4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
25 listed here are only a brief description.
26 The migration guide contains more detailed information related to new features,
27 breaking changes, and mappings for the large list of deprecated functions.
29 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
31 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
33 * Fixed a bug in the c_rehash script which was not properly sanitising shell
34 metacharacters to prevent command injection. This script is distributed by
35 some operating systems in a manner where it is automatically executed. On
36 such operating systems, an attacker could execute arbitrary commands with the
37 privileges of the script.
39 Use of the c_rehash script is considered obsolete and should be replaced
40 by the OpenSSL rehash command line tool.
45 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
46 certificate on an OCSP response. The bug caused the function in the case
47 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
48 response (meaning a successful verification) even in the case where the
49 response signing certificate fails to verify.
51 It is anticipated that most users of `OCSP_basic_verify` will not use the
52 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
53 a negative value (indicating a fatal error) in the case of a certificate
54 verification failure. The normal expected return value in this case would be
57 This issue also impacts the command line OpenSSL "ocsp" application. When
58 verifying an ocsp response with the "-no_cert_checks" option the command line
59 application will report that the verification is successful even though it
60 has in fact failed. In this case the incorrect successful response will also
61 be accompanied by error messages showing the failure and contradicting the
62 apparently successful result.
67 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
68 AAD data as the MAC key. This made the MAC key trivially predictable.
70 An attacker could exploit this issue by performing a man-in-the-middle attack
71 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
72 that the modified data would still pass the MAC integrity check.
74 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
75 endpoint will always be rejected by the recipient and the connection will
76 fail at that point. Many application protocols require data to be sent from
77 the client to the server first. Therefore, in such a case, only an OpenSSL
78 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
80 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
81 sent in both directions. In this case both clients and servers could be
82 affected, regardless of the application protocol.
84 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
85 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
86 the handshake when using this ciphersuite.
88 The confidentiality of data is not impacted by this issue, i.e. an attacker
89 cannot decrypt data that has been encrypted using this ciphersuite - they can
92 In order for this attack to work both endpoints must legitimately negotiate
93 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
94 OpenSSL 3.0, and is not available within the default provider or the default
95 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
96 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
97 following must have occurred:
99 1) OpenSSL must have been compiled with the (non-default) compile time option
100 enable-weak-ssl-ciphers
102 2) OpenSSL must have had the legacy provider explicitly loaded (either
103 through application code or via configuration)
105 3) The ciphersuite must have been explicitly added to the ciphersuite list
107 4) The libssl security level must have been set to 0 (default is 1)
109 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
111 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
112 others that both endpoints have in common
117 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
118 occuppied by the removed hash table entries.
120 This function is used when decoding certificates or keys. If a long lived
121 process periodically decodes certificates or keys its memory usage will
122 expand without bounds and the process might be terminated by the operating
123 system causing a denial of service. Also traversing the empty hash table
124 entries will take increasingly more time.
126 Typically such long lived processes might be TLS clients or TLS servers
127 configured to accept client certificate authentication.
130 *Hugo Landau, Aliaksei Levin*
132 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
133 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
134 statistics are no longer supported. For compatibility, these statistics are
135 still listed in the output but are now always reported as zero.
139 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
141 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
142 for non-prime moduli.
144 Internally this function is used when parsing certificates that contain
145 elliptic curve public keys in compressed form or explicit elliptic curve
146 parameters with a base point encoded in compressed form.
148 It is possible to trigger the infinite loop by crafting a certificate that
149 has invalid explicit curve parameters.
151 Since certificate parsing happens prior to verification of the certificate
152 signature, any process that parses an externally supplied certificate may thus
153 be subject to a denial of service attack. The infinite loop can also be
154 reached when parsing crafted private keys as they can contain explicit
155 elliptic curve parameters.
157 Thus vulnerable situations include:
159 - TLS clients consuming server certificates
160 - TLS servers consuming client certificates
161 - Hosting providers taking certificates or private keys from customers
162 - Certificate authorities parsing certification requests from subscribers
163 - Anything else which parses ASN.1 elliptic curve parameters
165 Also any other applications that use the BN_mod_sqrt() where the attacker
166 can control the parameter values are vulnerable to this DoS issue.
171 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
172 to the list of ciphersuites providing Perfect Forward Secrecy as
173 required by SECLEVEL >= 3.
175 *Dmitry Belyavskiy, Nicola Tuveri*
177 * Made the AES constant time code for no-asm configurations
178 optional due to the resulting 95% performance degradation.
179 The AES constant time code can be enabled, for no assembly
180 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
184 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
189 * The negative return value handling of the certificate verification callback
190 was reverted. The replacement is to set the verification retry state with
191 the SSL_set_retry_verify() function.
195 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
197 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
198 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
199 verify a certificate supplied by a server. That function may return a
200 negative return value to indicate an internal error (for example out of
201 memory). Such a negative return value is mishandled by OpenSSL and will cause
202 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
203 success and a subsequent call to SSL_get_error() to return the value
204 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
205 returned by OpenSSL if the application has previously called
206 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
207 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
208 totally unexpected and applications may not behave correctly as a result. The
209 exact behaviour will depend on the application but it could result in
210 crashes, infinite loops or other similar incorrect responses.
212 This issue is made more serious in combination with a separate bug in OpenSSL
213 3.0 that will cause X509_verify_cert() to indicate an internal error when
214 processing a certificate chain. This will occur where a certificate does not
215 include the Subject Alternative Name extension but where a Certificate
216 Authority has enforced name constraints. This issue can occur even with valid
222 * Corrected a few file name and file reference bugs in the build,
223 installation and setup scripts, which lead to installation verification
224 failures. Slightly enhanced the installation verification script.
228 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
233 * Fixed PVK encoder to properly query for the passphrase.
237 * Multiple fixes in the OSSL_HTTP API functions.
241 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
242 OSSL_PARAM_INTEGER data type and return error on negative numbers
243 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
244 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
248 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
252 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
256 * Multiple threading fixes.
260 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
264 * Allow fetching an operation from the provider that owns an unexportable key
265 as a fallback if that is still allowed by the property query.
269 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
271 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
276 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
277 S390X capability vector to zero. This simplifies testing of different code
278 paths on S390X architecture.
282 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
283 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
284 SP 800-38D". The communication will fail at this point.
288 * The EC_GROUP_clear_free() function is deprecated as there is nothing
289 confidential in EC_GROUP data.
293 * The byte order mark (BOM) character is ignored if encountered at the
294 beginning of a PEM-formatted file.
298 * Added CMS support for the Russian GOST algorithms.
302 * Due to move of the implementation of cryptographic operations
303 to the providers, validation of various operation parameters can
304 be postponed until the actual operation is executed where previously
305 it happened immediately when an operation parameter was set.
307 For example when setting an unsupported curve with
308 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
309 fail but later keygen operations with the EVP_PKEY_CTX will fail.
311 *OpenSSL team members and many third party contributors*
313 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
314 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
315 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
316 instead to retrieve these algorithms from a provider.
320 * On build targets where the multilib postfix is set in the build
321 configuration the libdir directory was changing based on whether
322 the lib directory with the multilib postfix exists on the system
323 or not. This unpredictable behavior was removed and eventual
324 multilib postfix is now always added to the default libdir. Use
325 `--libdir=lib` to override the libdir if adding the postfix is
330 * The triple DES key wrap functionality now conforms to RFC 3217 but is
331 no longer interoperable with OpenSSL 1.1.1.
335 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
336 function codes, this function can only cause problems for calling
341 * Add a configurable flag to output date formats as ISO 8601. Does not
342 change the default date format.
346 * Version of MSVC earlier than 1300 could get link warnings, which could
347 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
348 Support for this flag has been removed.
352 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
353 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
354 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
355 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
356 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
360 * The signatures of the functions to get and set options on SSL and
361 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
362 Some source code changes may be required.
366 * The public definitions of conf_method_st and conf_st have been
367 deprecated. They will be made opaque in a future release.
369 *Rich Salz and Tomáš Mráz*
371 * Client-initiated renegotiation is disabled by default. To allow it, use
372 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
373 flag, or the "ClientRenegotiation" config parameter as appropriate.
377 * Add "abspath" and "includedir" pragma's to config files, to prevent,
378 or modify relative pathname inclusion.
382 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
383 validated. Please consult the README-FIPS and
384 README-PROVIDERS files, as well as the migration guide.
386 *OpenSSL team members and many third party contributors*
388 * For the key types DH and DHX the allowed settable parameters are now different.
392 * The openssl commands that read keys, certificates, and CRLs now
393 automatically detect the PEM or DER format of the input files.
395 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
397 * Added enhanced PKCS#12 APIs which accept a library context.
401 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
405 * Added support for Kernel TLS (KTLS).
407 *Boris Pismenny, John Baldwin and Andrew Gallatin*
409 * Support for RFC 5746 secure renegotiation is now required by default for
410 SSL or TLS connections to succeed.
414 * The signature of the `copy` functional parameter of the
415 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
416 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
417 the signature of the `pub_decode` functional parameter of the
418 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
419 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
423 * The error return values from some control calls (ctrl) have changed.
