5729 initial ClientHello from many peers. It is common for user code to loop
5730 over the call to DTLSv1_listen until a valid ClientHello is received with
5731 an associated cookie. A defect in the implementation of DTLSv1_listen means
5732 that state is preserved in the SSL object from one invocation to the next
5733 that can lead to a segmentation fault. Errors processing the initial
5734 ClientHello can trigger this scenario. An example of such an error could be
5735 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5738 This issue was reported to OpenSSL by Per Allansson.
5743 * Segmentation fault in ASN1_TYPE_cmp fix
5745 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5746 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5747 certificate signature algorithm consistency this can be used to crash any
5748 certificate verification operation and exploited in a DoS attack. Any
5749 application which performs certificate verification is vulnerable including
5750 OpenSSL clients and servers which enable client authentication.
5755 * Segmentation fault for invalid PSS parameters fix
5757 The signature verification routines will crash with a NULL pointer
5758 dereference if presented with an ASN.1 signature using the RSA PSS
5759 algorithm and invalid parameters. Since these routines are used to verify
5760 certificate signature algorithms this can be used to crash any
5761 certificate verification operation and exploited in a DoS attack. Any
5762 application which performs certificate verification is vulnerable including
5763 OpenSSL clients and servers which enable client authentication.
5765 This issue was was reported to OpenSSL by Brian Carpenter.
5770 * ASN.1 structure reuse memory corruption fix
5772 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5773 memory corruption via an invalid write. Such reuse is and has been
5774 strongly discouraged and is believed to be rare.
5776 Applications that parse structures containing CHOICE or ANY DEFINED BY
5777 components may be affected. Certificate parsing (d2i_X509 and related
5778 functions) are however not affected. OpenSSL clients and servers are
5784 * PKCS7 NULL pointer dereferences fix
5786 The PKCS#7 parsing code does not handle missing outer ContentInfo
5787 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5788 missing content and trigger a NULL pointer dereference on parsing.
5790 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5791 otherwise parse PKCS#7 structures from untrusted sources are
5792 affected. OpenSSL clients and servers are not affected.
5794 This issue was reported to OpenSSL by Michal Zalewski (Google).
5799 * DoS via reachable assert in SSLv2 servers fix
5801 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5802 servers that both support SSLv2 and enable export cipher suites by sending
5803 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5805 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5806 (OpenSSL development team).
5811 * Empty CKE with client auth and DHE fix
5813 If client auth is used then a server can seg fault in the event of a DHE
5814 ciphersuite being selected and a zero length ClientKeyExchange message
5815 being sent by the client. This could be exploited in a DoS attack.
5820 * Handshake with unseeded PRNG fix
5822 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5823 with an unseeded PRNG. The conditions are:
5824 - The client is on a platform where the PRNG has not been seeded
5825 automatically, and the user has not seeded manually
5826 - A protocol specific client method version has been used (i.e. not
5827 SSL_client_methodv23)
5828 - A ciphersuite is used that does not require additional random data from
5829 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5831 If the handshake succeeds then the client random that has been used will
5832 have been generated from a PRNG with insufficient entropy and therefore the
5833 output may be predictable.
5835 For example using the following command with an unseeded openssl will
5836 succeed on an unpatched platform:
5838 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5843 * Use After Free following d2i_ECPrivatekey error fix
5845 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5846 could cause a use after free condition. This, in turn, could cause a double
5847 free in several private key parsing functions (such as d2i_PrivateKey
5848 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5849 for applications that receive EC private keys from untrusted
5850 sources. This scenario is considered rare.
5852 This issue was discovered by the BoringSSL project and fixed in their
5858 * X509_to_X509_REQ NULL pointer deref fix
5860 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5861 the certificate key is invalid. This function is rarely used in practice.
5863 This issue was discovered by Brian Carpenter.
5868 * Removed the export ciphers from the DEFAULT ciphers
5872 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5874 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5875 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5876 So far those who have to target multiple platforms would compromise
5877 and argue that binary targeting say ARMv5 would still execute on
5878 ARMv8. "Universal" build resolves this compromise by providing
5879 near-optimal performance even on newer platforms.
5883 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5884 (other platforms pending).
5886 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5888 * Add support for the SignedCertificateTimestampList certificate and
5889 OCSP response extensions from RFC6962.
5893 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5894 for corner cases. (Certain input points at infinity could lead to
5895 bogus results, with non-infinity inputs mapped to infinity too.)
5899 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5900 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5901 common cases are optimized and there still is room for further
5902 improvements. Vector Permutation AES for Altivec is also added.
5906 * Add support for little-endian ppc64 Linux target.
5908 *Marcelo Cerri (IBM)*
5910 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5911 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5912 are optimized and there still is room for further improvements.
5913 Both 32- and 64-bit modes are supported.
5915 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5917 * Improved ARMv7 NEON support.
5921 * Support for SPARC Architecture 2011 crypto extensions, first
5922 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5923 SHA256/512, MD5, GHASH and modular exponentiation.
5925 *Andy Polyakov, David Miller*
5927 * Accelerated modular exponentiation for Intel processors, a.k.a.
5930 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5932 * Support for new and upcoming Intel processors, including AVX2,
5933 BMI and SHA ISA extensions. This includes additional "stitched"
5934 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5937 This work was sponsored by Intel Corp.
5941 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5942 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5943 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5947 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5948 this fixes a limitation in previous versions of OpenSSL.
5952 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5953 MGF1 digest and OAEP label.
5957 * Add EVP support for key wrapping algorithms, to avoid problems with
5958 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5959 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5960 algorithms and include tests cases.
5964 * Add functions to allocate and set the fields of an ECDSA_METHOD
5967 *Douglas E. Engert, Steve Henson*
5969 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5970 difference in days and seconds between two tm or ASN1_TIME structures.
5974 * Add -rev test option to s_server to just reverse order of characters
5975 received by client and send back to server. Also prints an abbreviated
5976 summary of the connection parameters.
5980 * New option -brief for s_client and s_server to print out a brief summary
5981 of connection parameters.
5985 * Add callbacks for arbitrary TLS extensions.
5987 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5989 * New option -crl_download in several openssl utilities to download CRLs
5990 from CRLDP extension in certificates.
5994 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5998 * New function X509_CRL_diff to generate a delta CRL from the difference
5999 of two full CRLs. Add support to "crl" utility.
6003 * New functions to set lookup_crls function and to retrieve
6004 X509_STORE from X509_STORE_CTX.
6008 * Print out deprecated issuer and subject unique ID fields in
6013 * Extend OCSP I/O functions so they can be used for simple general purpose
6014 HTTP as well as OCSP. New wrapper function which can be used to download
6015 CRLs using the OCSP API.
6019 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6023 * `SSL_CONF*` functions. These provide a common framework for application
6024 configuration using configuration files or command lines.
6028 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6029 message callback and prints the results. Needs compile time option
6030 "enable-ssl-trace". New options to s_client and s_server to enable
6035 * New ctrl and macro to retrieve supported points extensions.
6036 Print out extension in s_server and s_client.
6040 * New functions to retrieve certificate signature and signature
6045 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6050 * New Suite B modes for TLS code. These use and enforce the requirements
6051 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6052 only use Suite B curves. The Suite B modes can be set by using the
6053 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6057 * New chain verification flags for Suite B levels of security. Check
6058 algorithms are acceptable when flags are set in X509_verify_cert.
6062 * Make tls1_check_chain return a set of flags indicating checks passed
6063 by a certificate chain. Add additional tests to handle client
6064 certificates: checks for matching certificate type and issuer name
6069 * If an attempt is made to use a signature algorithm not in the peer
6070 preference list abort the handshake. If client has no suitable
6071 signature algorithms in response to a certificate request do not
6072 use the certificate.
6076 * If server EC tmp key is not in client preference list abort handshake.
6080 * Add support for certificate stores in CERT structure. This makes it
6081 possible to have different stores per SSL structure or one store in
6082 the parent SSL_CTX. Include distinct stores for certificate chain
6083 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6084 to build and store a certificate chain in CERT structure: returning
6085 an error if the chain cannot be built: this will allow applications
6086 to test if a chain is correctly configured.
6088 Note: if the CERT based stores are not set then the parent SSL_CTX
6089 store is used to retain compatibility with existing behaviour.
6093 * New function ssl_set_client_disabled to set a ciphersuite disabled
6094 mask based on the current session, check mask when sending client
6095 hello and checking the requested ciphersuite.
6099 * New ctrls to retrieve and set certificate types in a certificate
6100 request message. Print out received values in s_client. If certificate
6101 types is not set with custom values set sensible values based on
6102 supported signature algorithms.
6106 * Support for distinct client and server supported signature algorithms.
6110 * Add certificate callback. If set this is called whenever a certificate
6111 is required by client or server. An application can decide which
6112 certificate chain to present based on arbitrary criteria: for example
6113 supported signature algorithms. Add very simple example to s_server.
6114 This fixes many of the problems and restrictions of the existing client
6115 certificate callback: for example you can now clear an existing
6116 certificate and specify the whole chain.
6120 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6121 the certificate can be used for (if anything). Set valid_flags field
6122 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6123 to have similar checks in it.
6125 Add new "cert_flags" field to CERT structure and include a "strict mode".
6126 This enforces some TLS certificate requirements (such as only permitting
6127 certificate signature algorithms contained in the supported algorithms
6128 extension) which some implementations ignore: this option should be used
6129 with caution as it could cause interoperability issues.
6133 * Update and tidy signature algorithm extension processing. Work out
6134 shared signature algorithms based on preferences and peer algorithms
6135 and print them out in s_client and s_server. Abort handshake if no
6136 shared signature algorithms.
6140 * Add new functions to allow customised supported signature algorithms
6141 for SSL and SSL_CTX structures. Add options to s_client and s_server
6146 * New function SSL_certs_clear() to delete all references to certificates
6147 from an SSL structure. Before this once a certificate had been added
6148 it couldn't be removed.
6152 * Integrate hostname, email address and IP address checking with certificate
6153 verification. New verify options supporting checking in openssl utility.
6157 * Fixes and wildcard matching support to hostname and email checking
6158 functions. Add manual page.
6160 *Florian Weimer (Red Hat Product Security Team)*
6162 * New functions to check a hostname email or IP address against a
6163 certificate. Add options x509 utility to print results of checks against
6168 * Fix OCSP checking.
6170 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6172 * Initial experimental support for explicitly trusted non-root CAs.
6173 OpenSSL still tries to build a complete chain to a root but if an
6174 intermediate CA has a trust setting included that is used. The first
6175 setting is used: whether to trust (e.g., -addtrust option to the x509
6180 * Add -trusted_first option which attempts to find certificates in the
6181 trusted store even if an untrusted chain is also supplied.
6185 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6186 platform support for Linux and Android.
6190 * Support for linux-x32, ILP32 environment in x86_64 framework.
6194 * Experimental multi-implementation support for FIPS capable OpenSSL.
6195 When in FIPS mode the approved implementations are used as normal,
6196 when not in FIPS mode the internal unapproved versions are used instead.
6197 This means that the FIPS capable OpenSSL isn't forced to use the
6198 (often lower performance) FIPS implementations outside FIPS mode.
6202 * Transparently support X9.42 DH parameters when calling
6203 PEM_read_bio_DHparameters. This means existing applications can handle
6204 the new parameter format automatically.
6208 * Initial experimental support for X9.42 DH parameter format: mainly
6209 to support use of 'q' parameter for RFC5114 parameters.
6213 * Add DH parameters from RFC5114 including test data to dhtest.
6217 * Support for automatic EC temporary key parameter selection. If enabled
6218 the most preferred EC parameters are automatically used instead of
6219 hardcoded fixed parameters. Now a server just has to call:
6220 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6221 support ECDH and use the most appropriate parameters.
6225 * Enhance and tidy EC curve and point format TLS extension code. Use
6226 static structures instead of allocation if default values are used.
6227 New ctrls to set curves we wish to support and to retrieve shared curves.
6228 Print out shared curves in s_server. New options to s_server and s_client
6229 to set list of supported curves.
6233 * New ctrls to retrieve supported signature algorithms and
6234 supported curve values as an array of NIDs. Extend openssl utility
6235 to print out received values.
6239 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6240 between NIDs and the more common NIST names such as "P-256". Enhance
6241 ecparam utility and ECC method to recognise the NIST names for curves.
6245 * Enhance SSL/TLS certificate chain handling to support different
6246 chains for each certificate instead of one chain in the parent SSL_CTX.
6250 * Support for fixed DH ciphersuite client authentication: where both
6251 server and client use DH certificates with common parameters.
6255 * Support for fixed DH ciphersuites: those requiring DH server
6260 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6262 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6263 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6264 X509_CINF_get_signature were reverted post internal team review.
6269 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6271 * OCSP Status Request extension unbounded memory growth
6273 A malicious client can send an excessively large OCSP Status Request
6274 extension. If that client continually requests renegotiation, sending a
6275 large OCSP Status Request extension each time, then there will be unbounded
6276 memory growth on the server. This will eventually lead to a Denial Of
6277 Service attack through memory exhaustion. Servers with a default
6278 configuration are vulnerable even if they do not support OCSP. Builds using
6279 the "no-ocsp" build time option are not affected.
6281 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6286 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6289 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6295 * OOB write in MDC2_Update()
6297 An overflow can occur in MDC2_Update() either if called directly or
6298 through the EVP_DigestUpdate() function using MDC2. If an attacker
6299 is able to supply very large amounts of input data after a previous
6300 call to EVP_EncryptUpdate() with a partial block then a length check
6301 can overflow resulting in a heap corruption.
6303 The amount of data needed is comparable to SIZE_MAX which is impractical
6306 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6311 * Malformed SHA512 ticket DoS
6313 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6314 DoS attack where a malformed ticket will result in an OOB read which will
6317 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6318 a custom server callback and ticket lookup mechanism.
6320 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6325 * OOB write in BN_bn2dec()
6327 The function BN_bn2dec() does not check the return value of BN_div_word().
6328 This can cause an OOB write if an application uses this function with an
6329 overly large BIGNUM. This could be a problem if an overly large certificate
6330 or CRL is printed out from an untrusted source. TLS is not affected because
6331 record limits will reject an oversized certificate before it is parsed.
6333 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6338 * OOB read in TS_OBJ_print_bio()
6340 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6341 the total length the OID text representation would use and not the amount
6342 of data written. This will result in OOB reads when large OIDs are
6345 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6350 * Pointer arithmetic undefined behaviour
6352 Avoid some undefined pointer arithmetic
6354 A common idiom in the codebase is to check limits in the following manner:
6357 Where "p" points to some malloc'd data of SIZE bytes and
6360 "len" here could be from some externally supplied data (e.g. from a TLS
6363 The rules of C pointer arithmetic are such that "p + len" is only well
6364 defined where len <= SIZE. Therefore the above idiom is actually
6365 undefined behaviour.
6367 For example this could cause problems if some malloc implementation
6368 provides an address for "p" such that "p + len" actually overflows for
6369 values of len that are too big and therefore p + len < limit.
6371 This issue was reported to OpenSSL by Guido Vranken
6376 * Constant time flag not preserved in DSA signing
6378 Operations in the DSA signing algorithm should run in constant time in
6379 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6380 implementation means that a non-constant time codepath is followed for
6381 certain operations. This has been demonstrated through a cache-timing
6382 attack to be sufficient for an attacker to recover the private DSA key.
6384 This issue was reported by César Pereida (Aalto University), Billy Brumley
6385 (Tampere University of Technology), and Yuval Yarom (The University of
6386 Adelaide and NICTA).
6391 * DTLS buffered message DoS
6393 In a DTLS connection where handshake messages are delivered out-of-order
6394 those messages that OpenSSL is not yet ready to process will be buffered
6395 for later use. Under certain circumstances, a flaw in the logic means that
6396 those messages do not get removed from the buffer even though the handshake
6397 has been completed. An attacker could force up to approx. 15 messages to
6398 remain in the buffer when they are no longer required. These messages will
6399 be cleared when the DTLS connection is closed. The default maximum size for
6400 a message is 100k. Therefore the attacker could force an additional 1500k
6401 to be consumed per connection. By opening many simulataneous connections an
6402 attacker could cause a DoS attack through memory exhaustion.
6404 This issue was reported to OpenSSL by Quan Luo.
6409 * DTLS replay protection DoS
6411 A flaw in the DTLS replay attack protection mechanism means that records
6412 that arrive for future epochs update the replay protection "window" before
6413 the MAC for the record has been validated. This could be exploited by an
6414 attacker by sending a record for the next epoch (which does not have to
6415 decrypt or have a valid MAC), with a very large sequence number. This means
6416 that all subsequent legitimate packets are dropped causing a denial of
6417 service for a specific DTLS connection.
6419 This issue was reported to OpenSSL by the OCAP audit team.
6424 * Certificate message OOB reads
6426 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6427 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6428 theoretical DoS risk but this has not been observed in practice on common
6431 The messages affected are client certificate, client certificate request
6432 and server certificate. As a result the attack can only be performed
6433 against a client or a server which enables client authentication.
6435 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6440 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6442 * Prevent padding oracle in AES-NI CBC MAC check
6444 A MITM attacker can use a padding oracle attack to decrypt traffic
6445 when the connection uses an AES CBC cipher and the server support
6448 This issue was introduced as part of the fix for Lucky 13 padding
6449 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6450 constant time by making sure that always the same bytes are read and
6451 compared against either the MAC or padding bytes. But it no longer
6452 checked that there was enough data to have both the MAC and padding
6455 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6460 * Fix EVP_EncodeUpdate overflow
6462 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6463 Base64 encoding of binary data. If an attacker is able to supply very large
6464 amounts of input data then a length check can overflow resulting in a heap
6467 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6468 the `PEM_write_bio*` family of functions. These are mainly used within the
6469 OpenSSL command line applications, so any application which processes data
6470 from an untrusted source and outputs it as a PEM file should be considered
6471 vulnerable to this issue. User applications that call these APIs directly
6472 with large amounts of untrusted data may also be vulnerable.
6474 This issue was reported by Guido Vranken.
6479 * Fix EVP_EncryptUpdate overflow
6481 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6482 is able to supply very large amounts of input data after a previous call to
6483 EVP_EncryptUpdate() with a partial block then a length check can overflow
6484 resulting in a heap corruption. Following an analysis of all OpenSSL
6485 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6486 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6487 the first called function after an EVP_EncryptInit(), and therefore that
6488 specific call must be safe. The second form is where the length passed to
6489 EVP_EncryptUpdate() can be seen from the code to be some small value and
6490 therefore there is no possibility of an overflow. Since all instances are
6491 one of these two forms, it is believed that there can be no overflows in
6492 internal code due to this problem. It should be noted that
6493 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6494 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6495 of these calls have also been analysed too and it is believed there are no
6496 instances in internal usage where an overflow could occur.
6498 This issue was reported by Guido Vranken.
6503 * Prevent ASN.1 BIO excessive memory allocation
6505 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6506 a short invalid encoding can casuse allocation of large amounts of memory
6507 potentially consuming excessive resources or exhausting memory.
6509 Any application parsing untrusted data through d2i BIO functions is
6510 affected. The memory based functions such as d2i_X509() are *not* affected.
6511 Since the memory based functions are used by the TLS library, TLS
6512 applications are not affected.
6514 This issue was reported by Brian Carpenter.
6521 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6522 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6523 in arbitrary stack data being returned in the buffer.
6525 This issue was reported by Guido Vranken.
6530 * Modify behavior of ALPN to invoke callback after SNI/servername
6531 callback, such that updates to the SSL_CTX affect ALPN.
6535 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6540 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6541 methods are enabled and ssl2 is disabled the methods return NULL.
6545 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6547 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6548 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6549 provide any "EXPORT" or "LOW" strength ciphers.
6553 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6554 is by default disabled at build-time. Builds that are not configured with
6555 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6556 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6557 will need to explicitly call either of:
6559 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6561 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6563 as appropriate. Even if either of those is used, or the application
6564 explicitly uses the version-specific SSLv2_method() or its client and
6565 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6566 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6567 ciphers, and SSLv2 56-bit DES are no longer available.
6572 * Fix a double-free in DSA code
6574 A double free bug was discovered when OpenSSL parses malformed DSA private
6575 keys and could lead to a DoS attack or memory corruption for applications
6576 that receive DSA private keys from untrusted sources. This scenario is
6579 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6585 * Disable SRP fake user seed to address a server memory leak.
6587 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6589 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6590 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6591 was changed to ignore the "fake user" SRP seed, even if the seed
6594 Users should use SRP_VBASE_get1_by_user instead. Note that in
6595 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6596 also that even though configuring the SRP seed attempts to hide
6597 invalid usernames by continuing the handshake with fake
6598 credentials, this behaviour is not constant time and no strong
6599 guarantees are made that the handshake is indistinguishable from
6600 that of a valid user.
6605 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6607 In the BN_hex2bn function the number of hex digits is calculated using an
6608 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6609 large values of `i` this can result in `bn_expand` not allocating any
6610 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6611 field as NULL leading to a subsequent NULL ptr deref. For very large values
6612 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6613 In this case memory is allocated to the internal BIGNUM data field, but it
6614 is insufficiently sized leading to heap corruption. A similar issue exists
6615 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6616 is ever called by user applications with very large untrusted hex/dec data.
6617 This is anticipated to be a rare occurrence.
6619 All OpenSSL internal usage of these functions use data that is not expected
6620 to be untrusted, e.g. config file data or application command line
6621 arguments. If user developed applications generate config file data based
6622 on untrusted data then it is possible that this could also lead to security
6623 consequences. This is also anticipated to be rare.
6625 This issue was reported to OpenSSL by Guido Vranken.
6630 * Fix memory issues in `BIO_*printf` functions
6632 The internal `fmtstr` function used in processing a "%s" format string in
6633 the `BIO_*printf` functions could overflow while calculating the length of a
6634 string and cause an OOB read when printing very long strings.
6636 Additionally the internal `doapr_outch` function can attempt to write to an
6637 OOB memory location (at an offset from the NULL pointer) in the event of a
6638 memory allocation failure. In 1.0.2 and below this could be caused where
6639 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6640 could be in processing a very long "%s" format string. Memory leaks can
6643 The first issue may mask the second issue dependent on compiler behaviour.
6644 These problems could enable attacks where large amounts of untrusted data
6645 is passed to the `BIO_*printf` functions. If applications use these functions
6646 in this way then they could be vulnerable. OpenSSL itself uses these
6647 functions when printing out human-readable dumps of ASN.1 data. Therefore
6648 applications that print this data could be vulnerable if the data is from
6649 untrusted sources. OpenSSL command line applications could also be
6650 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6651 as command line arguments.
6653 Libssl is not considered directly vulnerable. Additionally certificates etc
6654 received via remote connections via libssl are also unlikely to be able to
6655 trigger these issues because of message size limits enforced within libssl.
6657 This issue was reported to OpenSSL Guido Vranken.
6662 * Side channel attack on modular exponentiation
6664 A side-channel attack was found which makes use of cache-bank conflicts on
6665 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6666 of RSA keys. The ability to exploit this issue is limited as it relies on
6667 an attacker who has control of code in a thread running on the same
6668 hyper-threaded core as the victim thread which is performing decryptions.
6670 This issue was reported to OpenSSL by Yuval Yarom, The University of
6671 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6672 Nadia Heninger, University of Pennsylvania with more information at
6673 <http://cachebleed.info>.
6678 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6679 if no keysize is specified with default_bits. This fixes an
6680 omission in an earlier change that changed all RSA/DSA key generation
6681 commands to use 2048 bits by default.
6685 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6687 * Protection for DH small subgroup attacks
6689 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6690 switched on by default and cannot be disabled. This could have some
6695 * SSLv2 doesn't block disabled ciphers
6697 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6698 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6699 been disabled, provided that the SSLv2 protocol was not also disabled via
6702 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6703 and Sebastian Schinzel.
6708 * Reject DH handshakes with parameters shorter than 1024 bits.
6712 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6714 * Certificate verify crash with missing PSS parameter
6716 The signature verification routines will crash with a NULL pointer
6717 dereference if presented with an ASN.1 signature using the RSA PSS
6718 algorithm and absent mask generation function parameter. Since these
6719 routines are used to verify certificate signature algorithms this can be
6720 used to crash any certificate verification operation and exploited in a
6721 DoS attack. Any application which performs certificate verification is
6722 vulnerable including OpenSSL clients and servers which enable client
6725 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6730 * X509_ATTRIBUTE memory leak
6732 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6733 memory. This structure is used by the PKCS#7 and CMS routines so any
6734 application which reads PKCS#7 or CMS data from untrusted sources is
6735 affected. SSL/TLS is not affected.
6737 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6743 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6744 This changes the decoding behaviour for some invalid messages,
6745 though the change is mostly in the more lenient direction, and
6746 legacy behaviour is preserved as much as possible.
6750 * In DSA_generate_parameters_ex, if the provided seed is too short,
6751 use a random seed, as already documented.
6753 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6755 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6757 * Alternate chains certificate forgery
6759 During certificate verfification, OpenSSL will attempt to find an
6760 alternative certificate chain if the first attempt to build such a chain
6761 fails. An error in the implementation of this logic can mean that an
6762 attacker could cause certain checks on untrusted certificates to be
6763 bypassed, such as the CA flag, enabling them to use a valid leaf
6764 certificate to act as a CA and "issue" an invalid certificate.
6766 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6772 * Race condition handling PSK identify hint
6774 If PSK identity hints are received by a multi-threaded client then
6775 the values are wrongly updated in the parent SSL_CTX structure. This can
6776 result in a race condition potentially leading to a double free of the
6782 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6784 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6785 incompatibility in the handling of HMAC. The previous ABI has now been
6788 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6790 * Malformed ECParameters causes infinite loop
6792 When processing an ECParameters structure OpenSSL enters an infinite loop
6793 if the curve specified is over a specially malformed binary polynomial
6796 This can be used to perform denial of service against any
6797 system which processes public keys, certificate requests or
6798 certificates. This includes TLS clients and TLS servers with
6799 client authentication enabled.
6801 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6806 * Exploitable out-of-bounds read in X509_cmp_time
6808 X509_cmp_time does not properly check the length of the ASN1_TIME
6809 string and can read a few bytes out of bounds. In addition,
6810 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6813 An attacker can use this to craft malformed certificates and CRLs of
6814 various sizes and potentially cause a segmentation fault, resulting in
6815 a DoS on applications that verify certificates or CRLs. TLS clients
6816 that verify CRLs are affected. TLS clients and servers with client
6817 authentication enabled may be affected if they use custom verification
6820 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6821 independently by Hanno Böck.
6826 * PKCS7 crash with missing EnvelopedContent
6828 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6829 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6830 with missing content and trigger a NULL pointer dereference on parsing.
6832 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6833 structures from untrusted sources are affected. OpenSSL clients and
6834 servers are not affected.
6836 This issue was reported to OpenSSL by Michal Zalewski (Google).
6841 * CMS verify infinite loop with unknown hash function
6843 When verifying a signedData message the CMS code can enter an infinite loop
6844 if presented with an unknown hash function OID. This can be used to perform
6845 denial of service against any system which verifies signedData messages using
6847 This issue was reported to OpenSSL by Johannes Bauer.
6852 * Race condition handling NewSessionTicket
6854 If a NewSessionTicket is received by a multi-threaded client when attempting to
6855 reuse a previous ticket then a race condition can occur potentially leading to
6856 a double free of the ticket data.
6861 * Reject DH handshakes with parameters shorter than 768 bits.
6863 *Kurt Roeckx and Emilia Kasper*
6865 * dhparam: generate 2048-bit parameters by default.
6867 *Kurt Roeckx and Emilia Kasper*
6869 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6871 * Segmentation fault in ASN1_TYPE_cmp fix
6873 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6874 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6875 certificate signature algorithm consistency this can be used to crash any
6876 certificate verification operation and exploited in a DoS attack. Any
6877 application which performs certificate verification is vulnerable including
6878 OpenSSL clients and servers which enable client authentication.
6883 * ASN.1 structure reuse memory corruption fix
6885 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6886 memory corruption via an invalid write. Such reuse is and has been
6887 strongly discouraged and is believed to be rare.
6889 Applications that parse structures containing CHOICE or ANY DEFINED BY
6890 components may be affected. Certificate parsing (d2i_X509 and related
6891 functions) are however not affected. OpenSSL clients and servers are
6897 * PKCS7 NULL pointer dereferences fix
6899 The PKCS#7 parsing code does not handle missing outer ContentInfo
6900 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6901 missing content and trigger a NULL pointer dereference on parsing.
6903 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6904 otherwise parse PKCS#7 structures from untrusted sources are
6905 affected. OpenSSL clients and servers are not affected.
6907 This issue was reported to OpenSSL by Michal Zalewski (Google).
6912 * DoS via reachable assert in SSLv2 servers fix
6914 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6915 servers that both support SSLv2 and enable export cipher suites by sending
6916 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6918 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6919 (OpenSSL development team).
6924 * Use After Free following d2i_ECPrivatekey error fix
6926 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6927 could cause a use after free condition. This, in turn, could cause a double
6928 free in several private key parsing functions (such as d2i_PrivateKey
6929 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6930 for applications that receive EC private keys from untrusted
6931 sources. This scenario is considered rare.
6933 This issue was discovered by the BoringSSL project and fixed in their
6939 * X509_to_X509_REQ NULL pointer deref fix
6941 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6942 the certificate key is invalid. This function is rarely used in practice.
6944 This issue was discovered by Brian Carpenter.
6949 * Removed the export ciphers from the DEFAULT ciphers
6953 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6955 * Build fixes for the Windows and OpenVMS platforms
6957 *Matt Caswell and Richard Levitte*
6959 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6961 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6962 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6963 dereference. This could lead to a Denial Of Service attack. Thanks to
6964 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6969 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6970 dtls1_buffer_record function under certain conditions. In particular this
6971 could occur if an attacker sent repeated DTLS records with the same
6972 sequence number but for the next epoch. The memory leak could be exploited
6973 by an attacker in a Denial of Service attack through memory exhaustion.
6974 Thanks to Chris Mueller for reporting this issue.
6979 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6980 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6981 method would be set to NULL which could later result in a NULL pointer
6982 dereference. Thanks to Frank Schmirler for reporting this issue.
6987 * Abort handshake if server key exchange message is omitted for ephemeral
6990 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6991 reporting this issue.
6996 * Remove non-export ephemeral RSA code on client and server. This code
6997 violated the TLS standard by allowing the use of temporary RSA keys in
6998 non-export ciphersuites and could be used by a server to effectively
6999 downgrade the RSA key length used to a value smaller than the server
7000 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7001 INRIA or reporting this issue.
7006 * Fixed issue where DH client certificates are accepted without verification.
7007 An OpenSSL server will accept a DH certificate for client authentication
7008 without the certificate verify message. This effectively allows a client to
7009 authenticate without the use of a private key. This only affects servers
7010 which trust a client certificate authority which issues certificates
7011 containing DH keys: these are extremely rare and hardly ever encountered.
7012 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7018 * Ensure that the session ID context of an SSL is updated when its
7019 SSL_CTX is updated via SSL_set_SSL_CTX.
7021 The session ID context is typically set from the parent SSL_CTX,
7022 and can vary with the CTX.
7026 * Fix various certificate fingerprint issues.
7028 By using non-DER or invalid encodings outside the signed portion of a
7029 certificate the fingerprint can be changed without breaking the signature.
7030 Although no details of the signed portion of the certificate can be changed
7031 this can cause problems with some applications: e.g. those using the
7032 certificate fingerprint for blacklists.
7034 1. Reject signatures with non zero unused bits.
7036 If the BIT STRING containing the signature has non zero unused bits reject
7037 the signature. All current signature algorithms require zero unused bits.
7039 2. Check certificate algorithm consistency.
7041 Check the AlgorithmIdentifier inside TBS matches the one in the
7042 certificate signature. NB: this will result in signature failure
7043 errors for some broken certificates.
7045 Thanks to Konrad Kraszewski from Google for reporting this issue.
7047 3. Check DSA/ECDSA signatures use DER.
7049 Re-encode DSA/ECDSA signatures and compare with the original received
7050 signature. Return an error if there is a mismatch.
7052 This will reject various cases including garbage after signature
7053 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7054 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7055 (negative or with leading zeroes).
7057 Further analysis was conducted and fixes were developed by Stephen Henson
7058 of the OpenSSL core team.
7064 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7065 results on some platforms, including x86_64. This bug occurs at random
7066 with a very low probability, and is not known to be exploitable in any
7067 way, though its exact impact is difficult to determine. Thanks to Pieter
7068 Wuille (Blockstream) who reported this issue and also suggested an initial
7069 fix. Further analysis was conducted by the OpenSSL development team and
7070 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7071 the OpenSSL core team.
7076 * Do not resume sessions on the server if the negotiated protocol
7077 version does not match the session's version. Resuming with a different
7078 version, while not strictly forbidden by the RFC, is of questionable
7079 sanity and breaks all known clients.
7081 *David Benjamin, Emilia Käsper*
7083 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7084 early CCS messages during renegotiation. (Note that because
7085 renegotiation is encrypted, this early CCS was not exploitable.)
7089 * Tighten client-side session ticket handling during renegotiation:
7090 ensure that the client only accepts a session ticket if the server sends
7091 the extension anew in the ServerHello. Previously, a TLS client would
7092 reuse the old extension state and thus accept a session ticket if one was
7093 announced in the initial ServerHello.
7095 Similarly, ensure that the client requires a session ticket if one
7096 was advertised in the ServerHello. Previously, a TLS client would
7097 ignore a missing NewSessionTicket message.
7101 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7105 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7106 sends a carefully crafted handshake message, to cause OpenSSL to fail
7107 to free up to 64k of memory causing a memory leak. This could be
7108 exploited in a Denial Of Service attack. This issue affects OpenSSL
7109 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7110 whether SRTP is used or configured. Implementations of OpenSSL that
7111 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7113 The fix was developed by the OpenSSL team.
7118 * Session Ticket Memory Leak.
7120 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7121 integrity of that ticket is first verified. In the event of a session
7122 ticket integrity check failing, OpenSSL will fail to free memory
7123 causing a memory leak. By sending a large number of invalid session
7124 tickets an attacker could exploit this issue in a Denial Of Service
7130 * Build option no-ssl3 is incomplete.
7132 When OpenSSL is configured with "no-ssl3" as a build option, servers
7133 could accept and complete a SSL 3.0 handshake, and clients could be
7134 configured to send them.
7137 *Akamai and the OpenSSL team*
7139 * Add support for TLS_FALLBACK_SCSV.
7140 Client applications doing fallback retries should call
7141 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7144 *Adam Langley, Bodo Moeller*
7146 * Add additional DigestInfo checks.
7148 Re-encode DigestInto in DER and check against the original when
7149 verifying RSA signature: this will reject any improperly encoded
7150 DigestInfo structures.
7152 Note: this is a precautionary measure and no attacks are currently known.
7156 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7158 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7159 SRP code can be overrun an internal buffer. Add sanity check that
7160 g, A, B < N to SRP code.
7162 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7163 Group for discovering this issue.
7168 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7169 TLS 1.0 instead of higher protocol versions when the ClientHello message
7170 is badly fragmented. This allows a man-in-the-middle attacker to force a
7171 downgrade to TLS 1.0 even if both the server and the client support a
7172 higher protocol version, by modifying the client's TLS records.
7174 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7175 researching this issue.
7180 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7181 to a denial of service attack. A malicious server can crash the client
7182 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7183 ciphersuite and sending carefully crafted handshake messages.
7185 Thanks to Felix Gröbert (Google) for discovering and researching this
7191 * By sending carefully crafted DTLS packets an attacker could cause openssl
7192 to leak memory. This can be exploited through a Denial of Service attack.
7193 Thanks to Adam Langley for discovering and researching this issue.
7198 * An attacker can force openssl to consume large amounts of memory whilst
7199 processing DTLS handshake messages. This can be exploited through a
7200 Denial of Service attack.
7201 Thanks to Adam Langley for discovering and researching this issue.
7206 * An attacker can force an error condition which causes openssl to crash
7207 whilst processing DTLS packets due to memory being freed twice. This
7208 can be exploited through a Denial of Service attack.
7209 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7215 * If a multithreaded client connects to a malicious server using a resumed
7216 session and the server sends an ec point format extension it could write
7217 up to 255 bytes to freed memory.
7219 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7225 * A malicious server can crash an OpenSSL client with a null pointer
7226 dereference (read) by specifying an SRP ciphersuite even though it was not
7227 properly negotiated with the client. This can be exploited through a
7228 Denial of Service attack.
7230 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7231 discovering and researching this issue.
7236 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7237 X509_name_oneline, X509_name_print_ex et al. to leak some information
7238 from the stack. Applications may be affected if they echo pretty printing
7239 output to the attacker.
7241 Thanks to Ivan Fratric (Google) for discovering this issue.
7244 *Emilia Käsper, and Steve Henson*
7246 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7247 for corner cases. (Certain input points at infinity could lead to
7248 bogus results, with non-infinity inputs mapped to infinity too.)
7252 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7254 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7255 handshake can force the use of weak keying material in OpenSSL
7256 SSL/TLS clients and servers.
7258 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7259 researching this issue. ([CVE-2014-0224])
7261 *KIKUCHI Masashi, Steve Henson*
7263 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7264 OpenSSL DTLS client the code can be made to recurse eventually crashing
7267 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7270 *Imre Rad, Steve Henson*
7272 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7273 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7274 client or server. This is potentially exploitable to run arbitrary
7275 code on a vulnerable client or server.
7277 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7279 *Jüri Aedla, Steve Henson*
7281 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7282 are subject to a denial of service attack.
7284 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7285 this issue. ([CVE-2014-3470])
7287 *Felix Gröbert, Ivan Fratric, Steve Henson*
7289 * Harmonize version and its documentation. -f flag is used to display
7292 *mancha <mancha1@zoho.com>*
7294 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7295 in i2d_ECPrivateKey.
7297 *mancha <mancha1@zoho.com>*
7299 * Fix some double frees. These are not thought to be exploitable.
7301 *mancha <mancha1@zoho.com>*
7303 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7305 * A missing bounds check in the handling of the TLS heartbeat extension
7306 can be used to reveal up to 64k of memory to a connected client or
7309 Thanks for Neel Mehta of Google Security for discovering this bug and to
7310 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7311 preparing the fix ([CVE-2014-0160])
7313 *Adam Langley, Bodo Moeller*
7315 * Fix for the attack described in the paper "Recovering OpenSSL
7316 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7317 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7318 <http://eprint.iacr.org/2014/140>
7320 Thanks to Yuval Yarom and Naomi Benger for discovering this
7321 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7323 *Yuval Yarom and Naomi Benger*
7325 * TLS pad extension: draft-agl-tls-padding-03
7327 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7328 TLS client Hello record length value would otherwise be > 255 and
7329 less that 512 pad with a dummy extension containing zeroes so it
7330 is at least 512 bytes long.
7332 *Adam Langley, Steve Henson*
7334 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7336 * Fix for TLS record tampering bug. A carefully crafted invalid
7337 handshake could crash OpenSSL with a NULL pointer exception.
7338 Thanks to Anton Johansson for reporting this issues.
7341 * Keep original DTLS digest and encryption contexts in retransmission
7342 structures so we can use the previous session parameters if they need
7343 to be resent. ([CVE-2013-6450])
7347 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7348 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7349 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7350 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7351 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7352 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7354 *Rob Stradling, Adam Langley*
7356 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7358 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7359 supporting platforms or when small records were transferred.
7361 *Andy Polyakov, Steve Henson*
7363 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7365 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7367 This addresses the flaw in CBC record processing discovered by
7368 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7369 at: <http://www.isg.rhul.ac.uk/tls/>
7371 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7372 Security Group at Royal Holloway, University of London
7373 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7374 Emilia Käsper for the initial patch.
7377 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7379 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7380 ciphersuites which can be exploited in a denial of service attack.
7381 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7382 and detecting this bug and to Wolfgang Ettlinger
7383 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7388 * Return an error when checking OCSP signatures when key is NULL.
7389 This fixes a DoS attack. ([CVE-2013-0166])
7393 * Make openssl verify return errors.
7395 *Chris Palmer <palmer@google.com> and Ben Laurie*
7397 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7398 the right response is stapled. Also change SSL_get_certificate()
7399 so it returns the certificate actually sent.
7400 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7402 *Rob Stradling <rob.stradling@comodo.com>*
7404 * Fix possible deadlock when decoding public keys.
7408 * Don't use TLS 1.0 record version number in initial client hello
7413 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7415 * Sanity check record length before skipping explicit IV in TLS
7416 1.2, 1.1 and DTLS to fix DoS attack.
7418 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7419 fuzzing as a service testing platform.
7424 * Initialise tkeylen properly when encrypting CMS messages.
7425 Thanks to Solar Designer of Openwall for reporting this issue.
7429 * In FIPS mode don't try to use composite ciphers as they are not
7434 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7436 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7437 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7438 mean any application compiled against OpenSSL 1.0.0 headers setting
7439 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7440 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7441 0x10000000L Any application which was previously compiled against
7442 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7443 will need to be recompiled as a result. Letting be results in
7444 inability to disable specifically TLS 1.1 and in client context,
7445 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7449 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7450 disable just protocol X, but all protocols above X *if* there are
7451 protocols *below* X still enabled. In more practical terms it means
7452 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7453 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7454 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7459 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7461 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7462 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7463 in CRYPTO_realloc_clean.
7465 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7466 issue and to Adam Langley <agl@chromium.org> for fixing it.
7469 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7471 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7475 * Workarounds for some broken servers that "hang" if a client hello
7476 record length exceeds 255 bytes.
7478 1. Do not use record version number > TLS 1.0 in initial client
7479 hello: some (but not all) hanging servers will now work.