427 * A public key check is now performed during EVP_PKEY_derive_set_peer().
431 * Many functions in the EVP_ namespace that are getters of values from
432 implementations or contexts were renamed to include get or get0 in their
433 names. Old names are provided as macro aliases for compatibility and
438 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
439 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
440 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
445 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
448 * The output from the command line applications may have minor
453 * The output from numerous "printing" may have minor changes.
457 * Windows thread synchronization uses read/write primitives (SRWLock) when
458 supported by the OS, otherwise CriticalSection continues to be used.
462 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
463 work on read only BIO source/sinks that do not support these functions.
464 This allows piping or redirection of a file BIO using stdin to be buffered
465 into memory. This is used internally in OSSL_DECODER_from_bio().
469 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
470 this function would return one of the values OSSL_STORE_INFO_NAME,
471 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
472 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
473 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
474 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
475 using this function should be amended to handle the changed return value.
479 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
480 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
481 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
482 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
483 contains more than one certificate identifier: This means that all
484 certificates referenced there MUST be part of the validation chain.
488 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
489 RC5, DESX and DES have been moved to the legacy provider.
493 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
494 RIPEMD-160 have been moved to the legacy provider.
498 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
503 * The deprecated functions EVP_PKEY_get0_RSA(),
504 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
505 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
506 well as the similarly named "get1" functions behave differently in
511 * A number of functions handling low-level keys or engines were deprecated
512 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
513 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
514 EVP_PKEY_get0_siphash().
518 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
519 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
520 will need to load the legacy crypto provider. This includes these PBE
521 algorithms which use this KDF:
522 - NID_pbeWithMD2AndDES_CBC
523 - NID_pbeWithMD5AndDES_CBC
524 - NID_pbeWithSHA1AndRC2_CBC
525 - NID_pbeWithMD2AndRC2_CBC
526 - NID_pbeWithMD5AndRC2_CBC
527 - NID_pbeWithSHA1AndDES_CBC
531 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
532 BIO_debug_callback() functions.
536 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
537 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
541 * The RAND_METHOD APIs have been deprecated.
545 * The SRP APIs have been deprecated.
549 * Add a compile time option to prevent the caching of provider fetched
550 algorithms. This is enabled by including the no-cached-fetch option
551 at configuration time.
555 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
556 count of PKCS12_DEFAULT_ITER.
558 *Tomáš Mráz and Sahana Prasad*
560 * The openssl speed command does not use low-level API calls anymore.
564 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
567 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
569 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
573 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
574 providers may supply their own group implementations (using either the "key
575 exchange" or the "key encapsulation" methods) which will automatically be
576 detected and used by libssl.
578 *Matt Caswell, Nicola Tuveri*
580 * The undocumented function X509_certificate_type() has been deprecated;
584 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
588 * Removed RSA padding mode for SSLv23 (which was only used for
589 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
590 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
595 * Deprecated the obsolete X9.31 RSA key generation related functions.
597 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
598 is not allowed to return a value > 1, this is no more taken as failure.
600 *Viktor Dukhovni and David von Oheimb*
602 * Deprecated the obsolete X9.31 RSA key generation related functions
603 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
604 BN_X931_generate_prime_ex().
608 * The default key generation method for the regular 2-prime RSA keys was
609 changed to the FIPS 186-4 B.3.6 method.
613 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
617 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
621 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
622 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
624 *Rich Salz, Richard Levitte, and David von Oheimb*
626 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
630 * Deprecated `OCSP_parse_url()`.
634 * Validation of SM2 keys has been separated from the validation of regular EC
639 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
640 switches: a validation failure triggers an early exit, returning a failure
641 exit status to the parent process.
645 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
646 to ignore unknown ciphers.
650 * The `-cipher-commands` and `-digest-commands` options
651 of the command line utility `list` have been deprecated.
652 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
656 * Added convenience functions for generating asymmetric key pairs:
657 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
658 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
662 * All of the low level EC_KEY functions have been deprecated.
664 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
666 * Deprecated all the libcrypto and libssl error string loading
671 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
672 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
677 * The `-crypt` option to the `passwd` command line tool has been removed.
681 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
686 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
690 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
691 EVP_PKEY_get1_tls_encodedpoint().
695 * The security callback, which can be customised by application code, supports
696 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
697 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
701 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
702 interface. Their functionality remains unchanged.
706 * Added new option for 'openssl list', '-providers', which will display the
707 list of loaded providers, their names, version and status. It optionally
708 displays their gettable parameters.
712 * Removed EVP_PKEY_set_alias_type().
716 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
717 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
721 * Changed all "STACK" functions to be macros instead of inline functions. Macro
722 parameters are still checked for type safety at compile time via helper
727 * Remove the RAND_DRBG API
729 *Paul Dale and Matthias St. Pierre*
731 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
732 as well as actual hostnames.
736 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
737 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
738 conversely, silently ignore DTLS protocol version bounds when configuring
739 TLS-based contexts. The commands can be repeated to set bounds of both
740 types. The same applies with the corresponding "min_protocol" and
741 "max_protocol" command-line switches, in case some application uses both TLS
744 SSL_CTX instances that are created for a fixed protocol version (e.g.
745 `TLSv1_server_method()`) also silently ignore version bounds. Previously
746 attempts to apply bounds to these protocol versions would result in an
747 error. Now only the "version-flexible" SSL_CTX instances are subject to
748 limits in configuration files in command-line options.
752 * Deprecated the `ENGINE` API. Engines should be replaced with providers
757 * Reworked the recorded ERR codes to make better space for system errors.
758 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
759 given code is a system error (true) or an OpenSSL error (false).
763 * Reworked the test perl framework to better allow parallel testing.
765 *Nicola Tuveri and David von Oheimb*
767 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
768 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
772 * 'Configure' has been changed to figure out the configuration target if
773 none is given on the command line. Consequently, the 'config' script is
774 now only a mere wrapper. All documentation is changed to only mention
777 *Rich Salz and Richard Levitte*
779 * Added a library context `OSSL_LIB_CTX` that applications as well as
780 other libraries can use to form a separate context within which
781 libcrypto operations are performed.
785 * Added various `_ex` functions to the OpenSSL API that support using
786 a non-default `OSSL_LIB_CTX`.
790 * Handshake now fails if Extended Master Secret extension is dropped
795 * Dropped interactive mode from the `openssl` program.
799 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
801 *David von Oheimb and Shane Lontis*
803 * Deprecated `EC_METHOD_get_field_type()`.
807 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
808 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
809 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
813 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
817 * Add CAdES-BES signature verification support, mostly derived
818 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
820 *Filipe Raimundo da Silva*
822 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
826 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
827 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
831 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
835 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
836 EC_KEY_precompute_mult().
840 * Deprecated EC_POINTs_mul().
844 * Removed FIPS_mode() and FIPS_mode_set().
848 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
852 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
853 EC_POINT_get_Jprojective_coordinates_GFp().
857 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
858 arrays to be more easily constructed via a series of utility functions.
859 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
860 the various push functions and finally convert to a passable OSSL_PARAM
861 array using OSSL_PARAM_BLD_to_param().
865 * The security strength of SHA1 and MD5 based signatures in TLS has been
870 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
871 contain a provider side internal key.
875 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
879 * Project text documents not yet having a proper file name extension
880 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
881 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
885 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
886 have been converted to Markdown with the goal to produce documents
887 which not only look pretty when viewed online in the browser, but
888 remain well readable inside a plain text editor.
890 To achieve this goal, a 'minimalistic' Markdown style has been applied
891 which avoids formatting elements that interfere too much with the
892 reading flow in the text file. For example, it
894 * avoids [ATX headings][] and uses [setext headings][] instead
895 (which works for `<h1>` and `<h2>` headings only).
896 * avoids [inline links][] and uses [reference links][] instead.
897 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
899 [ATX headings]: https://github.github.com/gfm/#atx-headings
900 [setext headings]: https://github.github.com/gfm/#setext-headings
901 [inline links]: https://github.github.com/gfm/#inline-link
902 [reference links]: https://github.github.com/gfm/#reference-link
903 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
904 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
906 *Matthias St. Pierre*
908 * The test suite is changed to preserve results of each test recipe.
909 A new directory test-runs/ with subdirectories named like the
910 test recipes are created in the build tree for this purpose.
914 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
915 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
916 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
918 *David von Oheimb, Martin Peylo*
920 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
921 It supports arbitrary request and response content types, GET redirection,
922 TLS, connections via HTTP(S) proxies, connections and exchange via
923 user-defined BIOs (allowing implicit connections), persistent connections,
924 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
925 The legacy OCSP-focused (and only partly documented) API
926 is retained for backward compatibility, while most of it is deprecated.
930 * Added `util/check-format.pl`, a tool for checking adherence to the
931 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
932 The checks performed are incomplete and yield some false positives.
933 Still the tool should be useful for detecting most typical glitches.
937 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
938 If domain name resolution yields multiple IP addresses all of them are tried
939 after `connect()` failures.