7480 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7481 the number of ciphers sent in the client hello. This should be
7482 set to an even number, such as 50, for example by passing:
7483 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7484 Most broken servers should now work.
7485 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7486 TLS 1.2 client support entirely.
7490 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7494 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7496 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7497 STRING form instead of a DigestInfo.
7501 * The format used for MDC2 RSA signatures is inconsistent between EVP
7502 and the RSA_sign/RSA_verify functions. This was made more apparent when
7503 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7504 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7505 the correct format in RSA_verify so both forms transparently work.
7509 * Some servers which support TLS 1.0 can choke if we initially indicate
7510 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7511 encrypted premaster secret. As a workaround use the maximum permitted
7512 client version in client hello, this should keep such servers happy
7513 and still work with previous versions of OpenSSL.
7517 * Add support for TLS/DTLS heartbeats.
7519 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7521 * Add support for SCTP.
7523 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7525 * Improved PRNG seeding for VOS.
7527 *Paul Green <Paul.Green@stratus.com>*
7529 * Extensive assembler packs updates, most notably:
7531 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7532 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7533 - x86_64: bit-sliced AES implementation;
7534 - ARM: NEON support, contemporary platforms optimizations;
7535 - s390x: z196 support;
7536 - `*`: GHASH and GF(2^m) multiplication implementations;
7540 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7541 (removal of unnecessary code)
7543 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7545 * Add TLS key material exporter from RFC 5705.
7549 * Add DTLS-SRTP negotiation from RFC 5764.
7553 * Add Next Protocol Negotiation,
7554 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7555 disabled with a no-npn flag to config or Configure. Code donated
7558 *Adam Langley <agl@google.com> and Ben Laurie*
7560 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7561 NIST-P256, NIST-P521, with constant-time single point multiplication on
7562 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7563 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7564 Code made available under Apache License version 2.0.
7566 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7567 line to include this in your build of OpenSSL, and run "make depend" (or
7568 "make update"). This enables the following EC_METHODs:
7570 EC_GFp_nistp224_method()
7571 EC_GFp_nistp256_method()
7572 EC_GFp_nistp521_method()
7574 EC_GROUP_new_by_curve_name() will automatically use these (while
7575 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7578 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7580 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7581 all platforms. Move ssize_t definition from e_os.h to the public
7582 header file e_os2.h as it now appears in public header file cms.h
7586 * New -sigopt option to the ca, req and x509 utilities. Additional
7587 signature parameters can be passed using this option and in
7592 * Add RSA PSS signing function. This will generate and set the
7593 appropriate AlgorithmIdentifiers for PSS based on those in the
7594 corresponding EVP_MD_CTX structure. No application support yet.
7598 * Support for companion algorithm specific ASN1 signing routines.
7599 New function ASN1_item_sign_ctx() signs a pre-initialised
7600 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7601 the appropriate parameters.
7605 * Add new algorithm specific ASN1 verification initialisation function
7606 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7607 handling will be the same no matter what EVP_PKEY_METHOD is used.
7608 Add a PSS handler to support verification of PSS signatures: checked
7609 against a number of sample certificates.
7613 * Add signature printing for PSS. Add PSS OIDs.
7615 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7617 * Add algorithm specific signature printing. An individual ASN1 method
7618 can now print out signatures instead of the standard hex dump.
7620 More complex signatures (e.g. PSS) can print out more meaningful
7621 information. Include DSA version that prints out the signature
7626 * Password based recipient info support for CMS library: implementing
7631 * Split password based encryption into PBES2 and PBKDF2 functions. This
7632 neatly separates the code into cipher and PBE sections and is required
7633 for some algorithms that split PBES2 into separate pieces (such as
7634 password based CMS).
7638 * Session-handling fixes:
7639 - Fix handling of connections that are resuming with a session ID,
7640 but also support Session Tickets.
7641 - Fix a bug that suppressed issuing of a new ticket if the client
7642 presented a ticket with an expired session.
7643 - Try to set the ticket lifetime hint to something reasonable.
7644 - Make tickets shorter by excluding irrelevant information.
7645 - On the client side, don't ignore renewed tickets.
7647 *Adam Langley, Bodo Moeller (Google)*
7649 * Fix PSK session representation.
7653 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7655 This work was sponsored by Intel.
7659 * Add GCM support to TLS library. Some custom code is needed to split
7660 the IV between the fixed (from PRF) and explicit (from TLS record)
7661 portions. This adds all GCM ciphersuites supported by RFC5288 and
7662 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7663 add a special AESGCM string for GCM only.
7667 * Expand range of ctrls for AES GCM. Permit setting invocation
7668 field on decrypt and retrieval of invocation field only on encrypt.
7672 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7673 As required by RFC5289 these ciphersuites cannot be used if for
7674 versions of TLS earlier than 1.2.
7678 * For FIPS capable OpenSSL interpret a NULL default public key method
7679 as unset and return the appropriate default but do *not* set the default.
7680 This means we can return the appropriate method in applications that
7681 switch between FIPS and non-FIPS modes.
7685 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7686 ENGINE is used then we cannot handle that in the FIPS module so we
7687 keep original code iff non-FIPS operations are allowed.
7691 * Add -attime option to openssl utilities.
7693 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7695 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7699 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7700 FIPS EC methods unconditionally for now.
7704 * New build option no-ec2m to disable characteristic 2 code.
7708 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7709 all cases can be covered as some introduce binary incompatibilities.
7713 * Redirect RSA operations to FIPS module including keygen,
7714 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7718 * Add similar low-level API blocking to ciphers.
7722 * low-level digest APIs are not approved in FIPS mode: any attempt
7723 to use these will cause a fatal error. Applications that *really* want
7724 to use them can use the `private_*` version instead.
7728 * Redirect cipher operations to FIPS module for FIPS builds.
7732 * Redirect digest operations to FIPS module for FIPS builds.
7736 * Update build system to add "fips" flag which will link in fipscanister.o
7737 for static and shared library builds embedding a signature if needed.
7741 * Output TLS supported curves in preference order instead of numerical
7742 order. This is currently hardcoded for the highest order curves first.
7743 This should be configurable so applications can judge speed vs strength.
7747 * Add TLS v1.2 server support for client authentication.
7751 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7756 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7757 FIPS modules versions.
7761 * Add TLS v1.2 client side support for client authentication. Keep cache
7762 of handshake records longer as we don't know the hash algorithm to use
7763 until after the certificate request message is received.
7767 * Initial TLS v1.2 client support. Add a default signature algorithms
7768 extension including all the algorithms we support. Parse new signature
7769 format in client key exchange. Relax some ECC signing restrictions for
7770 TLS v1.2 as indicated in RFC5246.
7774 * Add server support for TLS v1.2 signature algorithms extension. Switch
7775 to new signature format when needed using client digest preference.
7776 All server ciphersuites should now work correctly in TLS v1.2. No client
7777 support yet and no support for client certificates.
7781 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7782 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7783 ciphersuites. At present only RSA key exchange ciphersuites work with
7784 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7785 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7786 and version checking.
7790 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7791 with this defined it will not be affected by any changes to ssl internal
7792 structures. Add several utility functions to allow openssl application
7793 to work with OPENSSL_NO_SSL_INTERN defined.
7797 * A long standing patch to add support for SRP from EdelWeb (Peter
7798 Sylvester and Christophe Renou) was integrated.
7799 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7800 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7803 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7807 * Permit abbreviated handshakes when renegotiating using the function
7808 SSL_renegotiate_abbreviated().
7810 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7812 * Add call to ENGINE_register_all_complete() to
7813 ENGINE_load_builtin_engines(), so some implementations get used
7814 automatically instead of needing explicit application support.
7818 * Add support for TLS key exporter as described in RFC5705.
7820 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7822 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7823 a few changes are required:
7825 Add SSL_OP_NO_TLSv1_1 flag.
7826 Add TLSv1_1 methods.
7827 Update version checking logic to handle version 1.1.
7828 Add explicit IV handling (ported from DTLS code).
7829 Add command line options to s_client/s_server.
7836 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7838 * X509_ATTRIBUTE memory leak
7840 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7841 memory. This structure is used by the PKCS#7 and CMS routines so any
7842 application which reads PKCS#7 or CMS data from untrusted sources is
7843 affected. SSL/TLS is not affected.
7845 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7851 * Race condition handling PSK identify hint
7853 If PSK identity hints are received by a multi-threaded client then
7854 the values are wrongly updated in the parent SSL_CTX structure. This can
7855 result in a race condition potentially leading to a double free of the
7861 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7863 * Malformed ECParameters causes infinite loop
7865 When processing an ECParameters structure OpenSSL enters an infinite loop
7866 if the curve specified is over a specially malformed binary polynomial
7869 This can be used to perform denial of service against any
7870 system which processes public keys, certificate requests or
7871 certificates. This includes TLS clients and TLS servers with
7872 client authentication enabled.
7874 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7879 * Exploitable out-of-bounds read in X509_cmp_time
7881 X509_cmp_time does not properly check the length of the ASN1_TIME
7882 string and can read a few bytes out of bounds. In addition,
7883 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7886 An attacker can use this to craft malformed certificates and CRLs of
7887 various sizes and potentially cause a segmentation fault, resulting in
7888 a DoS on applications that verify certificates or CRLs. TLS clients
7889 that verify CRLs are affected. TLS clients and servers with client
7890 authentication enabled may be affected if they use custom verification
7893 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7894 independently by Hanno Böck.
7899 * PKCS7 crash with missing EnvelopedContent
7901 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7902 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7903 with missing content and trigger a NULL pointer dereference on parsing.
7905 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7906 structures from untrusted sources are affected. OpenSSL clients and
7907 servers are not affected.
7909 This issue was reported to OpenSSL by Michal Zalewski (Google).
7914 * CMS verify infinite loop with unknown hash function
7916 When verifying a signedData message the CMS code can enter an infinite loop
7917 if presented with an unknown hash function OID. This can be used to perform
7918 denial of service against any system which verifies signedData messages using
7920 This issue was reported to OpenSSL by Johannes Bauer.
7925 * Race condition handling NewSessionTicket
7927 If a NewSessionTicket is received by a multi-threaded client when attempting to
7928 reuse a previous ticket then a race condition can occur potentially leading to
7929 a double free of the ticket data.
7934 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7936 * Segmentation fault in ASN1_TYPE_cmp fix
7938 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7939 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7940 certificate signature algorithm consistency this can be used to crash any
7941 certificate verification operation and exploited in a DoS attack. Any
7942 application which performs certificate verification is vulnerable including
7943 OpenSSL clients and servers which enable client authentication.
7948 * ASN.1 structure reuse memory corruption fix
7950 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7951 memory corruption via an invalid write. Such reuse is and has been
7952 strongly discouraged and is believed to be rare.
7954 Applications that parse structures containing CHOICE or ANY DEFINED BY
7955 components may be affected. Certificate parsing (d2i_X509 and related
7956 functions) are however not affected. OpenSSL clients and servers are
7962 * PKCS7 NULL pointer dereferences fix
7964 The PKCS#7 parsing code does not handle missing outer ContentInfo
7965 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7966 missing content and trigger a NULL pointer dereference on parsing.
7968 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7969 otherwise parse PKCS#7 structures from untrusted sources are
7970 affected. OpenSSL clients and servers are not affected.
7972 This issue was reported to OpenSSL by Michal Zalewski (Google).
7977 * DoS via reachable assert in SSLv2 servers fix
7979 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7980 servers that both support SSLv2 and enable export cipher suites by sending
7981 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7983 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7984 (OpenSSL development team).
7989 * Use After Free following d2i_ECPrivatekey error fix
7991 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7992 could cause a use after free condition. This, in turn, could cause a double
7993 free in several private key parsing functions (such as d2i_PrivateKey
7994 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7995 for applications that receive EC private keys from untrusted
7996 sources. This scenario is considered rare.
7998 This issue was discovered by the BoringSSL project and fixed in their
8004 * X509_to_X509_REQ NULL pointer deref fix
8006 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8007 the certificate key is invalid. This function is rarely used in practice.
8009 This issue was discovered by Brian Carpenter.
8014 * Removed the export ciphers from the DEFAULT ciphers
8018 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8020 * Build fixes for the Windows and OpenVMS platforms
8022 *Matt Caswell and Richard Levitte*
8024 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8026 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8027 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8028 dereference. This could lead to a Denial Of Service attack. Thanks to
8029 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8034 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8035 dtls1_buffer_record function under certain conditions. In particular this
8036 could occur if an attacker sent repeated DTLS records with the same
8037 sequence number but for the next epoch. The memory leak could be exploited
8038 by an attacker in a Denial of Service attack through memory exhaustion.
8039 Thanks to Chris Mueller for reporting this issue.
8044 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8045 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8046 method would be set to NULL which could later result in a NULL pointer
8047 dereference. Thanks to Frank Schmirler for reporting this issue.
8052 * Abort handshake if server key exchange message is omitted for ephemeral
8055 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8056 reporting this issue.
8061 * Remove non-export ephemeral RSA code on client and server. This code
8062 violated the TLS standard by allowing the use of temporary RSA keys in
8063 non-export ciphersuites and could be used by a server to effectively
8064 downgrade the RSA key length used to a value smaller than the server
8065 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8066 INRIA or reporting this issue.
8071 * Fixed issue where DH client certificates are accepted without verification.
8072 An OpenSSL server will accept a DH certificate for client authentication
8073 without the certificate verify message. This effectively allows a client to
8074 authenticate without the use of a private key. This only affects servers
8075 which trust a client certificate authority which issues certificates
8076 containing DH keys: these are extremely rare and hardly ever encountered.
8077 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8083 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8084 results on some platforms, including x86_64. This bug occurs at random
8085 with a very low probability, and is not known to be exploitable in any
8086 way, though its exact impact is difficult to determine. Thanks to Pieter
8087 Wuille (Blockstream) who reported this issue and also suggested an initial
8088 fix. Further analysis was conducted by the OpenSSL development team and
8089 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8090 the OpenSSL core team.
8095 * Fix various certificate fingerprint issues.
8097 By using non-DER or invalid encodings outside the signed portion of a
8098 certificate the fingerprint can be changed without breaking the signature.
8099 Although no details of the signed portion of the certificate can be changed
8100 this can cause problems with some applications: e.g. those using the
8101 certificate fingerprint for blacklists.
8103 1. Reject signatures with non zero unused bits.
8105 If the BIT STRING containing the signature has non zero unused bits reject
8106 the signature. All current signature algorithms require zero unused bits.
8108 2. Check certificate algorithm consistency.
8110 Check the AlgorithmIdentifier inside TBS matches the one in the
8111 certificate signature. NB: this will result in signature failure
8112 errors for some broken certificates.
8114 Thanks to Konrad Kraszewski from Google for reporting this issue.
8116 3. Check DSA/ECDSA signatures use DER.
8118 Reencode DSA/ECDSA signatures and compare with the original received
8119 signature. Return an error if there is a mismatch.
8121 This will reject various cases including garbage after signature
8122 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8123 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8124 (negative or with leading zeroes).
8126 Further analysis was conducted and fixes were developed by Stephen Henson
8127 of the OpenSSL core team.
8133 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8135 * Session Ticket Memory Leak.
8137 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8138 integrity of that ticket is first verified. In the event of a session
8139 ticket integrity check failing, OpenSSL will fail to free memory
8140 causing a memory leak. By sending a large number of invalid session
8141 tickets an attacker could exploit this issue in a Denial Of Service
8147 * Build option no-ssl3 is incomplete.
8149 When OpenSSL is configured with "no-ssl3" as a build option, servers
8150 could accept and complete a SSL 3.0 handshake, and clients could be
8151 configured to send them.
8154 *Akamai and the OpenSSL team*
8156 * Add support for TLS_FALLBACK_SCSV.
8157 Client applications doing fallback retries should call
8158 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8161 *Adam Langley, Bodo Moeller*
8163 * Add additional DigestInfo checks.
8165 Reencode DigestInto in DER and check against the original when
8166 verifying RSA signature: this will reject any improperly encoded
8167 DigestInfo structures.
8169 Note: this is a precautionary measure and no attacks are currently known.
8173 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8175 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8176 to a denial of service attack. A malicious server can crash the client
8177 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8178 ciphersuite and sending carefully crafted handshake messages.
8180 Thanks to Felix Gröbert (Google) for discovering and researching this
8186 * By sending carefully crafted DTLS packets an attacker could cause openssl
8187 to leak memory. This can be exploited through a Denial of Service attack.
8188 Thanks to Adam Langley for discovering and researching this issue.
8193 * An attacker can force openssl to consume large amounts of memory whilst
8194 processing DTLS handshake messages. This can be exploited through a
8195 Denial of Service attack.
8196 Thanks to Adam Langley for discovering and researching this issue.
8201 * An attacker can force an error condition which causes openssl to crash
8202 whilst processing DTLS packets due to memory being freed twice. This
8203 can be exploited through a Denial of Service attack.
8204 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8210 * If a multithreaded client connects to a malicious server using a resumed
8211 session and the server sends an ec point format extension it could write
8212 up to 255 bytes to freed memory.
8214 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8220 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8221 X509_name_oneline, X509_name_print_ex et al. to leak some information
8222 from the stack. Applications may be affected if they echo pretty printing
8223 output to the attacker.
8225 Thanks to Ivan Fratric (Google) for discovering this issue.
8228 *Emilia Käsper, and Steve Henson*
8230 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8231 for corner cases. (Certain input points at infinity could lead to
8232 bogus results, with non-infinity inputs mapped to infinity too.)
8236 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8238 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8239 handshake can force the use of weak keying material in OpenSSL
8240 SSL/TLS clients and servers.
8242 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8243 researching this issue. ([CVE-2014-0224])
8245 *KIKUCHI Masashi, Steve Henson*
8247 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8248 OpenSSL DTLS client the code can be made to recurse eventually crashing
8251 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8254 *Imre Rad, Steve Henson*
8256 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8257 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8258 client or server. This is potentially exploitable to run arbitrary
8259 code on a vulnerable client or server.
8261 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8263 *Jüri Aedla, Steve Henson*
8265 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8266 are subject to a denial of service attack.
8268 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8269 this issue. ([CVE-2014-3470])
8271 *Felix Gröbert, Ivan Fratric, Steve Henson*
8273 * Harmonize version and its documentation. -f flag is used to display
8276 *mancha <mancha1@zoho.com>*
8278 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8279 in i2d_ECPrivateKey.
8281 *mancha <mancha1@zoho.com>*
8283 * Fix some double frees. These are not thought to be exploitable.
8285 *mancha <mancha1@zoho.com>*
8287 * Fix for the attack described in the paper "Recovering OpenSSL
8288 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8289 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8290 <http://eprint.iacr.org/2014/140>
8292 Thanks to Yuval Yarom and Naomi Benger for discovering this
8293 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8295 *Yuval Yarom and Naomi Benger*
8297 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8299 * Keep original DTLS digest and encryption contexts in retransmission
8300 structures so we can use the previous session parameters if they need
8301 to be resent. ([CVE-2013-6450])
8305 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8306 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8307 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8308 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8309 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8310 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8312 *Rob Stradling, Adam Langley*
8314 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8316 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8318 This addresses the flaw in CBC record processing discovered by
8319 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8320 at: <http://www.isg.rhul.ac.uk/tls/>
8322 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8323 Security Group at Royal Holloway, University of London
8324 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8325 Emilia Käsper for the initial patch.
8328 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8330 * Return an error when checking OCSP signatures when key is NULL.
8331 This fixes a DoS attack. ([CVE-2013-0166])
8335 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8336 the right response is stapled. Also change SSL_get_certificate()
8337 so it returns the certificate actually sent.
8338 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8339 (This is a backport)
8341 *Rob Stradling <rob.stradling@comodo.com>*
8343 * Fix possible deadlock when decoding public keys.
8347 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8349 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8352 * Sanity check record length before skipping explicit IV in DTLS
8355 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8356 fuzzing as a service testing platform.
8361 * Initialise tkeylen properly when encrypting CMS messages.
8362 Thanks to Solar Designer of Openwall for reporting this issue.
8366 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8368 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8369 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8370 in CRYPTO_realloc_clean.
8372 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8373 issue and to Adam Langley <agl@chromium.org> for fixing it.
8376 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8378 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8380 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8381 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8382 content decryption and always return the same error. Note: this attack
8383 needs on average 2^20 messages so it only affects automated senders. The
8384 old behaviour can be re-enabled in the CMS code by setting the
8385 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8386 an MMA defence is not necessary.
8387 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8388 this issue. ([CVE-2012-0884])
8392 * Fix CVE-2011-4619: make sure we really are receiving a
8393 client hello before rejecting multiple SGC restarts. Thanks to
8394 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8398 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8400 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8401 Thanks to Antonio Martin, Enterprise Secure Access Research and
8402 Development, Cisco Systems, Inc. for discovering this bug and
8403 preparing a fix. ([CVE-2012-0050])
8407 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8409 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8410 of the Vaudenay padding oracle attack on CBC mode encryption
8411 which enables an efficient plaintext recovery attack against
8412 the OpenSSL implementation of DTLS. Their attack exploits timing
8413 differences arising during decryption processing. A research
8414 paper describing this attack can be found at:
8415 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8416 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8417 Security Group at Royal Holloway, University of London
8418 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8419 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8420 for preparing the fix. ([CVE-2011-4108])
8422 *Robin Seggelmann, Michael Tuexen*
8424 * Clear bytes used for block padding of SSL 3.0 records.
8427 *Adam Langley (Google)*
8429 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8430 Kadianakis <desnacked@gmail.com> for discovering this issue and
8431 Adam Langley for preparing the fix. ([CVE-2011-4619])
8433 *Adam Langley (Google)*
8435 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8437 *Andrey Kulikov <amdeich@gmail.com>*
8439 * Prevent malformed RFC3779 data triggering an assertion failure.
8440 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8441 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8443 *Rob Austein <sra@hactrn.net>*
8445 * Improved PRNG seeding for VOS.
8447 *Paul Green <Paul.Green@stratus.com>*
8449 * Fix ssl_ciph.c set-up race.
8451 *Adam Langley (Google)*
8453 * Fix spurious failures in ecdsatest.c.
8455 *Emilia Käsper (Google)*
8457 * Fix the BIO_f_buffer() implementation (which was mixing different
8458 interpretations of the `..._len` fields).
8460 *Adam Langley (Google)*
8462 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8463 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8464 threads won't reuse the same blinding coefficients.
8466 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8467 lock to call BN_BLINDING_invert_ex, and avoids one use of
8468 BN_BLINDING_update for each BN_BLINDING structure (previously,
8469 the last update always remained unused).
8471 *Emilia Käsper (Google)*
8473 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8475 *Bob Buckholz (Google)*
8477 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8479 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8480 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8482 *Kaspar Brand <ossl@velox.ch>*
8484 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8485 for multi-threaded use of ECDH. ([CVE-2011-3210])
8487 *Adam Langley (Google)*
8489 * Fix x509_name_ex_d2i memory leak on bad inputs.
8493 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8494 signature public key algorithm by using OID xref utilities instead.
8495 Before this you could only use some ECC ciphersuites with SHA1 only.
8499 * Add protection against ECDSA timing attacks as mentioned in the paper
8500 by Billy Bob Brumley and Nicola Tuveri, see:
8501 <http://eprint.iacr.org/2011/232.pdf>
8503 *Billy Bob Brumley and Nicola Tuveri*
8505 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8507 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8509 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8511 * Fix bug in string printing code: if *any* escaping is enabled we must
8512 escape the escape character (backslash) or the resulting string is
8517 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8519 * Disable code workaround for ancient and obsolete Netscape browsers
8520 and servers: an attacker can use it in a ciphersuite downgrade attack.
8521 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8525 * Fixed J-PAKE implementation error, originally discovered by
8526 Sebastien Martini, further info and confirmation from Stefan
8527 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8531 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8533 * Fix extension code to avoid race conditions which can result in a buffer
8534 overrun vulnerability: resumed sessions must not be modified as they can
8535 be shared by multiple threads. CVE-2010-3864
8539 * Fix WIN32 build system to correctly link an ENGINE directory into
8544 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8546 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8549 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8551 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8553 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8554 context. The operation can be customised via the ctrl mechanism in
8555 case ENGINEs want to include additional functionality.
8559 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8563 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8564 output hashes compatible with older versions of OpenSSL.
8566 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8568 * Fix compression algorithm handling: if resuming a session use the
8569 compression algorithm of the resumed session instead of determining
8570 it from client hello again. Don't allow server to change algorithm.
8574 * Add load_crls() function to commands tidying load_certs() too. Add option
8575 to verify utility to allow additional CRLs to be included.
8579 * Update OCSP request code to permit adding custom headers to the request:
8580 some responders need this.
8584 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8587 *Julia Lawall <julia@diku.dk>*
8589 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8590 needlessly dereferenced structures, used obsolete functions and
8591 didn't handle all updated verify codes correctly.
8595 * Disable MD2 in the default configuration.
8599 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8600 indicate the initial BIO being pushed or popped. This makes it possible
8601 to determine whether the BIO is the one explicitly called or as a result
8602 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8603 it handles reference counts correctly and doesn't zero out the I/O bio
8604 when it is not being explicitly popped. WARNING: applications which
8605 included workarounds for the old buggy behaviour will need to be modified
8606 or they could free up already freed BIOs.
8610 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8611 renaming to all platforms (within the 0.9.8 branch, this was
8612 done conditionally on Netware platforms to avoid a name clash).
8614 *Guenter <lists@gknw.net>*
8616 * Add ECDHE and PSK support to DTLS.
8618 *Michael Tuexen <tuexen@fh-muenster.de>*
8620 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8625 * Add "missing" function EVP_MD_flags() (without this the only way to
8626 retrieve a digest flags is by accessing the structure directly. Update
8627 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8628 or cipher is registered as in the "from" argument. Print out all
8629 registered digests in the dgst usage message instead of manually
8630 attempting to work them out.
8634 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8635 this allows the use of compression and extensions. Change default cipher
8636 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8637 by default unless an application cipher string requests it.
8641 * Alter match criteria in PKCS12_parse(). It used to try to use local
8642 key ids to find matching certificates and keys but some PKCS#12 files
8643 don't follow the (somewhat unwritten) rules and this strategy fails.
8644 Now just gather all certificates together and the first private key
8645 then look for the first certificate that matches the key.
8649 * Support use of registered digest and cipher names for dgst and cipher
8650 commands instead of having to add each one as a special case. So now
8657 openssl dgst -sha256 foo
8659 and this works for ENGINE based algorithms too.
8663 * Update Gost ENGINE to support parameter files.
8665 *Victor B. Wagner <vitus@cryptocom.ru>*
8667 * Support GeneralizedTime in ca utility.
8669 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8671 * Enhance the hash format used for certificate directory links. The new
8672 form uses the canonical encoding (meaning equivalent names will work
8673 even if they aren't identical) and uses SHA1 instead of MD5. This form
8674 is incompatible with the older format and as a result c_rehash should
8675 be used to rebuild symbolic links.
8679 * Make PKCS#8 the default write format for private keys, replacing the
8680 traditional format. This form is standardised, more secure and doesn't
8681 include an implicit MD5 dependency.
8685 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8686 committed to OpenSSL should pass this lot as a minimum.
8690 * Add session ticket override functionality for use by EAP-FAST.
8692 *Jouni Malinen <j@w1.fi>*
8694 * Modify HMAC functions to return a value. Since these can be implemented
8695 in an ENGINE errors can occur.
8699 * Type-checked OBJ_bsearch_ex.
8703 * Type-checked OBJ_bsearch. Also some constification necessitated
8704 by type-checking. Still to come: TXT_DB, bsearch(?),
8705 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8710 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8711 seconds to a tm structure directly, instead of going through OS
8712 specific date routines. This avoids any issues with OS routines such
8713 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8714 and X509_time_adj_ex() to cover the extended range. The existing
8715 X509_time_adj() is still usable and will no longer have any date issues.
8719 * Delta CRL support. New use deltas option which will attempt to locate
8720 and search any appropriate delta CRLs available.
8722 This work was sponsored by Google.
8726 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8727 code and add additional score elements. Validate alternate CRL paths
8728 as part of the CRL checking and indicate a new error "CRL path validation
8729 error" in this case. Applications wanting additional details can use
8730 the verify callback and check the new "parent" field. If this is not
8731 NULL CRL path validation is taking place. Existing applications won't
8732 see this because it requires extended CRL support which is off by
8735 This work was sponsored by Google.
8739 * Support for freshest CRL extension.
8741 This work was sponsored by Google.
8745 * Initial indirect CRL support. Currently only supported in the CRLs
8746 passed directly and not via lookup. Process certificate issuer
8747 CRL entry extension and lookup CRL entries by bother issuer name
8748 and serial number. Check and process CRL issuer entry in IDP extension.
8750 This work was sponsored by Google.
8754 * Add support for distinct certificate and CRL paths. The CRL issuer
8755 certificate is validated separately in this case. Only enabled if
8756 an extended CRL support flag is set: this flag will enable additional
8757 CRL functionality in future.
8759 This work was sponsored by Google.
8763 * Add support for policy mappings extension.
8765 This work was sponsored by Google.
8769 * Fixes to pathlength constraint, self issued certificate handling,
8770 policy processing to align with RFC3280 and PKITS tests.
8772 This work was sponsored by Google.
8776 * Support for name constraints certificate extension. DN, email, DNS
8777 and URI types are currently supported.
8779 This work was sponsored by Google.
8783 * To cater for systems that provide a pointer-based thread ID rather
8784 than numeric, deprecate the current numeric thread ID mechanism and
8785 replace it with a structure and associated callback type. This
8786 mechanism allows a numeric "hash" to be extracted from a thread ID in
8787 either case, and on platforms where pointers are larger than 'long',
8788 mixing is done to help ensure the numeric 'hash' is usable even if it
8789 can't be guaranteed unique. The default mechanism is to use "&errno"
8790 as a pointer-based thread ID to distinguish between threads.
8792 Applications that want to provide their own thread IDs should now use
8793 CRYPTO_THREADID_set_callback() to register a callback that will call
8794 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8796 Note that ERR_remove_state() is now deprecated, because it is tied
8797 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8798 to free the current thread's error state should be replaced by
8799 ERR_remove_thread_state(NULL).
8801 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8802 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8803 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8804 application was previously providing a numeric thread callback that
8805 was inappropriate for distinguishing threads, then uniqueness might
8806 have been obtained with &errno that happened immediately in the
8807 intermediate development versions of OpenSSL; this is no longer the
8808 case, the numeric thread callback will now override the automatic use
8811 *Geoff Thorpe, with help from Bodo Moeller*
8813 * Initial support for different CRL issuing certificates. This covers a
8814 simple case where the self issued certificates in the chain exist and
8815 the real CRL issuer is higher in the existing chain.
8817 This work was sponsored by Google.
8821 * Removed effectively defunct crypto/store from the build.
8825 * Revamp of STACK to provide stronger type-checking. Still to come:
8826 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8827 ASN1_STRING, CONF_VALUE.
8831 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8832 RAM on SSL connections. This option can save about 34k per idle SSL.
8836 * Revamp of LHASH to provide stronger type-checking. Still to come:
8837 STACK, TXT_DB, bsearch, qsort.
8841 * Initial support for Cryptographic Message Syntax (aka CMS) based
8842 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8843 support for data, signedData, compressedData, digestedData and
8844 encryptedData, envelopedData types included. Scripts to check against
8845 RFC4134 examples draft and interop and consistency checks of many
8846 content types and variants.
8850 * Add options to enc utility to support use of zlib compression BIO.
8854 * Extend mk1mf to support importing of options and assembly language
8855 files from Configure script, currently only included in VC-WIN32.
8856 The assembly language rules can now optionally generate the source
8857 files from the associated perl scripts.
8861 * Implement remaining functionality needed to support GOST ciphersuites.
8862 Interop testing has been performed using CryptoPro implementations.
8864 *Victor B. Wagner <vitus@cryptocom.ru>*
8866 * s390x assembler pack.
8870 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8875 * Implement Opaque PRF Input TLS extension as specified in
8876 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8877 official specification yet and no extension type assignment by
8878 IANA exists, this extension (for now) will have to be explicitly
8879 enabled when building OpenSSL by providing the extension number
8880 to use. For example, specify an option
8882 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8884 to the "config" or "Configure" script to enable the extension,
8885 assuming extension number 0x9527 (which is a completely arbitrary
8886 and unofficial assignment based on the MD5 hash of the Internet
8887 Draft). Note that by doing so, you potentially lose
8888 interoperability with other TLS implementations since these might
8889 be using the same extension number for other purposes.
8891 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8892 opaque PRF input value to use in the handshake. This will create
8893 an internal copy of the length-'len' string at 'src', and will
8894 return non-zero for success.
8896 To get more control and flexibility, provide a callback function
8899 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8900 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8904 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8907 Callback function 'cb' will be called in handshakes, and is
8908 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8909 Argument 'arg' is for application purposes (the value as given to
8910 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8911 be provided to the callback function). The callback function
8912 has to return non-zero to report success: usually 1 to use opaque
8913 PRF input just if possible, or 2 to enforce use of the opaque PRF
8914 input. In the latter case, the library will abort the handshake
8915 if opaque PRF input is not successfully negotiated.
8917 Arguments 'peerinput' and 'len' given to the callback function
8918 will always be NULL and 0 in the case of a client. A server will
8919 see the client's opaque PRF input through these variables if
8920 available (NULL and 0 otherwise). Note that if the server
8921 provides an opaque PRF input, the length must be the same as the
8922 length of the client's opaque PRF input.
8924 Note that the callback function will only be called when creating
8925 a new session (session resumption can resume whatever was
8926 previously negotiated), and will not be called in SSL 2.0
8927 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8928 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8929 for applications that need to enforce opaque PRF input.
8933 * Update ssl code to support digests other than SHA1+MD5 for handshake
8936 *Victor B. Wagner <vitus@cryptocom.ru>*
8938 * Add RFC4507 support to OpenSSL. This includes the corrections in
8939 RFC4507bis. The encrypted ticket format is an encrypted encoded
8940 SSL_SESSION structure, that way new session features are automatically
8943 If a client application caches session in an SSL_SESSION structure
8944 support is transparent because tickets are now stored in the encoded
8947 The SSL_CTX structure automatically generates keys for ticket
8948 protection in servers so again support should be possible
8949 with no application modification.
8951 If a client or server wishes to disable RFC4507 support then the option
8952 SSL_OP_NO_TICKET can be set.
8954 Add a TLS extension debugging callback to allow the contents of any client
8955 or server extensions to be examined.
8957 This work was sponsored by Google.
8961 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8962 OpenSSL should now compile cleanly on gcc 4.2
8964 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8966 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8967 support including streaming MAC support: this is required for GOST
8968 ciphersuite support.
8970 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8972 * Add option -stream to use PKCS#7 streaming in smime utility. New
8973 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8974 to output in BER and PEM format.
8978 * Experimental support for use of HMAC via EVP_PKEY interface. This
8979 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8980 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8981 ENGINE support for HMAC keys which are unextractable. New -mac and
8982 -macopt options to dgst utility.
8986 * New option -sigopt to dgst utility. Update dgst to use
8987 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8988 alternative signing parameters such as X9.31 or PSS in the dgst
8993 * Change ssl_cipher_apply_rule(), the internal function that does
8994 the work each time a ciphersuite string requests enabling
8995 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8996 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8997 the order of disabled ciphersuites such that those ciphersuites
8998 that most recently went from enabled to disabled not only stay
8999 in order with respect to each other, but also have higher priority
9000 than other disabled ciphersuites the next time ciphersuites are
9003 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9004 the same ciphersuites as with "HIGH" alone, but in a specific
9005 order where the PSK ciphersuites come first (since they are the
9006 most recently disabled ciphersuites when "HIGH" is parsed).
9008 Also, change ssl_create_cipher_list() (using this new
9009 functionality) such that between otherwise identical
9010 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9015 * Change ssl_create_cipher_list() so that it automatically
9016 arranges the ciphersuites in reasonable order before starting
9017 to process the rule string. Thus, the definition for "DEFAULT"
9018 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9019 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9020 This makes it much easier to arrive at a reasonable default order
9021 in applications for which anonymous ciphers are OK (meaning
9022 that you can't actually use DEFAULT).
9024 *Bodo Moeller; suggested by Victor Duchovni*
9026 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9027 processing) into multiple integers instead of setting
9028 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9029 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9030 (These masks as well as the individual bit definitions are hidden
9031 away into the non-exported interface ssl/ssl_locl.h, so this
9032 change to the definition of the SSL_CIPHER structure shouldn't
9033 affect applications.) This give us more bits for each of these
9034 categories, so there is no longer a need to coagulate AES128 and
9035 AES256 into a single algorithm bit, and to coagulate Camellia128
9036 and Camellia256 into a single algorithm bit, which has led to all
9039 Thus, among other things, the kludge introduced in 0.9.7m and
9040 0.9.8e for masking out AES256 independently of AES128 or masking
9041 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9043 With the change, we also introduce new ciphersuite aliases that
9044 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9049 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9050 Use the leftmost N bytes of the signature input if the input is
9051 larger than the prime q (with N being the size in bytes of q).
9055 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9056 it yet and it is largely untested.
9060 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9064 * Initial incomplete changes to avoid need for function casts in OpenSSL
9065 some compilers (gcc 4.2 and later) reject their use. Safestack is
9066 reimplemented. Update ASN1 to avoid use of legacy functions.
9070 * Win32/64 targets are linked with Winsock2.
9074 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9075 to external functions. This can be used to increase CRL handling
9076 efficiency especially when CRLs are very large by (for example) storing
9077 the CRL revoked certificates in a database.
9081 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9082 new CRLs added to a directory can be used. New command line option
9083 -verify_return_error to s_client and s_server. This causes real errors
9084 to be returned by the verify callback instead of carrying on no matter
9085 what. This reflects the way a "real world" verify callback would behave.
9089 * GOST engine, supporting several GOST algorithms and public key formats.
9090 Kindly donated by Cryptocom.
9094 * Partial support for Issuing Distribution Point CRL extension. CRLs
9095 partitioned by DP are handled but no indirect CRL or reason partitioning
9096 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9097 selected via a scoring technique which handles IDP and AKID in CRLs.
9101 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9102 will ultimately be used for all verify operations: this will remove the
9103 X509_STORE dependency on certificate verification and allow alternative
9104 lookup methods. X509_STORE based implementations of these two callbacks.
9108 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9109 Modify get_crl() to find a valid (unexpired) CRL if possible.
9113 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9114 this would be called X509_CRL_cmp() but that name is already used by
9115 a function that just compares CRL issuer names. Cache several CRL
9116 extensions in X509_CRL structure and cache CRLDP in X509.
9120 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9121 this maps equivalent X509_NAME structures into a consistent structure.
9122 Name comparison can then be performed rapidly using memcmp().
9126 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9131 * Allow digests to supply their own micalg string for S/MIME type using
9132 the ctrl EVP_MD_CTRL_MICALG.
9136 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9137 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9138 ctrl. It can then customise the structure before and/or after signing
9143 * New function OBJ_add_sigid() to allow application defined signature OIDs
9144 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9145 to free up any added signature OIDs.
9149 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9150 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9151 digest and cipher tables. New options added to openssl utility:
9152 list-message-digest-algorithms and list-cipher-algorithms.
9156 * Change the array representation of binary polynomials: the list
9157 of degrees of non-zero coefficients is now terminated with -1.
9158 Previously it was terminated with 0, which was also part of the
9159 value; thus, the array representation was not applicable to
9160 polynomials where t^0 has coefficient zero. This change makes
9161 the array representation useful in a more general context.
9165 * Various modifications and fixes to SSL/TLS cipher string
9166 handling. For ECC, the code now distinguishes between fixed ECDH
9167 with RSA certificates on the one hand and with ECDSA certificates
9168 on the other hand, since these are separate ciphersuites. The
9169 unused code for Fortezza ciphersuites has been removed.
9171 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9172 (not "ECDHE"). For consistency with the code for DH
9173 certificates, use of ECDH certificates is now considered ECDH
9174 authentication, not RSA or ECDSA authentication (the latter is
9175 merely the CA's signing algorithm and not actively used in the
9178 The temporary ciphersuite alias "ECCdraft" is no longer
9179 available, and ECC ciphersuites are no longer excluded from "ALL"
9180 and "DEFAULT". The following aliases now exist for RFC 4492
9181 ciphersuites, most of these by analogy with the DH case:
9183 kECDHr - ECDH cert, signed with RSA
9184 kECDHe - ECDH cert, signed with ECDSA
9185 kECDH - ECDH cert (signed with either RSA or ECDSA)
9186 kEECDH - ephemeral ECDH
9187 ECDH - ECDH cert or ephemeral ECDH
9193 AECDH - anonymous ECDH
9194 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9198 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9199 Use correct micalg parameters depending on digest(s) in signed message.
9203 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9204 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9208 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9209 an engine to register a method. Add ENGINE lookups for methods and
9210 functional reference processing.
9214 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9215 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9220 * New -resign option to smime utility. This adds one or more signers
9221 to an existing PKCS#7 signedData structure. Also -md option to use an
9222 alternative message digest algorithm for signing.
9226 * Tidy up PKCS#7 routines and add new functions to make it easier to
9227 create PKCS7 structures containing multiple signers. Update smime
9228 application to support multiple signers.
9232 * New -macalg option to pkcs12 utility to allow setting of an alternative
9237 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9238 Reorganize PBE internals to lookup from a static table using NIDs,
9239 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9240 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9241 PRF which will be automatically used with PBES2.