943 * All of the low level RSA functions have been deprecated.
947 * X509 certificates signed using SHA1 are no longer allowed at security
952 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
953 modified to use PKEY APIs. These commands are now in maintenance mode
954 and no new features will be added to them.
958 * The command line utility rsautl has been deprecated.
962 * The command line utilities genrsa and rsa have been modified to use PKEY
963 APIs. They now write PKCS#8 keys by default. These commands are now in
964 maintenance mode and no new features will be added to them.
968 * All of the low level DH functions have been deprecated.
970 *Paul Dale and Matt Caswell*
972 * All of the low level DSA functions have been deprecated.
976 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
977 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
981 * Deprecated low level ECDH and ECDSA functions.
985 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
989 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
990 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
991 a new formulation to include all the things it can be used for,
992 as well as words of caution.
996 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1000 * All of the low level HMAC functions have been deprecated.
1002 *Paul Dale and David von Oheimb*
1004 * Over two thousand fixes were made to the documentation, including:
1005 - Common options (such as -rand/-writerand, TLS version control, etc)
1006 were refactored and point to newly-enhanced descriptions in openssl.pod.
1007 - Added style conformance for all options (with help from Richard Levitte),
1008 documented all reported missing options, added a CI build to check
1009 that all options are documented and that no unimplemented options
1011 - Documented some internals, such as all use of environment variables.
1012 - Addressed all internal broken L<> references.
1016 * All of the low level CMAC functions have been deprecated.
1020 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1021 functions have been deprecated.
1023 *Paul Dale and David von Oheimb*
1025 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1026 set of functions. The documentation mentioned negative values for some
1027 errors, but this was never the case, so the mention of negative values
1030 Code that followed the documentation and thereby check with something
1031 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1035 * All of the low level cipher functions have been deprecated.
1037 *Matt Caswell and Paul Dale*
1039 * Removed include/openssl/opensslconf.h.in and replaced it with
1040 include/openssl/configuration.h.in, which differs in not including
1041 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1042 was added to include both.
1044 This allows internal hacks where one might need to modify the set
1045 of configured macros, for example this if deprecated symbols are
1046 still supposed to be available internally:
1048 #include <openssl/configuration.h>
1050 #undef OPENSSL_NO_DEPRECATED
1051 #define OPENSSL_SUPPRESS_DEPRECATED
1053 #include <openssl/macros.h>
1055 This should not be used by applications that use the exported
1056 symbols, as that will lead to linking errors.
1060 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1061 used in exponentiation with 512-bit moduli. No EC algorithms are
1062 affected. Analysis suggests that attacks against 2-prime RSA1024,
1063 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1064 difficult to perform and are not believed likely. Attacks against DH512
1065 are considered just feasible. However, for an attack the target would
1066 have to re-use the DH512 private key, which is not recommended anyway.
1067 Also applications directly using the low-level API BN_mod_exp may be
1068 affected if they use BN_FLG_CONSTTIME.
1073 * Most memory-debug features have been deprecated, and the functionality
1074 replaced with no-ops.
1078 * Added documentation for the STACK API.
1082 * Introduced a new method type and API, OSSL_ENCODER, to represent
1083 generic encoders. These do the same sort of job that PEM writers
1084 and d2i functions do, but with support for methods supplied by
1085 providers, and the possibility for providers to support other
1090 * Introduced a new method type and API, OSSL_DECODER, to represent
1091 generic decoders. These do the same sort of job that PEM readers
1092 and i2d functions do, but with support for methods supplied by
1093 providers, and the possibility for providers to support other
1098 * Added a .pragma directive to the syntax of configuration files, to
1099 allow varying behavior in a supported and predictable manner.
1100 Currently added pragma:
1104 This allows dollar signs to be a keyword character unless it's
1105 followed by a opening brace or parenthesis. This is useful for
1106 platforms where dollar signs are commonly used in names, such as
1107 volume names and system directory names on VMS.
1111 * Added functionality to create an EVP_PKEY from user data.
1115 * Change the interpretation of the '--api' configuration option to
1116 mean that this is a desired API compatibility level with no
1117 further meaning. The previous interpretation, that this would
1118 also mean to remove all deprecated symbols up to and including
1119 the given version, no requires that 'no-deprecated' is also used
1120 in the configuration.
1122 When building applications, the desired API compatibility level
1123 can be set with the OPENSSL_API_COMPAT macro like before. For
1124 API compatibility version below 3.0, the old style numerical
1125 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1126 For version 3.0 and on, the value is expected to be the decimal
1127 value calculated from the major and minor version like this:
1129 MAJOR * 10000 + MINOR * 100
1133 -DOPENSSL_API_COMPAT=30000 For 3.0
1134 -DOPENSSL_API_COMPAT=30200 For 3.2
1136 To hide declarations that are deprecated up to and including the
1137 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1138 given when building the application as well.
1142 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1143 access to certificate and CRL stores via URIs and OSSL_STORE
1146 This adds the following functions:
1148 - X509_LOOKUP_store()
1149 - X509_STORE_load_file()
1150 - X509_STORE_load_path()
1151 - X509_STORE_load_store()
1152 - SSL_add_store_cert_subjects_to_stack()
1153 - SSL_CTX_set_default_verify_store()
1154 - SSL_CTX_load_verify_file()
1155 - SSL_CTX_load_verify_dir()
1156 - SSL_CTX_load_verify_store()
1160 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1161 The presence of this system service is determined at run-time.
1165 * Added functionality to create an EVP_PKEY context based on data
1166 for methods from providers. This takes an algorithm name and a
1167 property query string and simply stores them, with the intent
1168 that any operation that uses this context will use those strings
1169 to fetch the needed methods implicitly, thereby making the port
1170 of application written for pre-3.0 OpenSSL easier.
1174 * The undocumented function NCONF_WIN32() has been deprecated; for
1175 conversion details see the HISTORY section of doc/man5/config.pod
1179 * Introduced the new functions EVP_DigestSignInit_ex() and
1180 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1181 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1182 pages for further details.
1186 * Over two thousand fixes were made to the documentation, including:
1187 adding missing command flags, better style conformance, documentation
1190 *Rich Salz, Richard Levitte*
1192 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1193 X25519, X448, Ed25519 and Ed448.
1197 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1202 * Deprecated the public definition of `ERR_STATE` as well as the function
1203 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1208 * Added ERR functionality to give callers access to the stored function
1209 names that have replaced the older function code based functions.
1211 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1212 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1213 ERR_peek_error_all() and ERR_peek_last_error_all().
1215 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1216 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1217 ERR_func_error_string().
1221 * Extended testing to be verbose for failing tests only. The make variables
1222 VERBOSE_FAILURE or VF can be used to enable this:
1224 $ make VF=1 test # Unix
1225 $ mms /macro=(VF=1) test ! OpenVMS
1226 $ nmake VF=1 test # Windows
1230 * Added the `-copy_extensions` option to the `x509` command for use with
1231 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1232 all extensions in the request are copied to the certificate or vice versa.
1234 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1236 * Added the `-copy_extensions` option to the `req` command for use with
1237 `-x509`. When given with the `copy` or `copyall` argument,
1238 all extensions in the certification request are copied to the certificate.
1242 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1243 they generate are by default RFC 5280 compliant in the following sense:
1244 There is a subjectKeyIdentifier extension with a hash value of the public key
1245 and for not self-signed certs there is an authorityKeyIdentifier extension
1246 with a keyIdentifier field or issuer information identifying the signing key.
1247 This is done unless some configuration overrides the new default behavior,
1248 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1252 * Added several checks to `X509_verify_cert()` according to requirements in
1253 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1254 (which may be done by using the CLI option `-x509_strict`):
1255 * The basicConstraints of CA certificates must be marked critical.
1256 * CA certificates must explicitly include the keyUsage extension.
1257 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1258 * The issuer name of any certificate must not be empty.
1259 * The subject name of CA certs, certs with keyUsage crlSign,
1260 and certs without subjectAlternativeName must not be empty.
1261 * If a subjectAlternativeName extension is given it must not be empty.
1262 * The signatureAlgorithm field and the cert signature must be consistent.
1263 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1264 must not be marked critical.
1265 * The authorityKeyIdentifier must be given for X.509v3 certs
1266 unless they are self-signed.
1267 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1271 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1272 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1276 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1277 used even when parsing explicit parameters, when loading a encoded key
1278 or calling `EC_GROUP_new_from_ecpkparameters()`/
1279 `EC_GROUP_new_from_ecparameters()`.
1280 This prevents bypass of security hardening and performance gains,
1281 especially for curves with specialized EC_METHODs.
1282 By default, if a key encoded with explicit parameters is loaded and later
1283 encoded, the output is still encoded with explicit parameters, even if
1284 internally a "named" EC_GROUP is used for computation.
1288 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1289 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1290 NULL. After this change, only the cofactor parameter can be NULL. It also
1291 does some minimal sanity checks on the passed order.
1296 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1297 An attack is simple, if the first CMS_recipientInfo is valid but the
1298 second CMS_recipientInfo is chosen ciphertext. If the second
1299 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1300 encryption key will be replaced by garbage, and the message cannot be
1301 decoded, but if the RSA decryption fails, the correct encryption key is
1302 used and the recipient will not notice the attack.
1303 As a work around for this potential attack the length of the decrypted
1304 key must be equal to the cipher default key length, in case the
1305 certifiate is not given and all recipientInfo are tried out.
1306 The old behaviour can be re-enabled in the CMS code by setting the
1307 CMS_DEBUG_DECRYPT flag.
1311 * Early start up entropy quality from the DEVRANDOM seed source has been
1312 improved for older Linux systems. The RAND subsystem will wait for
1313 /dev/random to be producing output before seeding from /dev/urandom.
1314 The seeded state is stored for future library initialisations using
1315 a system global shared memory segment. The shared memory identifier
1316 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1317 the desired value. The default identifier is 114.
1321 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1322 when primes for RSA keys are computed.
1323 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1324 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1325 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1326 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1327 This avoids possible fingerprinting of newly generated RSA modules.
1331 * Correct the extended master secret constant on EBCDIC systems. Without this
1332 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1333 negotiate EMS will fail. Unfortunately this also means that TLS connections
1334 between EBCDIC systems with this fix, and EBCDIC systems without this
1335 fix will fail if they negotiate EMS.
1339 * Changed the library initialisation so that the config file is now loaded
1340 by default. This was already the case for libssl. It now occurs for both
1341 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1342 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1346 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1347 where the former acts as a replacement for `ERR_put_error()`, and the
1348 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1349 `ERR_raise_data()` adds more flexibility by taking a format string and
1350 an arbitrary number of arguments following it, to be processed with
1355 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1356 to check if a named provider is loaded and available. When called, it
1357 will also activate all fallback providers if such are still present.
1361 * Enforce a minimum DH modulus size of 512 bits.
1365 * Changed DH parameters to generate the order q subgroup instead of 2q.
1366 Previously generated DH parameters are still accepted by DH_check
1367 but DH_generate_key works around that by clearing bit 0 of the
1368 private key for those. This avoids leaking bit 0 of the private key.
1372 * Significantly reduce secure memory usage by the randomness pools.
1376 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1381 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1382 algorithms. An implementation of a key exchange algorithm can be obtained
1383 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1384 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1385 the older EVP_PKEY_derive_init() function. See the man pages for the new
1386 functions for further details.
1390 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1394 * Removed the function names from error messages and deprecated the
1399 * Removed NextStep support and the macro OPENSSL_UNISTD
1403 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1404 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1405 Also removed "export var as function" capability; we do not export
1406 variables, only functions.
1410 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1411 an error and 1 indicating success. In previous versions of OpenSSL this
1412 was a void type. If a key was set longer than the maximum possible this
1417 * Support SM2 signing and verification schemes with X509 certificate.
1421 * Use SHA256 as the default digest for TS query in the `ts` app.
1425 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1429 * Default cipher lists/suites are now available via a function, the
1430 #defines are deprecated.
1434 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1435 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1436 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1440 * Join the directories crypto/x509 and crypto/x509v3
1444 * Added command 'openssl kdf' that uses the EVP_KDF API.
1448 * Added command 'openssl mac' that uses the EVP_MAC API.
1452 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1453 as default directories. Also added the command 'openssl info'
1454 for scripting purposes.
1458 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1463 * Add prediction resistance to the DRBG reseeding process.
1467 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1468 mandated by IEEE Std 1619-2018.
1472 * Added newline escaping functionality to a filename when using openssl dgst.
1473 This output format is to replicate the output format found in the `*sum`
1474 checksum programs. This aims to preserve backward compatibility.
1476 *Matt Eaton, Richard Levitte, and Paul Dale*
1478 * Removed the heartbeat message in DTLS feature, as it has very
1479 little usage and doesn't seem to fulfill a valuable purpose.
1480 The configuration option is now deprecated.
1484 * Changed the output of 'openssl {digestname} < file' to display the
1485 digest name in its output.
1489 * Added a new generic trace API which provides support for enabling
1490 instrumentation through trace output.
1492 *Richard Levitte & Matthias St. Pierre*
1494 * Added build tests for C++. These are generated files that only do one
1495 thing, to include one public OpenSSL head file each. This tests that
1496 the public header files can be usefully included in a C++ application.
1498 This test isn't enabled by default. It can be enabled with the option
1499 'enable-buildtest-c++'.
1503 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1507 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1511 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1515 * Added KMAC to EVP_MAC.
1519 * Added property based algorithm implementation selection framework to
1524 * Added SCA hardening for modular field inversion in EC_GROUP through
1525 a new dedicated field_inv() pointer in EC_METHOD.
1526 This also addresses a leakage affecting conversions from projective
1527 to affine coordinates.
1529 *Billy Bob Brumley, Nicola Tuveri*
1531 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1532 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1533 those algorithms that were already supported through the EVP_PKEY API
1534 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1535 and scrypt are now wrappers that call EVP_KDF.
1539 * Build devcrypto engine as a dynamic engine.
1541 *Eneas U de Queiroz*
1543 * Add keyed BLAKE2 to EVP_MAC.
1547 * Fix a bug in the computation of the endpoint-pair shared secret used
1548 by DTLS over SCTP. This breaks interoperability with older versions
1549 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1550 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1551 interoperability with such broken implementations. However, enabling
1552 this switch breaks interoperability with correct implementations.
1554 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1555 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1559 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1563 * Changed the license to the Apache License v2.0.
1567 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1569 - Major releases (indicated by incrementing the MAJOR release number)
1570 may introduce incompatible API/ABI changes.
1571 - Minor releases (indicated by incrementing the MINOR release number)
1572 may introduce new features but retain API/ABI compatibility.
1573 - Patch releases (indicated by incrementing the PATCH number)
1574 are intended for bug fixes and other improvements of existing
1575 features only (like improving performance or adding documentation)
1576 and retain API/ABI compatibility.
1580 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1584 * Remove the 'dist' target and add a tarball building script. The
1585 'dist' target has fallen out of use, and it shouldn't be
1586 necessary to configure just to create a source distribution.
1590 * Recreate the OS390-Unix config target. It no longer relies on a
1591 special script like it did for OpenSSL pre-1.1.0.
1595 * Instead of having the source directories listed in Configure, add
1596 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1601 * Add GMAC to EVP_MAC.
1605 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1609 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1610 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1611 to facilitate the continued use of MACs through raw private keys in
1612 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1616 * Deprecate ECDH_KDF_X9_62().
1620 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1621 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1622 are retained for backwards compatibility.
1626 * AES-XTS mode now enforces that its two keys are different to mitigate
1627 the attacked described in "Efficient Instantiations of Tweakable
1628 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1629 Details of this attack can be obtained from:
1630 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1634 * Rename the object files, i.e. give them other names than in previous
1635 versions. Their names now include the name of the final product, as
1636 well as its type mnemonic (bin, lib, shlib).
1640 * Added new option for 'openssl list', '-objects', which will display the
1641 list of built in objects, i.e. OIDs with names.
1645 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1646 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1651 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1652 improves application performance by removing data copies and providing
1653 applications with zero-copy system calls such as sendfile and splice.
1657 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1661 * `PKCS12_parse` now maintains the order of the parsed certificates
1662 when outputting them via `*ca` (rather than reversing it).
1666 * Deprecated pthread fork support methods.
1670 * Added support for FFDHE key exchange in TLS 1.3.
1674 * Added a new concept for OpenSSL plugability: providers. This
1675 functionality is designed to replace the ENGINE API and ENGINE
1676 implementations, and to be much more dynamic, allowing provider
1677 authors to introduce new algorithms among other things, as long as
1678 there's an API that supports the algorithm type.
1680 With this concept comes a new core API for interaction between
1681 libcrypto and provider implementations. Public libcrypto functions
1682 that want to use providers do so through this core API.
1684 The main documentation for this core API is found in
1685 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1686 refer to other manuals describing the API specific for supported
1687 algorithm types (also called operations).
1694 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
1696 * Avoid loading of a dynamic engine twice.
1700 * Prioritise DANE TLSA issuer certs over peer certs
1704 * Fixed random API for MacOS prior to 10.12
1706 These MacOS versions don't support the CommonCrypto APIs
1710 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1712 * Fixed an SM2 Decryption Buffer Overflow.
1714 In order to decrypt SM2 encrypted data an application is expected to
1715 call the API function EVP_PKEY_decrypt(). Typically an application will
1716 call this function twice. The first time, on entry, the "out" parameter
1717 can be NULL and, on exit, the "outlen" parameter is populated with the
1718 buffer size required to hold the decrypted plaintext. The application
1719 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1720 again, but this time passing a non-NULL value for the "out" parameter.