9245 * Replace the algorithm specific calls to generate keys in "req" with the
9250 * Update PKCS#7 enveloped data routines to use new API. This is now
9251 supported by any public key method supporting the encrypt operation. A
9252 ctrl is added to allow the public key algorithm to examine or modify
9253 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9258 * Add a ctrl to asn1 method to allow a public key algorithm to express
9259 a default digest type to use. In most cases this will be SHA1 but some
9260 algorithms (such as GOST) need to specify an alternative digest. The
9261 return value indicates how strong the preference is 1 means optional and
9262 2 is mandatory (that is it is the only supported type). Modify
9263 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9264 use the default md. Update openssl utilities to use the default digest
9265 type for signing if it is not explicitly indicated.
9269 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9270 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9271 signing method from the key type. This effectively removes the link
9272 between digests and public key types.
9276 * Add an OID cross reference table and utility functions. Its purpose is to
9277 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9278 rsaEncryption. This will allow some of the algorithm specific hackery
9279 needed to use the correct OID to be removed.
9283 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9284 structures for PKCS7_sign(). They are now set up by the relevant public
9289 * Add provisional EC pkey method with support for ECDSA and ECDH.
9293 * Add support for key derivation (agreement) in the API, DH method and
9298 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9299 public and private key formats. As a side effect these add additional
9300 command line functionality not previously available: DSA signatures can be
9301 generated and verified using pkeyutl and DH key support and generation in
9308 *Oliver Tappe <zooey@hirschkaefer.de>*
9310 * New make target "install_html_docs" installs HTML renditions of the
9313 *Oliver Tappe <zooey@hirschkaefer.de>*
9315 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9316 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9317 support key and parameter generation and add initial key generation
9318 functionality for RSA.
9322 * Add functions for main EVP_PKEY_method operations. The undocumented
9323 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9324 `EVP_PKEY_{encrypt,decrypt}_old`.
9328 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9329 key API, doesn't do much yet.
9333 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9334 public key algorithms. New option to openssl utility:
9335 "list-public-key-algorithms" to print out info.
9339 * Implement the Supported Elliptic Curves Extension for
9340 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9344 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9345 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9349 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9350 utilities such as rsa, dsa, dsaparam etc except they process any key
9355 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9356 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9357 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9362 * Initial support for pluggable public key ASN1.
9363 De-spaghettify the public key ASN1 handling. Move public and private
9364 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9365 algorithm specific handling to a single module within the relevant
9366 algorithm directory. Add functions to allow (near) opaque processing
9367 of public and private key structures.
9371 * Implement the Supported Point Formats Extension for
9372 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9376 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9377 for the psk identity [hint] and the psk callback functions to the
9378 SSL_SESSION, SSL and SSL_CTX structure.
9381 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9385 SSL_CTX_use_psk_identity_hint
9386 SSL_get_psk_identity_hint
9387 SSL_get_psk_identity
9388 SSL_use_psk_identity_hint
9390 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9392 * Add RFC 3161 compliant time stamp request creation, response generation
9393 and response verification functionality.
9395 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9397 * Add initial support for TLS extensions, specifically for the server_name
9398 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9399 have new members for a host name. The SSL data structure has an
9400 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9401 stored in that context to allow for session resumption, even after the
9402 SSL has been switched to a new SSL_CTX in reaction to a client's
9403 server_name extension.
9405 New functions (subject to change):
9407 SSL_get_servername()
9408 SSL_get_servername_type()
9411 New CTRL codes and macros (subject to change):
9413 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9414 - SSL_CTX_set_tlsext_servername_callback()
9415 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9416 - SSL_CTX_set_tlsext_servername_arg()
9417 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9419 openssl s_client has a new '-servername ...' option.
9421 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9422 '-key2 ...', '-servername_fatal' (subject to change). This allows
9423 testing the HostName extension for a specific single host name ('-cert'
9424 and '-key' remain fallbacks for handshakes without HostName
9425 negotiation). If the unrecognized_name alert has to be sent, this by
9426 default is a warning; it becomes fatal with the '-servername_fatal'
9429 *Peter Sylvester, Remy Allais, Christophe Renou*
9431 * Whirlpool hash implementation is added.
9435 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9436 bn(64,32). Because of instruction set limitations it doesn't have
9437 any negative impact on performance. This was done mostly in order
9438 to make it possible to share assembler modules, such as bn_mul_mont
9439 implementations, between 32- and 64-bit builds without hassle.
9443 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9444 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9449 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9450 dedicated Montgomery multiplication procedure, is introduced.
9451 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9452 "64-bit" performance on certain 32-bit targets.
9456 * New option SSL_OP_NO_COMP to disable use of compression selectively
9457 in SSL structures. New SSL ctrl to set maximum send fragment size.
9458 Save memory by setting the I/O buffer sizes dynamically instead of
9459 using the maximum available value.
9463 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9464 in addition to the text details.
9468 * Very, very preliminary EXPERIMENTAL support for printing of general
9469 ASN1 structures. This currently produces rather ugly output and doesn't
9470 handle several customised structures at all.
9474 * Integrated support for PVK file format and some related formats such
9475 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9476 these in the 'rsa' and 'dsa' utilities.
9480 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9484 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9485 place for the (very old) "NETSCAPE" format certificates which are now
9486 handled using new ASN1 code equivalents.
9490 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9491 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9492 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9496 * Modify CRL distribution points extension code to print out previously
9497 unsupported fields. Enhance extension setting code to allow setting of
9502 * Add print and set support for Issuing Distribution Point CRL extension.
9506 * Change 'Configure' script to enable Camellia by default.
9513 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9515 * When rejecting SSL/TLS records due to an incorrect version number, never
9516 update s->server with a new major version number. As of
9517 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9518 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9519 the previous behavior could result in a read attempt at NULL when
9520 receiving specific incorrect SSL/TLS records once record payload
9521 protection is active. ([CVE-2010-0740])
9523 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9525 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9526 could be crashed if the relevant tables were not present (e.g. chrooted).
9528 *Tomas Hoger <thoger@redhat.com>*
9530 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9532 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9534 *Martin Olsson, Neel Mehta*
9536 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9537 accommodate for stack sorting, always a write lock!).
9541 * On some versions of WIN32 Heap32Next is very slow. This can cause
9542 excessive delays in the RAND_poll(): over a minute. As a workaround
9543 include a time check in the inner Heap32Next loop too.
9547 * The code that handled flushing of data in SSL/TLS originally used the
9548 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9549 the problem outlined in PR#1949. The fix suggested there however can
9550 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9551 of Apache). So instead simplify the code to flush unconditionally.
9552 This should be fine since flushing with no data to flush is a no op.
9556 * Handle TLS versions 2.0 and later properly and correctly use the
9557 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9558 off ancient servers have a habit of sticking around for a while...
9562 * Modify compression code so it frees up structures without using the
9563 ex_data callbacks. This works around a problem where some applications
9564 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9565 restarting) then use compression (e.g. SSL with compression) later.
9566 This results in significant per-connection memory leaks and
9567 has caused some security issues including CVE-2008-1678 and
9572 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9573 change when encrypting or decrypting.
9577 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9578 connect and renegotiate with servers which do not support RI.
9579 Until RI is more widely deployed this option is enabled by default.
9583 * Add "missing" ssl ctrls to clear options and mode.
9587 * If client attempts to renegotiate and doesn't support RI respond with
9588 a no_renegotiation alert as required by RFC5746. Some renegotiating
9589 TLS clients will continue a connection gracefully when they receive
9590 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9591 waiting for a server hello which it will never receive. Now we treat a
9592 received no_renegotiation alert as a fatal error. This is because
9593 applications requesting a renegotiation might well expect it to succeed
9594 and would have no code in place to handle the server denying it so the
9595 only safe thing to do is to terminate the connection.
9599 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9600 peer supports secure renegotiation and 0 otherwise. Print out peer
9601 renegotiation support in s_client/s_server.
9605 * Replace the highly broken and deprecated SPKAC certification method with
9606 the updated NID creation version. This should correctly handle UTF8.
9610 * Implement RFC5746. Re-enable renegotiation but require the extension
9611 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9612 turns out to be a bad idea. It has been replaced by
9613 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9614 SSL_CTX_set_options(). This is really not recommended unless you
9615 know what you are doing.
9617 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9619 * Fixes to stateless session resumption handling. Use initial_ctx when
9620 issuing and attempting to decrypt tickets in case it has changed during
9621 servername handling. Use a non-zero length session ID when attempting
9622 stateless session resumption: this makes it possible to determine if
9623 a resumption has occurred immediately after receiving server hello
9624 (several places in OpenSSL subtly assume this) instead of later in
9629 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9630 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9631 fixes for a few places where the return code is not checked
9634 *Julia Lawall <julia@diku.dk>*
9636 * Add --strict-warnings option to Configure script to include devteam
9637 warnings in other configurations.
9641 * Add support for --libdir option and LIBDIR variable in makefiles. This
9642 makes it possible to install openssl libraries in locations which
9643 have names other than "lib", for example "/usr/lib64" which some
9646 *Steve Henson, based on patch from Jeremy Utley*
9648 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9649 X690 8.9.12 and can produce some misleading textual output of OIDs.
9651 *Steve Henson, reported by Dan Kaminsky*
9653 * Delete MD2 from algorithm tables. This follows the recommendation in
9654 several standards that it is not used in new applications due to
9655 several cryptographic weaknesses. For binary compatibility reasons
9656 the MD2 API is still compiled in by default.
9660 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9665 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9666 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9669 *Guenter <lists@gknw.net>*
9671 * Fix the server certificate chain building code to use X509_verify_cert(),
9672 it used to have an ad-hoc builder which was unable to cope with anything
9673 other than a simple chain.
9675 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9677 * Don't check self signed certificate signatures in X509_verify_cert()
9678 by default (a flag can override this): it just wastes time without
9679 adding any security. As a useful side effect self signed root CAs
9680 with non-FIPS digests are now usable in FIPS mode.
9684 * In dtls1_process_out_of_seq_message() the check if the current message
9685 is already buffered was missing. For every new message was memory
9686 allocated, allowing an attacker to perform an denial of service attack
9687 with sending out of seq handshake messages until there is no memory
9688 left. Additionally every future message was buffered, even if the
9689 sequence number made no sense and would be part of another handshake.
9690 So only messages with sequence numbers less than 10 in advance will be
9691 buffered. ([CVE-2009-1378])
9693 *Robin Seggelmann, discovered by Daniel Mentz*
9695 * Records are buffered if they arrive with a future epoch to be
9696 processed after finishing the corresponding handshake. There is
9697 currently no limitation to this buffer allowing an attacker to perform
9698 a DOS attack with sending records with future epochs until there is no
9699 memory left. This patch adds the pqueue_size() function to determine
9700 the size of a buffer and limits the record buffer to 100 entries.
9703 *Robin Seggelmann, discovered by Daniel Mentz*
9705 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9706 parent structure is freed. ([CVE-2009-1379])
9710 * Handle non-blocking I/O properly in SSL_shutdown() call.
9712 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9714 * Add `2.5.4.*` OIDs
9716 *Ilya O. <vrghost@gmail.com>*
9718 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9720 * Disable renegotiation completely - this fixes a severe security
9721 problem ([CVE-2009-3555]) at the cost of breaking all
9722 renegotiation. Renegotiation can be re-enabled by setting
9723 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9724 run-time. This is really not recommended unless you know what
9729 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9731 * Don't set val to NULL when freeing up structures, it is freed up by
9732 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9733 zeroing past the valid field. ([CVE-2009-0789])
9735 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9737 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9738 checked correctly. This would allow some invalid signed attributes to
9739 appear to verify correctly. ([CVE-2009-0591])
9741 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9743 * Reject UniversalString and BMPString types with invalid lengths. This
9744 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9745 a legal length. ([CVE-2009-0590])
9749 * Set S/MIME signing as the default purpose rather than setting it
9750 unconditionally. This allows applications to override it at the store
9755 * Permit restricted recursion of ASN1 strings. This is needed in practice
9756 to handle some structures.
9760 * Improve efficiency of mem_gets: don't search whole buffer each time
9763 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9765 * New -hex option for openssl rand.
9769 * Print out UTF8String and NumericString when parsing ASN1.
9773 * Support NumericString type for name components.
9777 * Allow CC in the environment to override the automatically chosen
9778 compiler. Note that nothing is done to ensure flags work with the
9783 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9785 * Properly check EVP_VerifyFinal() and similar return values
9788 *Ben Laurie, Bodo Moeller, Google Security Team*
9790 * Enable TLS extensions by default.
9794 * Allow the CHIL engine to be loaded, whether the application is
9795 multithreaded or not. (This does not release the developer from the
9796 obligation to set up the dynamic locking callbacks.)
9798 *Sander Temme <sander@temme.net>*
9800 * Use correct exit code if there is an error in dgst command.
9802 *Steve Henson; problem pointed out by Roland Dirlewanger*
9804 * Tweak Configure so that you need to say "experimental-jpake" to enable
9805 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9809 * Add experimental JPAKE support, including demo authentication in
9810 s_client and s_server.
9814 * Set the comparison function in v3_addr_canonize().
9816 *Rob Austein <sra@hactrn.net>*
9818 * Add support for XMPP STARTTLS in s_client.
9820 *Philip Paeps <philip@freebsd.org>*
9822 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9823 to ensure that even with this option, only ciphersuites in the
9824 server's preference list will be accepted. (Note that the option
9825 applies only when resuming a session, so the earlier behavior was
9826 just about the algorithm choice for symmetric cryptography.)
9830 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9832 * Fix NULL pointer dereference if a DTLS server received
9833 ChangeCipherSpec as first record ([CVE-2009-1386]).
9837 * Fix a state transition in s3_srvr.c and d1_srvr.c
9838 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9842 * The fix in 0.9.8c that supposedly got rid of unsafe
9843 double-checked locking was incomplete for RSA blinding,
9844 addressing just one layer of what turns out to have been
9845 doubly unsafe triple-checked locking.
9847 So now fix this for real by retiring the MONT_HELPER macro
9848 in crypto/rsa/rsa_eay.c.
9850 *Bodo Moeller; problem pointed out by Marius Schilder*
9852 * Various precautionary measures:
9854 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9856 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9857 (NB: This would require knowledge of the secret session ticket key
9858 to exploit, in which case you'd be SOL either way.)
9860 - Change bn_nist.c so that it will properly handle input BIGNUMs
9861 outside the expected range.
9863 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9866 *Neel Mehta, Bodo Moeller*
9868 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9869 the load fails. Useful for distros.
9871 *Ben Laurie and the FreeBSD team*
9873 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9877 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9881 * Expand ENGINE to support engine supplied SSL client certificate functions.
9883 This work was sponsored by Logica.
9887 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9888 keystores. Support for SSL/TLS client authentication too.
9889 Not compiled unless enable-capieng specified to Configure.
9891 This work was sponsored by Logica.
9895 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9896 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9897 attribute creation routines such as certificate requests and PKCS#12
9902 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9904 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9905 handshake which could lead to a client crash as found using the
9906 Codenomicon TLS test suite ([CVE-2008-1672])
9908 *Steve Henson, Mark Cox*
9910 * Fix double free in TLS server name extensions which could lead to
9911 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9915 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9917 Clear the error queue to ensure that error entries left from
9918 older function calls do not interfere with the correct operation.
9920 *Lutz Jaenicke, Erik de Castro Lopo*
9922 * Remove root CA certificates of commercial CAs:
9924 The OpenSSL project does not recommend any specific CA and does not
9925 have any policy with respect to including or excluding any CA.
9926 Therefore it does not make any sense to ship an arbitrary selection
9927 of root CA certificates with the OpenSSL software.
9931 * RSA OAEP patches to fix two separate invalid memory reads.
9932 The first one involves inputs when 'lzero' is greater than
9933 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9934 before the beginning of from). The second one involves inputs where
9935 the 'db' section contains nothing but zeroes (there is a one-byte
9936 invalid read after the end of 'db').
9938 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9940 * Partial backport from 0.9.9-dev:
9942 Introduce bn_mul_mont (dedicated Montgomery multiplication
9943 procedure) as a candidate for BIGNUM assembler implementation.
9944 While 0.9.9-dev uses assembler for various architectures, only
9945 x86_64 is available by default here in the 0.9.8 branch, and
9946 32-bit x86 is available through a compile-time setting.
9948 To try the 32-bit x86 assembler implementation, use Configure
9949 option "enable-montasm" (which exists only for this backport).
9951 As "enable-montasm" for 32-bit x86 disclaims code stability
9952 anyway, in this constellation we activate additional code
9953 backported from 0.9.9-dev for further performance improvements,
9954 namely BN_from_montgomery_word. (To enable this otherwise,
9955 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9957 *Andy Polyakov (backport partially by Bodo Moeller)*
9959 * Add TLS session ticket callback. This allows an application to set
9960 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9961 values. This is useful for key rollover for example where several key
9962 sets may exist with different names.
9966 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9967 This was broken until now in 0.9.8 releases, such that the only way
9968 a registered ENGINE could be used (assuming it initialises
9969 successfully on the host) was to explicitly set it as the default
9970 for the relevant algorithms. This is in contradiction with 0.9.7
9971 behaviour and the documentation. With this fix, when an ENGINE is
9972 registered into a given algorithm's table of implementations, the
9973 'uptodate' flag is reset so that auto-discovery will be used next
9974 time a new context for that algorithm attempts to select an
9977 *Ian Lister (tweaked by Geoff Thorpe)*
9979 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9980 implementation in the following ways:
9982 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9985 Lack of BER streaming support means one pass streaming processing is
9986 only supported if data is detached: setting the streaming flag is
9987 ignored for embedded content.
9989 CMS support is disabled by default and must be explicitly enabled
9990 with the enable-cms configuration option.
9994 * Update the GMP engine glue to do direct copies between BIGNUM and
9995 mpz_t when openssl and GMP use the same limb size. Otherwise the
9996 existing "conversion via a text string export" trick is still used.
9998 *Paul Sheer <paulsheer@gmail.com>*
10000 * Zlib compression BIO. This is a filter BIO which compressed and
10001 uncompresses any data passed through it.
10005 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10006 RFC3394 compatible AES key wrapping.
10010 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10011 sets string data without copying. X509_ALGOR_set0() and
10012 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10013 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10014 from an X509_ATTRIBUTE structure optionally checking it occurs only
10015 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10020 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10021 to get the expected BN_FLG_CONSTTIME behavior.
10023 *Bodo Moeller (Google)*
10027 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10028 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10029 - added some more tests to do_tests.pl
10030 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10031 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10032 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10033 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10034 - various changes to netware.pl to enable gcc-cross builds on Win32
10036 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10037 - various changes to fix missing prototype warnings
10038 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10039 - added AES, WHIRLPOOL and CPUID assembler code to build files
10040 - added missing AES assembler make rules to mk1mf.pl
10041 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10043 *Guenter Knauf <eflash@gmx.net>*
10045 * Implement certificate status request TLS extension defined in RFC3546.
10046 A client can set the appropriate parameters and receive the encoded
10047 OCSP response via a callback. A server can query the supplied parameters
10048 and set the encoded OCSP response in the callback. Add simplified examples
10049 to s_client and s_server.
10053 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10055 * Fix various bugs:
10056 + Binary incompatibility of ssl_ctx_st structure
10057 + DTLS interoperation with non-compliant servers
10058 + Don't call get_session_cb() without proposed session
10059 + Fix ia64 assembler code
10061 *Andy Polyakov, Steve Henson*
10063 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10065 * DTLS Handshake overhaul. There were longstanding issues with
10066 OpenSSL DTLS implementation, which were making it impossible for
10067 RFC 4347 compliant client to communicate with OpenSSL server.
10068 Unfortunately just fixing these incompatibilities would "cut off"
10069 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10070 server keeps tolerating non RFC compliant syntax. The opposite is
10071 not true, 0.9.8f client can not communicate with earlier server.
10072 This update even addresses CVE-2007-4995.
10076 * Changes to avoid need for function casts in OpenSSL: some compilers
10077 (gcc 4.2 and later) reject their use.
10078 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10081 * Add RFC4507 support to OpenSSL. This includes the corrections in
10082 RFC4507bis. The encrypted ticket format is an encrypted encoded
10083 SSL_SESSION structure, that way new session features are automatically
10086 If a client application caches session in an SSL_SESSION structure
10087 support is transparent because tickets are now stored in the encoded
10090 The SSL_CTX structure automatically generates keys for ticket
10091 protection in servers so again support should be possible
10092 with no application modification.
10094 If a client or server wishes to disable RFC4507 support then the option
10095 SSL_OP_NO_TICKET can be set.
10097 Add a TLS extension debugging callback to allow the contents of any client
10098 or server extensions to be examined.
10100 This work was sponsored by Google.
10104 * Add initial support for TLS extensions, specifically for the server_name
10105 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10106 have new members for a host name. The SSL data structure has an
10107 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10108 stored in that context to allow for session resumption, even after the
10109 SSL has been switched to a new SSL_CTX in reaction to a client's
10110 server_name extension.
10112 New functions (subject to change):
10114 SSL_get_servername()
10115 SSL_get_servername_type()
10118 New CTRL codes and macros (subject to change):
10120 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10121 - SSL_CTX_set_tlsext_servername_callback()
10122 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10123 - SSL_CTX_set_tlsext_servername_arg()
10124 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10126 openssl s_client has a new '-servername ...' option.
10128 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10129 '-key2 ...', '-servername_fatal' (subject to change). This allows
10130 testing the HostName extension for a specific single host name ('-cert'
10131 and '-key' remain fallbacks for handshakes without HostName
10132 negotiation). If the unrecognized_name alert has to be sent, this by
10133 default is a warning; it becomes fatal with the '-servername_fatal'
10136 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10138 * Add AES and SSE2 assembly language support to VC++ build.
10142 * Mitigate attack on final subtraction in Montgomery reduction.
10146 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10147 (which previously caused an internal error).
10151 * Squeeze another 10% out of IGE mode when in != out.
10155 * AES IGE mode speedup.
10157 *Dean Gaudet (Google)*
10159 * Add the Korean symmetric 128-bit cipher SEED (see
10160 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10161 add SEED ciphersuites from RFC 4162:
10163 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10164 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10165 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10166 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10168 To minimize changes between patchlevels in the OpenSSL 0.9.8
10169 series, SEED remains excluded from compilation unless OpenSSL
10170 is configured with 'enable-seed'.
10172 *KISA, Bodo Moeller*
10174 * Mitigate branch prediction attacks, which can be practical if a
10175 single processor is shared, allowing a spy process to extract
10176 information. For detailed background information, see
10177 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10178 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10179 and Necessary Software Countermeasures"). The core of the change
10180 are new versions BN_div_no_branch() and
10181 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10182 respectively, which are slower, but avoid the security-relevant
10183 conditional branches. These are automatically called by BN_div()
10184 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10185 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10186 remove a conditional branch.
10188 BN_FLG_CONSTTIME is the new name for the previous
10189 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10190 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10191 in the exponent causes BN_mod_exp_mont() to use the alternative
10192 implementation in BN_mod_exp_mont_consttime().) The old name
10193 remains as a deprecated alias.
10195 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10196 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10197 constant-time implementations for more than just exponentiation.
10198 Here too the old name is kept as a deprecated alias.
10200 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10201 the BN_BLINDING structure gets an independent copy of the
10202 modulus. This means that the previous `BIGNUM *m` argument to
10203 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10204 essentially becomes `const BIGNUM *m`, although we can't actually
10205 change this in the header file before 0.9.9. It allows
10206 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10207 enable BN_FLG_CONSTTIME.
10209 *Matthew D Wood (Intel Corp)*
10211 * In the SSL/TLS server implementation, be strict about session ID
10212 context matching (which matters if an application uses a single
10213 external cache for different purposes). Previously,
10214 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10215 set. This did ensure strict client verification, but meant that,
10216 with applications using a single external cache for quite
10217 different requirements, clients could circumvent ciphersuite
10218 restrictions for a given session ID context by starting a session
10219 in a different context.
10223 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10224 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10225 authentication-only ciphersuites.
10229 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10230 not complete and could lead to a possible single byte overflow
10231 ([CVE-2007-5135]) [Ben Laurie]
10233 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10235 * Since AES128 and AES256 (and similarly Camellia128 and
10236 Camellia256) share a single mask bit in the logic of
10237 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10238 kludge to work properly if AES128 is available and AES256 isn't
10239 (or if Camellia128 is available and Camellia256 isn't).
10243 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10244 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10245 When a point or a seed is encoded in a BIT STRING, we need to
10246 prevent the removal of trailing zero bits to get the proper DER
10247 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10248 of a NamedBitList, for which trailing 0 bits need to be removed.)
10252 * Have SSL/TLS server implementation tolerate "mismatched" record
10253 protocol version while receiving ClientHello even if the
10254 ClientHello is fragmented. (The server can't insist on the
10255 particular protocol version it has chosen before the ServerHello
10256 message has informed the client about his choice.)
10260 * Add RFC 3779 support.
10262 *Rob Austein for ARIN, Ben Laurie*
10264 * Load error codes if they are not already present instead of using a
10265 static variable. This allows them to be cleanly unloaded and reloaded.
10266 Improve header file function name parsing.
10270 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10271 or CAPABILITY handshake as required by RFCs.
10273 *Goetz Babin-Ebell*
10275 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10277 * Introduce limits to prevent malicious keys being able to
10278 cause a denial of service. ([CVE-2006-2940])
10280 *Steve Henson, Bodo Moeller*
10282 * Fix ASN.1 parsing of certain invalid structures that can result
10283 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10285 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10286 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10288 * Fix SSL client code which could crash if connecting to a
10289 malicious SSLv2 server. ([CVE-2006-4343])
10291 *Tavis Ormandy and Will Drewry, Google Security Team*
10293 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10294 match only those. Before that, "AES256-SHA" would be interpreted
10295 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10296 the same strength classification in 0.9.7h) as we currently only
10297 have a single AES bit in the ciphersuite description bitmap.
10298 That change, however, also applied to ciphersuite strings such as
10299 "RC4-MD5" that intentionally matched multiple ciphersuites --
10300 namely, SSL 2.0 ciphersuites in addition to the more common ones
10301 from SSL 3.0/TLS 1.0.
10303 So we change the selection algorithm again: Naming an explicit
10304 ciphersuite selects this one ciphersuite, and any other similar
10305 ciphersuite (same bitmap) from *other* protocol versions.
10306 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10307 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10309 Since SSL 2.0 does not have any ciphersuites for which the
10310 128/256 bit distinction would be relevant, this works for now.
10311 The proper fix will be to use different bits for AES128 and
10312 AES256, which would have avoided the problems from the beginning;
10313 however, bits are scarce, so we can only do this in a new release
10314 (not just a patchlevel) when we can change the SSL_CIPHER
10315 definition to split the single 'unsigned long mask' bitmap into
10316 multiple values to extend the available space.
10320 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10322 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10323 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10325 * Add AES IGE and biIGE modes.
10329 * Change the Unix randomness entropy gathering to use poll() when
10330 possible instead of select(), since the latter has some
10331 undesirable limitations.
10333 *Darryl Miles via Richard Levitte and Bodo Moeller*
10335 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10336 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10337 cannot be implicitly activated as part of, e.g., the "AES" alias.
10338 However, please upgrade to OpenSSL 0.9.9[-dev] for
10339 non-experimental use of the ECC ciphersuites to get TLS extension
10340 support, which is required for curve and point format negotiation
10341 to avoid potential handshake problems.
10345 * Disable rogue ciphersuites:
10347 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10348 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10349 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10351 The latter two were purportedly from
10352 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10355 Also deactivate the remaining ciphersuites from
10356 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10357 unofficial, and the ID has long expired.
10361 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10362 dual-core machines) and other potential thread-safety issues.
10366 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10367 versions), which is now available for royalty-free use
10368 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10369 Also, add Camellia TLS ciphersuites from RFC 4132.
10371 To minimize changes between patchlevels in the OpenSSL 0.9.8
10372 series, Camellia remains excluded from compilation unless OpenSSL
10373 is configured with 'enable-camellia'.
10377 * Disable the padding bug check when compression is in use. The padding
10378 bug check assumes the first packet is of even length, this is not
10379 necessarily true if compression is enabled and can result in false
10380 positives causing handshake failure. The actual bug test is ancient
10381 code so it is hoped that implementations will either have fixed it by
10382 now or any which still have the bug do not support compression.
10386 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10388 * When applying a cipher rule check to see if string match is an explicit
10389 cipher suite and only match that one cipher suite if it is.
10393 * Link in manifests for VC++ if needed.
10395 *Austin Ziegler <halostatue@gmail.com>*
10397 * Update support for ECC-based TLS ciphersuites according to
10398 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10399 TLS extensions, which are supported starting with the 0.9.9
10400 branch, not in the OpenSSL 0.9.8 branch).
10404 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10405 opaque EVP_CIPHER_CTX handling.
10409 * Fixes and enhancements to zlib compression code. We now only use
10410 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10411 to conform with the standards mentioned here:
10412 <http://www.zlib.net/DLL_FAQ.txt>
10413 Static zlib linking now works on Windows and the new --with-zlib-include
10414 --with-zlib-lib options to Configure can be used to supply the location
10415 of the headers and library. Gracefully handle case where zlib library
10420 * Several fixes and enhancements to the OID generation code. The old code
10421 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10422 handle numbers larger than ULONG_MAX, truncated printing and had a
10423 non standard OBJ_obj2txt() behaviour.
10427 * Add support for building of engines under engine/ as shared libraries
10428 under VC++ build system.
10432 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10433 Hopefully, we will not see any false combination of paths any more.
10437 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10439 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10440 (part of SSL_OP_ALL). This option used to disable the
10441 countermeasure against man-in-the-middle protocol-version
10442 rollback in the SSL 2.0 server implementation, which is a bad
10443 idea. ([CVE-2005-2969])
10445 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10446 for Information Security, National Institute of Advanced Industrial
10447 Science and Technology [AIST], Japan)*
10449 * Add two function to clear and return the verify parameter flags.
10453 * Keep cipherlists sorted in the source instead of sorting them at
10454 runtime, thus removing the need for a lock.
10458 * Avoid some small subgroup attacks in Diffie-Hellman.
10460 *Nick Mathewson and Ben Laurie*
10462 * Add functions for well-known primes.
10466 * Extended Windows CE support.
10468 *Satoshi Nakamura and Andy Polyakov*
10470 * Initialize SSL_METHOD structures at compile time instead of during
10471 runtime, thus removing the need for a lock.
10475 * Make PKCS7_decrypt() work even if no certificate is supplied by
10476 attempting to decrypt each encrypted key in turn. Add support to
10481 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10483 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10486 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10490 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10491 key into the same file any more.
10495 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10499 * Add -utf8 command line and config file option to 'ca'.
10501 *Stefan <stf@udoma.org*
10503 * Removed the macro des_crypt(), as it seems to conflict with some
10504 libraries. Use DES_crypt().
10508 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10509 involves renaming the source and generated shared-libs for
10510 both. The engines will accept the corrected or legacy ids
10511 ('ncipher' and '4758_cca' respectively) when binding. NB,
10512 this only applies when building 'shared'.
10514 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10516 * Add attribute functions to EVP_PKEY structure. Modify
10517 PKCS12_create() to recognize a CSP name attribute and
10518 use it. Make -CSP option work again in pkcs12 utility.
10522 * Add new functionality to the bn blinding code:
10523 - automatic re-creation of the BN_BLINDING parameters after
10524 a fixed number of uses (currently 32)
10525 - add new function for parameter creation
10526 - introduce flags to control the update behaviour of the
10527 BN_BLINDING parameters
10528 - hide BN_BLINDING structure
10529 Add a second BN_BLINDING slot to the RSA structure to improve
10530 performance when a single RSA object is shared among several
10535 * Add support for DTLS.
10537 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10539 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10540 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10544 * Remove buggy and incomplete DH cert support from
10545 ssl/ssl_rsa.c and ssl/s3_both.c
10549 * Use SHA-1 instead of MD5 as the default digest algorithm for
10550 the `apps/openssl` commands.
10554 * Compile clean with "-Wall -Wmissing-prototypes
10555 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10556 DEBUG_SAFESTACK must also be set.
10560 * Change ./Configure so that certain algorithms can be disabled by default.
10561 The new counterpiece to "no-xxx" is "enable-xxx".
10563 The patented RC5 and MDC2 algorithms will now be disabled unless
10564 "enable-rc5" and "enable-mdc2", respectively, are specified.
10566 (IDEA remains enabled despite being patented. This is because IDEA
10567 is frequently required for interoperability, and there is no license
10568 fee for non-commercial use. As before, "no-idea" can be used to
10569 avoid this algorithm.)
10573 * Add processing of proxy certificates (see RFC 3820). This work was
10574 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10575 EGEE (Enabling Grids for E-science in Europe).
10579 * RC4 performance overhaul on modern architectures/implementations, such
10580 as Intel P4, IA-64 and AMD64.
10584 * New utility extract-section.pl. This can be used specify an alternative
10585 section number in a pod file instead of having to treat each file as
10586 a separate case in Makefile. This can be done by adding two lines to the
10589 =for comment openssl_section:XXX
10591 The blank line is mandatory.
10595 * New arguments -certform, -keyform and -pass for s_client and s_server
10596 to allow alternative format key and certificate files and passphrase
10601 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10602 update associated structures and add various utility functions.
10604 Add new policy related verify parameters, include policy checking in
10605 standard verify code. Enhance 'smime' application with extra parameters
10606 to support policy checking and print out.
10610 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10611 Nehemiah processors. These extensions support AES encryption in hardware
10612 as well as RNG (though RNG support is currently disabled).
10614 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10616 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10620 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10622 *Andy Polyakov and a number of other people*
10624 * Improved PowerPC platform support. Most notably BIGNUM assembler
10625 implementation contributed by IBM.
10627 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10629 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10630 exponent rather than 'unsigned long'. There is a corresponding change to
10631 the new 'rsa_keygen' element of the RSA_METHOD structure.
10633 *Jelte Jansen, Geoff Thorpe*
10635 * Functionality for creating the initial serial number file is now
10636 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10638 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10639 number file to 1, which is bound to cause problems. To avoid
10640 the problems while respecting compatibility between different 0.9.7
10641 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10642 CA.pl for serial number initialization. With the new release 0.9.8,
10643 we can fix the problem directly in the 'ca' utility.)
10647 * Reduced header interdependencies by declaring more opaque objects in
10648 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10649 give fewer recursive includes, which could break lazy source code - so
10650 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10651 developers should define this symbol when building and using openssl to
10652 ensure they track the recommended behaviour, interfaces, [etc], but
10653 backwards-compatible behaviour prevails when this isn't defined.
10657 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10661 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10662 This will generate a random key of the appropriate length based on the
10663 cipher context. The EVP_CIPHER can provide its own random key generation
10664 routine to support keys of a specific form. This is used in the des and
10665 3des routines to generate a key of the correct parity. Update S/MIME
10666 code to use new functions and hence generate correct parity DES keys.
10667 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10668 valid (weak or incorrect parity).
10672 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10673 as looking them up. This is useful when the verified structure may contain
10674 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10675 present unless the new PKCS7_NO_CRL flag is asserted.
10679 * Extend ASN1 oid configuration module. It now additionally accepts the
10682 shortName = some long name, 1.2.3.4
10686 * Reimplemented the BN_CTX implementation. There is now no more static
10687 limitation on the number of variables it can handle nor the depth of the
10688 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10689 information can now expand as required, and rather than having a single
10690 static array of bignums, BN_CTX now uses a linked-list of such arrays
10691 allowing it to expand on demand whilst maintaining the usefulness of
10692 BN_CTX's "bundling".
10696 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10697 to allow all RSA operations to function using a single BN_CTX.
10701 * Preliminary support for certificate policy evaluation and checking. This
10702 is initially intended to pass the tests outlined in "Conformance Testing
10703 of Relying Party Client Certificate Path Processing Logic" v1.07.
10707 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10708 remained unused and not that useful. A variety of other little bignum
10709 tweaks and fixes have also been made continuing on from the audit (see
10714 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10715 associated ASN1, EVP and SSL functions and old ASN1 macros.
10719 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10720 and this should never fail. So the return value from the use of
10721 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10722 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10726 * BN_CTX_get() should return zero-valued bignums, providing the same
10727 initialised value as BN_new().
10729 *Geoff Thorpe, suggested by Ulf Möller*
10731 * Support for inhibitAnyPolicy certificate extension.
10735 * An audit of the BIGNUM code is underway, for which debugging code is
10736 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10737 is considered valid when processing BIGNUMs, and causes execution to
10738 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10739 further steps are taken to deliberately pollute unused data in BIGNUM
10740 structures to try and expose faulty code further on. For now, openssl will
10741 (in its default mode of operation) continue to tolerate the inconsistent
10742 forms that it has tolerated in the past, but authors and packagers should
10743 consider trying openssl and their own applications when compiled with
10744 these debugging symbols defined. It will help highlight potential bugs in
10745 their own code, and will improve the test coverage for OpenSSL itself. At
10746 some point, these tighter rules will become openssl's default to improve
10747 maintainability, though the assert()s and other overheads will remain only
10748 in debugging configurations. See bn.h for more details.
10750 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10752 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10753 that can only be obtained through BN_CTX_new() (which implicitly
10754 initialises it). The presence of this function only made it possible
10755 to overwrite an existing structure (and cause memory leaks).
10759 * Because of the callback-based approach for implementing LHASH as a
10760 template type, lh_insert() adds opaque objects to hash-tables and
10761 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10762 to clean up those corresponding objects before destroying the hash table
10763 (and losing the object pointers). So some over-zealous constifications in
10764 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10765 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10766 prototyped to have "const" restrictions on the object pointers they are
10767 given (and so aren't required to cast them away any more).
10771 * The tmdiff.h API was so ugly and minimal that our own timing utility
10772 (speed) prefers to use its own implementation. The two implementations
10773 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10774 its object type properly exposed (MS_TM) instead of casting to/from
10775 `char *`. This may still change yet if someone realises MS_TM and
10777 aren't necessarily the greatest nomenclatures - but this is what was used
10778 internally to the implementation so I've used that for now.
10782 * Ensure that deprecated functions do not get compiled when
10783 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10784 the self-tests were still using deprecated key-generation functions so
10785 these have been updated also.
10789 * Reorganise PKCS#7 code to separate the digest location functionality
10790 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10791 New function PKCS7_set_digest() to set the digest type for PKCS#7
10792 digestedData type. Add additional code to correctly generate the
10793 digestedData type and add support for this type in PKCS7 initialization
10798 * New function PKCS7_set0_type_other() this initializes a PKCS7
10799 structure of type "other".
10803 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10804 sure the loop does correctly stop and breaking ("division by zero")
10805 modulus operations are not performed. The (pre-generated) prime
10806 table crypto/bn/bn_prime.h was already correct, but it could not be
10807 re-generated on some platforms because of the "division by zero"
10808 situation in the script.
10810 *Ralf S. Engelschall*
10812 * Update support for ECC-based TLS ciphersuites according to
10813 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10814 SHA-1 now is only used for "small" curves (where the
10815 representation of a field element takes up to 24 bytes); for
10816 larger curves, the field element resulting from ECDH is directly
10817 used as premaster secret.
10819 *Douglas Stebila (Sun Microsystems Laboratories)*
10821 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10822 curve secp160r1 to the tests.
10824 *Douglas Stebila (Sun Microsystems Laboratories)*
10826 * Add the possibility to load symbols globally with DSO.
10828 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10830 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10831 control of the error stack.
10835 * Add support for STORE in ENGINE.
10839 * Add the STORE type. The intention is to provide a common interface
10840 to certificate and key stores, be they simple file-based stores, or
10841 HSM-type store, or LDAP stores, or...
10842 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10846 * Add a generic structure called OPENSSL_ITEM. This can be used to
10847 pass a list of arguments to any function as well as provide a way
10848 for a function to pass data back to the caller.
10852 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10853 works like BUF_strdup() but can be used to duplicate a portion of
10854 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10859 * Add the function sk_find_ex() which works like sk_find(), but will
10860 return an index to an element even if an exact match couldn't be
10861 found. The index is guaranteed to point at the element where the
10862 searched-for key would be inserted to preserve sorting order.
10866 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10867 takes an extra flags argument for optional functionality. Currently,
10868 the following flags are defined:
10870 OBJ_BSEARCH_VALUE_ON_NOMATCH
10871 This one gets OBJ_bsearch_ex() to return a pointer to the first
10872 element where the comparing function returns a negative or zero
10875 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10876 This one gets OBJ_bsearch_ex() to return a pointer to the first
10877 element where the comparing function returns zero. This is useful
10878 if there are more than one element where the comparing function
10883 * Make it possible to create self-signed certificates with 'openssl ca'
10884 in such a way that the self-signed certificate becomes part of the
10885 CA database and uses the same mechanisms for serial number generation
10886 as all other certificate signing. The new flag '-selfsign' enables
10887 this functionality. Adapt CA.sh and CA.pl.in.
10891 * Add functionality to check the public key of a certificate request
10892 against a given private. This is useful to check that a certificate
10893 request can be signed by that key (self-signing).
10897 * Make it possible to have multiple active certificates with the same
10898 subject in the CA index file. This is done only if the keyword
10899 'unique_subject' is set to 'no' in the main CA section (default
10900 if 'CA_default') of the configuration file. The value is saved
10901 with the database itself in a separate index attribute file,
10902 named like the index file with '.attr' appended to the name.
10906 * Generate multi-valued AVAs using '+' notation in config files for
10911 * Support for nameConstraints certificate extension.
10915 * Support for policyConstraints certificate extension.
10919 * Support for policyMappings certificate extension.
10923 * Make sure the default DSA_METHOD implementation only uses its
10924 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10925 and change its own handlers to be NULL so as to remove unnecessary
10926 indirection. This lets alternative implementations fallback to the
10927 default implementation more easily.
10931 * Support for directoryName in GeneralName related extensions
10936 * Make it possible to link applications using Makefile.shared.