1722 A bug in the implementation of the SM2 decryption code means that the
1723 calculation of the buffer size required to hold the plaintext returned
1724 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1725 size required by the second call. This can lead to a buffer overflow
1726 when EVP_PKEY_decrypt() is called by the application a second time with
1727 a buffer that is too small.
1729 A malicious attacker who is able present SM2 content for decryption to
1730 an application could cause attacker chosen data to overflow the buffer
1731 by up to a maximum of 62 bytes altering the contents of other data held
1732 after the buffer, possibly changing application behaviour or causing
1733 the application to crash. The location of the buffer is application
1734 dependent but is typically heap allocated.
1739 * Fixed various read buffer overruns processing ASN.1 strings
1741 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1742 structure which contains a buffer holding the string data and a field
1743 holding the buffer length. This contrasts with normal C strings which
1744 are repesented as a buffer for the string data which is terminated
1745 with a NUL (0) byte.
1747 Although not a strict requirement, ASN.1 strings that are parsed using
1748 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1749 well as any string whose value has been set with the ASN1_STRING_set()
1750 function will additionally NUL terminate the byte array in the
1751 ASN1_STRING structure.
1753 However, it is possible for applications to directly construct valid
1754 ASN1_STRING structures which do not NUL terminate the byte array by
1755 directly setting the "data" and "length" fields in the ASN1_STRING
1756 array. This can also happen by using the ASN1_STRING_set0() function.
1758 Numerous OpenSSL functions that print ASN.1 data have been found to
1759 assume that the ASN1_STRING byte array will be NUL terminated, even
1760 though this is not guaranteed for strings that have been directly
1761 constructed. Where an application requests an ASN.1 structure to be
1762 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1763 been directly constructed by the application without NUL terminating
1764 the "data" field, then a read buffer overrun can occur.
1766 The same thing can also occur during name constraints processing
1767 of certificates (for example if a certificate has been directly
1768 constructed by the application instead of loading it via the OpenSSL
1769 parsing functions, and the certificate contains non NUL terminated
1770 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1771 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1773 If a malicious actor can cause an application to directly construct an
1774 ASN1_STRING and then process it through one of the affected OpenSSL
1775 functions then this issue could be hit. This might result in a crash
1776 (causing a Denial of Service attack). It could also result in the
1777 disclosure of private memory contents (such as private keys, or
1778 sensitive plaintext).
1783 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1785 * Fixed a problem with verifying a certificate chain when using the
1786 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1787 the certificates present in a certificate chain. It is not set by default.
1789 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1790 the chain that have explicitly encoded elliptic curve parameters was added
1791 as an additional strict check.
1793 An error in the implementation of this check meant that the result of a
1794 previous check to confirm that certificates in the chain are valid CA
1795 certificates was overwritten. This effectively bypasses the check
1796 that non-CA certificates must not be able to issue other certificates.
1798 If a "purpose" has been configured then there is a subsequent opportunity
1799 for checks that the certificate is a valid CA. All of the named "purpose"
1800 values implemented in libcrypto perform this check. Therefore, where
1801 a purpose is set the certificate chain will still be rejected even when the
1802 strict flag has been used. A purpose is set by default in libssl client and
1803 server certificate verification routines, but it can be overridden or
1804 removed by an application.
1806 In order to be affected, an application must explicitly set the
1807 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1808 for the certificate verification or, in the case of TLS client or server
1809 applications, override the default purpose.
1814 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1815 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1816 renegotiation ClientHello omits the signature_algorithms extension (where it
1817 was present in the initial ClientHello), but includes a
1818 signature_algorithms_cert extension then a NULL pointer dereference will
1819 result, leading to a crash and a denial of service attack.
1821 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1822 (which is the default configuration). OpenSSL TLS clients are not impacted by
1826 *Peter Kästle and Samuel Sapalski*
1828 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1830 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1831 create a unique hash value based on the issuer and serial number data
1832 contained within an X509 certificate. However it was failing to correctly
1833 handle any errors that may occur while parsing the issuer field (which might
1834 occur if the issuer field is maliciously constructed). This may subsequently
1835 result in a NULL pointer deref and a crash leading to a potential denial of
1841 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1842 padding mode to correctly check for rollback attacks. This is considered a
1843 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1848 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1849 functions. Previously they could overflow the output length argument in some
1850 cases where the input length is close to the maximum permissable length for
1851 an integer on the platform. In such cases the return value from the function
1852 call would be 1 (indicating success), but the output length value would be
1853 negative. This could cause applications to behave incorrectly or crash.
1858 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1859 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1860 could be exploited in a side channel attack to recover the password. Since
1861 the attack is local host only this is outside of the current OpenSSL
1862 threat model and therefore no CVE is assigned.
1864 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1869 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1871 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1872 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1873 If an attacker can control both items being compared then this could lead
1874 to a possible denial of service attack. OpenSSL itself uses the
1875 GENERAL_NAME_cmp function for two purposes:
1876 1) Comparing CRL distribution point names between an available CRL and a
1877 CRL distribution point embedded in an X509 certificate
1878 2) When verifying that a timestamp response token signer matches the
1879 timestamp authority name (exposed via the API functions
1880 TS_RESP_verify_response and TS_RESP_verify_token)
1885 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1887 * Certificates with explicit curve parameters are now disallowed in
1888 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1892 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1893 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1894 conversely, silently ignore DTLS protocol version bounds when configuring
1895 TLS-based contexts. The commands can be repeated to set bounds of both
1896 types. The same applies with the corresponding "min_protocol" and
1897 "max_protocol" command-line switches, in case some application uses both TLS
1900 SSL_CTX instances that are created for a fixed protocol version (e.g.
1901 TLSv1_server_method()) also silently ignore version bounds. Previously
1902 attempts to apply bounds to these protocol versions would result in an
1903 error. Now only the "version-flexible" SSL_CTX instances are subject to
1904 limits in configuration files in command-line options.
1908 * Handshake now fails if Extended Master Secret extension is dropped
1913 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1915 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1917 * Fixed segmentation fault in SSL_check_chain()
1918 Server or client applications that call the SSL_check_chain() function
1919 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1920 dereference as a result of incorrect handling of the
1921 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1922 or unrecognised signature algorithm is received from the peer. This could
1923 be exploited by a malicious peer in a Denial of Service attack.
1928 * Added AES consttime code for no-asm configurations
1929 an optional constant time support for AES was added
1930 when building openssl for no-asm.
1931 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1932 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1933 At this time this feature is by default disabled.
1934 It will be enabled by default in 3.0.
1938 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1940 * Revert the change of EOF detection while reading in libssl to avoid
1941 regressions in applications depending on the current way of reporting
1942 the EOF. As the existing method is not fully accurate the change to
1943 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1944 branch and will be present in the 3.0 release.
1948 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1949 when primes for RSA keys are computed.
1950 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1951 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1952 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1953 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1954 This avoids possible fingerprinting of newly generated RSA modules.
1958 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1960 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1961 while reading in libssl then we would report an error back to the
1962 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1963 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1964 therefore give a hint as to what went wrong.
1968 * Check that ed25519 and ed448 are allowed by the security level. Previously
1969 signature algorithms not using an MD were not being checked that they were
1970 allowed by the security level.
1974 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1975 was not quite right. The behaviour was not consistent between resumption
1976 and normal handshakes, and also not quite consistent with historical
1977 behaviour. The behaviour in various scenarios has been clarified and
1978 it has been updated to make it match historical behaviour as closely as
1983 * *[VMS only]* The header files that the VMS compilers include automatically,
1984 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1985 that the C++ compiler doesn't understand. This is a shortcoming in the
1986 compiler, but can be worked around with `__cplusplus` guards.
1988 C++ applications that use OpenSSL libraries must be compiled using the
1989 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1990 functions. Otherwise, only functions with symbols of less than 31
1991 characters can be used, as the linker will not be able to successfully
1992 resolve symbols with longer names.
1996 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1997 The presence of this system service is determined at run-time.
2001 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2006 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2008 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2009 number generator (RNG). This was intended to include protection in the
2010 event of a fork() system call in order to ensure that the parent and child
2011 processes did not share the same RNG state. However this protection was not
2012 being used in the default case.
2014 A partial mitigation for this issue is that the output from a high
2015 precision timer is mixed into the RNG state so the likelihood of a parent
2016 and child process sharing state is significantly reduced.
2018 If an application already calls OPENSSL_init_crypto() explicitly using
2019 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2022 *Matthias St. Pierre*
2024 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2025 used even when parsing explicit parameters, when loading a encoded key
2026 or calling `EC_GROUP_new_from_ecpkparameters()`/
2027 `EC_GROUP_new_from_ecparameters()`.
2028 This prevents bypass of security hardening and performance gains,
2029 especially for curves with specialized EC_METHODs.
2030 By default, if a key encoded with explicit parameters is loaded and later
2031 encoded, the output is still encoded with explicit parameters, even if
2032 internally a "named" EC_GROUP is used for computation.
2036 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2037 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2038 NULL. After this change, only the cofactor parameter can be NULL. It also
2039 does some minimal sanity checks on the passed order.
2044 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2045 An attack is simple, if the first CMS_recipientInfo is valid but the
2046 second CMS_recipientInfo is chosen ciphertext. If the second
2047 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2048 encryption key will be replaced by garbage, and the message cannot be
2049 decoded, but if the RSA decryption fails, the correct encryption key is
2050 used and the recipient will not notice the attack.
2051 As a work around for this potential attack the length of the decrypted
2052 key must be equal to the cipher default key length, in case the
2053 certifiate is not given and all recipientInfo are tried out.
2054 The old behaviour can be re-enabled in the CMS code by setting the
2055 CMS_DEBUG_DECRYPT flag.
2060 * Early start up entropy quality from the DEVRANDOM seed source has been
2061 improved for older Linux systems. The RAND subsystem will wait for
2062 /dev/random to be producing output before seeding from /dev/urandom.
2063 The seeded state is stored for future library initialisations using
2064 a system global shared memory segment. The shared memory identifier
2065 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2066 the desired value. The default identifier is 114.
2070 * Correct the extended master secret constant on EBCDIC systems. Without this
2071 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2072 negotiate EMS will fail. Unfortunately this also means that TLS connections
2073 between EBCDIC systems with this fix, and EBCDIC systems without this
2074 fix will fail if they negotiate EMS.
2078 * Use Windows installation paths in the mingw builds
2080 Mingw isn't a POSIX environment per se, which means that Windows
2081 paths should be used for installation.
2086 * Changed DH_check to accept parameters with order q and 2q subgroups.
2087 With order 2q subgroups the bit 0 of the private key is not secret
2088 but DH_generate_key works around that by clearing bit 0 of the
2089 private key for those. This avoids leaking bit 0 of the private key.
2093 * Significantly reduce secure memory usage by the randomness pools.
2097 * Revert the DEVRANDOM_WAIT feature for Linux systems
2099 The DEVRANDOM_WAIT feature added a select() call to wait for the
2100 /dev/random device to become readable before reading from the
2101 /dev/urandom device.
2103 It turned out that this change had negative side effects on
2104 performance which were not acceptable. After some discussion it
2105 was decided to revert this feature and leave it up to the OS
2106 resp. the platform maintainer to ensure a proper initialization
2107 during early boot time.
2109 *Matthias St. Pierre*
2111 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2113 * Add build tests for C++. These are generated files that only do one
2114 thing, to include one public OpenSSL head file each. This tests that
2115 the public header files can be usefully included in a C++ application.
2117 This test isn't enabled by default. It can be enabled with the option
2118 'enable-buildtest-c++'.
2122 * Enable SHA3 pre-hashing for ECDSA and DSA.
2126 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2127 This changes the size when using the `genpkey` command when no size is given.
2128 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2129 generation commands to use 2048 bits by default.
2133 * Reorganize the manual pages to consistently have RETURN VALUES,
2134 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2135 util/fix-doc-nits accordingly.
2137 *Paul Yang, Joshua Lock*
2139 * Add the missing accessor EVP_PKEY_get0_engine()
2143 * Have commands like `s_client` and `s_server` output the signature scheme
2144 along with other cipher suite parameters when debugging.
2148 * Make OPENSSL_config() error agnostic again.
2152 * Do the error handling in RSA decryption constant time.
2156 * Prevent over long nonces in ChaCha20-Poly1305.
2158 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2159 for every encryption operation. RFC 7539 specifies that the nonce value
2160 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2161 and front pads the nonce with 0 bytes if it is less than 12
2162 bytes. However it also incorrectly allows a nonce to be set of up to 16
2163 bytes. In this case only the last 12 bytes are significant and any
2164 additional leading bytes are ignored.
2166 It is a requirement of using this cipher that nonce values are
2167 unique. Messages encrypted using a reused nonce value are susceptible to
2168 serious confidentiality and integrity attacks. If an application changes
2169 the default nonce length to be longer than 12 bytes and then makes a
2170 change to the leading bytes of the nonce expecting the new value to be a
2171 new unique nonce then such an application could inadvertently encrypt
2172 messages with a reused nonce.
2174 Additionally the ignored bytes in a long nonce are not covered by the
2175 integrity guarantee of this cipher. Any application that relies on the
2176 integrity of these ignored leading bytes of a long nonce may be further
2177 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2178 is safe because no such use sets such a long nonce value. However user
2179 applications that use this cipher directly and set a non-default nonce
2180 length to be longer than 12 bytes may be vulnerable.
2182 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2188 * Add DEVRANDOM_WAIT feature for Linux systems
2190 On older Linux systems where the getrandom() system call is not available,
2191 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2192 Contrary to getrandom(), the /dev/urandom device will not block during
2193 early boot when the kernel CSPRNG has not been seeded yet.
2195 To mitigate this known weakness, use select() to wait for /dev/random to
2196 become readable before reading from /dev/urandom.
2198 * Ensure that SM2 only uses SM3 as digest algorithm
2202 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2204 * Change the info callback signals for the start and end of a post-handshake
2205 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2206 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2207 confused by this and assume that a TLSv1.2 renegotiation has started. This
2208 can break KeyUpdate handling. Instead we no longer signal the start and end
2209 of a post handshake message exchange (although the messages themselves are
2210 still signalled). This could break some applications that were expecting
2211 the old signals. However without this KeyUpdate is not usable for many
2216 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2218 * Timing vulnerability in DSA signature generation
2220 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2221 timing side channel attack. An attacker could use variations in the signing
2222 algorithm to recover the private key.
2224 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2229 * Timing vulnerability in ECDSA signature generation
2231 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2232 timing side channel attack. An attacker could use variations in the signing
2233 algorithm to recover the private key.
2235 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2240 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2241 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2242 of two gigabytes and the error handling improved.
2244 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2245 categorized as a normal bug, not a security issue, because the DRBG reseeds
2246 automatically and is fully functional even without additional randomness
2247 provided by the application.
2249 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2251 * Add a new ClientHello callback. Provides a callback interface that gives
2252 the application the ability to adjust the nascent SSL object at the
2253 earliest stage of ClientHello processing, immediately after extensions have
2254 been collected but before they have been processed. In particular, this
2255 callback can adjust the supported TLS versions in response to the contents
2260 * Add SM2 base algorithm support.
2264 * s390x assembly pack: add (improved) hardware-support for the following
2265 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2266 aes-cfb/cfb8, aes-ecb.
2270 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2271 parameter is no longer accepted, as it leads to a corrupt table. NULL
2272 pem_str is reserved for alias entries only.
2276 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2277 step for prime curves. The new implementation is based on formulae from
2278 differential addition-and-doubling in homogeneous projective coordinates
2279 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2280 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2281 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2282 to work in projective coordinates.
2284 *Billy Bob Brumley, Nicola Tuveri*
2286 * Change generating and checking of primes so that the error rate of not
2287 being prime depends on the intended use based on the size of the input.
2288 For larger primes this will result in more rounds of Miller-Rabin.
2289 The maximal error rate for primes with more than 1080 bits is lowered
2292 *Kurt Roeckx, Annie Yousar*
2294 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2298 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2299 moving between systems, and to avoid confusion when a Windows build is
2300 done with mingw vs with MSVC. For POSIX installs, there's still a
2301 symlink or copy named 'tsget' to avoid that confusion as well.
2305 * Revert blinding in ECDSA sign and instead make problematic addition
2306 length-invariant. Switch even to fixed-length Montgomery multiplication.
2310 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2311 step for binary curves. The new implementation is based on formulae from
2312 differential addition-and-doubling in mixed Lopez-Dahab projective
2313 coordinates, modified to independently blind the operands.
2315 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2317 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2318 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2319 EC_METHODs to implement their own specialized "ladder step", to take
2320 advantage of more favorable coordinate systems or more efficient
2321 differential addition-and-doubling algorithms.
2323 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2325 * Modified the random device based seed sources to keep the relevant
2326 file descriptors open rather than reopening them on each access.
2327 This allows such sources to operate in a chroot() jail without
2328 the associated device nodes being available. This behaviour can be
2329 controlled using RAND_keep_random_devices_open().
2333 * Numerous side-channel attack mitigations have been applied. This may have
2334 performance impacts for some algorithms for the benefit of improved
2335 security. Specific changes are noted in this change log by their respective
2340 * AIX shared library support overhaul. Switch to AIX "natural" way of
2341 handling shared libraries, which means collecting shared objects of
2342 different versions and bitnesses in one common archive. This allows to
2343 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2344 doesn't affect the way 3rd party applications are linked, only how
2345 multi-version installation is managed.
2349 * Make ec_group_do_inverse_ord() more robust and available to other
2350 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2351 mitigations are applied to the fallback BN_mod_inverse().
2352 When using this function rather than BN_mod_inverse() directly, new
2353 EC cryptosystem implementations are then safer-by-default.
2357 * Add coordinate blinding for EC_POINT and implement projective
2358 coordinate blinding for generic prime curves as a countermeasure to
2359 chosen point SCA attacks.