10937 Make that possible even when linking against static libraries!
10941 * Support for single pass processing for S/MIME signing. This now
10942 means that S/MIME signing can be done from a pipe, in addition
10943 cleartext signing (multipart/signed type) is effectively streaming
10944 and the signed data does not need to be all held in memory.
10946 This is done with a new flag PKCS7_STREAM. When this flag is set
10947 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10948 is done after the data is output (and digests calculated) in
10949 SMIME_write_PKCS7().
10953 * Add full support for -rpath/-R, both in shared libraries and
10954 applications, at least on the platforms where it's known how
10959 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10960 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10961 will now compute a table of multiples of the generator that
10962 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10963 faster (notably in the case of a single point multiplication,
10964 scalar * generator).
10966 *Nils Larsch, Bodo Moeller*
10968 * IPv6 support for certificate extensions. The various extensions
10969 which use the IP:a.b.c.d can now take IPv6 addresses using the
10970 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10975 * Added an ENGINE that implements RSA by performing private key
10976 exponentiations with the GMP library. The conversions to and from
10977 GMP's mpz_t format aren't optimised nor are any montgomery forms
10978 cached, and on x86 it appears OpenSSL's own performance has caught up.
10979 However there are likely to be other architectures where GMP could
10980 provide a boost. This ENGINE is not built in by default, but it can be
10981 specified at Configure time and should be accompanied by the necessary
10982 linker additions, eg;
10983 ./config -DOPENSSL_USE_GMP -lgmp
10987 * "openssl engine" will not display ENGINE/DSO load failure errors when
10988 testing availability of engines with "-t" - the old behaviour is
10989 produced by increasing the feature's verbosity with "-tt".
10993 * ECDSA routines: under certain error conditions uninitialized BN objects
10994 could be freed. Solution: make sure initialization is performed early
10995 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11000 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11001 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11002 software implementations. For DSA and DH, parameter generation can
11003 also be overridden by providing the appropriate method callbacks.
11007 * Change the "progress" mechanism used in key-generation and
11008 primality testing to functions that take a new BN_GENCB pointer in
11009 place of callback/argument pairs. The new API functions have `_ex`
11010 postfixes and the older functions are reimplemented as wrappers for
11011 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11012 declarations of the old functions to help (graceful) attempts to
11013 migrate to the new functions. Also, the new key-generation API
11014 functions operate on a caller-supplied key-structure and return
11015 success/failure rather than returning a key or NULL - this is to
11016 help make "keygen" another member function of RSA_METHOD etc.
11018 Example for using the new callback interface:
11020 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11021 void *my_arg = ...;
11024 BN_GENCB_set(&my_cb, my_callback, my_arg);
11026 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11027 /* For the meaning of a, b in calls to my_callback(), see the
11028 * documentation of the function that calls the callback.
11029 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11030 * my_callback should return 1 if it wants BN_is_prime_ex()
11031 * to continue, or 0 to stop.
11036 * Change the ZLIB compression method to be stateful, and make it
11037 available to TLS with the number defined in
11038 draft-ietf-tls-compression-04.txt.
11042 * Add the ASN.1 structures and functions for CertificatePair, which
11043 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11045 CertificatePair ::= SEQUENCE {
11046 forward [0] Certificate OPTIONAL,
11047 reverse [1] Certificate OPTIONAL,
11048 -- at least one of the pair shall be present -- }
11050 Also implement the PEM functions to read and write certificate
11051 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11053 This needed to be defined, mostly for the sake of the LDAP
11054 attribute crossCertificatePair, but may prove useful elsewhere as
11059 * Make it possible to inhibit symlinking of shared libraries in
11060 Makefile.shared, for Cygwin's sake.
11064 * Extend the BIGNUM API by creating a function
11065 void BN_set_negative(BIGNUM *a, int neg);
11066 and a macro that behave like
11067 int BN_is_negative(const BIGNUM *a);
11069 to avoid the need to access 'a->neg' directly in applications.
11073 * Implement fast modular reduction for pseudo-Mersenne primes
11074 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11075 EC_GROUP_new_curve_GFp() will now automatically use this
11078 *Nils Larsch <nla@trustcenter.de>*
11080 * Add new lock type (CRYPTO_LOCK_BN).
11084 * Change the ENGINE framework to automatically load engines
11085 dynamically from specific directories unless they could be
11086 found to already be built in or loaded. Move all the
11087 current engines except for the cryptodev one to a new
11088 directory engines/.
11089 The engines in engines/ are built as shared libraries if
11090 the "shared" options was given to ./Configure or ./config.
11091 Otherwise, they are inserted in libcrypto.a.
11092 /usr/local/ssl/engines is the default directory for dynamic
11093 engines, but that can be overridden at configure time through
11094 the usual use of --prefix and/or --openssldir, and at run
11095 time with the environment variable OPENSSL_ENGINES.
11097 *Geoff Thorpe and Richard Levitte*
11099 * Add Makefile.shared, a helper makefile to build shared
11100 libraries. Adapt Makefile.org.
11104 * Add version info to Win32 DLLs.
11106 *Peter 'Luna' Runestig" <peter@runestig.com>*
11108 * Add new 'medium level' PKCS#12 API. Certificates and keys
11109 can be added using this API to created arbitrary PKCS#12
11110 files while avoiding the low-level API.
11112 New options to PKCS12_create(), key or cert can be NULL and
11113 will then be omitted from the output file. The encryption
11114 algorithm NIDs can be set to -1 for no encryption, the mac
11115 iteration count can be set to 0 to omit the mac.
11117 Enhance pkcs12 utility by making the -nokeys and -nocerts
11118 options work when creating a PKCS#12 file. New option -nomac
11119 to omit the mac, NONE can be set for an encryption algorithm.
11120 New code is modified to use the enhanced PKCS12_create()
11121 instead of the low-level API.
11125 * Extend ASN1 encoder to support indefinite length constructed
11126 encoding. This can output sequences tags and octet strings in
11127 this form. Modify pk7_asn1.c to support indefinite length
11128 encoding. This is experimental and needs additional code to
11129 be useful, such as an ASN1 bio and some enhanced streaming
11132 Extend template encode functionality so that tagging is passed
11133 down to the template encoder.
11137 * Let 'openssl req' fail if an argument to '-newkey' is not
11138 recognized instead of using RSA as a default.
11142 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11143 As these are not official, they are not included in "ALL";
11144 the "ECCdraft" ciphersuite group alias can be used to select them.
11146 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11148 * Add ECDH engine support.
11150 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11152 * Add ECDH in new directory crypto/ecdh/.
11154 *Douglas Stebila (Sun Microsystems Laboratories)*
11156 * Let BN_rand_range() abort with an error after 100 iterations
11157 without success (which indicates a broken PRNG).
11161 * Change BN_mod_sqrt() so that it verifies that the input value
11162 is really the square of the return value. (Previously,
11163 BN_mod_sqrt would show GIGO behaviour.)
11167 * Add named elliptic curves over binary fields from X9.62, SECG,
11168 and WAP/WTLS; add OIDs that were still missing.
11170 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11172 * Extend the EC library for elliptic curves over binary fields
11173 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11176 EC_GF2m_simple_method
11180 EC_GROUP_new_curve_GF2m
11181 EC_GROUP_set_curve_GF2m
11182 EC_GROUP_get_curve_GF2m
11183 EC_POINT_set_affine_coordinates_GF2m
11184 EC_POINT_get_affine_coordinates_GF2m
11185 EC_POINT_set_compressed_coordinates_GF2m
11187 Point compression for binary fields is disabled by default for
11188 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11191 As binary polynomials are represented as BIGNUMs, various members
11192 of the EC_GROUP and EC_POINT data structures can be shared
11193 between the implementations for prime fields and binary fields;
11194 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11195 are essentially identical to their `..._GFp` counterparts.
11196 (For simplicity, the `..._GFp` prefix has been dropped from
11197 various internal method names.)
11199 An internal 'field_div' method (similar to 'field_mul' and
11200 'field_sqr') has been added; this is used only for binary fields.
11202 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11204 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11205 through methods ('mul', 'precompute_mult').
11207 The generic implementations (now internally called 'ec_wNAF_mul'
11208 and 'ec_wNAF_precomputed_mult') remain the default if these
11209 methods are undefined.
11211 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11213 * New function EC_GROUP_get_degree, which is defined through
11214 EC_METHOD. For curves over prime fields, this returns the bit
11215 length of the modulus.
11217 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11219 * New functions EC_GROUP_dup, EC_POINT_dup.
11220 (These simply call ..._new and ..._copy).
11222 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11224 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11225 Polynomials are represented as BIGNUMs (where the sign bit is not
11226 used) in the following functions [macros]:
11229 BN_GF2m_sub [= BN_GF2m_add]
11230 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11231 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11232 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11234 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11235 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11236 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11237 BN_GF2m_cmp [= BN_ucmp]
11239 (Note that only the 'mod' functions are actually for fields GF(2^m).
11240 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11242 For some functions, an the irreducible polynomial defining a
11243 field can be given as an 'unsigned int[]' with strictly
11244 decreasing elements giving the indices of those bits that are set;
11245 i.e., p[] represents the polynomial
11246 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11248 p[0] > p[1] > ... > p[k] = 0.
11249 This applies to the following functions:
11252 BN_GF2m_mod_mul_arr
11253 BN_GF2m_mod_sqr_arr
11254 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11255 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11256 BN_GF2m_mod_exp_arr
11257 BN_GF2m_mod_sqrt_arr
11258 BN_GF2m_mod_solve_quad_arr
11262 Conversion can be performed by the following functions:
11267 bntest.c has additional tests for binary polynomial arithmetic.
11269 Two implementations for BN_GF2m_mod_div() are available.
11270 The default algorithm simply uses BN_GF2m_mod_inv() and
11271 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11272 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11273 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11275 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11277 * Add new error code 'ERR_R_DISABLED' that can be used when some
11278 functionality is disabled at compile-time.
11280 *Douglas Stebila <douglas.stebila@sun.com>*
11282 * Change default behaviour of 'openssl asn1parse' so that more
11283 information is visible when viewing, e.g., a certificate:
11285 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11286 mode the content of non-printable OCTET STRINGs is output in a
11287 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11288 avoid the appearance of a printable string.
11290 *Nils Larsch <nla@trustcenter.de>*
11292 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11294 EC_GROUP_set_asn1_flag()
11295 EC_GROUP_get_asn1_flag()
11296 EC_GROUP_set_point_conversion_form()
11297 EC_GROUP_get_point_conversion_form()
11298 These control ASN1 encoding details:
11299 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11300 has been set to OPENSSL_EC_NAMED_CURVE.
11301 - Points are encoded in uncompressed form by default; options for
11302 asn1_for are as for point2oct, namely
11303 POINT_CONVERSION_COMPRESSED
11304 POINT_CONVERSION_UNCOMPRESSED
11305 POINT_CONVERSION_HYBRID
11307 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11309 EC_GROUP_set_seed()
11310 EC_GROUP_get0_seed()
11311 EC_GROUP_get_seed_len()
11312 This is used only for ASN1 purposes (so far).
11314 *Nils Larsch <nla@trustcenter.de>*
11316 * Add 'field_type' member to EC_METHOD, which holds the NID
11317 of the appropriate field type OID. The new function
11318 EC_METHOD_get_field_type() returns this value.
11320 *Nils Larsch <nla@trustcenter.de>*
11323 EC_POINT_point2bn()
11324 EC_POINT_bn2point()
11325 EC_POINT_point2hex()
11326 EC_POINT_hex2point()
11327 providing useful interfaces to EC_POINT_point2oct() and
11328 EC_POINT_oct2point().
11330 *Nils Larsch <nla@trustcenter.de>*
11332 * Change internals of the EC library so that the functions
11333 EC_GROUP_set_generator()
11334 EC_GROUP_get_generator()
11335 EC_GROUP_get_order()
11336 EC_GROUP_get_cofactor()
11337 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11338 to methods, which would lead to unnecessary code duplication when
11339 adding different types of curves.
11341 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11343 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11344 arithmetic, and such that modified wNAFs are generated
11345 (which avoid length expansion in many cases).
11349 * Add a function EC_GROUP_check_discriminant() (defined via
11350 EC_METHOD) that verifies that the curve discriminant is non-zero.
11352 Add a function EC_GROUP_check() that makes some sanity tests
11353 on a EC_GROUP, its generator and order. This includes
11354 EC_GROUP_check_discriminant().
11356 *Nils Larsch <nla@trustcenter.de>*
11358 * Add ECDSA in new directory crypto/ecdsa/.
11360 Add applications 'openssl ecparam' and 'openssl ecdsa'
11361 (these are based on 'openssl dsaparam' and 'openssl dsa').
11363 ECDSA support is also included in various other files across the
11364 library. Most notably,
11365 - 'openssl req' now has a '-newkey ecdsa:file' option;
11366 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11367 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11368 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11369 them suitable for ECDSA where domain parameters must be
11370 extracted before the specific public key;
11371 - ECDSA engine support has been added.
11373 *Nils Larsch <nla@trustcenter.de>*
11375 * Include some named elliptic curves, and add OIDs from X9.62,
11376 SECG, and WAP/WTLS. Each curve can be obtained from the new
11378 EC_GROUP_new_by_curve_name(),
11379 and the list of available named curves can be obtained with
11380 EC_get_builtin_curves().
11381 Also add a 'curve_name' member to EC_GROUP objects, which can be
11383 EC_GROUP_set_curve_name()
11384 EC_GROUP_get_curve_name()
11386 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11388 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11389 was actually never needed) and in BN_mul(). The removal in BN_mul()
11390 required a small change in bn_mul_part_recursive() and the addition
11391 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11392 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11393 bn_sub_words() and bn_add_words() except they take arrays with
11398 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11400 * Cleanse PEM buffers before freeing them since they may contain
11403 *Benjamin Bennett <ben@psc.edu>*
11405 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11406 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11407 authentication-only ciphersuites.
11411 * Since AES128 and AES256 share a single mask bit in the logic of
11412 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11413 kludge to work properly if AES128 is available and AES256 isn't.
11417 * Expand security boundary to match 1.1.1 module.
11421 * Remove redundant features: hash file source, editing of test vectors
11422 modify fipsld to use external fips_premain.c signature.
11426 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11427 run algorithm test programs.
11431 * Make algorithm test programs more tolerant of whitespace.
11435 * Have SSL/TLS server implementation tolerate "mismatched" record
11436 protocol version while receiving ClientHello even if the
11437 ClientHello is fragmented. (The server can't insist on the
11438 particular protocol version it has chosen before the ServerHello
11439 message has informed the client about his choice.)
11443 * Load error codes if they are not already present instead of using a
11444 static variable. This allows them to be cleanly unloaded and reloaded.
11448 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11450 * Introduce limits to prevent malicious keys being able to
11451 cause a denial of service. ([CVE-2006-2940])
11453 *Steve Henson, Bodo Moeller*
11455 * Fix ASN.1 parsing of certain invalid structures that can result
11456 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11458 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11459 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11461 * Fix SSL client code which could crash if connecting to a
11462 malicious SSLv2 server. ([CVE-2006-4343])
11464 *Tavis Ormandy and Will Drewry, Google Security Team*
11466 * Change ciphersuite string processing so that an explicit
11467 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11468 will no longer include "AES128-SHA"), and any other similar
11469 ciphersuite (same bitmap) from *other* protocol versions (so that
11470 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11471 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11472 changes from 0.9.8b and 0.9.8d.
11476 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11478 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11479 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11481 * Change the Unix randomness entropy gathering to use poll() when
11482 possible instead of select(), since the latter has some
11483 undesirable limitations.
11485 *Darryl Miles via Richard Levitte and Bodo Moeller*
11487 * Disable rogue ciphersuites:
11489 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11490 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11491 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11493 The latter two were purportedly from
11494 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11497 Also deactivate the remaining ciphersuites from
11498 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11499 unofficial, and the ID has long expired.
11503 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11504 dual-core machines) and other potential thread-safety issues.
11508 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11510 * Adapt fipsld and the build system to link against the validated FIPS
11511 module in FIPS mode.
11515 * Fixes for VC++ 2005 build under Windows.
11519 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11520 from a Windows bash shell such as MSYS. It is autodetected from the
11521 "config" script when run from a VC++ environment. Modify standard VC++
11522 build to use fipscanister.o from the GNU make build.
11526 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11528 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11529 The value now differs depending on if you build for FIPS or not.
11530 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11531 safely run with a non-FIPSed libcrypto, as it may crash because of
11532 the difference induced by this change.
11536 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11538 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11539 (part of SSL_OP_ALL). This option used to disable the
11540 countermeasure against man-in-the-middle protocol-version
11541 rollback in the SSL 2.0 server implementation, which is a bad
11542 idea. ([CVE-2005-2969])
11544 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11545 for Information Security, National Institute of Advanced Industrial
11546 Science and Technology [AIST, Japan)]*
11548 * Minimal support for X9.31 signatures and PSS padding modes. This is
11549 mainly for FIPS compliance and not fully integrated at this stage.
11553 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11554 the exponentiation using a fixed-length exponent. (Otherwise,
11555 the information leaked through timing could expose the secret key
11556 after many signatures; cf. Bleichenbacher's attack on DSA with
11561 * Make a new fixed-window mod_exp implementation the default for
11562 RSA, DSA, and DH private-key operations so that the sequence of
11563 squares and multiplies and the memory access pattern are
11564 independent of the particular secret key. This will mitigate
11565 cache-timing and potential related attacks.
11567 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11568 and this is automatically used by BN_mod_exp_mont() if the new flag
11569 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11570 will use this BN flag for private exponents unless the flag
11571 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11572 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11574 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11576 * Change the client implementation for SSLv23_method() and
11577 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11578 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11579 (Previously, the SSL 2.0 backwards compatible Client Hello
11580 message format would be used even with SSL_OP_NO_SSLv2.)
11584 * Add support for smime-type MIME parameter in S/MIME messages which some
11589 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11590 a threadsafe manner. Modify rsa code to use new function and add calls
11591 to dsa and dh code (which had race conditions before).
11595 * Include the fixed error library code in the C error file definitions
11596 instead of fixing them up at runtime. This keeps the error code
11597 structures constant.
11601 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11603 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11606 * Fixes for newer kerberos headers. NB: the casts are needed because
11607 the 'length' field is signed on one version and unsigned on another
11608 with no (?) obvious way to tell the difference, without these VC++
11609 complains. Also the "definition" of FAR (blank) is no longer included
11610 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11611 some needed definitions.
11615 * Undo Cygwin change.
11619 * Added support for proxy certificates according to RFC 3820.
11620 Because they may be a security thread to unaware applications,
11621 they must be explicitly allowed in run-time. See
11622 docs/HOWTO/proxy_certificates.txt for further information.
11626 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11628 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11629 server and client random values. Previously
11630 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11631 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11633 This change has negligible security impact because:
11635 1. Server and client random values still have 24 bytes of pseudo random
11638 2. Server and client random values are sent in the clear in the initial
11641 3. The master secret is derived using the premaster secret (48 bytes in
11642 size for static RSA ciphersuites) as well as client server and random
11645 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11648 *Stephen Henson, reported by UK NISCC*
11650 * Use Windows randomness collection on Cygwin.
11654 * Fix hang in EGD/PRNGD query when communication socket is closed
11655 prematurely by EGD/PRNGD.
11657 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11659 * Prompt for pass phrases when appropriate for PKCS12 input format.
11663 * Back-port of selected performance improvements from development
11664 branch, as well as improved support for PowerPC platforms.
11668 * Add lots of checks for memory allocation failure, error codes to indicate
11669 failure and freeing up memory if a failure occurs.
11671 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11673 * Add new -passin argument to dgst.
11677 * Perform some character comparisons of different types in X509_NAME_cmp:
11678 this is needed for some certificates that re-encode DNs into UTF8Strings
11679 (in violation of RFC3280) and can't or won't issue name rollover
11684 * Make an explicit check during certificate validation to see that
11685 the CA setting in each certificate on the chain is correct. As a
11686 side effect always do the following basic checks on extensions,
11687 not just when there's an associated purpose to the check:
11689 - if there is an unhandled critical extension (unless the user
11690 has chosen to ignore this fault)
11691 - if the path length has been exceeded (if one is set at all)
11692 - that certain extensions fit the associated purpose (if one has
11697 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11699 * Avoid a race condition when CRLs are checked in a multi threaded
11700 environment. This would happen due to the reordering of the revoked
11701 entries during signature checking and serial number lookup. Now the
11702 encoding is cached and the serial number sort performed under a lock.
11703 Add new STACK function sk_is_sorted().
11707 * Add Delta CRL to the extension code.
11711 * Various fixes to s3_pkt.c so alerts are sent properly.
11713 *David Holmes <d.holmes@f5.com>*
11715 * Reduce the chances of duplicate issuer name and serial numbers (in
11716 violation of RFC3280) using the OpenSSL certificate creation utilities.
11717 This is done by creating a random 64 bit value for the initial serial
11718 number when a serial number file is created or when a self signed
11719 certificate is created using 'openssl req -x509'. The initial serial
11720 number file is created using 'openssl x509 -next_serial' in CA.pl
11721 rather than being initialized to 1.
11725 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11727 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11728 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11730 *Joe Orton, Steve Henson*
11732 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11735 *Joe Orton, Steve Henson*
11737 * Make it possible to have multiple active certificates with the same
11738 subject in the CA index file. This is done only if the keyword
11739 'unique_subject' is set to 'no' in the main CA section (default
11740 if 'CA_default') of the configuration file. The value is saved
11741 with the database itself in a separate index attribute file,
11742 named like the index file with '.attr' appended to the name.
11746 * X509 verify fixes. Disable broken certificate workarounds when
11747 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11748 keyUsage extension present. Don't accept CRLs with unhandled critical
11749 extensions: since verify currently doesn't process CRL extensions this
11750 rejects a CRL with *any* critical extensions. Add new verify error codes
11755 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11756 A clarification of RFC2560 will require the use of OCTET STRINGs and
11757 some implementations cannot handle the current raw format. Since OpenSSL
11758 copies and compares OCSP nonces as opaque blobs without any attempt at
11759 parsing them this should not create any compatibility issues.
11763 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11764 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11765 this HMAC (and other) operations are several times slower than OpenSSL
11770 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11772 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11774 * Use the correct content when signing type "other".
11778 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11780 * Fix various bugs revealed by running the NISCC test suite:
11782 Stop out of bounds reads in the ASN1 code when presented with
11783 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11785 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11787 If verify callback ignores invalid public key errors don't try to check
11788 certificate signature with the NULL public key.
11792 * New -ignore_err option in ocsp application to stop the server
11793 exiting on the first error in a request.
11797 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11798 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11803 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11804 extra data after the compression methods not only for TLS 1.0
11805 but also for SSL 3.0 (as required by the specification).
11807 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11809 * Change X509_certificate_type() to mark the key as exported/exportable
11810 when it's 512 *bits* long, not 512 bytes.
11814 * Change AES_cbc_encrypt() so it outputs exact multiple of
11815 blocks during encryption.
11819 * Various fixes to base64 BIO and non blocking I/O. On write
11820 flushes were not handled properly if the BIO retried. On read
11821 data was not being buffered properly and had various logic bugs.
11822 This also affects blocking I/O when the data being decoded is a
11827 * Various S/MIME bugfixes and compatibility changes:
11828 output correct application/pkcs7 MIME type if
11829 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11830 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11831 of files as .eml work). Correctly handle very long lines in MIME
11836 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11838 * Countermeasure against the Klima-Pokorny-Rosa extension of
11839 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11840 a protocol version number mismatch like a decryption error
11841 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11845 * Turn on RSA blinding by default in the default implementation
11846 to avoid a timing attack. Applications that don't want it can call
11847 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11848 They would be ill-advised to do so in most cases.
11850 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11852 * Change RSA blinding code so that it works when the PRNG is not
11853 seeded (in this case, the secret RSA exponent is abused as
11854 an unpredictable seed -- if it is not unpredictable, there
11855 is no point in blinding anyway). Make RSA blinding thread-safe
11856 by remembering the creator's thread ID in rsa->blinding and
11857 having all other threads use local one-time blinding factors
11858 (this requires more computation than sharing rsa->blinding, but
11859 avoids excessive locking; and if an RSA object is not shared
11860 between threads, blinding will still be very fast).
11864 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11865 ENGINE as defaults for all supported algorithms irrespective of
11866 the 'flags' parameter. 'flags' is now honoured, so applications
11867 should make sure they are passing it correctly.
11871 * Target "mingw" now allows native Windows code to be generated in
11872 the Cygwin environment as well as with the MinGW compiler.
11876 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11878 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11879 via timing by performing a MAC computation even if incorrect
11880 block cipher padding has been found. This is a countermeasure
11881 against active attacks where the attacker has to distinguish
11882 between bad padding and a MAC verification error. ([CVE-2003-0078])
11884 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11885 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11886 Martin Vuagnoux (EPFL, Ilion)*
11888 * Make the no-err option work as intended. The intention with no-err
11889 is not to have the whole error stack handling routines removed from
11890 libcrypto, it's only intended to remove all the function name and
11891 reason texts, thereby removing some of the footprint that may not
11892 be interesting if those errors aren't displayed anyway.
11894 NOTE: it's still possible for any application or module to have its
11895 own set of error texts inserted. The routines are there, just not
11896 used by default when no-err is given.
11900 * Add support for FreeBSD on IA64.
11902 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11904 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11905 Kerberos function mit_des_cbc_cksum(). Before this change,
11906 the value returned by DES_cbc_cksum() was like the one from
11907 mit_des_cbc_cksum(), except the bytes were swapped.
11909 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11911 * Allow an application to disable the automatic SSL chain building.
11912 Before this a rather primitive chain build was always performed in
11913 ssl3_output_cert_chain(): an application had no way to send the
11914 correct chain if the automatic operation produced an incorrect result.
11916 Now the chain builder is disabled if either:
11918 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11920 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11922 The reasoning behind this is that an application would not want the
11923 auto chain building to take place if extra chain certificates are
11924 present and it might also want a means of sending no additional
11925 certificates (for example the chain has two certificates and the
11930 * Add the possibility to build without the ENGINE framework.
11932 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11934 * Under Win32 gmtime() can return NULL: check return value in
11935 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11939 * DSA routines: under certain error conditions uninitialized BN objects
11940 could be freed. Solution: make sure initialization is performed early
11941 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11942 Nils Larsch <nla@trustcenter.de> via PR#459)
11946 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11947 checked on reconnect on the client side, therefore session resumption
11948 could still fail with a "ssl session id is different" error. This
11949 behaviour is masked when SSL_OP_ALL is used due to
11950 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11951 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11952 followup to PR #377.
11956 * IA-32 assembler support enhancements: unified ELF targets, support
11957 for SCO/Caldera platforms, fix for Cygwin shared build.
11961 * Add support for FreeBSD on sparc64. As a consequence, support for
11962 FreeBSD on non-x86 processors is separate from x86 processors on
11963 the config script, much like the NetBSD support.
11965 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11967 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11969 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11972 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11973 code (06) was taken as the first octet of the session ID and the last
11974 octet was ignored consequently. As a result SSLv2 client side session
11975 caching could not have worked due to the session ID mismatch between
11977 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11982 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11983 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11988 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11989 seems that in spite of existing for more than a year, many application
11990 author have done nothing to provide the necessary callbacks, which
11991 means that this particular engine will not work properly anywhere.
11992 This is a very unfortunate situation which forces us, in the name
11993 of usability, to give the hw_ncipher.c a static lock, which is part
11995 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11996 appear in 0.9.8 or later. We EXPECT application authors to have
11997 dealt properly with this when 0.9.8 is released (unless we actually
11998 make such changes in the libcrypto locking code that changes will
11999 have to be made anyway).
12003 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12004 octets have been read, EOF or an error occurs. Without this change
12005 some truncated ASN1 structures will not produce an error.
12009 * Disable Heimdal support, since it hasn't been fully implemented.
12010 Still give the possibility to force the use of Heimdal, but with
12011 warnings and a request that patches get sent to openssl-dev.
12015 * Add the VC-CE target, introduce the WINCE sysname, and add
12016 INSTALL.WCE and appropriate conditionals to make it build.
12018 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12020 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12021 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12022 edit numbers of the version.
12024 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12026 * Introduce safe string copy and catenation functions
12027 (BUF_strlcpy() and BUF_strlcat()).
12029 *Ben Laurie (CHATS) and Richard Levitte*
12031 * Avoid using fixed-size buffers for one-line DNs.
12033 *Ben Laurie (CHATS)*
12035 * Add BUF_MEM_grow_clean() to avoid information leakage when
12036 resizing buffers containing secrets, and use where appropriate.
12038 *Ben Laurie (CHATS)*
12040 * Avoid using fixed size buffers for configuration file location.
12042 *Ben Laurie (CHATS)*
12044 * Avoid filename truncation for various CA files.
12046 *Ben Laurie (CHATS)*
12048 * Use sizeof in preference to magic numbers.
12050 *Ben Laurie (CHATS)*
12052 * Avoid filename truncation in cert requests.
12054 *Ben Laurie (CHATS)*
12056 * Add assertions to check for (supposedly impossible) buffer
12059 *Ben Laurie (CHATS)*
12061 * Don't cache truncated DNS entries in the local cache (this could
12062 potentially lead to a spoofing attack).
12064 *Ben Laurie (CHATS)*
12066 * Fix various buffers to be large enough for hex/decimal
12067 representations in a platform independent manner.
12069 *Ben Laurie (CHATS)*
12071 * Add CRYPTO_realloc_clean() to avoid information leakage when
12072 resizing buffers containing secrets, and use where appropriate.
12074 *Ben Laurie (CHATS)*
12076 * Add BIO_indent() to avoid much slightly worrying code to do
12079 *Ben Laurie (CHATS)*
12081 * Convert sprintf()/BIO_puts() to BIO_printf().
12083 *Ben Laurie (CHATS)*
12085 * buffer_gets() could terminate with the buffer only half
12088 *Ben Laurie (CHATS)*
12090 * Add assertions to prevent user-supplied crypto functions from
12091 overflowing internal buffers by having large block sizes, etc.
12093 *Ben Laurie (CHATS)*
12095 * New OPENSSL_assert() macro (similar to assert(), but enabled
12098 *Ben Laurie (CHATS)*
12100 * Eliminate unused copy of key in RC4.
12102 *Ben Laurie (CHATS)*
12104 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12106 *Ben Laurie (CHATS)*
12108 * Fix off-by-one error in EGD path.
12110 *Ben Laurie (CHATS)*
12112 * If RANDFILE path is too long, ignore instead of truncating.
12114 *Ben Laurie (CHATS)*
12116 * Eliminate unused and incorrectly sized X.509 structure
12119 *Ben Laurie (CHATS)*
12121 * Eliminate unused and dangerous function knumber().
12123 *Ben Laurie (CHATS)*
12125 * Eliminate unused and dangerous structure, KSSL_ERR.
12127 *Ben Laurie (CHATS)*
12129 * Protect against overlong session ID context length in an encoded
12130 session object. Since these are local, this does not appear to be
12133 *Ben Laurie (CHATS)*
12135 * Change from security patch (see 0.9.6e below) that did not affect
12136 the 0.9.6 release series:
12138 Remote buffer overflow in SSL3 protocol - an attacker could
12139 supply an oversized master key in Kerberos-enabled versions.
12142 *Ben Laurie (CHATS)*
12144 * Change the SSL kerb5 codes to match RFC 2712.
12148 * Make -nameopt work fully for req and add -reqopt switch.
12150 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12152 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12154 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12156 * Make sure tests can be performed even if the corresponding algorithms
12157 have been removed entirely. This was also the last step to make
12158 OpenSSL compilable with DJGPP under all reasonable conditions.
12160 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12162 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12163 to allow version independent disabling of normally unselected ciphers,
12164 which may be activated as a side-effect of selecting a single cipher.
12166 (E.g., cipher list string "RSA" enables ciphersuites that are left
12167 out of "ALL" because they do not provide symmetric encryption.
12168 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12170 *Lutz Jaenicke, Bodo Moeller*
12172 * Add appropriate support for separate platform-dependent build
12173 directories. The recommended way to make a platform-dependent
12174 build directory is the following (tested on Linux), maybe with
12177 # Place yourself outside of the OpenSSL source tree. In
12178 # this example, the environment variable OPENSSL_SOURCE
12179 # is assumed to contain the absolute OpenSSL source directory.
12180 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12181 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12182 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12183 mkdir -p `dirname $F`
12184 ln -s $OPENSSL_SOURCE/$F $F
12187 To be absolutely sure not to disturb the source tree, a "make clean"
12188 is a good thing. If it isn't successful, don't worry about it,
12189 it probably means the source directory is very clean.
12193 * Make sure any ENGINE control commands make local copies of string
12194 pointers passed to them whenever necessary. Otherwise it is possible
12195 the caller may have overwritten (or deallocated) the original string
12196 data when a later ENGINE operation tries to use the stored values.
12198 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12200 * Improve diagnostics in file reading and command-line digests.
12202 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12204 * Add AES modes CFB and OFB to the object database. Correct an
12205 error in AES-CFB decryption.
12209 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12210 allows existing EVP_CIPHER_CTX structures to be reused after
12211 calling `EVP_*Final()`. This behaviour is used by encryption
12212 BIOs and some applications. This has the side effect that
12213 applications must explicitly clean up cipher contexts with
12214 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12218 * Check the values of dna and dnb in bn_mul_recursive before calling
12219 bn_mul_comba (a non zero value means the a or b arrays do not contain
12220 n2 elements) and fallback to bn_mul_normal if either is not zero.
12224 * Fix escaping of non-ASCII characters when using the -subj option
12225 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12229 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12230 form for "surname", serialNumber has no short form.
12231 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12232 therefore remove "mail" short name for "internet 7".
12233 The OID for unique identifiers in X509 certificates is
12234 x500UniqueIdentifier, not uniqueIdentifier.
12235 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12239 * Add an "init" command to the ENGINE config module and auto initialize
12240 ENGINEs. Without any "init" command the ENGINE will be initialized
12241 after all ctrl commands have been executed on it. If init=1 the
12242 ENGINE is initialized at that point (ctrls before that point are run
12243 on the uninitialized ENGINE and after on the initialized one). If
12244 init=0 then the ENGINE will not be initialized at all.
12248 * Fix the 'app_verify_callback' interface so that the user-defined
12249 argument is actually passed to the callback: In the
12250 SSL_CTX_set_cert_verify_callback() prototype, the callback
12251 declaration has been changed from
12254 int (*cb)(X509_STORE_CTX *,void *);
12255 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12256 i=s->ctx->app_verify_callback(&ctx)
12257 has been changed into
12258 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12260 To update applications using SSL_CTX_set_cert_verify_callback(),
12261 a dummy argument can be added to their callback functions.
12263 *D. K. Smetters <smetters@parc.xerox.com>*
12265 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12267 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12269 * Add and OPENSSL_LOAD_CONF define which will cause
12270 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12271 This allows older applications to transparently support certain
12272 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12273 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12274 load the config file and OPENSSL_add_all_algorithms_conf() which will
12275 always load it have also been added.
12279 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12280 Adjust NIDs and EVP layer.
12282 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12284 * Config modules support in openssl utility.
12286 Most commands now load modules from the config file,
12287 though in a few (such as version) this isn't done
12288 because it couldn't be used for anything.
12290 In the case of ca and req the config file used is
12291 the same as the utility itself: that is the -config
12292 command line option can be used to specify an
12297 * Move default behaviour from OPENSSL_config(). If appname is NULL
12298 use "openssl_conf" if filename is NULL use default openssl config file.
12302 * Add an argument to OPENSSL_config() to allow the use of an alternative
12303 config section name. Add a new flag to tolerate a missing config file
12304 and move code to CONF_modules_load_file().
12308 * Support for crypto accelerator cards from Accelerated Encryption
12309 Processing, www.aep.ie. (Use engine 'aep')
12310 The support was copied from 0.9.6c [engine] and adapted/corrected
12311 to work with the new engine framework.
12313 *AEP Inc. and Richard Levitte*
12315 * Support for SureWare crypto accelerator cards from Baltimore
12316 Technologies. (Use engine 'sureware')
12317 The support was copied from 0.9.6c [engine] and adapted
12318 to work with the new engine framework.
12322 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12323 make the newer ENGINE framework commands for the CHIL engine work.
12325 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12327 * Make it possible to produce shared libraries on ReliantUNIX.
12329 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12331 * Add the configuration target debug-linux-ppro.
12332 Make 'openssl rsa' use the general key loading routines
12333 implemented in `apps.c`, and make those routines able to
12334 handle the key format FORMAT_NETSCAPE and the variant
12337 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12339 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12341 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12343 * Add -keyform to rsautl, and document -engine.
12345 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12347 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12348 BIO_R_NO_SUCH_FILE error code rather than the generic
12349 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12353 * Add new functions
12354 ERR_peek_last_error
12355 ERR_peek_last_error_line
12356 ERR_peek_last_error_line_data.
12357 These are similar to
12359 ERR_peek_error_line
12360 ERR_peek_error_line_data,
12361 but report on the latest error recorded rather than the first one
12362 still in the error queue.
12364 *Ben Laurie, Bodo Moeller*
12366 * default_algorithms option in ENGINE config module. This allows things
12368 default_algorithms = ALL
12369 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12373 * Preliminary ENGINE config module.
12377 * New experimental application configuration code.
12381 * Change the AES code to follow the same name structure as all other
12382 symmetric ciphers, and behave the same way. Move everything to
12383 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12385 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12387 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12389 *Ben Laurie and Theo de Raadt*
12391 * Add option to output public keys in req command.
12393 *Massimiliano Pala madwolf@openca.org*
12395 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12396 (up to about 10% better than before for P-192 and P-224).
12400 * New functions/macros
12402 SSL_CTX_set_msg_callback(ctx, cb)
12403 SSL_CTX_set_msg_callback_arg(ctx, arg)
12404 SSL_set_msg_callback(ssl, cb)
12405 SSL_set_msg_callback_arg(ssl, arg)
12407 to request calling a callback function
12409 void cb(int write_p, int version, int content_type,
12410 const void *buf, size_t len, SSL *ssl, void *arg)
12412 whenever a protocol message has been completely received
12413 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12414 protocol version according to which the SSL library interprets
12415 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12416 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12417 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12418 specification (change_cipher_spec(20), alert(21), handshake(22)).
12419 'buf' and 'len' point to the actual message, 'ssl' to the
12420 SSL object, and 'arg' is the application-defined value set by
12421 SSL[_CTX]_set_msg_callback_arg().
12423 'openssl s_client' and 'openssl s_server' have new '-msg' options
12424 to enable a callback that displays all protocol messages.
12428 * Change the shared library support so shared libraries are built as
12429 soon as the corresponding static library is finished, and thereby get
12430 openssl and the test programs linked against the shared library.
12431 This still only happens when the keyword "shard" has been given to
12432 the configuration scripts.
12434 NOTE: shared library support is still an experimental thing, and
12435 backward binary compatibility is still not guaranteed.
12437 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12439 * Add support for Subject Information Access extension.
12441 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12443 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12444 additional bytes when new memory had to be allocated, not just
12445 when reusing an existing buffer.
12449 * New command line and configuration option 'utf8' for the req command.
12450 This allows field values to be specified as UTF8 strings.
12454 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12455 runs for the former and machine-readable output for the latter.
12459 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12460 of the e-mail address in the DN (i.e., it will go into a certificate
12461 extension only). The new configuration file option 'email_in_dn = no'
12462 has the same effect.
12464 *Massimiliano Pala madwolf@openca.org*
12466 * Change all functions with names starting with `des_` to be starting
12467 with `DES_` instead. Add wrappers that are compatible with libdes,
12468 but are named `_ossl_old_des_*`. Finally, add macros that map the
12469 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12470 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12471 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12474 Since we provide two compatibility mappings, the user needs to
12475 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12476 compatibility is desired. The default (i.e., when that macro
12477 isn't defined) is OpenSSL 0.9.6c compatibility.
12479 There are also macros that enable and disable the support of old
12480 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12481 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12482 are defined, the default will apply: to support the old des routines.
12484 In either case, one must include openssl/des.h to get the correct
12485 definitions. Do not try to just include openssl/des_old.h, that
12488 NOTE: This is a major break of an old API into a new one. Software
12489 authors are encouraged to switch to the `DES_` style functions. Some
12490 time in the future, des_old.h and the libdes compatibility functions
12491 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12492 default), and then completely removed.
12496 * Test for certificates which contain unsupported critical extensions.
12497 If such a certificate is found during a verify operation it is
12498 rejected by default: this behaviour can be overridden by either
12499 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12500 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12501 X509_supported_extension() has also been added which returns 1 if a
12502 particular extension is supported.
12506 * Modify the behaviour of EVP cipher functions in similar way to digests
12507 to retain compatibility with existing code.
12511 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12512 compatibility with existing code. In particular the 'ctx' parameter does
12513 not have to be to be initialized before the call to EVP_DigestInit() and
12514 it is tidied up after a call to EVP_DigestFinal(). New function
12515 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12516 EVP_MD_CTX_copy() changed to not require the destination to be
12517 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12518 requires the destination to be valid.
12520 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12521 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12525 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12526 so that complete 'Handshake' protocol structures are kept in memory
12527 instead of overwriting 'msg_type' and 'length' with 'body' data.
12531 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12533 *Massimo Santin via Richard Levitte*
12535 * Major restructuring to the underlying ENGINE code. This includes
12536 reduction of linker bloat, separation of pure "ENGINE" manipulation
12537 (initialisation, etc) from functionality dealing with implementations
12538 of specific crypto interfaces. This change also introduces integrated
12539 support for symmetric ciphers and digest implementations - so ENGINEs
12540 can now accelerate these by providing EVP_CIPHER and EVP_MD
12541 implementations of their own. This is detailed in
12542 [crypto/engine/README.md](crypto/engine/README.md)
12543 as it couldn't be adequately described here. However, there are a few
12544 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12545 were changed in the original introduction of ENGINE code have now
12546 reverted back - the hooking from this code to ENGINE is now a good
12547 deal more passive and at run-time, operations deal directly with
12548 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12549 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12550 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12551 they were not being used by the framework as there is no concept of a
12552 BIGNUM_METHOD and they could not be generalised to the new
12553 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12554 ENGINE_cpy() has been removed as it cannot be consistently defined in
12559 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12563 * Change mkdef.pl to sort symbols that get the same entry number,
12564 and make sure the automatically generated functions `ERR_load_*`
12565 become part of libeay.num as well.