2361 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2363 * Add blinding to ECDSA and DSA signatures to protect against side channel
2364 attacks discovered by Keegan Ryan (NCC Group).
2368 * Enforce checking in the `pkeyutl` command to ensure that the input
2369 length does not exceed the maximum supported digest length when performing
2370 a sign, verify or verifyrecover operation.
2374 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2375 I/O in combination with something like select() or poll() will hang. This
2376 can be turned off again using SSL_CTX_clear_mode().
2377 Many applications do not properly handle non-application data records, and
2378 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2379 around the problems in those applications, but can also break some.
2380 It's recommended to read the manpages about SSL_read(), SSL_write(),
2381 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2382 SSL_CTX_set_read_ahead() again.
2386 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2387 now allow empty (zero character) pass phrases.
2391 * Apply blinding to binary field modular inversion and remove patent
2392 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2396 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2397 binary and prime elliptic curves.
2401 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2402 constant time fixed point multiplication.
2406 * Revise elliptic curve scalar multiplication with timing attack
2407 defenses: ec_wNAF_mul redirects to a constant time implementation
2408 when computing fixed point and variable point multiplication (which
2409 in OpenSSL are mostly used with secret scalars in keygen, sign,
2410 ECDH derive operations).
2411 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2414 * Updated CONTRIBUTING
2418 * Updated DRBG / RAND to request nonce and additional low entropy
2419 randomness from the system.
2421 *Matthias St. Pierre*
2423 * Updated 'openssl rehash' to use OpenSSL consistent default.
2427 * Moved the load of the ssl_conf module to libcrypto, which helps
2428 loading engines that libssl uses before libssl is initialised.
2432 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2436 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2438 *Ingo Schwarze, Rich Salz*
2440 * Added output of accepting IP address and port for 'openssl s_server'
2444 * Added a new API for TLSv1.3 ciphersuites:
2445 SSL_CTX_set_ciphersuites()
2446 SSL_set_ciphersuites()
2450 * Memory allocation failures consistently add an error to the error
2455 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2456 in libcrypto when run as setuid/setgid.
2460 * Load any config file by default when libssl is used.
2464 * Added new public header file <openssl/rand_drbg.h> and documentation
2465 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2467 *Matthias St. Pierre*
2469 * QNX support removed (cannot find contributors to get their approval
2470 for the license change).
2474 * TLSv1.3 replay protection for early data has been implemented. See the
2475 SSL_read_early_data() man page for further details.
2479 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2480 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2481 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2482 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2483 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2484 configuration has been separated out. See the ciphers man page or the
2485 SSL_CTX_set_ciphersuites() man page for more information.
2489 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2490 in responder mode now supports the new "-multi" option, which
2491 spawns the specified number of child processes to handle OCSP
2492 requests. The "-timeout" option now also limits the OCSP
2493 responder's patience to wait to receive the full client request
2494 on a newly accepted connection. Child processes are respawned
2495 as needed, and the CA index file is automatically reloaded
2496 when changed. This makes it possible to run the "ocsp" responder
2497 as a long-running service, making the OpenSSL CA somewhat more
2498 feature-complete. In this mode, most diagnostic messages logged
2499 after entering the event loop are logged via syslog(3) rather than
2504 * Added support for X448 and Ed448. Heavily based on original work by
2509 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2510 objects loaded. This adds the functions OSSL_STORE_expect() and
2511 OSSL_STORE_find() as well as needed tools to construct searches and
2512 get the search data out of them.
2516 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2517 version of OpenSSL should review their configuration settings to ensure
2518 that they are still appropriate for TLSv1.3. For further information see:
2519 <https://wiki.openssl.org/index.php/TLS1.3>
2523 * Grand redesign of the OpenSSL random generator
2525 The default RAND method now utilizes an AES-CTR DRBG according to
2526 NIST standard SP 800-90Ar1. The new random generator is essentially
2527 a port of the default random generator from the OpenSSL FIPS 2.0
2528 object module. It is a hybrid deterministic random bit generator
2529 using an AES-CTR bit stream and which seeds and reseeds itself
2530 automatically using trusted system entropy sources.
2532 Some of its new features are:
2533 - Support for multiple DRBG instances with seed chaining.
2534 - The default RAND method makes use of a DRBG.
2535 - There is a public and private DRBG instance.
2536 - The DRBG instances are fork-safe.
2537 - Keep all global DRBG instances on the secure heap if it is enabled.
2538 - The public and private DRBG instance are per thread for lock free
2541 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2543 * Changed Configure so it only says what it does and doesn't dump
2544 so much data. Instead, ./configdata.pm should be used as a script
2545 to display all sorts of configuration data.
2549 * Added processing of "make variables" to Configure.
2553 * Added SHA512/224 and SHA512/256 algorithm support.
2557 * The last traces of Netware support, first removed in 1.1.0, have
2562 * Get rid of Makefile.shared, and in the process, make the processing
2563 of certain files (rc.obj, or the .def/.map/.opt files produced from
2564 the ordinal files) more visible and hopefully easier to trace and
2565 debug (or make silent).
2569 * Make it possible to have environment variable assignments as
2570 arguments to config / Configure.
2574 * Add multi-prime RSA (RFC 8017) support.
2578 * Add SM3 implemented according to GB/T 32905-2016
2579 *Jack Lloyd <jack.lloyd@ribose.com>,*
2580 *Ronald Tse <ronald.tse@ribose.com>,*
2581 *Erick Borsboom <erick.borsboom@ribose.com>*
2583 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2584 as documented in RFC6066.
2585 Based on a patch from Tomasz Moń
2587 *Filipe Raimundo da Silva*
2589 * Add SM4 implemented according to GB/T 32907-2016.
2590 *Jack Lloyd <jack.lloyd@ribose.com>,*
2591 *Ronald Tse <ronald.tse@ribose.com>,*
2592 *Erick Borsboom <erick.borsboom@ribose.com>*
2594 * Reimplement -newreq-nodes and ERR_error_string_n; the
2595 original author does not agree with the license change.
2599 * Add ARIA AEAD TLS support.
2603 * Some macro definitions to support VS6 have been removed. Visual
2604 Studio 6 has not worked since 1.1.0
2608 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2609 without clearing the errors.
2613 * Add "atfork" functions. If building on a system that without
2614 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2615 requirements. The RAND facility now uses/requires this.
2623 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2624 not possible to disable entirely. However, it's still possible to
2625 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2628 To disable, configure with 'no-ui-console'. 'no-ui' is still
2629 possible to use as an alias. Check at compile time with the
2630 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2631 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2635 * Add a STORE module, which implements a uniform and URI based reader of
2636 stores that can contain keys, certificates, CRLs and numerous other
2637 objects. The main API is loosely based on a few stdio functions,
2638 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2639 OSSL_STORE_error and OSSL_STORE_close.
2640 The implementation uses backends called "loaders" to implement arbitrary
2641 URI schemes. There is one built in "loader" for the 'file' scheme.
2645 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2646 then adjusted to work on FreeBSD 8.4 as well.
2647 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2648 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2652 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2653 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2654 error code calls like this:
2656 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2658 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2659 that can be encoded in C. For the foreseeable future, this will only
2662 *Richard Levitte and Tim Hudson*
2664 * Removed BSD cryptodev engine.
2668 * Add a build target 'build_all_generated', to build all generated files
2669 and only that. This can be used to prepare everything that requires
2670 things like perl for a system that lacks perl and then move everything
2671 to that system and do the rest of the build there.
2675 * In the UI interface, make it possible to duplicate the user data. This
2676 can be used by engines that need to retain the data for a longer time
2677 than just the call where this user data is passed.
2681 * Ignore the '-named_curve auto' value for compatibility of applications
2684 *Tomáš Mráz <tmraz@fedoraproject.org>*
2686 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2687 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2688 alerts across multiple records (some of which could be empty). In practice
2689 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2690 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2691 support this at all. Supporting it adds significant complexity to the
2692 record layer, and its removal is unlikely to cause interoperability
2697 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2698 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2699 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2704 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2705 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2707 *Richard Levitte, Andy Polyakov*
2709 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2714 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2715 platform rather than 'mingw'.
2719 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2720 success if they are asked to add an object which already exists
2721 in the store. This change cascades to other functions which load
2722 certificates and CRLs.
2726 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2727 facilitate stack unwinding even from assembly subroutines.
2731 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2732 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2736 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2737 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2738 which is the minimum version we support.
2742 * Certificate time validation (X509_cmp_time) enforces stricter
2743 compliance with RFC 5280. Fractional seconds and timezone offsets
2744 are no longer allowed.
2748 * Add support for ARIA
2752 * s_client will now send the Server Name Indication (SNI) extension by
2753 default unless the new "-noservername" option is used. The server name is
2754 based on the host provided to the "-connect" option unless overridden by
2755 using "-servername".