12569 * New function SSL_renegotiate_pending(). This returns true once
12570 renegotiation has been requested (either SSL_renegotiate() call
12571 or HelloRequest/ClientHello received from the peer) and becomes
12572 false once a handshake has been completed.
12573 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12574 sends a HelloRequest, but does not ensure that a handshake takes
12575 place. SSL_renegotiate_pending() is useful for checking if the
12576 client has followed the request.)
12580 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12581 By default, clients may request session resumption even during
12582 renegotiation (if session ID contexts permit); with this option,
12583 session resumption is possible only in the first handshake.
12585 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12586 more bits available for options that should not be part of
12587 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12591 * Add some demos for certificate and certificate request creation.
12595 * Make maximum certificate chain size accepted from the peer application
12596 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12597 "Douglas E. Engert" <deengert@anl.gov>.
12601 * Add support for shared libraries for Unixware-7
12602 (Boyd Lynn Gerber <gerberb@zenez.com>).
12606 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12607 be done prior to destruction. Use this to unload error strings from
12608 ENGINEs that load their own error strings. NB: This adds two new API
12609 functions to "get" and "set" this destroy handler in an ENGINE.
12613 * Alter all existing ENGINE implementations (except "openssl" and
12614 "openbsd") to dynamically instantiate their own error strings. This
12615 makes them more flexible to be built both as statically-linked ENGINEs
12616 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12617 Also, add stub code to each that makes building them as self-contained
12618 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12622 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12623 implementations into applications that are completely implemented in
12624 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12625 commands that can be used to configure what shared-library to load and
12626 to control aspects of the way it is handled. Also, made an update to
12627 the [README-Engine.md](README-Engine.md) file
12628 that brings its information up-to-date and
12629 provides some information and instructions on the "dynamic" ENGINE
12630 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12634 * Make it possible to unload ranges of ERR strings with a new
12635 "ERR_unload_strings" function.
12639 * Add a copy() function to EVP_MD.
12643 * Make EVP_MD routines take a context pointer instead of just the
12644 md_data void pointer.
12648 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12649 that the digest can only process a single chunk of data
12650 (typically because it is provided by a piece of
12651 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12652 is only going to provide a single chunk of data, and hence the
12653 framework needn't accumulate the data for oneshot drivers.
12657 * As with "ERR", make it possible to replace the underlying "ex_data"
12658 functions. This change also alters the storage and management of global
12659 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12660 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12661 index counters. The API functions that use this state have been changed
12662 to take a "class_index" rather than pointers to the class's local STACK
12663 and counter, and there is now an API function to dynamically create new
12664 classes. This centralisation allows us to (a) plug a lot of the
12665 thread-safety problems that existed, and (b) makes it possible to clean
12666 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12667 such data would previously have always leaked in application code and
12668 workarounds were in place to make the memory debugging turn a blind eye
12669 to it. Application code that doesn't use this new function will still
12670 leak as before, but their memory debugging output will announce it now
12671 rather than letting it slide.
12673 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12674 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12675 has a return value to indicate success or failure.
12679 * Make it possible to replace the underlying "ERR" functions such that the
12680 global state (2 LHASH tables and 2 locks) is only used by the "default"
12681 implementation. This change also adds two functions to "get" and "set"
12682 the implementation prior to it being automatically set the first time
12683 any other ERR function takes place. Ie. an application can call "get",
12684 pass the return value to a module it has just loaded, and that module
12685 can call its own "set" function using that value. This means the
12686 module's "ERR" operations will use (and modify) the error state in the
12687 application and not in its own statically linked copy of OpenSSL code.
12691 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12692 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12693 the operation, and provides a more encapsulated way for external code
12694 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12695 to use these functions rather than manually incrementing the counts.
12697 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12701 * Add EVP test program.
12705 * Add symmetric cipher support to ENGINE. Expect the API to change!
12709 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12710 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12711 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12712 These allow a CRL to be built without having to access X509_CRL fields
12713 directly. Modify 'ca' application to use new functions.
12717 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12718 bug workarounds. Rollback attack detection is a security feature.
12719 The problem will only arise on OpenSSL servers when TLSv1 is not
12720 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12721 Software authors not wanting to support TLSv1 will have special reasons
12722 for their choice and can explicitly enable this option.
12724 *Bodo Moeller, Lutz Jaenicke*
12726 * Rationalise EVP so it can be extended: don't include a union of
12727 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12728 (similar to those existing for EVP_CIPHER_CTX).
12733 EVP_MD_CTX_init(&md); /* new function call */
12734 EVP_DigestInit(&md, EVP_sha1());
12735 EVP_DigestUpdate(&md, in, len);
12736 EVP_DigestFinal(&md, out, NULL);
12737 EVP_MD_CTX_cleanup(&md); /* new function call */
12741 * Make DES key schedule conform to the usual scheme, as well as
12742 correcting its structure. This means that calls to DES functions
12743 now have to pass a pointer to a des_key_schedule instead of a
12744 plain des_key_schedule (which was actually always a pointer
12747 des_key_schedule ks;
12749 des_set_key_checked(..., &ks);
12750 des_ncbc_encrypt(..., &ks, ...);
12752 (Note that a later change renames 'des_...' into 'DES_...'.)
12756 * Initial reduction of linker bloat: the use of some functions, such as
12757 PEM causes large amounts of unused functions to be linked in due to
12758 poor organisation. For example pem_all.c contains every PEM function
12759 which has a knock on effect of linking in large amounts of (unused)
12760 ASN1 code. Grouping together similar functions and splitting unrelated
12761 functions prevents this.
12765 * Cleanup of EVP macros.
12769 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12770 correct `_ecb suffix`.
12774 * Add initial OCSP responder support to ocsp application. The
12775 revocation information is handled using the text based index
12776 use by the ca application. The responder can either handle
12777 requests generated internally, supplied in files (for example
12778 via a CGI script) or using an internal minimal server.
12782 * Add configuration choices to get zlib compression for TLS.
12786 * Changes to Kerberos SSL for RFC 2712 compliance:
12787 1. Implemented real KerberosWrapper, instead of just using
12788 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12789 2. Implemented optional authenticator field of KerberosWrapper.
12791 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12792 and authenticator structs; see crypto/krb5/.
12794 Generalized Kerberos calls to support multiple Kerberos libraries.
12795 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12796 via Richard Levitte*
12798 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12799 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12800 values for each of the key sizes rather than having just
12801 parameters (and 'speed' generating keys each time).
12805 * Speed up EVP routines.
12808 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12809 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12810 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12811 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12813 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12814 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12815 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12818 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12820 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12824 * Added the OS2-EMX target.
12826 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12828 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12829 New functions to support `NCONF` routines in extension code.
12830 New function `CONF_set_nconf()`
12831 to allow functions which take an `NCONF` to also handle the old `LHASH`
12832 structure: this means that the old `CONF` compatible routines can be
12833 retained (in particular w.rt. extensions) without having to duplicate the
12834 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12838 * Enhance the general user interface with mechanisms for inner control
12839 and with possibilities to have yes/no kind of prompts.
12843 * Change all calls to low-level digest routines in the library and
12844 applications to use EVP. Add missing calls to HMAC_cleanup() and
12845 don't assume HMAC_CTX can be copied using memcpy().
12847 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12849 * Add the possibility to control engines through control names but with
12850 arbitrary arguments instead of just a string.
12851 Change the key loaders to take a UI_METHOD instead of a callback
12852 function pointer. NOTE: this breaks binary compatibility with earlier
12853 versions of OpenSSL [engine].
12854 Adapt the nCipher code for these new conditions and add a card insertion
12859 * Enhance the general user interface with mechanisms to better support
12860 dialog box interfaces, application-defined prompts, the possibility
12861 to use defaults (for example default passwords from somewhere else)
12862 and interrupts/cancellations.
12866 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12867 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12871 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12872 tidy up some unnecessarily weird code in 'sk_new()').
12874 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12876 * Change the key loading routines for ENGINEs to use the same kind
12877 callback (pem_password_cb) as all other routines that need this
12882 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12883 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12884 than this minimum value is recommended.
12888 * New random seeder for OpenVMS, using the system process statistics
12889 that are easily reachable.
12893 * Windows apparently can't transparently handle global
12894 variables defined in DLLs. Initialisations such as:
12896 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12898 won't compile. This is used by the any applications that need to
12899 declare their own ASN1 modules. This was fixed by adding the option
12900 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12901 needed for static libraries under Win32.
12905 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12906 setting of purpose and trust fields. New X509_STORE trust and
12907 purpose functions and tidy up setting in other SSL functions.
12911 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12912 structure. These are inherited by X509_STORE_CTX when it is
12913 initialised. This allows various defaults to be set in the
12914 X509_STORE structure (such as flags for CRL checking and custom
12915 purpose or trust settings) for functions which only use X509_STORE_CTX
12916 internally such as S/MIME.
12918 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12919 trust settings if they are not set in X509_STORE. This allows X509_STORE
12920 purposes and trust (in S/MIME for example) to override any set by default.
12922 Add command line options for CRL checking to smime, s_client and s_server
12927 * Initial CRL based revocation checking. If the CRL checking flag(s)
12928 are set then the CRL is looked up in the X509_STORE structure and
12929 its validity and signature checked, then if the certificate is found
12930 in the CRL the verify fails with a revoked error.
12932 Various new CRL related callbacks added to X509_STORE_CTX structure.
12934 Command line options added to 'verify' application to support this.
12936 This needs some additional work, such as being able to handle multiple
12937 CRLs with different times, extension based lookup (rather than just
12938 by subject name) and ultimately more complete V2 CRL extension
12943 * Add a general user interface API (crypto/ui/). This is designed
12944 to replace things like des_read_password and friends (backward
12945 compatibility functions using this new API are provided).
12946 The purpose is to remove prompting functions from the DES code
12947 section as well as provide for prompting through dialog boxes in
12948 a window system and the like.
12952 * Add "ex_data" support to ENGINE so implementations can add state at a
12953 per-structure level rather than having to store it globally.
12957 * Make it possible for ENGINE structures to be copied when retrieved by
12958 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12959 This causes the "original" ENGINE structure to act like a template,
12960 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12961 operational state can be localised to each ENGINE structure, despite the
12962 fact they all share the same "methods". New ENGINE structures returned in
12963 this case have no functional references and the return value is the single
12964 structural reference. This matches the single structural reference returned
12965 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12970 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12971 needs to match any other type at all we need to manually clear the
12976 * Changes to the "openssl engine" utility to include;
12977 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12978 about an ENGINE's available control commands.
12979 - executing control commands from command line arguments using the
12980 '-pre' and '-post' switches. '-post' is only used if '-t' is
12981 specified and the ENGINE is successfully initialised. The syntax for
12982 the individual commands are colon-separated, for example;
12983 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12987 * New dynamic control command support for ENGINEs. ENGINEs can now
12988 declare their own commands (numbers), names (strings), descriptions,
12989 and input types for run-time discovery by calling applications. A
12990 subset of these commands are implicitly classed as "executable"
12991 depending on their input type, and only these can be invoked through
12992 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12993 can be based on user input, config files, etc). The distinction is
12994 that "executable" commands cannot return anything other than a boolean
12995 result and can only support numeric or string input, whereas some
12996 discoverable commands may only be for direct use through
12997 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12998 pointers, or other custom uses. The "executable" commands are to
12999 support parameterisations of ENGINE behaviour that can be
13000 unambiguously defined by ENGINEs and used consistently across any
13001 OpenSSL-based application. Commands have been added to all the
13002 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13003 control over shared-library paths without source code alterations.
13007 * Changed all ENGINE implementations to dynamically allocate their
13008 ENGINEs rather than declaring them statically. Apart from this being
13009 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13010 this also allows the implementations to compile without using the
13011 internal engine_int.h header.
13015 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13016 'const' value. Any code that should be able to modify a RAND_METHOD
13017 should already have non-const pointers to it (ie. they should only
13018 modify their own ones).
13022 * Made a variety of little tweaks to the ENGINE code.
13023 - "atalla" and "ubsec" string definitions were moved from header files
13024 to C code. "nuron" string definitions were placed in variables
13025 rather than hard-coded - allowing parameterisation of these values
13026 later on via ctrl() commands.
13027 - Removed unused "#if 0"'d code.
13028 - Fixed engine list iteration code so it uses ENGINE_free() to release
13029 structural references.
13030 - Constified the RAND_METHOD element of ENGINE structures.
13031 - Constified various get/set functions as appropriate and added
13032 missing functions (including a catch-all ENGINE_cpy that duplicates
13033 all ENGINE values onto a new ENGINE except reference counts/state).
13034 - Removed NULL parameter checks in get/set functions. Setting a method
13035 or function to NULL is a way of cancelling out a previously set
13036 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13037 and doesn't justify the extra error symbols and code.
13038 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13039 flags from engine_int.h to engine.h.
13040 - Changed prototypes for ENGINE handler functions (init(), finish(),
13041 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13045 * Implement binary inversion algorithm for BN_mod_inverse in addition
13046 to the algorithm using long division. The binary algorithm can be
13047 used only if the modulus is odd. On 32-bit systems, it is faster
13048 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13049 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13050 up to 450 bits. In 64-bit environments, the binary algorithm
13051 appears to be advantageous for much longer moduli; here we use it
13052 for moduli up to 2048 bits.
13056 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13057 could not support the combine flag in choice fields.
13061 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13062 extensions from a certificate request to the certificate.
13066 * Allow multiple 'certopt' and 'nameopt' options to be separated
13067 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13068 file: this allows the display of the certificate about to be
13069 signed to be customised, to allow certain fields to be included
13070 or excluded and extension details. The old system didn't display
13071 multicharacter strings properly, omitted fields not in the policy
13072 and couldn't display additional details such as extensions.
13076 * Function EC_POINTs_mul for multiple scalar multiplication
13077 of an arbitrary number of elliptic curve points
13078 \sum scalars[i]*points[i],
13079 optionally including the generator defined for the EC_GROUP:
13080 scalar*generator + \sum scalars[i]*points[i].
13082 EC_POINT_mul is a simple wrapper function for the typical case
13083 that the point list has just one item (besides the optional
13088 * First EC_METHODs for curves over GF(p):
13090 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13091 operations and provides various method functions that can also
13092 operate with faster implementations of modular arithmetic.
13094 EC_GFp_mont_method() reuses most functions that are part of
13095 EC_GFp_simple_method, but uses Montgomery arithmetic.
13097 *Bodo Moeller; point addition and point doubling
13098 implementation directly derived from source code provided by
13099 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13101 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13102 crypto/ec/ec_lib.c):
13104 Curves are EC_GROUP objects (with an optional group generator)
13105 based on EC_METHODs that are built into the library.
13107 Points are EC_POINT objects based on EC_GROUP objects.
13109 Most of the framework would be able to handle curves over arbitrary
13110 finite fields, but as there are no obvious types for fields other
13111 than GF(p), some functions are limited to that for now.
13115 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13116 that the file contains a complete HTTP response.
13120 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13121 change the def and num file printf format specifier from "%-40sXXX"
13122 to "%-39s XXX". The latter will always guarantee a space after the
13123 field while the former will cause them to run together if the field
13124 is 40 of more characters long.
13128 * Constify the cipher and digest 'method' functions and structures
13129 and modify related functions to take constant EVP_MD and EVP_CIPHER
13134 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13135 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13139 * Modify `EVP_Digest*()` routines so they now return values. Although the
13140 internal software routines can never fail additional hardware versions
13145 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13147 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13148 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13151 ERR_R_NESTED_ASN1_ERROR
13153 ERR_R_MISSING_ASN1_EOS
13154 were 4 .. 9, conflicting with
13155 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13157 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13158 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13160 Add new error code 'ERR_R_INTERNAL_ERROR'.
13164 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13169 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13170 sets the subject name for a new request or supersedes the
13171 subject name in a given request. Formats that can be parsed are
13172 'CN=Some Name, OU=myOU, C=IT'
13174 'CN=Some Name/OU=myOU/C=IT'.
13176 Add options '-batch' and '-verbose' to 'openssl req'.
13178 *Massimiliano Pala <madwolf@hackmasters.net>*
13180 * Introduce the possibility to access global variables through
13181 functions on platform were that's the best way to handle exporting
13182 global variables in shared libraries. To enable this functionality,
13183 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13184 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13185 is normally done by Configure or something similar).
13187 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13188 in the source file (foo.c) like this:
13190 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13191 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13193 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13194 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13196 OPENSSL_DECLARE_GLOBAL(int,foo);
13197 #define foo OPENSSL_GLOBAL_REF(foo)
13198 OPENSSL_DECLARE_GLOBAL(double,bar);
13199 #define bar OPENSSL_GLOBAL_REF(bar)
13201 The #defines are very important, and therefore so is including the
13202 header file everywhere where the defined globals are used.
13204 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13205 of ASN.1 items, but that structure is a bit different.
13207 The largest change is in util/mkdef.pl which has been enhanced with
13208 better and easier to understand logic to choose which symbols should
13209 go into the Windows .def files as well as a number of fixes and code
13210 cleanup (among others, algorithm keywords are now sorted
13211 lexicographically to avoid constant rewrites).
13215 * In BN_div() keep a copy of the sign of 'num' before writing the
13216 result to 'rm' because if rm==num the value will be overwritten
13217 and produce the wrong result if 'num' is negative: this caused
13218 problems with BN_mod() and BN_nnmod().
13222 * Function OCSP_request_verify(). This checks the signature on an
13223 OCSP request and verifies the signer certificate. The signer
13224 certificate is just checked for a generic purpose and OCSP request
13229 * Add OCSP_check_validity() function to check the validity of OCSP
13230 responses. OCSP responses are prepared in real time and may only
13231 be a few seconds old. Simply checking that the current time lies
13232 between thisUpdate and nextUpdate max reject otherwise valid responses
13233 caused by either OCSP responder or client clock inaccuracy. Instead
13234 we allow thisUpdate and nextUpdate to fall within a certain period of
13235 the current time. The age of the response can also optionally be
13236 checked. Two new options -validity_period and -status_age added to
13241 * If signature or public key algorithm is unrecognized print out its
13242 OID rather that just UNKNOWN.
13246 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13247 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13248 ID to be generated from the issuer certificate alone which can then be
13249 passed to OCSP_id_issuer_cmp().
13253 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13254 ASN1 modules to export functions returning ASN1_ITEM pointers
13255 instead of the ASN1_ITEM structures themselves. This adds several
13256 new macros which allow the underlying ASN1 function/structure to
13257 be accessed transparently. As a result code should not use ASN1_ITEM
13258 references directly (such as &X509_it) but instead use the relevant
13259 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13260 use of the new ASN1 code on platforms where exporting structures
13261 is problematical (for example in shared libraries) but exporting
13262 functions returning pointers to structures is not.
13266 * Add support for overriding the generation of SSL/TLS session IDs.
13267 These callbacks can be registered either in an SSL_CTX or per SSL.
13268 The purpose of this is to allow applications to control, if they wish,
13269 the arbitrary values chosen for use as session IDs, particularly as it
13270 can be useful for session caching in multiple-server environments. A
13271 command-line switch for testing this (and any client code that wishes
13272 to use such a feature) has been added to "s_server".
13274 *Geoff Thorpe, Lutz Jaenicke*
13276 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13277 of the form `#if defined(...) || defined(...) || ...` and
13278 `#if !defined(...) && !defined(...) && ...`. This also avoids
13279 the growing number of special cases it was previously handling.
13283 * Make all configuration macros available for application by making
13284 sure they are available in opensslconf.h, by giving them names starting
13285 with `OPENSSL_` to avoid conflicts with other packages and by making
13286 sure e_os2.h will cover all platform-specific cases together with
13288 Additionally, it is now possible to define configuration/platform-
13289 specific names (called "system identities"). In the C code, these
13290 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13291 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13292 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13297 * New option -set_serial to 'req' and 'x509' this allows the serial
13298 number to use to be specified on the command line. Previously self
13299 signed certificates were hard coded with serial number 0 and the
13300 CA options of 'x509' had to use a serial number in a file which was
13305 * New options to 'ca' utility to support V2 CRL entry extensions.
13306 Currently CRL reason, invalidity date and hold instruction are
13307 supported. Add new CRL extensions to V3 code and some new objects.
13311 * New function EVP_CIPHER_CTX_set_padding() this is used to
13312 disable standard block padding (aka PKCS#5 padding) in the EVP
13313 API, which was previously mandatory. This means that the data is
13314 not padded in any way and so the total length much be a multiple
13315 of the block size, otherwise an error occurs.
13319 * Initial (incomplete) OCSP SSL support.
13323 * New function OCSP_parse_url(). This splits up a URL into its host,
13324 port and path components: primarily to parse OCSP URLs. New -url
13325 option to ocsp utility.
13329 * New nonce behavior. The return value of OCSP_check_nonce() now
13330 reflects the various checks performed. Applications can decide
13331 whether to tolerate certain situations such as an absent nonce
13332 in a response when one was present in a request: the ocsp application
13333 just prints out a warning. New function OCSP_add1_basic_nonce()
13334 this is to allow responders to include a nonce in a response even if
13335 the request is nonce-less.
13339 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13340 skipped when using openssl x509 multiple times on a single input file,
13341 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13345 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13346 set string type: to handle setting ASN1_TIME structures. Fix ca
13347 utility to correctly initialize revocation date of CRLs.
13351 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13352 the clients preferred ciphersuites and rather use its own preferences.
13353 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13354 Internet Explorer by ensuring unchanged hash method during stepup.
13355 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13359 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13360 to aes and add a new 'exist' option to print out symbols that don't
13365 * Additional options to ocsp utility to allow flags to be set and
13366 additional certificates supplied.
13370 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13371 OCSP client a number of certificate to only verify the response
13376 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13377 handle the new API. Currently only ECB, CBC modes supported. Add new
13380 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13381 Encryption Standard (AES) Ciphersuites for Transport Layer
13382 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13383 not enabled by default and were not part of the "ALL" ciphersuite
13384 alias because they were not yet official; they could be
13385 explicitly requested by specifying the "AESdraft" ciphersuite
13386 group alias. In the final release of OpenSSL 0.9.7, the group
13387 alias is called "AES" and is part of "ALL".)
13389 *Ben Laurie, Steve Henson, Bodo Moeller*
13391 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13392 request to response.
13396 * Functions for OCSP responders. OCSP_request_onereq_count(),
13397 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13398 extract information from a certificate request. OCSP_response_create()
13399 creates a response and optionally adds a basic response structure.
13400 OCSP_basic_add1_status() adds a complete single response to a basic
13401 response and returns the OCSP_SINGLERESP structure just added (to allow
13402 extensions to be included for example). OCSP_basic_add1_cert() adds a
13403 certificate to a basic response and OCSP_basic_sign() signs a basic
13404 response with various flags. New helper functions ASN1_TIME_check()
13405 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13406 (converts ASN1_TIME to GeneralizedTime).
13410 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13411 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13412 structure from a certificate. X509_pubkey_digest() digests the public_key
13413 contents: this is used in various key identifiers.
13417 * Make sk_sort() tolerate a NULL argument.
13419 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13421 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13422 passed by the function are trusted implicitly. If any of them signed the
13423 response then it is assumed to be valid and is not verified.
13427 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13428 to data. This was previously part of the PKCS7 ASN1 code. This
13429 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13430 *Steve Henson, reported by Kenneth R. Robinette
13431 <support@securenetterm.com>*
13433 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13434 routines: without these tracing memory leaks is very painful.
13435 Fix leaks in PKCS12 and PKCS7 routines.
13439 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13440 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13441 effectively meant GeneralizedTime would never be used. Now it
13442 is initialised to -1 but X509_time_adj() now has to check the value
13443 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13444 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13445 *Steve Henson, reported by Kenneth R. Robinette
13446 <support@securenetterm.com>*
13448 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13449 result in a zero length in the ASN1_INTEGER structure which was
13450 not consistent with the structure when d2i_ASN1_INTEGER() was used
13451 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13452 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13453 where it did not print out a minus for negative ASN1_INTEGER.
13457 * Add summary printout to ocsp utility. The various functions which
13458 convert status values to strings have been renamed to:
13459 OCSP_response_status_str(), OCSP_cert_status_str() and
13460 OCSP_crl_reason_str() and are no longer static. New options
13461 to verify nonce values and to disable verification. OCSP response
13462 printout format cleaned up.
13466 * Add additional OCSP certificate checks. These are those specified
13467 in RFC2560. This consists of two separate checks: the CA of the
13468 certificate being checked must either be the OCSP signer certificate
13469 or the issuer of the OCSP signer certificate. In the latter case the
13470 OCSP signer certificate must contain the OCSP signing extended key
13471 usage. This check is performed by attempting to match the OCSP
13472 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13473 in the OCSP_CERTID structures of the response.
13477 * Initial OCSP certificate verification added to OCSP_basic_verify()
13478 and related routines. This uses the standard OpenSSL certificate
13479 verify routines to perform initial checks (just CA validity) and
13480 to obtain the certificate chain. Then additional checks will be
13481 performed on the chain. Currently the root CA is checked to see
13482 if it is explicitly trusted for OCSP signing. This is used to set
13483 a root CA as a global signing root: that is any certificate that
13484 chains to that CA is an acceptable OCSP signing certificate.
13488 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13489 extensions from a separate configuration file.
13490 As when reading extensions from the main configuration file,
13491 the '-extensions ...' option may be used for specifying the
13494 *Massimiliano Pala <madwolf@comune.modena.it>*
13496 * New OCSP utility. Allows OCSP requests to be generated or
13497 read. The request can be sent to a responder and the output
13498 parsed, outputted or printed in text form. Not complete yet:
13499 still needs to check the OCSP response validity.
13503 * New subcommands for 'openssl ca':
13504 `openssl ca -status <serial>` prints the status of the cert with
13505 the given serial number (according to the index file).
13506 `openssl ca -updatedb` updates the expiry status of certificates
13509 *Massimiliano Pala <madwolf@comune.modena.it>*
13511 * New '-newreq-nodes' command option to CA.pl. This is like
13512 '-newreq', but calls 'openssl req' with the '-nodes' option
13513 so that the resulting key is not encrypted.
13515 *Damien Miller <djm@mindrot.org>*
13517 * New configuration for the GNU Hurd.
13519 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13521 * Initial code to implement OCSP basic response verify. This
13522 is currently incomplete. Currently just finds the signer's
13523 certificate and verifies the signature on the response.
13527 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13528 value of OPENSSLDIR. This is available via the new '-d' option
13529 to 'openssl version', and is also included in 'openssl version -a'.
13533 * Allowing defining memory allocation callbacks that will be given
13534 file name and line number information in additional arguments
13535 (a `const char*` and an int). The basic functionality remains, as
13536 well as the original possibility to just replace malloc(),
13537 realloc() and free() by functions that do not know about these
13538 additional arguments. To register and find out the current
13539 settings for extended allocation functions, the following
13540 functions are provided:
13542 CRYPTO_set_mem_ex_functions
13543 CRYPTO_set_locked_mem_ex_functions
13544 CRYPTO_get_mem_ex_functions
13545 CRYPTO_get_locked_mem_ex_functions
13547 These work the same way as CRYPTO_set_mem_functions and friends.
13548 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13549 extended allocation function is enabled.
13550 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13551 a conventional allocation function is enabled.
13553 *Richard Levitte, Bodo Moeller*
13555 * Finish off removing the remaining LHASH function pointer casts.
13556 There should no longer be any prototype-casting required when using
13557 the LHASH abstraction, and any casts that remain are "bugs". See
13558 the callback types and macros at the head of lhash.h for details
13559 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13563 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13564 If /dev/[u]random devices are not available or do not return enough
13565 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13567 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13568 /etc/entropy will be queried once each in this sequence, querying stops
13569 when enough entropy was collected without querying more sockets.
13573 * Change the Unix RAND_poll() variant to be able to poll several
13574 random devices, as specified by DEVRANDOM, until a sufficient amount
13575 of data has been collected. We spend at most 10 ms on each file
13576 (select timeout) and read in non-blocking mode. DEVRANDOM now
13577 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13578 (previously it was just the string "/dev/urandom"), so on typical
13579 platforms the 10 ms delay will never occur.
13580 Also separate out the Unix variant to its own file, rand_unix.c.
13581 For VMS, there's a currently-empty rand_vms.c.
13585 * Move OCSP client related routines to ocsp_cl.c. These
13586 provide utility functions which an application needing
13587 to issue a request to an OCSP responder and analyse the
13588 response will typically need: as opposed to those which an
13589 OCSP responder itself would need which will be added later.
13591 OCSP_request_sign() signs an OCSP request with an API similar
13592 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13593 response. OCSP_response_get1_basic() extracts basic response
13594 from response. OCSP_resp_find_status(): finds and extracts status
13595 information from an OCSP_CERTID structure (which will be created
13596 when the request structure is built). These are built from lower
13597 level functions which work on OCSP_SINGLERESP structures but
13598 won't normally be used unless the application wishes to examine
13599 extensions in the OCSP response for example.
13601 Replace nonce routines with a pair of functions.
13602 OCSP_request_add1_nonce() adds a nonce value and optionally
13603 generates a random value. OCSP_check_nonce() checks the
13604 validity of the nonce in an OCSP response.
13608 * Change function OCSP_request_add() to OCSP_request_add0_id().
13609 This doesn't copy the supplied OCSP_CERTID and avoids the
13610 need to free up the newly created id. Change return type
13611 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13612 This can then be used to add extensions to the request.
13613 Deleted OCSP_request_new(), since most of its functionality
13614 is now in OCSP_REQUEST_new() (and the case insensitive name
13615 clash) apart from the ability to set the request name which
13616 will be added elsewhere.
13620 * Update OCSP API. Remove obsolete extensions argument from
13621 various functions. Extensions are now handled using the new
13622 OCSP extension code. New simple OCSP HTTP function which
13623 can be used to send requests and parse the response.
13627 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13628 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13629 uses the special reorder version of SET OF to sort the attributes
13630 and reorder them to match the encoded order. This resolves a long
13631 standing problem: a verify on a PKCS7 structure just after signing
13632 it used to fail because the attribute order did not match the
13633 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13634 it uses the received order. This is necessary to tolerate some broken
13635 software that does not order SET OF. This is handled by encoding
13636 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13637 to produce the required SET OF.
13641 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13642 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13643 files to get correct declarations of the ASN.1 item variables.
13647 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13648 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13649 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13650 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13651 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13652 ASN1_ITEM and no wrapper functions.
13656 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13657 replace the old function pointer based I/O routines. Change most of
13658 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13662 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13663 lines, recognize more "algorithms" that can be deselected, and make
13664 it complain about algorithm deselection that isn't recognised.
13668 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13669 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13670 to use new functions. Add NO_ASN1_OLD which can be set to remove
13671 some old style ASN1 functions: this can be used to determine if old
13672 code will still work when these eventually go away.
13676 * New extension functions for OCSP structures, these follow the
13677 same conventions as certificates and CRLs.
13681 * New function X509V3_add1_i2d(). This automatically encodes and
13682 adds an extension. Its behaviour can be customised with various
13683 flags to append, replace or delete. Various wrappers added for
13684 certificates and CRLs.
13688 * Fix to avoid calling the underlying ASN1 print routine when
13689 an extension cannot be parsed. Correct a typo in the
13690 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13694 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13695 entries for variables.
13699 * Add functionality to `apps/openssl.c` for detecting locking
13700 problems: As the program is single-threaded, all we have
13701 to do is register a locking callback using an array for
13702 storing which locks are currently held by the program.
13706 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13707 SSL_get_ex_data_X509_STORE_idx(), which is used in
13708 ssl_verify_cert_chain() and thus can be called at any time
13709 during TLS/SSL handshakes so that thread-safety is essential.
13710 Unfortunately, the ex_data design is not at all suited
13711 for multi-threaded use, so it probably should be abolished.
13715 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13717 *Broadcom, tweaked and integrated by Geoff Thorpe*
13719 * Move common extension printing code to new function
13720 X509V3_print_extensions(). Reorganise OCSP print routines and
13721 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13725 * New function X509_signature_print() to remove duplication in some
13730 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13731 set (this was treated exactly the same as SET OF previously). This
13732 is used to reorder the STACK representing the structure to match the
13733 encoding. This will be used to get round a problem where a PKCS7
13734 structure which was signed could not be verified because the STACK
13735 order did not reflect the encoded order.
13739 * Reimplement the OCSP ASN1 module using the new code.
13743 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13744 for its ASN1 operations. The old style function pointers still exist
13745 for now but they will eventually go away.
13749 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13750 completely replaces the old ASN1 functionality with a table driven
13751 encoder and decoder which interprets an ASN1_ITEM structure describing
13752 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13753 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13754 has also been converted to the new form.
13758 * Change BN_mod_exp_recp so that negative moduli are tolerated
13759 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13760 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13761 for negative moduli.
13765 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13766 of not touching the result's sign bit.
13770 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13775 * Changed the LHASH code to use prototypes for callbacks, and created
13776 macros to declare and implement thin (optionally static) functions
13777 that provide type-safety and avoid function pointer casting for the
13778 type-specific callbacks.
13782 * Added Kerberos Cipher Suites to be used with TLS, as written in
13784 *Veers Staats <staatsvr@asc.hpc.mil>,
13785 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13787 * Reformat the FAQ so the different questions and answers can be divided
13788 in sections depending on the subject.
13792 * Have the zlib compression code load ZLIB.DLL dynamically under
13797 * New function BN_mod_sqrt for computing square roots modulo a prime
13798 (using the probabilistic Tonelli-Shanks algorithm unless
13799 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13800 be handled deterministically).
13802 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13804 * Make BN_mod_inverse faster by explicitly handling small quotients
13805 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13806 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13810 * New function BN_kronecker.
13814 * Fix BN_gcd so that it works on negative inputs; the result is
13815 positive unless both parameters are zero.
13816 Previously something reasonably close to an infinite loop was
13817 possible because numbers could be growing instead of shrinking
13818 in the implementation of Euclid's algorithm.
13822 * Fix BN_is_word() and BN_is_one() macros to take into account the
13823 sign of the number in question.
13825 Fix BN_is_word(a,w) to work correctly for w == 0.
13827 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13828 because its test if the absolute value of 'a' equals 'w'.
13829 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13830 it exists mostly for use in the implementations of BN_is_zero(),
13831 BN_is_one(), and BN_is_word().
13835 * New function BN_swap.
13839 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13840 the exponentiation functions are more likely to produce reasonable
13841 results on negative inputs.
13845 * Change BN_mod_mul so that the result is always non-negative.
13846 Previously, it could be negative if one of the factors was negative;
13847 I don't think anyone really wanted that behaviour.
13851 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13852 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13853 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13854 and add new functions:
13863 BN_mod_lshift1_quick
13865 BN_mod_lshift_quick
13867 These functions always generate non-negative results.
13869 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13870 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13872 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13873 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13874 be reduced modulo `m`.
13876 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13879 The following entry accidentally appeared in the CHANGES file
13880 distributed with OpenSSL 0.9.7. The modifications described in
13881 it do *not* apply to OpenSSL 0.9.7.
13883 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13884 was actually never needed) and in BN_mul(). The removal in BN_mul()
13885 required a small change in bn_mul_part_recursive() and the addition
13886 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13887 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13888 bn_sub_words() and bn_add_words() except they take arrays with
13894 * In 'openssl passwd', verify passwords read from the terminal
13895 unless the '-salt' option is used (which usually means that
13896 verification would just waste user's time since the resulting
13897 hash is going to be compared with some given password hash)
13898 or the new '-noverify' option is used.
13900 This is an incompatible change, but it does not affect
13901 non-interactive use of 'openssl passwd' (passwords on the command
13902 line, '-stdin' option, '-in ...' option) and thus should not
13903 cause any problems.
13907 * Remove all references to RSAref, since there's no more need for it.
13911 * Make DSO load along a path given through an environment variable
13912 (SHLIB_PATH) with shl_load().
13916 * Constify the ENGINE code as a result of BIGNUM constification.
13917 Also constify the RSA code and most things related to it. In a
13918 few places, most notable in the depth of the ASN.1 code, ugly
13919 casts back to non-const were required (to be solved at a later
13924 * Make it so the openssl application has all engines loaded by default.
13928 * Constify the BIGNUM routines a little more.
13932 * Add the following functions:
13934 ENGINE_load_cswift()
13936 ENGINE_load_atalla()
13937 ENGINE_load_nuron()
13938 ENGINE_load_builtin_engines()
13940 That way, an application can itself choose if external engines that
13941 are built-in in OpenSSL shall ever be used or not. The benefit is
13942 that applications won't have to be linked with libdl or other dso
13943 libraries unless it's really needed.
13945 Changed 'openssl engine' to load all engines on demand.
13946 Changed the engine header files to avoid the duplication of some
13947 declarations (they differed!).
13951 * 'openssl engine' can now list capabilities.
13955 * Better error reporting in 'openssl engine'.
13959 * Never call load_dh_param(NULL) in s_server.
13963 * Add engine application. It can currently list engines by name and
13964 identity, and test if they are actually available.
13968 * Improve RPM specification file by forcing symbolic linking and making
13969 sure the installed documentation is also owned by root.root.
13971 *Damien Miller <djm@mindrot.org>*
13973 * Give the OpenSSL applications more possibilities to make use of
13974 keys (public as well as private) handled by engines.
13978 * Add OCSP code that comes from CertCo.
13982 * Add VMS support for the Rijndael code.
13986 * Added untested support for Nuron crypto accelerator.
13990 * Add support for external cryptographic devices. This code was
13991 previously distributed separately as the "engine" branch.
13993 *Geoff Thorpe, Richard Levitte*
13995 * Rework the filename-translation in the DSO code. It is now possible to
13996 have far greater control over how a "name" is turned into a filename
13997 depending on the operating environment and any oddities about the
13998 different shared library filenames on each system.
14002 * Support threads on FreeBSD-elf in Configure.
14006 * Fix for SHA1 assembly problem with MASM: it produces
14007 warnings about corrupt line number information when assembling
14008 with debugging information. This is caused by the overlapping
14011 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14014 NCONF_get_number() has no error checking at all. As a replacement,
14015 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14016 promoted strongly. The old NCONF_get_number is kept around for
14017 binary backward compatibility.
14018 Make it possible for methods to load from something other than a BIO,
14019 by providing a function pointer that is given a name instead of a BIO.
14020 For example, this could be used to load configuration data from an
14025 * Fix for non blocking accept BIOs. Added new I/O special reason
14026 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14027 with non blocking I/O was not possible because no retry code was
14028 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14033 * Added the beginnings of Rijndael support.
14037 * Fix for bug in DirectoryString mask setting. Add support for
14038 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14039 to allow certificate printing to more controllable, additional
14040 'certopt' option to 'x509' to allow new printing options to be
14045 * Clean old EAY MD5 hack from e_os.h.
14049 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14051 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14052 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14054 *Joe Orton, Steve Henson*
14056 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14058 * Fix additional bug revealed by the NISCC test suite:
14060 Stop bug triggering large recursion when presented with
14061 certain ASN.1 tags ([CVE-2003-0851])
14065 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14067 * Fix various bugs revealed by running the NISCC test suite:
14069 Stop out of bounds reads in the ASN1 code when presented with
14070 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14072 If verify callback ignores invalid public key errors don't try to check
14073 certificate signature with the NULL public key.
14077 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14078 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14083 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14084 extra data after the compression methods not only for TLS 1.0
14085 but also for SSL 3.0 (as required by the specification).
14087 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14089 * Change X509_certificate_type() to mark the key as exported/exportable
14090 when it's 512 *bits* long, not 512 bytes.
14094 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14096 * Countermeasure against the Klima-Pokorny-Rosa extension of
14097 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14098 a protocol version number mismatch like a decryption error
14099 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14103 * Turn on RSA blinding by default in the default implementation
14104 to avoid a timing attack. Applications that don't want it can call
14105 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14106 They would be ill-advised to do so in most cases.
14108 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14110 * Change RSA blinding code so that it works when the PRNG is not
14111 seeded (in this case, the secret RSA exponent is abused as
14112 an unpredictable seed -- if it is not unpredictable, there
14113 is no point in blinding anyway). Make RSA blinding thread-safe
14114 by remembering the creator's thread ID in rsa->blinding and
14115 having all other threads use local one-time blinding factors
14116 (this requires more computation than sharing rsa->blinding, but
14117 avoids excessive locking; and if an RSA object is not shared
14118 between threads, blinding will still be very fast).
14122 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14124 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14125 via timing by performing a MAC computation even if incorrect
14126 block cipher padding has been found. This is a countermeasure
14127 against active attacks where the attacker has to distinguish
14128 between bad padding and a MAC verification error. ([CVE-2003-0078])
14130 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14131 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14132 Martin Vuagnoux (EPFL, Ilion)*
14134 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14136 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14137 memory from its contents. This is done with a counter that will
14138 place alternating values in each byte. This can be used to solve
14139 two issues: 1) the removal of calls to memset() by highly optimizing
14140 compilers, and 2) cleansing with other values than 0, since those can
14141 be read through on certain media, for example a swap space on disk.
14145 * Bugfix: client side session caching did not work with external caching,
14146 because the session->cipher setting was not restored when reloading
14147 from the external cache. This problem was masked, when
14148 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14149 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14153 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14154 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14156 *Zeev Lieber <zeev-l@yahoo.com>*
14158 * Undo an undocumented change introduced in 0.9.6e which caused
14159 repeated calls to OpenSSL_add_all_ciphers() and
14160 OpenSSL_add_all_digests() to be ignored, even after calling
14165 * Change the default configuration reader to deal with last line not
14166 being properly terminated.
14170 * Change X509_NAME_cmp() so it applies the special rules on handling
14171 DN values that are of type PrintableString, as well as RDNs of type
14172 emailAddress where the value has the type ia5String.
14174 *stefank@valicert.com via Richard Levitte*
14176 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14177 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14178 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14179 the bitwise-OR of the two for use by the majority of applications
14180 wanting this behaviour, and update the docs. The documented
14181 behaviour and actual behaviour were inconsistent and had been
14182 changing anyway, so this is more a bug-fix than a behavioural
14185 *Geoff Thorpe, diagnosed by Nadav Har'El*
14187 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14188 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14192 * Fix initialization code race conditions in
14193 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14194 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14195 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14196 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14197 ssl2_get_cipher_by_char(),
14198 ssl3_get_cipher_by_char().
14200 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14202 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14203 the cached sessions are flushed, as the remove_cb() might use ex_data
14204 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14205 (see [openssl.org #212]).
14207 *Geoff Thorpe, Lutz Jaenicke*
14209 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14210 length, instead of the encoding length to d2i_ASN1_OBJECT.
14214 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14216 * [In 0.9.6g-engine release:]
14217 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14219 *Lynn Gazis <lgazis@rainbow.com>*
14221 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14223 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14224 and get fix the header length calculation.
14225 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14226 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14228 * Use proper error handling instead of 'assertions' in buffer
14229 overflow checks added in 0.9.6e. This prevents DoS (the
14230 assertions could call abort()).
14232 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14234 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14236 * Add various sanity checks to asn1_get_length() to reject
14237 the ASN1 length bytes if they exceed sizeof(long), will appear
14238 negative or the content length exceeds the length of the
14241 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14243 * Fix cipher selection routines: ciphers without encryption had no flags
14244 for the cipher strength set and where therefore not handled correctly
14245 by the selection routines (PR #130).
14249 * Fix EVP_dsa_sha macro.
14254 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14255 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14256 that was added in OpenSSL 0.9.6d.
14258 As the countermeasure turned out to be incompatible with some
14259 broken SSL implementations, the new option is part of SSL_OP_ALL.
14260 SSL_OP_ALL is usually employed when compatibility with weird SSL
14261 implementations is desired (e.g. '-bugs' option to 's_client' and
14262 's_server'), so the new option is automatically set in many
14267 * Changes in security patch:
14269 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14270 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14271 Air Force Materiel Command, USAF, under agreement number
14274 * Add various sanity checks to asn1_get_length() to reject
14275 the ASN1 length bytes if they exceed sizeof(long), will appear
14276 negative or the content length exceeds the length of the
14277 supplied buffer. ([CVE-2002-0659])
14279 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14281 * Assertions for various potential buffer overflows, not known to
14282 happen in practice.
14284 *Ben Laurie (CHATS)*
14286 * Various temporary buffers to hold ASCII versions of integers were
14287 too small for 64 bit platforms. ([CVE-2002-0655])
14288 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14290 * Remote buffer overflow in SSL3 protocol - an attacker could
14291 supply an oversized session ID to a client. ([CVE-2002-0656])
14293 *Ben Laurie (CHATS)*
14295 * Remote buffer overflow in SSL2 protocol - an attacker could
14296 supply an oversized client master key. ([CVE-2002-0656])
14298 *Ben Laurie (CHATS)*
14300 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14302 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14303 encoded as NULL) with id-dsa-with-sha1.
14305 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14307 * Check various `X509_...()` return values in `apps/req.c`.
14309 *Nils Larsch <nla@trustcenter.de>*
14311 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14312 an end-of-file condition would erroneously be flagged, when the CRLF
14313 was just at the end of a processed block. The bug was discovered when
14314 processing data through a buffering memory BIO handing the data to a
14315 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14316 <ptsekov@syntrex.com> and Nedelcho Stanev.
14320 * Implement a countermeasure against a vulnerability recently found
14321 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14322 before application data chunks to avoid the use of known IVs
14323 with data potentially chosen by the attacker.
14327 * Fix length checks in ssl3_get_client_hello().
14331 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14332 to prevent ssl3_read_internal() from incorrectly assuming that
14333 ssl3_read_bytes() found application data while handshake
14334 processing was enabled when in fact s->s3->in_read_app_data was
14335 merely automatically cleared during the initial handshake.
14337 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14339 * Fix object definitions for Private and Enterprise: they were not
14340 recognized in their shortname (=lowercase) representation. Extend
14341 obj_dat.pl to issue an error when using undefined keywords instead
14342 of silently ignoring the problem (Svenning Sorensen
14343 <sss@sss.dnsalias.net>).
14347 * Fix DH_generate_parameters() so that it works for 'non-standard'
14348 generators, i.e. generators other than 2 and 5. (Previously, the
14349 code did not properly initialise the 'add' and 'rem' values to
14350 BN_generate_prime().)
14352 In the new general case, we do not insist that 'generator' is
14353 actually a primitive root: This requirement is rather pointless;
14354 a generator of the order-q subgroup is just as good, if not
14359 * Map new X509 verification errors to alerts. Discovered and submitted by
14360 Tom Wu <tom@arcot.com>.
14364 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14365 returning non-zero before the data has been completely received
14366 when using non-blocking I/O.
14368 *Bodo Moeller; problem pointed out by John Hughes*
14370 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14372 *Ben Laurie, Lutz Jaenicke*
14374 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14375 Yoram Zahavi <YoramZ@gilian.com>).
14379 * Add information about CygWin 1.3 and on, and preserve proper
14380 configuration for the versions before that.
14382 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14384 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14385 check whether we deal with a copy of a session and do not delete from
14386 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14387 <izhar@checkpoint.com>.
14391 * Do not store session data into the internal session cache, if it
14392 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14393 flag is set). Proposed by Aslam <aslam@funk.com>.
14397 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14402 * [In 0.9.6d-engine release:]
14403 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14405 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14407 * Add the configuration target linux-s390x.
14409 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14411 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14412 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14413 variable as an indication that a ClientHello message has been
14414 received. As the flag value will be lost between multiple
14415 invocations of ssl3_accept when using non-blocking I/O, the
14416 function may not be aware that a handshake has actually taken
14417 place, thus preventing a new session from being added to the
14420 To avoid this problem, we now set s->new_session to 2 instead of
14421 using a local variable.
14423 *Lutz Jaenicke, Bodo Moeller*
14425 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14426 if the SSL_R_LENGTH_MISMATCH error is detected.
14428 *Geoff Thorpe, Bodo Moeller*
14430 * New 'shared_ldflag' column in Configure platform table.
14434 * Fix EVP_CIPHER_mode macro.
14436 *"Dan S. Camper" <dan@bti.net>*
14438 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14439 type, we must throw them away by setting rr->length to 0.
14441 *D P Chang <dpc@qualys.com>*
14443 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14445 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14446 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14447 worked incorrectly for those cases where range = `10..._2` and
14448 `3*range` is two bits longer than range.)
14452 * Only add signing time to PKCS7 structures if it is not already
14457 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14458 OBJ_ld_ce should be OBJ_id_ce.
14459 Also some ip-pda OIDs in crypto/objects/objects.txt were
14460 incorrect (cf. RFC 3039).
14462 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14464 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14465 returns early because it has nothing to do.
14467 *Andy Schneider <andy.schneider@bjss.co.uk>*
14469 * [In 0.9.6c-engine release:]
14470 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14472 *Andy Schneider <andy.schneider@bjss.co.uk>*
14474 * [In 0.9.6c-engine release:]
14475 Add support for Cryptographic Appliance's keyserver technology.
14476 (Use engine 'keyclient')
14478 *Cryptographic Appliances and Geoff Thorpe*
14480 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14481 is called via tools/c89.sh because arguments have to be
14482 rearranged (all '-L' options must appear before the first object
14485 *Richard Shapiro <rshapiro@abinitio.com>*
14487 * [In 0.9.6c-engine release:]
14488 Add support for Broadcom crypto accelerator cards, backported
14491 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14493 * [In 0.9.6c-engine release:]
14494 Add support for SureWare crypto accelerator cards from
14495 Baltimore Technologies. (Use engine 'sureware')
14497 *Baltimore Technologies and Mark Cox*
14499 * [In 0.9.6c-engine release:]
14500 Add support for crypto accelerator cards from Accelerated
14501 Encryption Processing, www.aep.ie. (Use engine 'aep')
14503 *AEP Inc. and Mark Cox*
14505 * Add a configuration entry for gcc on UnixWare.
14507 *Gary Benson <gbenson@redhat.com>*
14509 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14510 messages are stored in a single piece (fixed-length part and
14511 variable-length part combined) and fix various bugs found on the way.
14515 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14516 instead. BIO_gethostbyname() does not know what timeouts are
14517 appropriate, so entries would stay in cache even when they have
14519 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14521 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14522 faced with a pathologically small ClientHello fragment that does
14523 not contain client_version: Instead of aborting with an error,
14524 simply choose the highest available protocol version (i.e.,
14525 TLS 1.0 unless it is disabled). In practice, ClientHello
14526 messages are never sent like this, but this change gives us
14527 strictly correct behaviour at least for TLS.
14531 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14532 never resets s->method to s->ctx->method when called from within
14533 one of the SSL handshake functions.
14535 *Bodo Moeller; problem pointed out by Niko Baric*
14537 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14538 (sent using the client's version number) if client_version is
14539 smaller than the protocol version in use. Also change
14540 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14541 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14542 the client will at least see that alert.
14546 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14551 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14552 client receives HelloRequest while in a handshake.
14554 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14556 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14557 should end in 'break', not 'goto end' which circumvents various
14558 cleanups done in state SSL_ST_OK. But session related stuff
14559 must be disabled for SSL_ST_OK in the case that we just sent a
14562 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14563 before just sending a HelloRequest.
14565 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14567 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14568 reveal whether illegal block cipher padding was found or a MAC
14569 verification error occurred. (Neither SSLerr() codes nor alerts
14570 are directly visible to potential attackers, but the information
14571 may leak via logfiles.)
14573 Similar changes are not required for the SSL 2.0 implementation
14574 because the number of padding bytes is sent in clear for SSL 2.0,
14575 and the extra bytes are just ignored. However ssl/s2_pkt.c
14576 failed to verify that the purported number of padding bytes is in
14581 * Add OpenUNIX-8 support including shared libraries
14582 (Boyd Lynn Gerber <gerberb@zenez.com>).
14586 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14587 'wristwatch attack' using huge encoding parameters (cf.
14588 James H. Manger's CRYPTO 2001 paper). Note that the
14589 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14590 encoding parameters and hence was not vulnerable.
14594 * BN_sqr() bug fix.
14596 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14598 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14599 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14600 followed by modular reduction.
14602 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14604 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14605 equivalent based on BN_pseudo_rand() instead of BN_rand().
14609 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14610 This function was broken, as the check for a new client hello message
14611 to handle SGC did not allow these large messages.
14612 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14616 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14620 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14621 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14625 * Rework the configuration and shared library support for Tru64 Unix.
14626 The configuration part makes use of modern compiler features and
14627 still retains old compiler behavior for those that run older versions
14628 of the OS. The shared library support part includes a variant that
14629 uses the RPATH feature, and is available through the special
14630 configuration target "alpha-cc-rpath", which will never be selected
14633 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14635 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14636 with the same message size as in ssl3_get_certificate_request().
14637 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14638 messages might inadvertently be reject as too long.
14640 *Petr Lampa <lampa@fee.vutbr.cz>*
14642 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14646 * Modified SSL library such that the verify_callback that has been set
14647 specifically for an SSL object with SSL_set_verify() is actually being
14648 used. Before the change, a verify_callback set with this function was
14649 ignored and the verify_callback() set in the SSL_CTX at the time of
14650 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14651 to allow the necessary settings.
14655 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14656 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14657 done automatically (in contradiction to the requirements of the C
14658 standard). This made problems when used from OpenSSH.
14662 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14663 dh->length and always used
14665 BN_rand_range(priv_key, dh->p).
14667 BN_rand_range() is not necessary for Diffie-Hellman, and this
14668 specific range makes Diffie-Hellman unnecessarily inefficient if
14669 dh->length (recommended exponent length) is much smaller than the
14670 length of dh->p. We could use BN_rand_range() if the order of
14671 the subgroup was stored in the DH structure, but we only have
14676 BN_rand(priv_key, l, ...)
14678 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14685 RSA_eay_public_encrypt
14686 RSA_eay_private_decrypt
14687 RSA_eay_private_encrypt (signing)
14688 RSA_eay_public_decrypt (signature verification)
14690 (default implementations for RSA_public_encrypt,
14691 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14692 always reject numbers >= n.
14696 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14697 to synchronize access to 'locking_thread'. This is necessary on
14698 systems where access to 'locking_thread' (an 'unsigned long'
14699 variable) is not atomic.
14703 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14704 *before* setting the 'crypto_lock_rand' flag. The previous code had
14705 a race condition if 0 is a valid thread ID.
14707 *Travis Vitek <vitek@roguewave.com>*
14709 * Add support for shared libraries under Irix.
14711 *Albert Chin-A-Young <china@thewrittenword.com>*
14713 * Add configuration option to build on Linux on both big-endian and
14714 little-endian MIPS.
14716 *Ralf Baechle <ralf@uni-koblenz.de>*
14718 * Add the possibility to create shared libraries on HP-UX.
14722 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14724 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14725 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14726 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14727 PRNG state recovery was possible based on the output of
14728 one PRNG request appropriately sized to gain knowledge on
14729 'md' followed by enough consecutive 1-byte PRNG requests
14730 to traverse all of 'state'.
14732 1. When updating 'md_local' (the current thread's copy of 'md')
14733 during PRNG output generation, hash all of the previous
14734 'md_local' value, not just the half used for PRNG output.
14736 2. Make the number of bytes from 'state' included into the hash
14737 independent from the number of PRNG bytes requested.
14739 The first measure alone would be sufficient to avoid
14740 Markku-Juhani's attack. (Actually it had never occurred
14741 to me that the half of 'md_local' used for chaining was the
14742 half from which PRNG output bytes were taken -- I had always
14743 assumed that the secret half would be used.) The second
14744 measure makes sure that additional data from 'state' is never
14745 mixed into 'md_local' in small portions; this heuristically
14746 further strengthens the PRNG.
14750 * Fix crypto/bn/asm/mips3.s.
14754 * When only the key is given to "enc", the IV is undefined. Print out
14755 an error message in this case.
14759 * Handle special case when X509_NAME is empty in X509 printing routines.
14763 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14764 positive and less than q.
14768 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14769 used: it isn't thread safe and the add_lock_callback should handle
14772 *Paul Rose <Paul.Rose@bridge.com>*
14774 * Verify that incoming data obeys the block size in
14775 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14781 *Ulf Möller, Bodo Möller*
14783 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14784 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14785 when fixing the server behaviour for backwards-compatible 'client
14786 hello' messages. (Note that the attack is impractical against
14787 SSL 3.0 and TLS 1.0 anyway because length and version checking
14788 means that the probability of guessing a valid ciphertext is
14789 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14792 Before 0.9.5, the countermeasure (hide the error by generating a
14793 random 'decryption result') did not work properly because
14794 ERR_clear_error() was missing, meaning that SSL_get_error() would
14795 detect the supposedly ignored error.
14797 Both problems are now fixed.
14801 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14802 (previously it was 1024).
14806 * Fix for compatibility mode trust settings: ignore trust settings
14807 unless some valid trust or reject settings are present.
14811 * Fix for blowfish EVP: its a variable length cipher.
14815 * Fix various bugs related to DSA S/MIME verification. Handle missing
14816 parameters in DSA public key structures and return an error in the
14817 DSA routines if parameters are absent.
14821 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14822 in the current directory if neither $RANDFILE nor $HOME was set.
14823 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14824 caused some confusion to Windows users who haven't defined $HOME.
14825 Thus RAND_file_name() is changed again: e_os.h can define a
14826 DEFAULT_HOME, which will be used if $HOME is not set.
14827 For Windows, we use "C:"; on other platforms, we still require
14828 environment variables.
14830 * Move 'if (!initialized) RAND_poll()' into regions protected by
14831 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14832 having multiple threads call RAND_poll() concurrently.
14836 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14837 combination of a flag and a thread ID variable.
14838 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14839 flag), *other* threads can enter ssleay_add_bytes without obeying
14840 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14841 that they do not hold after the first thread unsets add_do_not_lock).
14845 * Change bctest again: '-x' expressions are not available in all
14846 versions of 'test'.
14850 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14852 * Fix a couple of memory leaks in PKCS7_dataDecode()
14854 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14856 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14857 the default extension for executables, if any. Also, make the perl
14858 scripts that use symlink() to test if it really exists and use "cp"
14859 if it doesn't. All this made OpenSSL compilable and installable in
14864 * Fix for asn1_GetSequence() for indefinite length constructed data.
14865 If SEQUENCE is length is indefinite just set c->slen to the total
14866 amount of data available.
14868 *Steve Henson, reported by shige@FreeBSD.org*
14870 *This change does not apply to 0.9.7.*
14872 * Change bctest to avoid here-documents inside command substitution
14873 (workaround for FreeBSD /bin/sh bug).
14874 For compatibility with Ultrix, avoid shell functions (introduced
14875 in the bctest version that searches along $PATH).
14879 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14880 with des_encrypt() defined on some operating systems, like Solaris
14885 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14886 On the Importance of Eliminating Errors in Cryptographic
14887 Computations, J. Cryptology 14 (2001) 2, 101-119,
14888 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14892 * MIPS assembler BIGNUM division bug fix.
14896 * Disabled incorrect Alpha assembler code.
14900 * Fix PKCS#7 decode routines so they correctly update the length
14901 after reading an EOC for the EXPLICIT tag.
14905 *This change does not apply to 0.9.7.*
14907 * Fix bug in PKCS#12 key generation routines. This was triggered
14908 if a 3DES key was generated with a 0 initial byte. Include
14909 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14910 (but broken) behaviour.
14914 * Enhance bctest to search for a working bc along $PATH and print
14917 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14919 * Fix memory leaks in err.c: free err_data string if necessary;
14920 don't write to the wrong index in ERR_set_error_data.
14924 * Implement ssl23_peek (analogous to ssl23_read), which previously
14929 * Replace rdtsc with `_emit` statements for VC++ version 5.
14931 *Jeremy Cooper <jeremy@baymoo.org>*
14933 * Make it possible to reuse SSLv2 sessions.
14937 * In copy_email() check for >= 0 as a return value for
14938 X509_NAME_get_index_by_NID() since 0 is a valid index.
14940 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14942 * Avoid coredump with unsupported or invalid public keys by checking if
14943 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14944 PKCS7_verify() fails with non detached data.
14948 * Don't use getenv in library functions when run as setuid/setgid.
14949 New function OPENSSL_issetugid().
14953 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14954 due to incorrect handling of multi-threading:
14956 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14958 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14960 3. Count how many times MemCheck_off() has been called so that
14961 nested use can be treated correctly. This also avoids
14962 inband-signalling in the previous code (which relied on the
14963 assumption that thread ID 0 is impossible).
14967 * Add "-rand" option also to s_client and s_server.
14971 * Fix CPU detection on Irix 6.x.
14972 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14973 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14975 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14980 *This change does not apply to 0.9.7.*
14982 * Use the cached encoding of an X509_NAME structure rather than
14983 copying it. This is apparently the reason for the libsafe "errors"
14984 but the code is actually correct.
14988 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14989 Bleichenbacher's DSA attack.
14990 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14991 to be set and top=0 forces the highest bit to be set; top=-1 is new
14992 and leaves the highest bit random.
14994 *Ulf Moeller, Bodo Moeller*
14996 * In the `NCONF_...`-based implementations for `CONF_...` queries
14997 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14998 a temporary CONF structure with the data component set to NULL
14999 (which gives segmentation faults in lh_retrieve).
15000 Instead, use NULL for the CONF pointer in CONF_get_string and
15001 CONF_get_number (which may use environment variables) and directly
15002 return NULL from CONF_get_section.
15006 * Fix potential buffer overrun for EBCDIC.
15010 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15011 keyUsage if basicConstraints absent for a CA.
15015 * Make SMIME_write_PKCS7() write mail header values with a format that
15016 is more generally accepted (no spaces before the semicolon), since
15017 some programs can't parse those values properly otherwise. Also make
15018 sure BIO's that break lines after each write do not create invalid
15023 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15024 macros previously used would not encode an empty SEQUENCE OF
15025 and break the signature.
15029 *This change does not apply to 0.9.7.*
15031 * Zero the premaster secret after deriving the master secret in
15036 * Add some EVP_add_digest_alias registrations (as found in
15037 OpenSSL_add_all_digests()) to SSL_library_init()
15038 aka OpenSSL_add_ssl_algorithms(). This provides improved
15039 compatibility with peers using X.509 certificates
15040 with unconventional AlgorithmIdentifier OIDs.
15044 * Fix for Irix with NO_ASM.
15046 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15048 * ./config script fixes.
15050 *Ulf Moeller, Richard Levitte*
15052 * Fix 'openssl passwd -1'.
15056 * Change PKCS12_key_gen_asc() so it can cope with non null
15057 terminated strings whose length is passed in the passlen
15058 parameter, for example from PEM callbacks. This was done
15059 by adding an extra length parameter to asc2uni().
15061 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15063 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15064 call failed, free the DSA structure.
15068 * Fix to uni2asc() to cope with zero length Unicode strings.
15069 These are present in some PKCS#12 files.
15073 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15074 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15075 when writing a 32767 byte record.
15077 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15079 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15080 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15082 (RSA objects have a reference count access to which is protected
15083 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15084 so they are meant to be shared between threads.)
15085 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15086 "Reddie, Steven" <Steven.Reddie@ca.com>*
15088 * Fix a deadlock in CRYPTO_mem_leaks().
15092 * Use better test patterns in bntest.
15096 * rand_win.c fix for Borland C.
15100 * BN_rshift bugfix for n == 0.
15104 * Add a 'bctest' script that checks for some known 'bc' bugs
15105 so that 'make test' does not abort just because 'bc' is broken.
15109 * Store verify_result within SSL_SESSION also for client side to
15110 avoid potential security hole. (Re-used sessions on the client side
15111 always resulted in verify_result==X509_V_OK, not using the original
15112 result of the server certificate verification.)
15116 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15117 SSL3_RT_APPLICATION_DATA, return 0.
15118 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15123 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15124 releases, have been re-implemented by renaming the previous
15125 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15126 and ssl3_read_internal, respectively, and adding 'peek' parameters
15127 to them. The new ssl[23]_{read,peek} functions are calls to
15128 ssl[23]_read_internal with the 'peek' flag set appropriately.
15129 A 'peek' parameter has also been added to ssl3_read_bytes, which
15130 does the actual work for ssl3_read_internal.
15134 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15135 the method-specific "init()" handler. Also clean up ex_data after
15136 calling the method-specific "finish()" handler. Previously, this was
15137 happening the other way round.
15141 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15142 The previous value, 12, was not always sufficient for BN_mod_exp().
15146 * Make sure that shared libraries get the internal name engine with
15147 the full version number and not just 0. This should mark the
15148 shared libraries as not backward compatible. Of course, this should
15149 be changed again when we can guarantee backward binary compatibility.
15153 * Fix typo in get_cert_by_subject() in by_dir.c
15155 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15157 * Rework the system to generate shared libraries:
15159 - Make note of the expected extension for the shared libraries and
15160 if there is a need for symbolic links from for example libcrypto.so.0
15161 to libcrypto.so.0.9.7. There is extended info in Configure for
15164 - Make as few rebuilds of the shared libraries as possible.
15166 - Still avoid linking the OpenSSL programs with the shared libraries.
15168 - When installing, install the shared libraries separately from the
15173 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15175 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15176 and not in SSL_clear because the latter is also used by the
15177 accept/connect functions; previously, the settings made by
15178 SSL_set_read_ahead would be lost during the handshake.
15180 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15182 * Correct util/mkdef.pl to be selective about disabled algorithms.
15183 Previously, it would create entries for disabled algorithms no
15188 * Added several new manual pages for SSL_* function.
15192 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15194 * In ssl23_get_client_hello, generate an error message when faced
15195 with an initial SSL 3.0/TLS record that is too small to contain the
15196 first two bytes of the ClientHello message, i.e. client_version.
15197 (Note that this is a pathologic case that probably has never happened
15198 in real life.) The previous approach was to use the version number
15199 from the record header as a substitute; but our protocol choice
15200 should not depend on that one because it is not authenticated
15201 by the Finished messages.
15205 * More robust randomness gathering functions for Windows.
15207 *Jeffrey Altman <jaltman@columbia.edu>*
15209 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15210 not set then we don't setup the error code for issuer check errors
15211 to avoid possibly overwriting other errors which the callback does
15212 handle. If an application does set the flag then we assume it knows
15213 what it is doing and can handle the new informational codes
15218 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15219 a general "ANY" type, as such it should be able to decode anything
15220 including tagged types. However it didn't check the class so it would
15221 wrongly interpret tagged types in the same way as their universal
15222 counterpart and unknown types were just rejected. Changed so that the
15223 tagged and unknown types are handled in the same way as a SEQUENCE:
15224 that is the encoding is stored intact. There is also a new type
15225 "V_ASN1_OTHER" which is used when the class is not universal, in this
15226 case we have no idea what the actual type is so we just lump them all
15231 * On VMS, stdout may very well lead to a file that is written to
15232 in a record-oriented fashion. That means that every write() will
15233 write a separate record, which will be read separately by the
15234 programs trying to read from it. This can be very confusing.
15236 The solution is to put a BIO filter in the way that will buffer
15237 text until a linefeed is reached, and then write everything a
15238 line at a time, so every record written will be an actual line,
15239 not chunks of lines and not (usually doesn't happen, but I've
15240 seen it once) several lines in one record. BIO_f_linebuffer() is
15243 Currently, it's a VMS-only method, because that's where it has
15244 been tested well enough.
15248 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15249 it can return incorrect results.
15250 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15251 but it was in 0.9.6-beta[12].)
15255 * Disable the check for content being present when verifying detached
15256 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15257 include zero length content when signing messages.
15261 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15262 BIO_ctrl (for BIO pairs).
15266 * Add DSO method for VMS.
15270 * Bug fix: Montgomery multiplication could produce results with the
15275 * Add RPM specification openssl.spec and modify it to build three
15276 packages. The default package contains applications, application
15277 documentation and run-time libraries. The devel package contains
15278 include files, static libraries and function documentation. The
15279 doc package contains the contents of the doc directory. The original
15280 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15284 * Add a large number of documentation files for many SSL routines.
15286 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15288 * Add a configuration entry for Sony News 4.
15290 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15292 * Don't set the two most significant bits to one when generating a
15293 random number < q in the DSA library.
15297 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15298 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15299 the underlying transport is blocking) if a handshake took place.
15300 (The default behaviour is needed by applications such as s_client
15301 and s_server that use select() to determine when to use SSL_read;
15302 but for applications that know in advance when to expect data, it
15303 just makes things more complicated.)
15307 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15312 * Add a few more EBCDIC conditionals that make `req` and `x509`
15313 work better on such systems.
15315 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15317 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15318 Update PKCS12_parse() so it copies the friendlyName and the
15319 keyid to the certificates aux info.
15323 * Fix bug in PKCS7_verify() which caused an infinite loop
15324 if there was more than one signature.
15326 *Sven Uszpelkat <su@celocom.de>*
15328 * Major change in util/mkdef.pl to include extra information
15329 about each symbol, as well as presenting variables as well
15330 as functions. This change means that there's n more need
15331 to rebuild the .num files when some algorithms are excluded.
15335 * Allow the verify time to be set by an application,
15336 rather than always using the current time.
15340 * Phase 2 verify code reorganisation. The certificate
15341 verify code now looks up an issuer certificate by a
15342 number of criteria: subject name, authority key id
15343 and key usage. It also verifies self signed certificates
15344 by the same criteria. The main comparison function is
15345 X509_check_issued() which performs these checks.
15347 Lot of changes were necessary in order to support this
15348 without completely rewriting the lookup code.
15350 Authority and subject key identifier are now cached.
15352 The LHASH 'certs' is X509_STORE has now been replaced
15353 by a STACK_OF(X509_OBJECT). This is mainly because an
15354 LHASH can't store or retrieve multiple objects with
15355 the same hash value.
15357 As a result various functions (which were all internal
15358 use only) have changed to handle the new X509_STORE
15359 structure. This will break anything that messed round
15360 with X509_STORE internally.
15362 The functions X509_STORE_add_cert() now checks for an
15363 exact match, rather than just subject name.
15365 The X509_STORE API doesn't directly support the retrieval
15366 of multiple certificates matching a given criteria, however
15367 this can be worked round by performing a lookup first
15368 (which will fill the cache with candidate certificates)
15369 and then examining the cache for matches. This is probably
15370 the best we can do without throwing out X509_LOOKUP
15371 entirely (maybe later...).
15373 The X509_VERIFY_CTX structure has been enhanced considerably.
15375 All certificate lookup operations now go via a get_issuer()
15376 callback. Although this currently uses an X509_STORE it
15377 can be replaced by custom lookups. This is a simple way
15378 to bypass the X509_STORE hackery necessary to make this
15379 work and makes it possible to use more efficient techniques
15380 in future. A very simple version which uses a simple
15381 STACK for its trusted certificate store is also provided
15382 using X509_STORE_CTX_trusted_stack().
15384 The verify_cb() and verify() callbacks now have equivalents
15385 in the X509_STORE_CTX structure.
15387 X509_STORE_CTX also has a 'flags' field which can be used
15388 to customise the verify behaviour.
15392 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15393 excludes S/MIME capabilities.
15397 * When a certificate request is read in keep a copy of the
15398 original encoding of the signed data and use it when outputting
15399 again. Signatures then use the original encoding rather than
15400 a decoded, encoded version which may cause problems if the
15401 request is improperly encoded.
15405 * For consistency with other BIO_puts implementations, call
15406 buffer_write(b, ...) directly in buffer_puts instead of calling
15409 In BIO_puts, increment b->num_write as in BIO_write.
15411 *Peter.Sylvester@EdelWeb.fr*
15413 * Fix BN_mul_word for the case where the word is 0. (We have to use
15414 BN_zero, we may not return a BIGNUM with an array consisting of
15415 words set to zero.)
15419 * Avoid calling abort() from within the library when problems are
15420 detected, except if preprocessor symbols have been defined
15421 (such as REF_CHECK, BN_DEBUG etc.).
15425 * New openssl application 'rsautl'. This utility can be
15426 used for low-level RSA operations. DER public key
15427 BIO/fp routines also added.
15431 * New Configure entry and patches for compiling on QNX 4.
15433 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15435 * A demo state-machine implementation was sponsored by
15436 Nuron (<http://www.nuron.com/>) and is now available in
15437 demos/state_machine.
15441 * New options added to the 'dgst' utility for signature
15442 generation and verification.
15446 * Unrecognized PKCS#7 content types are now handled via a
15447 catch all ASN1_TYPE structure. This allows unsupported
15448 types to be stored as a "blob" and an application can
15449 encode and decode it manually.
15453 * Fix various signed/unsigned issues to make a_strex.c
15454 compile under VC++.
15456 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15458 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15459 length if passed a buffer. ASN1_INTEGER_to_BN failed
15460 if passed a NULL BN and its argument was negative.
15462 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15464 * Modification to PKCS#7 encoding routines to output definite
15465 length encoding. Since currently the whole structures are in
15466 memory there's not real point in using indefinite length
15467 constructed encoding. However if OpenSSL is compiled with
15468 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15472 * Added BIO_vprintf() and BIO_vsnprintf().
15476 * Added more prefixes to parse for in the strings written
15477 through a logging bio, to cover all the levels that are available
15478 through syslog. The prefixes are now:
15480 PANIC, EMERG, EMR => LOG_EMERG
15481 ALERT, ALR => LOG_ALERT
15482 CRIT, CRI => LOG_CRIT
15483 ERROR, ERR => LOG_ERR
15484 WARNING, WARN, WAR => LOG_WARNING
15485 NOTICE, NOTE, NOT => LOG_NOTICE
15486 INFO, INF => LOG_INFO
15487 DEBUG, DBG => LOG_DEBUG
15489 and as before, if none of those prefixes are present at the
15490 beginning of the string, LOG_ERR is chosen.
15492 On Win32, the `LOG_*` levels are mapped according to this:
15494 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15495 LOG_WARNING => EVENTLOG_WARNING_TYPE
15496 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15500 * Made it possible to reconfigure with just the configuration
15501 argument "reconf" or "reconfigure". The command line arguments
15502 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15503 and are retrieved from there when reconfiguring.
15509 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15511 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15515 * The obj_dat.pl script was messing up the sorting of object
15516 names. The reason was that it compared the quoted version
15517 of strings as a result "OCSP" > "OCSP Signing" because
15518 " > SPACE. Changed script to store unquoted versions of
15519 names and add quotes on output. It was also omitting some
15520 names from the lookup table if they were given a default
15521 value (that is if SN is missing it is given the same
15522 value as LN and vice versa), these are now added on the
15523 grounds that if an object has a name we should be able to
15524 look it up. Finally added warning output when duplicate
15525 short or long names are found.
15529 * Changes needed for Tandem NSK.
15531 *Scott Uroff <scott@xypro.com>*
15533 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15534 RSA_padding_check_SSLv23(), special padding was never detected
15535 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15536 version rollback attacks was not effective.
15538 In s23_clnt.c, don't use special rollback-attack detection padding
15539 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15540 client; similarly, in s23_srvr.c, don't do the rollback check if
15541 SSL 2.0 is the only protocol enabled in the server.
15545 * Make it possible to get hexdumps of unprintable data with 'openssl
15546 asn1parse'. By implication, the functions ASN1_parse_dump() and
15547 BIO_dump_indent() are added.
15551 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15552 these print out strings and name structures based on various
15553 flags including RFC2253 support and proper handling of
15554 multibyte characters. Added options to the 'x509' utility
15555 to allow the various flags to be set.
15559 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15560 Also change the functions X509_cmp_current_time() and
15561 X509_gmtime_adj() work with an ASN1_TIME structure,
15562 this will enable certificates using GeneralizedTime in validity
15563 dates to be checked.
15567 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15568 negative public key encodings) on by default,
15569 NO_NEG_PUBKEY_BUG can be set to disable it.
15573 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15574 content octets. An i2c_ASN1_OBJECT is unnecessary because
15575 the encoding can be trivially obtained from the structure.
15579 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15580 not read locks (`CRYPTO_r_[un]lock`).
15584 * A first attempt at creating official support for shared
15585 libraries through configuration. I've kept it so the
15586 default is static libraries only, and the OpenSSL programs
15587 are always statically linked for now, but there are
15588 preparations for dynamic linking in place.
15589 This has been tested on Linux and Tru64.
15593 * Randomness polling function for Win9x, as described in:
15594 Peter Gutmann, Software Generation of Practically Strong
15599 * Fix so PRNG is seeded in req if using an already existing
15604 * New options to smime application. -inform and -outform
15605 allow alternative formats for the S/MIME message including
15606 PEM and DER. The -content option allows the content to be
15607 specified separately. This should allow things like Netscape
15608 form signing output easier to verify.
15612 * Fix the ASN1 encoding of tags using the 'long form'.
15616 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15617 STRING types. These convert content octets to and from the
15618 underlying type. The actual tag and length octets are
15619 already assumed to have been read in and checked. These
15620 are needed because all other string types have virtually
15621 identical handling apart from the tag. By having versions
15622 of the ASN1 functions that just operate on content octets
15623 IMPLICIT tagging can be handled properly. It also allows
15624 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15625 and ASN1_INTEGER are identical apart from the tag.
15629 * Change the handling of OID objects as follows:
15631 - New object identifiers are inserted in objects.txt, following
15632 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15633 - objects.pl is used to process obj_mac.num and create a new
15635 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15638 This is currently kind of a hack, and the perl code in objects.pl
15639 isn't very elegant, but it works as I intended. The simplest way
15640 to check that it worked correctly is to look in obj_dat.h and
15641 check the array nid_objs and make sure the objects haven't moved
15642 around (this is important!). Additions are OK, as well as
15643 consistent name changes.
15647 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15651 * Addition of the command line parameter '-rand file' to 'openssl req'.
15652 The given file adds to whatever has already been seeded into the
15653 random pool through the RANDFILE configuration file option or
15654 environment variable, or the default random state file.
15658 * mkstack.pl now sorts each macro group into lexical order.
15659 Previously the output order depended on the order the files
15660 appeared in the directory, resulting in needless rewriting
15665 * Patches to make OpenSSL compile under Win32 again. Mostly
15666 work arounds for the VC++ problem that it treats func() as
15667 func(void). Also stripped out the parts of mkdef.pl that
15668 added extra typesafe functions: these no longer exist.
15672 * Reorganisation of the stack code. The macros are now all
15673 collected in safestack.h . Each macro is defined in terms of
15674 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15675 DEBUG_SAFESTACK is now handled in terms of function casts,
15676 this has the advantage of retaining type safety without the
15677 use of additional functions. If DEBUG_SAFESTACK is not defined
15678 then the non typesafe macros are used instead. Also modified the
15679 mkstack.pl script to handle the new form. Needs testing to see
15680 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15681 the default if no major problems. Similar behaviour for ASN1_SET_OF
15682 and PKCS12_STACK_OF.
15686 * When some versions of IIS use the 'NET' form of private key the
15687 key derivation algorithm is different. Normally MD5(password) is
15688 used as a 128 bit RC4 key. In the modified case
15689 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15690 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15691 as the old Netscape_RSA functions except they have an additional
15692 'sgckey' parameter which uses the modified algorithm. Also added
15693 an -sgckey command line option to the rsa utility. Thanks to
15694 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15695 algorithm to openssl-dev.
15699 * The evp_local.h macros were using 'c.##kname' which resulted in
15700 invalid expansion on some systems (SCO 5.0.5 for example).
15701 Corrected to 'c.kname'.
15703 *Phillip Porch <root@theporch.com>*
15705 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15706 a STACK of email addresses from a certificate or request, these look
15707 in the subject name and the subject alternative name extensions and
15708 omit any duplicate addresses.
15712 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15713 This makes DSA verification about 2 % faster.
15717 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15718 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15719 plus overhead for 1024 bit moduli).
15720 This makes exponentiations about 0.5 % faster for 1024 bit
15721 exponents (as measured by "openssl speed rsa2048").
15725 * Rename memory handling macros to avoid conflicts with other
15727 Malloc => OPENSSL_malloc
15728 Malloc_locked => OPENSSL_malloc_locked
15729 Realloc => OPENSSL_realloc
15730 Free => OPENSSL_free
15734 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15735 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15739 * CygWin32 support.
15741 *John Jarvie <jjarvie@newsguy.com>*
15743 * The type-safe stack code has been rejigged. It is now only compiled
15744 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15745 by default all type-specific stack functions are "#define"d back to
15746 standard stack functions. This results in more streamlined output
15747 but retains the type-safety checking possibilities of the original
15752 * The STACK code has been cleaned up, and certain type declarations
15753 that didn't make a lot of sense have been brought in line. This has
15754 also involved a cleanup of sorts in safestack.h to more correctly
15755 map type-safe stack functions onto their plain stack counterparts.
15756 This work has also resulted in a variety of "const"ifications of
15757 lots of the code, especially `_cmp` operations which should normally
15758 be prototyped with "const" parameters anyway.
15762 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15763 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15764 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15765 where all of 'md' is used each time the PRNG is used, but 'state'
15766 is used only indexed by a cyclic counter. As entropy may not be
15767 well distributed from the beginning, 'md' is important as a
15768 chaining variable. However, the output function chains only half
15769 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15770 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15771 in all of 'state' being rewritten, with the new values depending
15772 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15776 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15777 the handshake is continued after ssl_verify_cert_chain();
15778 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15779 can lead to 'unexplainable' connection aborts later.
15781 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15783 * Major EVP API cipher revision.
15784 Add hooks for extra EVP features. This allows various cipher
15785 parameters to be set in the EVP interface. Support added for variable
15786 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15787 setting of RC2 and RC5 parameters.
15789 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15792 Remove lots of duplicated code from the EVP library. For example *every*
15793 cipher init() function handles the 'iv' in the same way according to the
15794 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15795 for CFB and OFB modes they zero ctx->num.
15797 New functionality allows removal of S/MIME code RC2 hack.
15799 Most of the routines have the same form and so can be declared in terms
15802 By shifting this to the top level EVP_CipherInit() it can be removed from
15803 all individual ciphers. If the cipher wants to handle IVs or keys
15804 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15807 Change lots of functions like EVP_EncryptUpdate() to now return a
15808 value: although software versions of the algorithms cannot fail
15809 any installed hardware versions can.
15813 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15814 this option is set, tolerate broken clients that send the negotiated
15815 protocol version number instead of the requested protocol version
15820 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15821 i.e. non-zero for export ciphersuites, zero otherwise.
15822 Previous versions had this flag inverted, inconsistent with
15823 rsa_tmp_cb (..._TMP_RSA_CB).
15825 *Bodo Moeller; problem reported by Amit Chopra*
15827 * Add missing DSA library text string. Work around for some IIS
15828 key files with invalid SEQUENCE encoding.
15832 * Add a document (doc/standards.txt) that list all kinds of standards
15833 and so on that are implemented in OpenSSL.
15837 * Enhance c_rehash script. Old version would mishandle certificates
15838 with the same subject name hash and wouldn't handle CRLs at all.
15839 Added -fingerprint option to crl utility, to support new c_rehash
15844 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15848 * Fix for SSL server purpose checking. Server checking was
15849 rejecting certificates which had extended key usage present
15850 but no ssl client purpose.
15852 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15854 * Make PKCS#12 code work with no password. The PKCS#12 spec
15855 is a little unclear about how a blank password is handled.
15856 Since the password in encoded as a BMPString with terminating
15857 double NULL a zero length password would end up as just the
15858 double NULL. However no password at all is different and is
15859 handled differently in the PKCS#12 key generation code. NS
15860 treats a blank password as zero length. MSIE treats it as no
15861 password on export: but it will try both on import. We now do
15862 the same: PKCS12_parse() tries zero length and no password if
15863 the password is set to "" or NULL (NULL is now a valid password:
15864 it wasn't before) as does the pkcs12 application.
15868 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15869 perror when PEM_read_bio_X509_REQ fails, the error message must
15870 be obtained from the error queue.
15874 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15875 it in ERR_remove_state if appropriate, and change ERR_get_state
15876 accordingly to avoid race conditions (this is necessary because
15877 thread_hash is no longer constant once set).
15881 * Bugfix for linux-elf makefile.one.
15885 * RSA_get_default_method() will now cause a default
15886 RSA_METHOD to be chosen if one doesn't exist already.
15887 Previously this was only set during a call to RSA_new()
15888 or RSA_new_method(NULL) meaning it was possible for
15889 RSA_get_default_method() to return NULL.
15893 * Added native name translation to the existing DSO code
15894 that will convert (if the flag to do so is set) filenames
15895 that are sufficiently small and have no path information
15896 into a canonical native form. Eg. "blah" converted to
15897 "libblah.so" or "blah.dll" etc.
15901 * New function ERR_error_string_n(e, buf, len) which is like
15902 ERR_error_string(e, buf), but writes at most 'len' bytes
15903 including the 0 terminator. For ERR_error_string_n, 'buf'
15906 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15908 * CONF library reworked to become more general. A new CONF
15909 configuration file reader "class" is implemented as well as a
15910 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15911 old `CONF_*` functions are still there, but are reimplemented to
15912 work in terms of the new functions. Also, a set of functions
15913 to handle the internal storage of the configuration data is
15914 provided to make it easier to write new configuration file
15915 reader "classes" (I can definitely see something reading a
15916 configuration file in XML format, for example), called `_CONF_*`,
15917 or "the configuration storage API"...
15919 The new configuration file reading functions are:
15921 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15922 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15924 NCONF_default, NCONF_WIN32
15926 NCONF_dump_fp, NCONF_dump_bio
15928 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15929 NCONF_new creates a new CONF object. This works in the same way
15930 as other interfaces in OpenSSL, like the BIO interface.
15931 `NCONF_dump_*` dump the internal storage of the configuration file,
15932 which is useful for debugging. All other functions take the same
15933 arguments as the old `CONF_*` functions with the exception of the
15934 first that must be a `CONF *` instead of a `LHASH *`.
15936 To make it easier to use the new classes with the old `CONF_*` functions,
15937 the function CONF_set_default_method is provided.
15941 * Add '-tls1' option to 'openssl ciphers', which was already
15942 mentioned in the documentation but had not been implemented.
15943 (This option is not yet really useful because even the additional
15944 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15948 * Initial DSO code added into libcrypto for letting OpenSSL (and
15949 OpenSSL-based applications) load shared libraries and bind to
15950 them in a portable way.
15952 *Geoff Thorpe, with contributions from Richard Levitte*
15954 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15956 * Make sure _lrotl and _lrotr are only used with MSVC.
15958 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15959 (the default implementation of RAND_status).
15961 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15962 to '-clrext' (= clear extensions), as intended and documented.
15963 *Bodo Moeller; inconsistency pointed out by Michael Attili
15964 <attili@amaxo.com>*
15966 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15967 was larger than the MD block size.
15969 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15971 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15972 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15973 using the passed key: if the passed key was a private key the result
15974 of X509_print(), for example, would be to print out all the private key
15979 * des_quad_cksum() byte order bug fix.
15980 *Ulf Möller, using the problem description in krb4-0.9.7, where
15981 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15983 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15986 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15988 * For easily testing in shell scripts whether some command
15989 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15990 returns with exit code 0 iff no command of the given name is available.
15991 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15992 the output goes to stdout and nothing is printed to stderr.
15993 Additional arguments are always ignored.
15995 Since for each cipher there is a command of the same name,
15996 the 'no-cipher' compilation switches can be tested this way.
15998 ('openssl no-XXX' is not able to detect pseudo-commands such
15999 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16003 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16007 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16008 is set; it will be thrown away anyway because each handshake creates
16010 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16011 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16012 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16013 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16017 * New s_client option -ign_eof: EOF at stdin is ignored, and
16018 'Q' and 'R' lose their special meanings (quit/renegotiate).
16019 This is part of what -quiet does; unlike -quiet, -ign_eof
16020 does not suppress any output.
16024 * Add compatibility options to the purpose and trust code. The
16025 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16026 accepts a certificate or CA, this was the previous behaviour,
16027 with all the associated security issues.
16029 X509_TRUST_COMPAT is the old trust behaviour: only and
16030 automatically trust self signed roots in certificate store. A
16031 new trust setting X509_TRUST_DEFAULT is used to specify that
16032 a purpose has no associated trust setting and it should instead
16033 use the value in the default purpose.
16037 * Fix the PKCS#8 DSA private key code so it decodes keys again
16038 and fix a memory leak.
16042 * In util/mkerr.pl (which implements 'make errors'), preserve
16043 reason strings from the previous version of the .c file, as
16044 the default to have only downcase letters (and digits) in
16045 automatically generated reasons codes is not always appropriate.
16049 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16050 using strerror. Previously, ERR_reason_error_string() returned
16051 library names as reason strings for SYSerr; but SYSerr is a special
16052 case where small numbers are errno values, not library numbers.
16056 * Add '-dsaparam' option to 'openssl dhparam' application. This
16057 converts DSA parameters into DH parameters. (When creating parameters,
16058 DSA_generate_parameters is used.)
16062 * Include 'length' (recommended exponent length) in C code generated
16063 by 'openssl dhparam -C'.
16067 * The second argument to set_label in perlasm was already being used
16068 so couldn't be used as a "file scope" flag. Moved to third argument
16073 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16074 instead of RAND_bytes for encryption IVs and salts.
16078 * Include RAND_status() into RAND_METHOD instead of implementing
16079 it only for md_rand.c Otherwise replacing the PRNG by calling
16080 RAND_set_rand_method would be impossible.
16084 * Don't let DSA_generate_key() enter an infinite loop if the random
16085 number generation fails.
16089 * New 'rand' application for creating pseudo-random output.
16093 * Added configuration support for Linux/IA64
16095 *Rolf Haberrecker <rolf@suse.de>*
16097 * Assembler module support for Mingw32.
16101 * Shared library support for HPUX (in shlib/).
16103 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16105 * Shared library support for Solaris gcc.
16107 *Lutz Behnke <behnke@trustcenter.de>*
16109 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16111 * PKCS7_encrypt() was adding text MIME headers twice because they
16112 were added manually and by SMIME_crlf_copy().
16116 * In bntest.c don't call BN_rand with zero bits argument.
16118 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16120 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16121 case was implemented. This caused BN_div_recp() to fail occasionally.
16125 * Add an optional second argument to the set_label() in the perl
16126 assembly language builder. If this argument exists and is set
16127 to 1 it signals that the assembler should use a symbol whose
16128 scope is the entire file, not just the current function. This
16129 is needed with MASM which uses the format label:: for this scope.
16131 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16133 * Change the ASN1 types so they are typedefs by default. Before
16134 almost all types were #define'd to ASN1_STRING which was causing
16135 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16140 * Change names of new functions to the new get1/get0 naming
16141 convention: After 'get1', the caller owns a reference count
16142 and has to call `..._free`; 'get0' returns a pointer to some
16143 data structure without incrementing reference counters.
16144 (Some of the existing 'get' functions increment a reference
16145 counter, some don't.)
16146 Similarly, 'set1' and 'add1' functions increase reference
16147 counters or duplicate objects.
16151 * Allow for the possibility of temp RSA key generation failure:
16152 the code used to assume it always worked and crashed on failure.
16156 * Fix potential buffer overrun problem in BIO_printf().
16157 *Ulf Möller, using public domain code by Patrick Powell; problem
16158 pointed out by David Sacerdote <das33@cornell.edu>*
16160 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16161 RAND_egd() and RAND_status(). In the command line application,
16162 the EGD socket can be specified like a seed file using RANDFILE
16167 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16168 Some CAs (e.g. Verisign) distribute certificates in this form.
16172 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16173 list to exclude them. This means that no special compilation option
16174 is needed to use anonymous DH: it just needs to be included in the
16179 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16180 EVP_MD_type. The old functionality is available in a new macro called
16181 EVP_MD_md(). Change code that uses it and update docs.
16185 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16186 where the `void *` argument is replaced by a function pointer argument.
16187 Previously `void *` was abused to point to functions, which works on
16188 many platforms, but is not correct. As these functions are usually
16189 called by macros defined in OpenSSL header files, most source code
16190 should work without changes.
16194 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16195 sections with information on -D... compiler switches used for
16196 compiling the library so that applications can see them. To enable
16197 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16198 must be defined. E.g.,
16199 #define OPENSSL_ALGORITHM_DEFINES
16200 #include <openssl/opensslconf.h>
16201 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16203 *Richard Levitte, Ulf and Bodo Möller*
16205 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16210 * Change the 'other' type in certificate aux info to a STACK_OF
16211 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16212 the required ASN1 format: arbitrary types determined by an OID.
16216 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16217 argument to 'req'. This is not because the function is newer or
16218 better than others it just uses the work 'NEW' in the certificate
16219 request header lines. Some software needs this.
16223 * Reorganise password command line arguments: now passwords can be
16224 obtained from various sources. Delete the PEM_cb function and make
16225 it the default behaviour: i.e. if the callback is NULL and the
16226 usrdata argument is not NULL interpret it as a null terminated pass
16227 phrase. If usrdata and the callback are NULL then the pass phrase
16228 is prompted for as usual.
16232 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16233 the support is automatically enabled. The resulting binaries will
16234 autodetect the card and use it if present.
16236 *Ben Laurie and Compaq Inc.*
16238 * Work around for Netscape hang bug. This sends certificate request
16239 and server done in one record. Since this is perfectly legal in the
16240 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16241 the bugs/SSLv3 entry for more info.
16245 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16249 * Add -rand argument to smime and pkcs12 applications and read/write
16254 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16258 * Add command line password options to the remaining applications.
16262 * Bug fix for BN_div_recp() for numerators with an even number of
16267 * More tests in bntest.c, and changed test_bn output.
16271 * ./config recognizes MacOS X now.
16275 * Bug fix for BN_div() when the first words of num and divisor are
16276 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16280 * Add support for various broken PKCS#8 formats, and command line
16281 options to produce them.
16285 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16286 get temporary BIGNUMs from a BN_CTX.
16290 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16295 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16296 include a #define from the old name to the new. The original intent
16297 was that statically linked binaries could for example just call
16298 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16299 link with digests. This never worked because SSLeay_add_all_digests()
16300 and SSLeay_add_all_ciphers() were in the same source file so calling
16301 one would link with the other. They are now in separate source files.
16305 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16309 * Use a less unusual form of the Miller-Rabin primality test (it used
16310 a binary algorithm for exponentiation integrated into the Miller-Rabin
16311 loop, our standard modexp algorithms are faster).
16315 * Support for the EBCDIC character set completed.
16317 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16319 * Source code cleanups: use const where appropriate, eliminate casts,
16320 use `void *` instead of `char *` in lhash.
16324 * Bugfix: ssl3_send_server_key_exchange was not restartable
16325 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16326 this the server could overwrite ephemeral keys that the client
16331 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16332 using 50 iterations of the Rabin-Miller test.
16334 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16335 iterations of the Rabin-Miller test as required by the appendix
16336 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16337 As BN_is_prime_fasttest includes trial division, DSA parameter
16338 generation becomes much faster.
16340 This implies a change for the callback functions in DSA_is_prime
16341 and DSA_generate_parameters: The callback function is called once
16342 for each positive witness in the Rabin-Miller test, not just
16343 occasionally in the inner loop; and the parameters to the
16344 callback function now provide an iteration count for the outer
16345 loop rather than for the current invocation of the inner loop.
16346 DSA_generate_parameters additionally can call the callback
16347 function with an 'iteration count' of -1, meaning that a
16348 candidate has passed the trial division test (when q is generated
16349 from an application-provided seed, trial division is skipped).
16353 * New function BN_is_prime_fasttest that optionally does trial
16354 division before starting the Rabin-Miller test and has
16355 an additional BN_CTX * argument (whereas BN_is_prime always
16356 has to allocate at least one BN_CTX).
16357 'callback(1, -1, cb_arg)' is called when a number has passed the
16358 trial division stage.
16362 * Fix for bug in CRL encoding. The validity dates weren't being handled
16367 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16371 * New function BN_pseudo_rand().
16375 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16376 bignum version of BN_from_montgomery() with the working code from
16377 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16382 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16383 made it impossible to use the same SSL_SESSION data structure in
16384 SSL2 clients in multiple threads.
16388 * The return value of RAND_load_file() no longer counts bytes obtained
16389 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16390 to seed the PRNG (previously an explicit byte count was required).
16392 *Ulf Möller, Bodo Möller*
16394 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16395 used `char *` instead of `void *` and had casts all over the place.
16399 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16403 * Retain source code compatibility for BN_prime_checks macro:
16404 BN_is_prime(..., BN_prime_checks, ...) now uses
16405 BN_prime_checks_for_size to determine the appropriate number of
16406 Rabin-Miller iterations.
16410 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16411 DH_CHECK_P_NOT_SAFE_PRIME.
16412 (Check if this is true? OpenPGP calls them "strong".)
16416 * Merge the functionality of "dh" and "gendh" programs into a new program
16417 "dhparam". The old programs are retained for now but will handle DH keys
16418 (instead of parameters) in future.
16422 * Make the ciphers, s_server and s_client programs check the return values
16423 when a new cipher list is set.
16427 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16428 ciphers. Before when the 56bit ciphers were enabled the sorting was
16431 The syntax for the cipher sorting has been extended to support sorting by
16432 cipher-strength (using the strength_bits hard coded in the tables).
16433 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16435 Fix a bug in the cipher-command parser: when supplying a cipher command
16436 string with an "undefined" symbol (neither command nor alphanumeric
16437 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16438 an error is flagged.
16440 Due to the strength-sorting extension, the code of the
16441 ssl_create_cipher_list() function was completely rearranged. I hope that
16442 the readability was also increased :-)
16444 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16446 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16447 for the first serial number and places 2 in the serial number file. This
16448 avoids problems when the root CA is created with serial number zero and
16449 the first user certificate has the same issuer name and serial number
16454 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16455 the new code. Add documentation for this stuff.
16459 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16460 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16461 structures and behave in an analogous way to the X509v3 functions:
16462 they shouldn't be called directly but wrapper functions should be used
16465 So we also now have some wrapper functions that call the X509at functions
16466 when passed certificate requests. (TO DO: similar things can be done with
16467 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16468 things. Some of these need some d2i or i2d and print functionality
16469 because they handle more complex structures.)
16473 * Add missing #ifndefs that caused missing symbols when building libssl
16474 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16475 NO_RSA in `ssl/s2*.c`.
16477 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16479 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16480 has a return value which indicates the quality of the random data
16481 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16482 error queue. New function RAND_pseudo_bytes() generates output that is
16483 guaranteed to be unique but not unpredictable. RAND_add is like
16484 RAND_seed, but takes an extra argument for an entropy estimate
16485 (RAND_seed always assumes full entropy).
16489 * Do more iterations of Rabin-Miller probable prime test (specifically,
16490 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16491 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16492 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16493 false-positive rate of at most 2^-80 for random input.
16497 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16501 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16502 in the 0.9.5 release), this returns the chain
16503 from an X509_CTX structure with a dup of the stack and all
16504 the X509 reference counts upped: so the stack will exist
16505 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16508 Also make SSL_SESSION_print() print out the verify return
16513 * Add manpage for the pkcs12 command. Also change the default
16514 behaviour so MAC iteration counts are used unless the new
16515 -nomaciter option is used. This improves file security and
16516 only older versions of MSIE (4.0 for example) need it.
16520 * Honor the no-xxx Configure options when creating .DEF files.
16524 * Add PKCS#10 attributes to field table: challengePassword,
16525 unstructuredName and unstructuredAddress. These are taken from
16526 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16527 international characters are used.
16529 More changes to X509_ATTRIBUTE code: allow the setting of types
16530 based on strings. Remove the 'loc' parameter when adding
16531 attributes because these will be a SET OF encoding which is sorted
16536 * Initial changes to the 'req' utility to allow request generation
16537 automation. This will allow an application to just generate a template
16538 file containing all the field values and have req construct the
16541 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16542 used all over the place including certificate requests and PKCS#7
16543 structures. They are currently handled manually where necessary with
16544 some primitive wrappers for PKCS#7. The new functions behave in a
16545 manner analogous to the X509 extension functions: they allow
16546 attributes to be looked up by NID and added.
16548 Later something similar to the X509V3 code would be desirable to
16549 automatically handle the encoding, decoding and printing of the
16550 more complex types. The string types like challengePassword can
16551 be handled by the string table functions.
16553 Also modified the multi byte string table handling. Now there is
16554 a 'global mask' which masks out certain types. The table itself
16555 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16556 is useful when for example there is only one permissible type
16557 (as in countryName) and using the mask might result in no valid
16562 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16563 SSL_get_peer_finished to allow applications to obtain the latest
16564 Finished messages sent to the peer or expected from the peer,
16565 respectively. (SSL_get_peer_finished is usually the Finished message
16566 actually received from the peer, otherwise the protocol will be aborted.)
16568 As the Finished message are message digests of the complete handshake
16569 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16570 be used for external authentication procedures when the authentication
16571 provided by SSL/TLS is not desired or is not enough.
16575 * Enhanced support for Alpha Linux is added. Now ./config checks if
16576 the host supports BWX extension and if Compaq C is present on the
16577 $PATH. Just exploiting of the BWX extension results in 20-30%
16578 performance kick for some algorithms, e.g. DES and RC4 to mention
16579 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16584 * Add support for MS "fast SGC". This is arguably a violation of the
16585 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16586 weak crypto and after checking the certificate is SGC a second one
16587 with strong crypto. MS SGC stops the first handshake after receiving
16588 the server certificate message and sends a second client hello. Since
16589 a server will typically do all the time consuming operations before
16590 expecting any further messages from the client (server key exchange
16591 is the most expensive) there is little difference between the two.
16593 To get OpenSSL to support MS SGC we have to permit a second client
16594 hello message after we have sent server done. In addition we have to
16595 reset the MAC if we do get this second client hello.
16599 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16600 if a DER encoded private key is RSA or DSA traditional format. Changed
16601 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16602 format DER encoded private key. Newer code should use PKCS#8 format which
16603 has the key type encoded in the ASN1 structure. Added DER private key
16604 support to pkcs8 application.
16608 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16609 ciphersuites has been selected (as required by the SSL 3/TLS 1
16610 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16611 is set, we interpret this as a request to violate the specification
16612 (the worst that can happen is a handshake failure, and 'correct'
16613 behaviour would result in a handshake failure anyway).
16617 * In SSL_CTX_add_session, take into account that there might be multiple
16618 SSL_SESSION structures with the same session ID (e.g. when two threads
16619 concurrently obtain them from an external cache).
16620 The internal cache can handle only one SSL_SESSION with a given ID,
16621 so if there's a conflict, we now throw out the old one to achieve
16626 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16627 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16628 some routines that use cipher OIDs: some ciphers do not have OIDs
16629 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16634 * Simplify the trust setting structure and code. Now we just have
16635 two sequences of OIDs for trusted and rejected settings. These will
16636 typically have values the same as the extended key usage extension
16637 and any application specific purposes.
16639 The trust checking code now has a default behaviour: it will just
16640 check for an object with the same NID as the passed id. Functions can
16641 be provided to override either the default behaviour or the behaviour
16642 for a given id. SSL client, server and email already have functions
16643 in place for compatibility: they check the NID and also return "trusted"
16644 if the certificate is self signed.
16648 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16649 traditional format into an EVP_PKEY structure.
16653 * Add a password callback function PEM_cb() which either prompts for
16654 a password if usr_data is NULL or otherwise assumes it is a null
16655 terminated password. Allow passwords to be passed on command line
16656 environment or config files in a few more utilities.
16660 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16661 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16662 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16663 Update documentation.
16667 * Support for ASN1 "NULL" type. This could be handled before by using
16668 ASN1_TYPE but there wasn't any function that would try to read a NULL
16669 and produce an error if it couldn't. For compatibility we also have
16670 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16671 don't allocate anything because they don't need to.
16675 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16678 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16680 * Rebuild of the memory allocation routines used by OpenSSL code and
16681 possibly others as well. The purpose is to make an interface that
16682 provide hooks so anyone can build a separate set of allocation and
16683 deallocation routines to be used by OpenSSL, for example memory
16684 pool implementations, or something else, which was previously hard
16685 since Malloc(), Realloc() and Free() were defined as macros having
16686 the values malloc, realloc and free, respectively (except for Win32
16687 compilations). The same is provided for memory debugging code.
16688 OpenSSL already comes with functionality to find memory leaks, but
16689 this gives people a chance to debug other memory problems.
16691 With these changes, a new set of functions and macros have appeared:
16693 CRYPTO_set_mem_debug_functions() [F]
16694 CRYPTO_get_mem_debug_functions() [F]
16695 CRYPTO_dbg_set_options() [F]
16696 CRYPTO_dbg_get_options() [F]
16697 CRYPTO_malloc_debug_init() [M]
16699 The memory debug functions are NULL by default, unless the library
16700 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16701 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16702 gives the standard debugging functions that come with OpenSSL) or
16703 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16704 provided by the library user) must be used. When the standard
16705 debugging functions are used, CRYPTO_dbg_set_options can be used to
16706 request additional information:
16707 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16708 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16710 Also, things like CRYPTO_set_mem_functions will always give the
16711 expected result (the new set of functions is used for allocation
16712 and deallocation) at all times, regardless of platform and compiler
16715 To finish it up, some functions that were never use in any other
16716 way than through macros have a new API and new semantic:
16718 CRYPTO_dbg_malloc()
16719 CRYPTO_dbg_realloc()
16722 All macros of value have retained their old syntax.
16724 *Richard Levitte and Bodo Moeller*
16726 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16727 ordering of SMIMECapabilities wasn't in "strength order" and there
16728 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16733 * Some ASN1 types with illegal zero length encoding (INTEGER,
16734 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16736 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16738 * Merge in my S/MIME library for OpenSSL. This provides a simple
16739 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16740 functionality to handle multipart/signed properly) and a utility
16741 called 'smime' to call all this stuff. This is based on code I
16742 originally wrote for Celo who have kindly allowed it to be
16743 included in OpenSSL.
16747 * Add variants des_set_key_checked and des_set_key_unchecked of
16748 des_set_key (aka des_key_sched). Global variable des_check_key
16749 decides which of these is called by des_set_key; this way
16750 des_check_key behaves as it always did, but applications and
16751 the library itself, which was buggy for des_check_key == 1,
16752 have a cleaner way to pick the version they need.
16756 * New function PKCS12_newpass() which changes the password of a
16761 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16762 dynamic mix. In both cases the ids can be used as an index into the
16763 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16764 functions so they accept a list of the field values and the
16765 application doesn't need to directly manipulate the X509_TRUST
16770 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16775 * Modify the way the V3 extension code looks up extensions. This now
16776 works in a similar way to the object code: we have some "standard"
16777 extensions in a static table which is searched with OBJ_bsearch()
16778 and the application can add dynamic ones if needed. The file
16779 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16780 updated whenever a new extension is added to the core code and kept
16781 in ext_nid order. There is a simple program 'tabtest.c' which checks
16782 this. New extensions are not added too often so this file can readily
16783 be maintained manually.
16785 There are two big advantages in doing things this way. The extensions
16786 can be looked up immediately and no longer need to be "added" using
16787 X509V3_add_standard_extensions(): this function now does nothing.
16788 Side note: I get *lots* of email saying the extension code doesn't
16789 work because people forget to call this function.
16790 Also no dynamic allocation is done unless new extensions are added:
16791 so if we don't add custom extensions there is no need to call
16792 X509V3_EXT_cleanup().
16796 * Modify enc utility's salting as follows: make salting the default. Add a
16797 magic header, so unsalted files fail gracefully instead of just decrypting
16798 to garbage. This is because not salting is a big security hole, so people
16799 should be discouraged from doing it.
16803 * Fixes and enhancements to the 'x509' utility. It allowed a message
16804 digest to be passed on the command line but it only used this
16805 parameter when signing a certificate. Modified so all relevant
16806 operations are affected by the digest parameter including the
16807 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16808 DSA key was used because it didn't fix the digest.
16812 * Initial certificate chain verify code. Currently tests the untrusted
16813 certificates for consistency with the verify purpose (which is set
16814 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16816 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16817 this is because it will reject chains with invalid extensions whereas
16818 every previous version of OpenSSL and SSLeay made no checks at all.
16820 Trust code: checks the root CA for the relevant trust settings. Trust
16821 settings have an initial value consistent with the verify purpose: e.g.
16822 if the verify purpose is for SSL client use it expects the CA to be
16823 trusted for SSL client use. However the default value can be changed to
16824 permit custom trust settings: one example of this would be to only trust
16825 certificates from a specific "secure" set of CAs.
16827 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16828 which should be used for version portability: especially since the
16829 verify structure is likely to change more often now.
16831 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16832 to set them. If not set then assume SSL clients will verify SSL servers
16835 Two new options to the verify program: -untrusted allows a set of
16836 untrusted certificates to be passed in and -purpose which sets the
16837 intended purpose of the certificate. If a purpose is set then the
16838 new chain verify code is used to check extension consistency.
16842 * Support for the authority information access extension.
16846 * Modify RSA and DSA PEM read routines to transparently handle
16847 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16848 public keys in a format compatible with certificate
16849 SubjectPublicKeyInfo structures. Unfortunately there were already
16850 functions called *_PublicKey_* which used various odd formats so
16851 these are retained for compatibility: however the DSA variants were
16852 never in a public release so they have been deleted. Changed dsa/rsa
16853 utilities to handle the new format: note no releases ever handled public
16854 keys so we should be OK.
16856 The primary motivation for this change is to avoid the same fiasco
16857 that dogs private keys: there are several incompatible private key
16858 formats some of which are standard and some OpenSSL specific and
16859 require various evil hacks to allow partial transparent handling and
16860 even then it doesn't work with DER formats. Given the option anything
16861 other than PKCS#8 should be dumped: but the other formats have to
16862 stay in the name of compatibility.
16864 With public keys and the benefit of hindsight one standard format
16865 is used which works with EVP_PKEY, RSA or DSA structures: though
16866 it clearly returns an error if you try to read the wrong kind of key.
16868 Added a -pubkey option to the 'x509' utility to output the public key.
16869 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16870 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16871 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16872 that do the same as the `EVP_PKEY_assign_*()` except they up the
16873 reference count of the added key (they don't "swallow" the
16878 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16879 CRLs would fail if the file contained no certificates or no CRLs:
16880 added a new function to read in both types and return the number
16881 read: this means that if none are read it will be an error. The
16882 DER versions of the certificate and CRL reader would always fail
16883 because it isn't possible to mix certificates and CRLs in DER format
16884 without choking one or the other routine. Changed this to just read
16885 a certificate: this is the best we can do. Also modified the code
16886 in `apps/verify.c` to take notice of return codes: it was previously
16887 attempting to read in certificates from NULL pointers and ignoring
16888 any errors: this is one reason why the cert and CRL reader seemed
16889 to work. It doesn't check return codes from the default certificate
16890 routines: these may well fail if the certificates aren't installed.
16894 * Code to support otherName option in GeneralName.
16898 * First update to verify code. Change the verify utility
16899 so it warns if it is passed a self signed certificate:
16900 for consistency with the normal behaviour. X509_verify
16901 has been modified to it will now verify a self signed
16902 certificate if *exactly* the same certificate appears
16903 in the store: it was previously impossible to trust a
16904 single self signed certificate. This means that:
16905 openssl verify ss.pem
16906 now gives a warning about a self signed certificate but
16907 openssl verify -CAfile ss.pem ss.pem
16912 * For servers, store verify_result in SSL_SESSION data structure
16913 (and add it to external session representation).
16914 This is needed when client certificate verifications fails,
16915 but an application-provided verification callback (set by
16916 SSL_CTX_set_cert_verify_callback) allows accepting the session
16917 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16918 but returns 1): When the session is reused, we have to set
16919 ssl->verify_result to the appropriate error code to avoid
16922 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16924 * Fix a bug in the new PKCS#7 code: it didn't consider the
16925 case in PKCS7_dataInit() where the signed PKCS7 structure
16926 didn't contain any existing data because it was being created.
16928 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16930 * Add a salt to the key derivation routines in enc.c. This
16931 forms the first 8 bytes of the encrypted file. Also add a
16932 -S option to allow a salt to be input on the command line.
16936 * New function X509_cmp(). Oddly enough there wasn't a function
16937 to compare two certificates. We do this by working out the SHA1
16938 hash and comparing that. X509_cmp() will be needed by the trust
16943 * SSL_get1_session() is like SSL_get_session(), but increments
16944 the reference count in the SSL_SESSION returned.
16946 *Geoff Thorpe <geoff@eu.c2.net>*
16948 * Fix for 'req': it was adding a null to request attributes.
16949 Also change the X509_LOOKUP and X509_INFO code to handle
16950 certificate auxiliary information.
16954 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16959 * Add the possibility to add extra information to the memory leak
16960 detecting output, to form tracebacks, showing from where each
16961 allocation was originated: CRYPTO_push_info("constant string") adds
16962 the string plus current file name and line number to a per-thread
16963 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16964 is like calling CYRPTO_pop_info() until the stack is empty.
16965 Also updated memory leak detection code to be multi-thread-safe.
16969 * Add options -text and -noout to pkcs7 utility and delete the
16970 encryption options which never did anything. Update docs.
16974 * Add options to some of the utilities to allow the pass phrase
16975 to be included on either the command line (not recommended on
16976 OSes like Unix) or read from the environment. Update the
16977 manpages and fix a few bugs.
16981 * Add a few manpages for some of the openssl commands.
16985 * Fix the -revoke option in ca. It was freeing up memory twice,
16986 leaking and not finding already revoked certificates.
16990 * Extensive changes to support certificate auxiliary information.
16991 This involves the use of X509_CERT_AUX structure and X509_AUX
16992 functions. An X509_AUX function such as PEM_read_X509_AUX()
16993 can still read in a certificate file in the usual way but it
16994 will also read in any additional "auxiliary information". By
16995 doing things this way a fair degree of compatibility can be
16996 retained: existing certificates can have this information added
16997 using the new 'x509' options.
16999 Current auxiliary information includes an "alias" and some trust
17000 settings. The trust settings will ultimately be used in enhanced
17001 certificate chain verification routines: currently a certificate
17002 can only be trusted if it is self signed and then it is trusted
17007 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17008 The problem was that one of the replacement routines had not been working
17009 since SSLeay releases. For now the offending routine has been replaced
17010 with non-optimised assembler. Even so, this now gives around 95%
17011 performance improvement for 1024 bit RSA signs.
17015 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17016 handling. Most clients have the effective key size in bits equal to
17017 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17018 A few however don't do this and instead use the size of the decrypted key
17019 to determine the RC2 key length and the AlgorithmIdentifier to determine
17020 the effective key length. In this case the effective key length can still
17021 be 40 bits but the key length can be 168 bits for example. This is fixed
17022 by manually forcing an RC2 key into the EVP_PKEY structure because the
17023 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17024 the key length and effective key length are equal.
17028 * Add a bunch of functions that should simplify the creation of
17029 X509_NAME structures. Now you should be able to do:
17030 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17031 and have it automatically work out the correct field type and fill in
17032 the structures. The more adventurous can try:
17033 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17034 and it will (hopefully) work out the correct multibyte encoding.
17038 * Change the 'req' utility to use the new field handling and multibyte
17039 copy routines. Before the DN field creation was handled in an ad hoc
17040 way in req, ca, and x509 which was rather broken and didn't support
17041 BMPStrings or UTF8Strings. Since some software doesn't implement
17042 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17043 using the dirstring_type option. See the new comment in the default
17044 openssl.cnf for more info.
17048 * Make crypto/rand/md_rand.c more robust:
17049 - Assure unique random numbers after fork().
17050 - Make sure that concurrent threads access the global counter and
17051 md serializably so that we never lose entropy in them
17052 or use exactly the same state in multiple threads.
17053 Access to the large state is not always serializable because
17054 the additional locking could be a performance killer, and
17055 md should be large enough anyway.
17059 * New file `apps/app_rand.c` with commonly needed functionality
17060 for handling the random seed file.
17062 Use the random seed file in some applications that previously did not:
17064 dsaparam -genkey (which also ignored its '-rand' option),
17067 x509 (when signing).
17068 Except on systems with /dev/urandom, it is crucial to have a random
17069 seed file at least for key creation, DSA signing, and for DH exchanges;
17070 for RSA signatures we could do without one.
17072 gendh and gendsa (unlike genrsa) used to read only the first byte
17073 of each file listed in the '-rand' option. The function as previously
17074 found in genrsa is now in app_rand.c and is used by all programs
17075 that support '-rand'.
17079 * In RAND_write_file, use mode 0600 for creating files;
17080 don't just chmod when it may be too late.
17084 * Report an error from X509_STORE_load_locations
17085 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17089 * New function ASN1_mbstring_copy() this copies a string in either
17090 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17091 into an ASN1_STRING type. A mask of permissible types is passed
17092 and it chooses the "minimal" type to use or an error if not type
17097 * Add function equivalents to the various macros in asn1.h. The old
17098 macros are retained with an `M_` prefix. Code inside the library can
17099 use the `M_` macros. External code (including the openssl utility)
17100 should *NOT* in order to be "shared library friendly".
17104 * Add various functions that can check a certificate's extensions
17105 to see if it usable for various purposes such as SSL client,
17106 server or S/MIME and CAs of these types. This is currently
17107 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17108 verification. Also added a -purpose flag to x509 utility to
17109 print out all the purposes.
17113 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17118 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17119 for, obtain and decode and extension and obtain its critical flag.
17120 This allows all the necessary extension code to be handled in a
17121 single function call.
17125 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17126 platforms. See crypto/rc4/rc4_enc.c for further details.
17130 * New -noout option to asn1parse. This causes no output to be produced
17131 its main use is when combined with -strparse and -out to extract data
17132 from a file (which may not be in ASN.1 format).
17136 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17137 when producing the local key id.
17139 *Richard Levitte <levitte@stacken.kth.se>*
17141 * New option -dhparam in s_server. This allows a DH parameter file to be
17142 stated explicitly. If it is not stated then it tries the first server
17143 certificate file. The previous behaviour hard coded the filename
17148 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17149 a public key to be input or output. For example:
17150 openssl rsa -in key.pem -pubout -out pubkey.pem
17151 Also added necessary DSA public key functions to handle this.
17155 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17156 in the message. This was handled by allowing
17157 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17159 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17161 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17162 to the end of the strings whereas this didn't. This would cause problems
17163 if strings read with d2i_ASN1_bytes() were later modified.
17165 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17167 * Fix for base64 decode bug. When a base64 bio reads only one line of
17168 data and it contains EOF it will end up returning an error. This is
17169 caused by input 46 bytes long. The cause is due to the way base64
17170 BIOs find the start of base64 encoded data. They do this by trying a
17171 trial decode on each line until they find one that works. When they
17172 do a flag is set and it starts again knowing it can pass all the
17173 data directly through the decoder. Unfortunately it doesn't reset
17174 the context it uses. This means that if EOF is reached an attempt
17175 is made to pass two EOFs through the context and this causes the
17176 resulting error. This can also cause other problems as well. As is
17177 usual with these problems it takes *ages* to find and the fix is
17178 trivial: move one line.
17180 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17182 * Ugly workaround to get s_client and s_server working under Windows. The
17183 old code wouldn't work because it needed to select() on sockets and the
17184 tty (for keypresses and to see if data could be written). Win32 only
17185 supports select() on sockets so we select() with a 1s timeout on the
17186 sockets and then see if any characters are waiting to be read, if none
17187 are present then we retry, we also assume we can always write data to
17188 the tty. This isn't nice because the code then blocks until we've
17189 received a complete line of data and it is effectively polling the
17190 keyboard at 1s intervals: however it's quite a bit better than not
17191 working at all :-) A dedicated Windows application might handle this
17192 with an event loop for example.
17196 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17197 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17198 will be called when RSA_sign() and RSA_verify() are used. This is useful
17199 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17200 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17201 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17202 This necessitated the support of an extra signature type NID_md5_sha1
17203 for SSL signatures and modifications to the SSL library to use it instead
17204 of calling RSA_public_decrypt() and RSA_private_encrypt().
17208 * Add new -verify -CAfile and -CApath options to the crl program, these
17209 will lookup a CRL issuers certificate and verify the signature in a
17210 similar way to the verify program. Tidy up the crl program so it
17211 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17212 less strict. It will now permit CRL extensions even if it is not
17213 a V2 CRL: this will allow it to tolerate some broken CRLs.
17217 * Initialize all non-automatic variables each time one of the openssl
17218 sub-programs is started (this is necessary as they may be started
17219 multiple times from the "OpenSSL>" prompt).
17221 *Lennart Bang, Bodo Moeller*
17223 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17224 removing all other RSA functionality (this is what NO_RSA does). This
17225 is so (for example) those in the US can disable those operations covered
17226 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17231 * Non-copying interface to BIO pairs.
17232 (still largely untested)
17236 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17237 ASCII string. This was handled independently in various places before.
17241 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17242 UTF8 strings a character at a time.
17246 * Use client_version from client hello to select the protocol
17247 (s23_srvr.c) and for RSA client key exchange verification
17248 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17252 * Add various utility functions to handle SPKACs, these were previously
17253 handled by poking round in the structure internals. Added new function
17254 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17255 print, verify and generate SPKACs. Based on an original idea from
17256 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17260 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17264 * Allow the config file extension section to be overwritten on the
17265 command line. Based on an original idea from Massimiliano Pala
17266 <madwolf@comune.modena.it>. The new option is called -extensions
17267 and can be applied to ca, req and x509. Also -reqexts to override
17268 the request extensions in req and -crlexts to override the crl extensions
17273 * Add new feature to the SPKAC handling in ca. Now you can include
17274 the same field multiple times by preceding it by "XXXX." for example:
17277 this is the same syntax as used in the req config file.
17281 * Allow certificate extensions to be added to certificate requests. These
17282 are specified in a 'req_extensions' option of the req section of the
17283 config file. They can be printed out with the -text option to req but
17284 are otherwise ignored at present.
17288 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17289 data read consists of only the final block it would not decrypted because
17290 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17291 A misplaced 'break' also meant the decrypted final block might not be
17292 copied until the next read.
17296 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17297 a few extra parameters to the DH structure: these will be useful if
17298 for example we want the value of 'q' or implement X9.42 DH.
17302 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17303 provides hooks that allow the default DSA functions or functions on a
17304 "per key" basis to be replaced. This allows hardware acceleration and
17305 hardware key storage to be handled without major modification to the
17306 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17307 associated functions.
17311 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17312 as "read only": it can't be written to and the buffer it points to will
17313 not be freed. Reading from a read only BIO is much more efficient than
17314 a normal memory BIO. This was added because there are several times when
17315 an area of memory needs to be read from a BIO. The previous method was
17316 to create a memory BIO and write the data to it, this results in two
17317 copies of the data and an O(n^2) reading algorithm. There is a new
17318 function BIO_new_mem_buf() which creates a read only memory BIO from
17319 an area of memory. Also modified the PKCS#7 routines to use read only
17324 * Bugfix: ssl23_get_client_hello did not work properly when called in
17325 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17326 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17327 but a retry condition occurred while trying to read the rest.
17331 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17332 NID_pkcs7_encrypted by default: this was wrong since this should almost
17333 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17334 the encrypted data type: this is a more sensible place to put it and it
17335 allows the PKCS#12 code to be tidied up that duplicated this
17340 * Changed obj_dat.pl script so it takes its input and output files on
17341 the command line. This should avoid shell escape redirection problems
17346 * Initial support for certificate extension requests, these are included
17347 in things like Xenroll certificate requests. Included functions to allow
17348 extensions to be obtained and added.
17352 * -crlf option to s_client and s_server for sending newlines as
17353 CRLF (as required by many protocols).
17357 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17359 * Install libRSAglue.a when OpenSSL is built with RSAref.
17361 *Ralf S. Engelschall*
17363 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17365 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17367 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17372 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17373 DH parameters/keys (q is lost during that conversion, but the resulting
17374 DH parameters contain its length).
17376 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17377 much faster than DH_generate_parameters (which creates parameters
17378 where `p = 2*q + 1`), and also the smaller q makes DH computations
17379 much more efficient (160-bit exponentiation instead of 1024-bit
17380 exponentiation); so this provides a convenient way to support DHE
17381 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17382 utter importance to use
17383 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17385 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17386 when such DH parameters are used, because otherwise small subgroup
17387 attacks may become possible!
17391 * Avoid memory leak in i2d_DHparams.
17395 * Allow the -k option to be used more than once in the enc program:
17396 this allows the same encrypted message to be read by multiple recipients.
17400 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17401 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17402 it will always use the numerical form of the OID, even if it has a short
17407 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17408 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17409 otherwise bn_mod_exp was called. In the case of hardware keys for example
17410 no private key components need be present and it might store extra data
17411 in the RSA structure, which cannot be accessed from bn_mod_exp.
17412 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17413 private key operations.
17417 * Added support for SPARC Linux.
17421 * pem_password_cb function type incompatibly changed from
17422 typedef int pem_password_cb(char *buf, int size, int rwflag);
17424 ....(char *buf, int size, int rwflag, void *userdata);
17425 so that applications can pass data to their callbacks:
17426 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17427 additional void * argument, which is just handed through whenever
17428 the password callback is called.
17430 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17432 New function SSL_CTX_set_default_passwd_cb_userdata.
17434 Compatibility note: As many C implementations push function arguments
17435 onto the stack in reverse order, the new library version is likely to
17436 interoperate with programs that have been compiled with the old
17437 pem_password_cb definition (PEM_whatever takes some data that
17438 happens to be on the stack as its last argument, and the callback
17439 just ignores this garbage); but there is no guarantee whatsoever that
17442 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17443 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17444 problems not only on Windows, but also on some Unix platforms.
17445 To avoid problematic command lines, these definitions are now in an
17446 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17447 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17451 * MIPS III/IV assembler module is reimplemented.
17455 * More DES library cleanups: remove references to srand/rand and
17456 delete an unused file.
17460 * Add support for the free Netwide assembler (NASM) under Win32,
17461 since not many people have MASM (ml) and it can be hard to obtain.
17462 This is currently experimental but it seems to work OK and pass all
17463 the tests. Check out INSTALL.W32 for info.
17467 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17468 without temporary keys kept an extra copy of the server key,
17469 and connections with temporary keys did not free everything in case
17474 * New function RSA_check_key and new openssl rsa option -check
17475 for verifying the consistency of RSA keys.
17477 *Ulf Moeller, Bodo Moeller*
17479 * Various changes to make Win32 compile work:
17480 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17481 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17482 comparison" warnings.
17483 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17487 * Add a debugging option to PKCS#5 v2 key generation function: when
17488 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17489 derived keys are printed to stderr.
17493 * Copy the flags in ASN1_STRING_dup().
17495 *Roman E. Pavlov <pre@mo.msk.ru>*
17497 * The x509 application mishandled signing requests containing DSA
17498 keys when the signing key was also DSA and the parameters didn't match.
17500 It was supposed to omit the parameters when they matched the signing key:
17501 the verifying software was then supposed to automatically use the CA's
17502 parameters if they were absent from the end user certificate.
17504 Omitting parameters is no longer recommended. The test was also
17505 the wrong way round! This was probably due to unusual behaviour in
17506 EVP_cmp_parameters() which returns 1 if the parameters match.
17507 This meant that parameters were omitted when they *didn't* match and
17508 the certificate was useless. Certificates signed with 'ca' didn't have
17511 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17513 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17514 The interface is as follows:
17515 Applications can use
17516 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17517 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17518 "off" is now the default.
17519 The library internally uses
17520 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17521 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17522 to disable memory-checking temporarily.
17524 Some inconsistent states that previously were possible (and were
17525 even the default) are now avoided.
17527 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17528 with each memory chunk allocated; this is occasionally more helpful
17529 than just having a counter.
17531 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17533 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17538 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17539 which largely parallels "options", but is for changing API behaviour,
17540 whereas "options" are about protocol behaviour.
17541 Initial "mode" flags are:
17543 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17544 a single record has been written.
17545 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17546 retries use the same buffer location.
17547 (But all of the contents must be
17552 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17555 * Fix problems with no-hmac etc.
17557 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17559 * New functions RSA_get_default_method(), RSA_set_method() and
17560 RSA_get_method(). These allows replacement of RSA_METHODs without having
17561 to mess around with the internals of an RSA structure.
17565 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17566 Also really enable memory leak checks in openssl.c and in some
17569 *Chad C. Mulligan, Bodo Moeller*
17571 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17572 up the length of negative integers. This has now been simplified to just
17573 store the length when it is first determined and use it later, rather
17574 than trying to keep track of where data is copied and updating it to
17576 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17578 * Add a new function PKCS7_signatureVerify. This allows the verification
17579 of a PKCS#7 signature but with the signing certificate passed to the
17580 function itself. This contrasts with PKCS7_dataVerify which assumes the
17581 certificate is present in the PKCS#7 structure. This isn't always the
17582 case: certificates can be omitted from a PKCS#7 structure and be
17583 distributed by "out of band" means (such as a certificate database).
17587 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17588 function prototypes in pem.h, also change util/mkdef.pl to add the
17589 necessary function names.
17593 * mk1mf.pl (used by Windows builds) did not properly read the
17594 options set by Configure in the top level Makefile, and Configure
17595 was not even able to write more than one option correctly.
17596 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17600 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17601 file to be loaded from a BIO or FILE pointer. The BIO version will
17602 for example allow memory BIOs to contain config info.
17606 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17607 Whoever hopes to achieve shared-library compatibility across versions
17608 must use this, not the compile-time macro.
17609 (Exercise 0.9.4: Which is the minimum library version required by
17611 Note: All this applies only to multi-threaded programs, others don't
17616 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17617 through a BIO pair triggered the default case, i.e.
17618 SSLerr(...,SSL_R_UNKNOWN_STATE).
17622 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17623 can use the SSL library even if none of the specific BIOs is
17628 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17629 for the encoded length.
17631 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17633 * Add initial documentation of the X509V3 functions.
17637 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17638 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17639 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17640 secure PKCS#8 private key format with a high iteration count.
17644 * Fix determination of Perl interpreter: A perl or perl5
17645 *directory* in $PATH was also accepted as the interpreter.
17647 *Ralf S. Engelschall*
17649 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17650 wrong with it but it was very old and did things like calling
17651 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17652 unusual formatting.
17656 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17657 to use the new extension code.
17661 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17662 with macros. This should make it easier to change their form, add extra
17663 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17668 * Add to configuration table a new entry that can specify an alternative
17669 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17670 according to Mark Crispin <MRC@Panda.COM>.
17674 * DES CBC did not update the IV. Weird.
17678 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17679 Changing the behaviour of the former might break existing programs --
17680 where IV updating is needed, des_ncbc_encrypt can be used.
17683 * When bntest is run from "make test" it drives bc to check its
17684 calculations, as well as internally checking them. If an internal check
17685 fails, it needs to cause bc to give a non-zero result or make test carries
17686 on without noticing the failure. Fixed.
17690 * DES library cleanups.
17694 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17695 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17696 ciphers. NOTE: although the key derivation function has been verified
17697 against some published test vectors it has not been extensively tested
17698 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17703 * Instead of "mkdir -p", which is not fully portable, use new
17704 Perl script "util/mkdir-p.pl".
17708 * Rewrite the way password based encryption (PBE) is handled. It used to
17709 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17710 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17711 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17712 the 'parameter' field of the AlgorithmIdentifier is passed to the
17713 underlying key generation function so it must do its own ASN1 parsing.
17714 This has also changed the EVP_PBE_CipherInit() function which now has a
17715 'parameter' argument instead of literal salt and iteration count values
17716 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17720 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17721 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17722 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17723 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17724 value was just used as a "magic string" and not used directly its
17725 value doesn't matter.
17729 * Introduce some semblance of const correctness to BN. Shame C doesn't
17734 * "linux-sparc64" configuration (ultrapenguin).
17736 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17737 "linux-sparc" configuration.
17739 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17741 * config now generates no-xxx options for missing ciphers.
17745 * Support the EBCDIC character set (work in progress).
17746 File ebcdic.c not yet included because it has a different license.
17748 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17750 * Support BS2000/OSD-POSIX.
17752 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17754 * Make callbacks for key generation use `void *` instead of `char *`.
17758 * Make S/MIME samples compile (not yet tested).
17762 * Additional typesafe stacks.
17766 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17770 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17772 * New configuration variant "sco5-gcc".
17774 * Updated some demos.
17776 *Sean O Riordain, Wade Scholine*
17778 * Add missing BIO_free at exit of pkcs12 application.
17782 * Fix memory leak in conf.c.
17786 * Updates for Win32 to assembler version of MD5.
17790 * Set #! path to perl in `apps/der_chop` to where we found it
17791 instead of using a fixed path.
17795 * SHA library changes for irix64-mips4-cc.
17799 * Improvements for VMS support.
17803 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17805 * Bignum library bug fix. IRIX 6 passes "make test" now!
17806 This also avoids the problems with SC4.2 and unpatched SC5.
17808 *Andy Polyakov <appro@fy.chalmers.se>*
17810 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17811 These are required because of the typesafe stack would otherwise break
17812 existing code. If old code used a structure member which used to be STACK
17813 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17814 sk_num or sk_value it would produce an error because the num, data members
17815 are not present in STACK_OF. Now it just produces a warning. sk_set
17816 replaces the old method of assigning a value to sk_value
17817 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17818 that does this will no longer work (and should use sk_set instead) but
17819 this could be regarded as a "questionable" behaviour anyway.
17823 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17824 correctly handle encrypted S/MIME data.
17828 * Change type of various DES function arguments from des_cblock
17829 (which means, in function argument declarations, pointer to char)
17830 to des_cblock * (meaning pointer to array with 8 char elements),
17831 which allows the compiler to do more typechecking; it was like
17832 that back in SSLeay, but with lots of ugly casts.
17834 Introduce new type const_des_cblock.
17838 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17839 problems: find RecipientInfo structure that matches recipient certificate
17840 and initialise the ASN1 structures properly based on passed cipher.
17844 * Belatedly make the BN tests actually check the results.
17848 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17849 to and from BNs: it was completely broken. New compilation option
17850 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17851 key elements as negative integers.
17855 * Reorganize and speed up MD5.
17857 *Andy Polyakov <appro@fy.chalmers.se>*
17861 *Richard Levitte <richard@levitte.org>*
17863 * New option -out to asn1parse to allow the parsed structure to be
17864 output to a file. This is most useful when combined with the -strparse
17865 option to examine the output of things like OCTET STRINGS.
17869 * Make SSL library a little more fool-proof by not requiring any longer
17870 that `SSL_set_{accept,connect}_state` be called before
17871 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17872 in many applications because usually everything *appeared* to work as
17873 intended anyway -- now it really works as intended).
17877 * Move openssl.cnf out of lib/.
17881 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17882 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17883 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17885 *Ralf S. Engelschall*
17887 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17888 handle PKCS#7 enveloped data properly.
17890 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17892 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17893 copying pointers. The cert_st handling is changed by this in
17894 various ways (and thus what used to be known as ctx->default_cert
17895 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17896 any longer when s->cert does not give us what we need).
17897 ssl_cert_instantiate becomes obsolete by this change.
17898 As soon as we've got the new code right (possibly it already is?),
17899 we have solved a couple of bugs of the earlier code where s->cert
17900 was used as if it could not have been shared with other SSL structures.
17902 Note that using the SSL API in certain dirty ways now will result
17903 in different behaviour than observed with earlier library versions:
17904 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17905 does not influence s as it used to.
17907 In order to clean up things more thoroughly, inside SSL_SESSION
17908 we don't use CERT any longer, but a new structure SESS_CERT
17909 that holds per-session data (if available); currently, this is
17910 the peer's certificate chain and, for clients, the server's certificate
17911 and temporary key. CERT holds only those values that can have
17912 meaningful defaults in an SSL_CTX.
17916 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17917 from the internal representation. Various PKCS#7 fixes: remove some
17918 evil casts and set the enc_dig_alg field properly based on the signing
17923 * Allow PKCS#12 password to be set from the command line or the
17924 environment. Let 'ca' get its config file name from the environment
17925 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17930 * Allow certificate policies extension to use an IA5STRING for the
17931 organization field. This is contrary to the PKIX definition but
17932 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17937 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17938 without disallowing inline assembler and the like for non-pedantic builds.
17942 * Support Borland C++ builder.
17944 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17950 * SHA-1 cleanups and performance enhancements.
17952 *Andy Polyakov <appro@fy.chalmers.se>*
17954 * Sparc v8plus assembler for the bignum library.
17956 *Andy Polyakov <appro@fy.chalmers.se>*
17958 * Accept any -xxx and +xxx compiler options in Configure.
17962 * Update HPUX configuration.
17966 * Add missing `sk_<type>_unshift()` function to safestack.h
17968 *Ralf S. Engelschall*
17970 * New function SSL_CTX_use_certificate_chain_file that sets the
17971 "extra_cert"s in addition to the certificate. (This makes sense
17972 only for "PEM" format files, as chains as a whole are not
17977 * Support verify_depth from the SSL API.
17978 x509_vfy.c had what can be considered an off-by-one-error:
17979 Its depth (which was not part of the external interface)
17980 was actually counting the number of certificates in a chain;
17981 now it really counts the depth.
17985 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17986 instead of X509err, which often resulted in confusing error
17987 messages since the error codes are not globally unique
17988 (e.g. an alleged error in ssl3_accept when a certificate
17989 didn't match the private key).
17991 * New function SSL_CTX_set_session_id_context that allows to set a default
17992 value (so that you don't need SSL_set_session_id_context for each
17993 connection using the SSL_CTX).
17997 * OAEP decoding bug fix.
18001 * Support INSTALL_PREFIX for package builders, as proposed by
18006 * New Configure options "threads" and "no-threads". For systems
18007 where the proper compiler options are known (currently Solaris
18008 and Linux), "threads" is the default.
18012 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18016 * Install various scripts to $(OPENSSLDIR)/misc, not to
18017 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18018 such as /usr/local/bin.
18022 * "make linux-shared" to build shared libraries.
18024 *Niels Poppe <niels@netbox.org>*
18026 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18030 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18031 extension adding in x509 utility.
18035 * Remove NOPROTO sections and error code comments.
18039 * Partial rewrite of the DEF file generator to now parse the ANSI
18044 * New Configure options --prefix=DIR and --openssldir=DIR.
18048 * Complete rewrite of the error code script(s). It is all now handled
18049 by one script at the top level which handles error code gathering,
18050 header rewriting and C source file generation. It should be much better
18051 than the old method: it now uses a modified version of Ulf's parser to
18052 read the ANSI prototypes in all header files (thus the old K&R definitions
18053 aren't needed for error creation any more) and do a better job of
18054 translating function codes into names. The old 'ASN1 error code embedded
18055 in a comment' is no longer necessary and it doesn't use .err files which
18056 have now been deleted. Also the error code call doesn't have to appear all
18057 on one line (which resulted in some large lines...).
18061 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18065 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18066 0 (which usually indicates a closed connection), but continue reading.
18070 * Fix some race conditions.
18074 * Add support for CRL distribution points extension. Add Certificate
18075 Policies and CRL distribution points documentation.
18079 * Move the autogenerated header file parts to crypto/opensslconf.h.
18083 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18084 8 of keying material. Merlin has also confirmed interop with this fix
18085 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18087 *Merlin Hughes <merlin@baltimore.ie>*
18089 * Fix lots of warnings.
18091 *Richard Levitte <levitte@stacken.kth.se>*
18093 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18094 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18096 *Richard Levitte <levitte@stacken.kth.se>*
18098 * Fix problems with sizeof(long) == 8.
18100 *Andy Polyakov <appro@fy.chalmers.se>*
18102 * Change functions to ANSI C.
18106 * Fix typos in error codes.
18108 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18110 * Remove defunct assembler files from Configure.
18114 * SPARC v8 assembler BIGNUM implementation.
18116 *Andy Polyakov <appro@fy.chalmers.se>*
18118 * Support for Certificate Policies extension: both print and set.
18119 Various additions to support the r2i method this uses.
18123 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18124 return a const string when you are expecting an allocated buffer.
18128 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18129 types DirectoryString and DisplayText.
18133 * Add code to allow r2i extensions to access the configuration database,
18134 add an LHASH database driver and add several ctx helper functions.
18138 * Fix an evil bug in bn_expand2() which caused various BN functions to
18139 fail when they extended the size of a BIGNUM.
18143 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18144 support typesafe stack.
18148 * Fix typo in SSL_[gs]et_options().
18150 *Nils Frostberg <nils@medcom.se>*
18152 * Delete various functions and files that belonged to the (now obsolete)
18153 old X509V3 handling code.
18157 * New Configure option "rsaref".
18161 * Don't auto-generate pem.h.
18165 * Introduce type-safe ASN.1 SETs.
18169 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18171 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18173 * Introduce type-safe STACKs. This will almost certainly break lots of code
18174 that links with OpenSSL (well at least cause lots of warnings), but fear
18175 not: the conversion is trivial, and it eliminates loads of evil casts. A
18176 few STACKed things have been converted already. Feel free to convert more.
18177 In the fullness of time, I'll do away with the STACK type altogether.
18181 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18182 specified in `<certfile>` by updating the entry in the index.txt file.
18183 This way one no longer has to edit the index.txt file manually for
18184 revoking a certificate. The -revoke option does the gory details now.
18186 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18188 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18189 `-text` option at all and this way the `-noout -text` combination was
18190 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18192 *Ralf S. Engelschall*
18194 * Make sure a corresponding plain text error message exists for the
18195 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18196 verify callback function determined that a certificate was revoked.
18198 *Ralf S. Engelschall*
18200 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18201 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18202 all available ciphers including rc5, which was forgotten until now.
18203 In order to let the testing shell script know which algorithms
18204 are available, a new (up to now undocumented) command
18205 `openssl list-cipher-commands` is used.
18209 * Bugfix: s_client occasionally would sleep in select() when
18210 it should have checked SSL_pending() first.
18214 * New functions DSA_do_sign and DSA_do_verify to provide access to
18215 the raw DSA values prior to ASN.1 encoding.
18219 * Tweaks to Configure
18221 *Niels Poppe <niels@netbox.org>*
18223 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18228 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18232 * New config option to avoid instructions that are illegal on the 80386.
18233 The default code is faster, but requires at least a 486.
18237 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18238 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18239 same as SSL2_VERSION anyway.
18243 * New "-showcerts" option for s_client.
18247 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18248 application. Various cleanups and fixes.
18252 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18253 modify error routines to work internally. Add error codes and PBE init
18254 to library startup routines.
18258 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18259 packing functions to asn1 and evp. Changed function names and error
18260 codes along the way.
18264 * PKCS12 integration: and so it begins... First of several patches to
18265 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18266 objects to objects.h
18270 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18271 and display support for Thawte strong extranet extension.
18275 * Add LinuxPPC support.
18277 *Jeff Dubrule <igor@pobox.org>*
18279 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18280 bn_div_words in alpha.s.
18282 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18284 * Make sure the RSA OAEP test is skipped under -DRSAref because
18285 OAEP isn't supported when OpenSSL is built with RSAref.
18287 *Ulf Moeller <ulf@fitug.de>*
18289 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18290 so they no longer are missing under -DNOPROTO.
18292 *Soren S. Jorvang <soren@t.dk>*
18294 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18296 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18297 doesn't work when the session is reused. Coming soon!
18301 * Fix a security hole, that allows sessions to be reused in the wrong
18302 context thus bypassing client cert protection! All software that uses
18303 client certs and session caches in multiple contexts NEEDS PATCHING to
18304 allow session reuse! A fuller solution is in the works.
18306 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18308 * Some more source tree cleanups (removed obsolete files
18309 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18310 permission on "config" script to be executable) and a fix for the INSTALL
18313 *Ulf Moeller <ulf@fitug.de>*
18315 * Remove some legacy and erroneous uses of malloc, free instead of
18318 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18320 * Make rsa_oaep_test return non-zero on error.
18322 *Ulf Moeller <ulf@fitug.de>*
18324 * Add support for native Solaris shared libraries. Configure
18325 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18326 if someone would make that last step automatic.
18328 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18330 * ctx_size was not built with the right compiler during "make links". Fixed.
18334 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18335 except NULL ciphers". This means the default cipher list will no longer
18336 enable NULL ciphers. They need to be specifically enabled e.g. with
18337 the string "DEFAULT:eNULL".
18341 * Fix to RSA private encryption routines: if p < q then it would
18342 occasionally produce an invalid result. This will only happen with
18343 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18347 * Be less restrictive and allow also `perl util/perlpath.pl
18348 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18349 because this way one can also use an interpreter named `perl5` (which is
18350 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18351 installed as `perl`).
18353 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18355 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18357 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18359 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18360 advapi32.lib to Win32 build and change the pem test comparison
18361 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18362 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18363 and crypto/des/ede_cbcm_enc.c.
18367 * DES quad checksum was broken on big-endian architectures. Fixed.
18371 * Comment out two functions in bio.h that aren't implemented. Fix up the
18372 Win32 test batch file so it (might) work again. The Win32 test batch file
18373 is horrible: I feel ill....
18377 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18378 in e_os.h. Audit of header files to check ANSI and non ANSI
18379 sections: 10 functions were absent from non ANSI section and not exported
18380 from Windows DLLs. Fixed up libeay.num for new functions.
18384 * Make `openssl version` output lines consistent.
18386 *Ralf S. Engelschall*
18388 * Fix Win32 symbol export lists for BIO functions: Added
18389 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18390 to ms/libeay{16,32}.def.
18392 *Ralf S. Engelschall*
18394 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18395 fine under Unix and passes some trivial tests I've now added. But the
18396 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18397 added to make sure no one expects that this stuff really works in the
18398 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18399 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18402 *Ralf S. Engelschall*
18404 * Fix the generation of two part addresses in perl.
18406 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18408 * Add config entry for Linux on MIPS.
18410 *John Tobey <jtobey@channel1.com>*
18412 * Make links whenever Configure is run, unless we are on Windoze.
18416 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18417 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18422 * Add a useful kludge to allow package maintainers to specify compiler and
18423 other platforms details on the command line without having to patch the
18424 Configure script every time: One now can use
18425 `perl Configure <id>:<details>`,
18426 i.e. platform ids are allowed to have details appended
18427 to them (separated by colons). This is treated as there would be a static
18428 pre-configured entry in Configure's %table under key `<id>` with value
18429 `<details>` and `perl Configure <id>` is called. So, when you want to
18430 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18431 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18432 now, which overrides the FreeBSD-elf entry on-the-fly.
18434 *Ralf S. Engelschall*
18436 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18440 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18441 on the `perl Configure ...` command line. This way one can compile
18442 OpenSSL libraries with Position Independent Code (PIC) which is needed
18443 for linking it into DSOs.
18445 *Ralf S. Engelschall*
18447 * Remarkably, export ciphers were totally broken and no-one had noticed!
18452 * Cleaned up the LICENSE document: The official contact for any license
18453 questions now is the OpenSSL core team under openssl-core@openssl.org.
18454 And add a paragraph about the dual-license situation to make sure people
18455 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18456 to the OpenSSL toolkit.
18458 *Ralf S. Engelschall*
18460 * General source tree makefile cleanups: Made `making xxx in yyy...`
18461 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18462 Additionally cleaned up the `make links` target: Remove unnecessary
18463 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18464 to speed processing and no longer clutter the display with confusing
18465 stuff. Instead only the actually done links are displayed.
18467 *Ralf S. Engelschall*
18469 * Permit null encryption ciphersuites, used for authentication only. It used
18470 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18471 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18476 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18477 signed attributes when verifying signatures (this would break them),
18478 the detached data encoding was wrong and public keys obtained using
18479 X509_get_pubkey() weren't freed.
18483 * Add text documentation for the BUFFER functions. Also added a work around
18484 to a Win95 console bug. This was triggered by the password read stuff: the
18485 last character typed gets carried over to the next fread(). If you were
18486 generating a new cert request using 'req' for example then the last
18487 character of the passphrase would be CR which would then enter the first
18492 * Added the new 'Includes OpenSSL Cryptography Software' button as
18493 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18494 button and can be used by applications based on OpenSSL to show the
18495 relationship to the OpenSSL project.
18497 *Ralf S. Engelschall*
18499 * Remove confusing variables in function signatures in files
18500 ssl/ssl_lib.c and ssl/ssl.h.
18502 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18504 * Don't install bss_file.c under PREFIX/include/
18506 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18508 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18509 functions that return function pointers and has support for NT specific
18510 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18511 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18512 unsigned to signed types: this was killing the Win32 compile.
18516 * Add new certificate file to stack functions,
18517 SSL_add_dir_cert_subjects_to_stack() and
18518 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18519 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18520 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18521 This means that Apache-SSL and similar packages don't have to mess around
18522 to add as many CAs as they want to the preferred list.
18526 * Experiment with doxygen documentation. Currently only partially applied to
18528 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18529 openssl.doxy as the configuration file.
18533 * Get rid of remaining C++-style comments which strict C compilers hate.
18535 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18537 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18538 compiled in by default: it has problems with large keys.
18542 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18543 DH private keys and/or callback functions which directly correspond to
18544 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18545 is needed for applications which have to configure certificates on a
18546 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18548 For the RSA certificate situation is makes no difference, but
18549 for the DSA certificate situation this fixes the "no shared cipher"
18550 problem where the OpenSSL cipher selection procedure failed because the
18551 temporary keys were not overtaken from the context and the API provided
18552 no way to reconfigure them.
18553 The new functions now let applications reconfigure the stuff and they
18554 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18555 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18556 non-public-API function ssl_cert_instantiate() is used as a helper
18557 function and also to reduce code redundancy inside ssl_rsa.c.
18559 *Ralf S. Engelschall*
18561 * Move s_server -dcert and -dkey options out of the undocumented feature
18562 area because they are useful for the DSA situation and should be
18563 recognized by the users.
18565 *Ralf S. Engelschall*
18567 * Fix the cipher decision scheme for export ciphers: the export bits are
18568 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18569 SSL_EXP_MASK. So, the original variable has to be used instead of the
18570 already masked variable.
18572 *Richard Levitte <levitte@stacken.kth.se>*
18574 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18576 *Richard Levitte <levitte@stacken.kth.se>*
18578 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18579 from `int` to `unsigned int` because it is a length and initialized by
18580 EVP_DigestFinal() which expects an `unsigned int *`.
18582 *Richard Levitte <levitte@stacken.kth.se>*
18584 * Don't hard-code path to Perl interpreter on shebang line of Configure
18585 script. Instead use the usual Shell->Perl transition trick.
18587 *Ralf S. Engelschall*
18589 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18590 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18591 -noout -modulus` as it's already the case for `openssl rsa -noout
18592 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18593 currently the public key is printed (a decision which was already done by
18594 `openssl dsa -modulus` in the past) which serves a similar purpose.
18595 Additionally the NO_RSA no longer completely removes the whole -modulus
18596 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18599 *Ralf S. Engelschall*
18601 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18602 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18604 *Arne Ansper <arne@ats.cyber.ee>*
18606 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18607 to be added. Now both 'req' and 'ca' can use new objects defined in the
18612 * Add cool BIO that does syslog (or event log on NT).
18614 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18616 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18617 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18618 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18619 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18623 * Add preliminary config info for new extension code.
18627 * Make RSA_NO_PADDING really use no padding.
18629 *Ulf Moeller <ulf@fitug.de>*
18631 * Generate errors when private/public key check is done.
18635 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18636 for some CRL extensions and new objects added.
18640 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18641 key usage extension and fuller support for authority key id.
18645 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18646 padding method for RSA, which is recommended for new applications in PKCS
18647 #1 v2.0 (RFC 2437, October 1998).
18648 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18649 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18650 against Bleichbacher's attack on RSA.
18651 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18654 * Updates to the new SSL compression code
18656 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18658 * Fix so that the version number in the master secret, when passed
18659 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18660 (because the server will not accept higher), that the version number
18661 is 0x03,0x01, not 0x03,0x00
18663 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18665 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18666 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18667 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18671 * Support for RAW extensions where an arbitrary extension can be
18672 created by including its DER encoding. See `apps/openssl.cnf` for
18677 * Make sure latest Perl versions don't interpret some generated C array
18678 code as Perl array code in the crypto/err/err_genc.pl script.
18680 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18682 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18683 not many people have the assembler. Various Win32 compilation fixes and
18684 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18685 build instructions.
18689 * Modify configure script 'Configure' to automatically create crypto/date.h
18690 file under Win32 and also build pem.h from pem.org. New script
18691 util/mkfiles.pl to create the MINFO file on environments that can't do a
18692 'make files': perl util/mkfiles.pl >MINFO should work.
18696 * Major rework of DES function declarations, in the pursuit of correctness
18697 and purity. As a result, many evil casts evaporated, and some weirdness,
18698 too. You may find this causes warnings in your code. Zapping your evil
18699 casts will probably fix them. Mostly.
18703 * Fix for a typo in asn1.h. Bug fix to object creation script
18704 obj_dat.pl. It considered a zero in an object definition to mean
18705 "end of object": none of the objects in objects.h have any zeros
18706 so it wasn't spotted.
18708 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18710 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18711 Masking (CBCM). In the absence of test vectors, the best I have been able
18712 to do is check that the decrypt undoes the encrypt, so far. Send me test
18713 vectors if you have them.
18717 * Correct calculation of key length for export ciphers (too much space was
18718 allocated for null ciphers). This has not been tested!
18722 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18723 message is now correct (it understands "crypto" and "ssl" on its
18724 command line). There is also now an "update" option. This will update
18725 the util/ssleay.num and util/libeay.num files with any new functions.
18727 perl util/mkdef.pl crypto ssl update
18728 it will update them.
18732 * Overhauled the Perl interface:
18733 - ported BN stuff to OpenSSL's different BN library
18734 - made the perl/ source tree CVS-aware
18735 - renamed the package from SSLeay to OpenSSL (the files still contain
18736 their history because I've copied them in the repository)
18737 - removed obsolete files (the test scripts will be replaced
18738 by better Test::Harness variants in the future)
18740 *Ralf S. Engelschall*
18742 * First cut for a very conservative source tree cleanup:
18743 1. merge various obsolete readme texts into doc/ssleay.txt
18744 where we collect the old documents and readme texts.
18745 2. remove the first part of files where I'm already sure that we no
18746 longer need them because of three reasons: either they are just temporary
18747 files which were left by Eric or they are preserved original files where
18748 I've verified that the diff is also available in the CVS via "cvs diff
18749 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18750 the crypto/md/ stuff).
18752 *Ralf S. Engelschall*
18754 * More extension code. Incomplete support for subject and issuer alt
18755 name, issuer and authority key id. Change the i2v function parameters
18756 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18757 what that's for :-) Fix to ASN1 macro which messed up
18758 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18762 * Preliminary support for ENUMERATED type. This is largely copied from the
18767 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18769 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18771 * Make sure `make rehash` target really finds the `openssl` program.
18773 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18775 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18776 like to hear about it if this slows down other processors.
18780 * Add CygWin32 platform information to Configure script.
18782 *Alan Batie <batie@aahz.jf.intel.com>*
18784 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18786 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18788 * New program nseq to manipulate netscape certificate sequences
18792 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18797 * Fixes to BN code. Previously the default was to define BN_RECURSION
18798 but the BN code had some problems that would cause failures when
18799 doing certificate verification and some other functions.
18801 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18803 * Add ASN1 and PEM code to support netscape certificate sequences.
18807 * Add ASN1 and PEM code to support netscape certificate sequences.
18811 * Add several PKIX and private extended key usage OIDs.
18815 * Modify the 'ca' program to handle the new extension code. Modify
18816 openssl.cnf for new extension format, add comments.
18820 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18821 and add a sample to openssl.cnf so req -x509 now adds appropriate
18826 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18827 error code, add initial support to X509_print() and x509 application.
18831 * Takes a deep breath and start adding X509 V3 extension support code. Add
18832 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18833 stuff is currently isolated and isn't even compiled yet.
18837 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18838 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18839 Removed the versions check from X509 routines when loading extensions:
18840 this allows certain broken certificates that don't set the version
18841 properly to be processed.
18845 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18846 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18847 can still be regenerated with "make depend".
18851 * Spelling mistake in C version of CAST-128.
18853 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18855 * Changes to the error generation code. The perl script err-code.pl
18856 now reads in the old error codes and retains the old numbers, only
18857 adding new ones if necessary. It also only changes the .err files if new
18858 codes are added. The makefiles have been modified to only insert errors
18859 when needed (to avoid needlessly modifying header files). This is done
18860 by only inserting errors if the .err file is newer than the auto generated
18861 C file. To rebuild all the error codes from scratch (the old behaviour)
18862 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18863 or delete all the .err files.
18867 * CAST-128 was incorrectly implemented for short keys. The C version has
18868 been fixed, but is untested. The assembler versions are also fixed, but
18869 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18870 to regenerate it if needed.
18871 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18872 Hagino <itojun@kame.net>*
18874 * File was opened incorrectly in randfile.c.
18876 *Ulf Möller <ulf@fitug.de>*
18878 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18879 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18880 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18881 al: it's just almost always a UTCTime. Note this patch adds new error
18882 codes so do a "make errors" if there are problems.
18886 * Correct Linux 1 recognition in config.
18888 *Ulf Möller <ulf@fitug.de>*
18890 * Remove pointless MD5 hash when using DSA keys in ca.
18892 *Anonymous <nobody@replay.com>*
18894 * Generate an error if given an empty string as a cert directory. Also
18895 generate an error if handed NULL (previously returned 0 to indicate an
18896 error, but didn't set one).
18898 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18900 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18904 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18905 parameters. This was causing a warning which killed off the Win32 compile.
18909 * Remove C++ style comments from crypto/bn/bn_local.h.
18911 *Neil Costigan <neil.costigan@celocom.com>*
18913 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18914 based on a text string, looking up short and long names and finally
18915 "dot" format. The "dot" format stuff didn't work. Added new function
18916 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18917 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18918 OID is not part of the table.
18922 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18923 X509_LOOKUP_by_alias().
18927 * Sort openssl functions by name.
18931 * Get the `gendsa` command working and add it to the `list` command. Remove
18932 encryption from sample DSA keys (in case anyone is interested the password
18937 * Make *all* `*_free` functions accept a NULL pointer.
18939 *Frans Heymans <fheymans@isaserver.be>*
18941 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18944 *Anonymous <nobody@replay.com>*
18946 * s_server should send the CAfile as acceptable CAs, not its own cert.
18948 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18950 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18952 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18954 * Temp key "for export" tests were wrong in s3_srvr.c.
18956 *Anonymous <nobody@replay.com>*
18958 * Add prototype for temp key callback functions
18959 SSL_CTX_set_tmp_{rsa,dh}_callback().
18963 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18964 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18968 * X509_name_add_entry() freed the wrong thing after an error.
18970 *Arne Ansper <arne@ats.cyber.ee>*
18972 * rsa_eay.c would attempt to free a NULL context.
18974 *Arne Ansper <arne@ats.cyber.ee>*
18976 * BIO_s_socket() had a broken should_retry() on Windoze.
18978 *Arne Ansper <arne@ats.cyber.ee>*
18980 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18982 *Arne Ansper <arne@ats.cyber.ee>*
18984 * Make sure the already existing X509_STORE->depth variable is initialized
18985 in X509_STORE_new(), but document the fact that this variable is still
18986 unused in the certificate verification process.
18988 *Ralf S. Engelschall*
18990 * Fix the various library and `apps/` files to free up pkeys obtained from
18991 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18995 * Fix reference counting in X509_PUBKEY_get(). This makes
18996 demos/maurice/example2.c work, amongst others, probably.
18998 *Steve Henson and Ben Laurie*
19000 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19001 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19002 are no longer created. This way we have a single and consistent command
19003 line interface `openssl <command>`, similar to `cvs <command>`.
19005 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19007 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19008 BIT STRING wrapper always have zero unused bits.
19012 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19016 * Make the top-level INSTALL documentation easier to understand.
19020 * Makefiles updated to exit if an error occurs in a sub-directory
19021 make (including if user presses ^C) [Paul Sutton]
19023 * Make Montgomery context stuff explicit in RSA data structure.
19027 * Fix build order of pem and err to allow for generated pem.h.
19031 * Fix renumbering bug in X509_NAME_delete_entry().
19035 * Enhanced the err-ins.pl script so it makes the error library number
19036 global and can add a library name. This is needed for external ASN1 and
19037 other error libraries.
19041 * Fixed sk_insert which never worked properly.
19045 * Fix ASN1 macros so they can handle indefinite length constructed
19046 EXPLICIT tags. Some non standard certificates use these: they can now
19051 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19052 into a single doc/ssleay.txt bundle. This way the information is still
19053 preserved but no longer messes up this directory. Now it's new room for
19054 the new set of documentation files.
19056 *Ralf S. Engelschall*
19058 * SETs were incorrectly DER encoded. This was a major pain, because they
19059 shared code with SEQUENCEs, which aren't coded the same. This means that
19060 almost everything to do with SETs or SEQUENCEs has either changed name or
19061 number of arguments.
19063 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19065 * Fix test data to work with the above.
19069 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19070 was already fixed by Eric for 0.9.1 it seems.
19072 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19074 * Autodetect FreeBSD3.
19078 * Fix various bugs in Configure. This affects the following platforms:
19082 unixware-2.0-pentium
19087 * Eliminate generated files from CVS. Reorder tests to regenerate files
19088 before they are needed.
19092 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19096 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19098 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19099 changed SSLeay to OpenSSL in version strings.
19101 *Ralf S. Engelschall*
19103 * Some fixups to the top-level documents.
19107 * Fixed the nasty bug where rsaref.h was not found under compile-time
19108 because the symlink to include/ was missing.
19110 *Ralf S. Engelschall*
19112 * Incorporated the popular no-RSA/DSA-only patches
19113 which allow to compile a RSA-free SSLeay.
19115 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19117 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19118 when "ssleay" is still not found.
19120 *Ralf S. Engelschall*
19122 * Added more platforms to Configure: Cray T3E, HPUX 11,
19124 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19126 * Updated the README file.
19128 *Ralf S. Engelschall*
19130 * Added various .cvsignore files in the CVS repository subdirs
19131 to make a "cvs update" really silent.
19133 *Ralf S. Engelschall*
19135 * Recompiled the error-definition header files and added
19136 missing symbols to the Win32 linker tables.
19138 *Ralf S. Engelschall*
19140 * Cleaned up the top-level documents;
19141 o new files: CHANGES and LICENSE
19142 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19143 o merged COPYRIGHT into LICENSE
19144 o removed obsolete TODO file
19145 o renamed MICROSOFT to INSTALL.W32
19147 *Ralf S. Engelschall*
19149 * Removed dummy files from the 0.9.1b source tree:
19150 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19151 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19152 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19153 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19154 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19156 *Ralf S. Engelschall*
19158 * Added various platform portability fixes.
19162 * The Genesis of the OpenSSL rpject:
19163 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19164 Young and Tim J. Hudson created while they were working for C2Net until
19167 *The OpenSSL Project*
19169 ### Changes between 0.9.0b and 0.9.1b [not released]
19171 * Updated a few CA certificates under certs/
19175 * Changed some BIGNUM api stuff.
19179 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19180 DGUX x86, Linux Alpha, etc.
19184 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19185 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19190 * Add -strparse option to asn1pars program which parses nested
19193 *Dr Stephen Henson <shenson@bigfoot.com>*
19195 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19199 * DSA fix for "ca" program.
19203 * Added "-genkey" option to "dsaparam" program.
19207 * Added RIPE MD160 (rmd160) message digest.
19211 * Added -a (all) option to "ssleay version" command.
19215 * Added PLATFORM define which is the id given to Configure.
19219 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19223 * Extended the ASN.1 parser routines.
19227 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19231 * Added a BN_CTX to the BN library.
19235 * Fixed the weak key values in DES library
19239 * Changed API in EVP library for cipher aliases.
19243 * Added support for RC2/64bit cipher.
19247 * Converted the lhash library to the crypto/mem.c functions.
19251 * Added more recognized ASN.1 object ids.
19255 * Added more RSA padding checks for SSL/TLS.
19259 * Added BIO proxy/filter functionality.
19263 * Added extra_certs to SSL_CTX which can be used
19264 send extra CA certificates to the client in the CA cert chain sending
19265 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19269 * Now Fortezza is denied in the authentication phase because
19270 this is key exchange mechanism is not supported by SSLeay at all.
19274 * Additional PKCS1 checks.
19278 * Support the string "TLSv1" for all TLS v1 ciphers.
19282 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19283 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19287 * Fixed a few memory leaks.
19291 * Fixed various code and comment typos.
19295 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19296 bytes sent in the client random.
19298 *Edward Bishop <ebishop@spyglass.com>*
19302 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19303 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19304 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19305 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19306 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19307 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19308 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19309 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19310 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19311 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19312 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19313 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19314 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19315 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19316 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19317 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19318 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19319 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19320 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19321 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19322 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19323 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19324 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19325 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19326 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19327 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19328 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19329 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19330 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19331 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19332 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19333 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19334 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19335 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19336 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19337 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19338 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19339 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19340 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19341 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19342 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19343 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19344 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19345 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19346 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19347 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19348 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19349 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19350 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19351 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19352 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19353 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19354 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19355 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19356 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19357 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19358 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19359 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19360 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19361 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19362 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19363 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19364 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19365 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19366 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19367 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19368 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19369 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19370 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19371 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19372 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19373 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19374 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19375 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19376 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19377 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19378 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19379 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19380 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19381 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19382 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19383 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19384 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19385 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19386 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19387 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19388 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19389 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19390 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19391 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19392 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19393 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19394 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19395 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19396 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19397 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19398 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19399 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19400 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19401 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19402 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19403 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19404 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19405 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19406 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19407 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19408 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19409 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19410 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19411 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19412 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19413 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19414 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19415 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19416 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19417 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19418 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19419 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19420 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19421 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19422 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19423 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19424 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19425 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19426 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19427 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19428 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19429 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19430 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19431 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19432 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19433 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19434 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19435 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19436 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19437 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19438 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19439 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19440 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19441 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19442 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19443 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19444 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19445 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19446 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19447 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19448 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19449 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19450 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19451 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19452 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19453 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19454 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19455 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19456 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19457 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19458 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19459 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19460 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19461 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19462 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19463 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655