2759 * Add support for SipHash
2763 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2764 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2765 prevent issues where no progress is being made and the peer continually
2766 sends unrecognised record types, using up resources processing them.
2770 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2771 using the algorithm defined in
2772 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2776 * Heartbeat support has been removed; the ABI is changed for now.
2778 *Richard Levitte, Rich Salz*
2780 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2784 * The RSA "null" method, which was partially supported to avoid patent
2785 issues, has been replaced to always returns NULL.
2792 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2794 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2795 used even when parsing explicit parameters, when loading a encoded key
2796 or calling `EC_GROUP_new_from_ecpkparameters()`/
2797 `EC_GROUP_new_from_ecparameters()`.
2798 This prevents bypass of security hardening and performance gains,
2799 especially for curves with specialized EC_METHODs.
2800 By default, if a key encoded with explicit parameters is loaded and later
2801 encoded, the output is still encoded with explicit parameters, even if
2802 internally a "named" EC_GROUP is used for computation.
2806 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2807 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2808 NULL. After this change, only the cofactor parameter can be NULL. It also
2809 does some minimal sanity checks on the passed order.
2814 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2815 An attack is simple, if the first CMS_recipientInfo is valid but the
2816 second CMS_recipientInfo is chosen ciphertext. If the second
2817 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2818 encryption key will be replaced by garbage, and the message cannot be
2819 decoded, but if the RSA decryption fails, the correct encryption key is
2820 used and the recipient will not notice the attack.
2821 As a work around for this potential attack the length of the decrypted
2822 key must be equal to the cipher default key length, in case the
2823 certifiate is not given and all recipientInfo are tried out.
2824 The old behaviour can be re-enabled in the CMS code by setting the
2825 CMS_DEBUG_DECRYPT flag.
2830 * Use Windows installation paths in the mingw builds
2832 Mingw isn't a POSIX environment per se, which means that Windows
2833 paths should be used for installation.
2838 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2840 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2841 This changes the size when using the `genpkey` command when no size is given.
2842 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2843 generation commands to use 2048 bits by default.
2847 * Prevent over long nonces in ChaCha20-Poly1305.
2849 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2850 for every encryption operation. RFC 7539 specifies that the nonce value
2851 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2852 and front pads the nonce with 0 bytes if it is less than 12
2853 bytes. However it also incorrectly allows a nonce to be set of up to 16
2854 bytes. In this case only the last 12 bytes are significant and any
2855 additional leading bytes are ignored.
2857 It is a requirement of using this cipher that nonce values are
2858 unique. Messages encrypted using a reused nonce value are susceptible to
2859 serious confidentiality and integrity attacks. If an application changes
2860 the default nonce length to be longer than 12 bytes and then makes a
2861 change to the leading bytes of the nonce expecting the new value to be a
2862 new unique nonce then such an application could inadvertently encrypt
2863 messages with a reused nonce.
2865 Additionally the ignored bytes in a long nonce are not covered by the
2866 integrity guarantee of this cipher. Any application that relies on the
2867 integrity of these ignored leading bytes of a long nonce may be further
2868 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2869 is safe because no such use sets such a long nonce value. However user
2870 applications that use this cipher directly and set a non-default nonce
2871 length to be longer than 12 bytes may be vulnerable.
2873 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2879 * Added SCA hardening for modular field inversion in EC_GROUP through
2880 a new dedicated field_inv() pointer in EC_METHOD.
2881 This also addresses a leakage affecting conversions from projective
2882 to affine coordinates.
2884 *Billy Bob Brumley, Nicola Tuveri*
2886 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2887 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2891 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2895 * Remove the 'dist' target and add a tarball building script. The
2896 'dist' target has fallen out of use, and it shouldn't be
2897 necessary to configure just to create a source distribution.
2901 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2903 * Timing vulnerability in DSA signature generation
2905 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2906 timing side channel attack. An attacker could use variations in the signing
2907 algorithm to recover the private key.
2909 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2914 * Timing vulnerability in ECDSA signature generation
2916 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2917 timing side channel attack. An attacker could use variations in the signing
2918 algorithm to recover the private key.
2920 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2925 * Add coordinate blinding for EC_POINT and implement projective
2926 coordinate blinding for generic prime curves as a countermeasure to
2927 chosen point SCA attacks.
2929 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2931 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2933 * Client DoS due to large DH parameter
2935 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2936 malicious server can send a very large prime value to the client. This will
2937 cause the client to spend an unreasonably long period of time generating a
2938 key for this prime resulting in a hang until the client has finished. This
2939 could be exploited in a Denial Of Service attack.
2941 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2946 * Cache timing vulnerability in RSA Key Generation
2948 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2949 a cache timing side channel attack. An attacker with sufficient access to
2950 mount cache timing attacks during the RSA key generation process could
2951 recover the private key.
2953 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2954 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2959 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2960 parameter is no longer accepted, as it leads to a corrupt table. NULL
2961 pem_str is reserved for alias entries only.
2965 * Revert blinding in ECDSA sign and instead make problematic addition
2966 length-invariant. Switch even to fixed-length Montgomery multiplication.
2970 * Change generating and checking of primes so that the error rate of not
2971 being prime depends on the intended use based on the size of the input.
2972 For larger primes this will result in more rounds of Miller-Rabin.
2973 The maximal error rate for primes with more than 1080 bits is lowered
2976 *Kurt Roeckx, Annie Yousar*
2978 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2982 * Add blinding to ECDSA and DSA signatures to protect against side channel
2983 attacks discovered by Keegan Ryan (NCC Group).
2987 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2988 now allow empty (zero character) pass phrases.
2992 * Certificate time validation (X509_cmp_time) enforces stricter
2993 compliance with RFC 5280. Fractional seconds and timezone offsets
2994 are no longer allowed.
2998 * Fixed a text canonicalisation bug in CMS
3000 Where a CMS detached signature is used with text content the text goes
3001 through a canonicalisation process first prior to signing or verifying a
3002 signature. This process strips trailing space at the end of lines, converts
3003 line terminators to CRLF and removes additional trailing line terminators
3004 at the end of a file. A bug in the canonicalisation process meant that
3005 some characters, such as form-feed, were incorrectly treated as whitespace
3006 and removed. This is contrary to the specification (RFC5485). This fix
3007 could mean that detached text data signed with an earlier version of
3008 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3009 signed with a fixed OpenSSL may fail to verify with an earlier version of
3010 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3011 and use the "-binary" flag (for the "cms" command line application) or set
3012 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3016 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3018 * Constructed ASN.1 types with a recursive definition could exceed the stack
3020 Constructed ASN.1 types with a recursive definition (such as can be found
3021 in PKCS7) could eventually exceed the stack given malicious input with
3022 excessive recursion. This could result in a Denial Of Service attack. There
3023 are no such structures used within SSL/TLS that come from untrusted sources
3024 so this is considered safe.
3026 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3032 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3034 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3035 effectively reduced to only comparing the least significant bit of each
3036 byte. This allows an attacker to forge messages that would be considered as
3037 authenticated in an amount of tries lower than that guaranteed by the
3038 security claims of the scheme. The module can only be compiled by the
3039 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3041 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3047 * Add a build target 'build_all_generated', to build all generated files
3048 and only that. This can be used to prepare everything that requires
3049 things like perl for a system that lacks perl and then move everything
3050 to that system and do the rest of the build there.
3054 * Backport SSL_OP_NO_RENGOTIATION
3056 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3057 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3058 changes this is no longer possible in 1.1.0. Therefore the new
3059 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3060 1.1.0 to provide equivalent functionality.
3062 Note that if an application built against 1.1.0h headers (or above) is run
3063 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3064 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3068 * Removed the OS390-Unix config target. It relied on a script that doesn't
3073 * rsaz_1024_mul_avx2 overflow bug on x86_64
3075 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3076 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3077 Analysis suggests that attacks against RSA and DSA as a result of this
3078 defect would be very difficult to perform and are not believed likely.
3079 Attacks against DH1024 are considered just feasible, because most of the
3080 work necessary to deduce information about a private key may be performed
3081 offline. The amount of resources required for such an attack would be
3082 significant. However, for an attack on TLS to be meaningful, the server
3083 would have to share the DH1024 private key among multiple clients, which is
3084 no longer an option since CVE-2016-0701.
3086 This only affects processors that support the AVX2 but not ADX extensions
3087 like Intel Haswell (4th generation).
3089 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3090 was originally found via the OSS-Fuzz project.
3095 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3097 * bn_sqrx8x_internal carry bug on x86_64
3099 There is a carry propagating bug in the x86_64 Montgomery squaring
3100 procedure. No EC algorithms are affected. Analysis suggests that attacks
3101 against RSA and DSA as a result of this defect would be very difficult to
3102 perform and are not believed likely. Attacks against DH are considered just
3103 feasible (although very difficult) because most of the work necessary to
3104 deduce information about a private key may be performed offline. The amount
3105 of resources required for such an attack would be very significant and
3106 likely only accessible to a limited number of attackers. An attacker would
3107 additionally need online access to an unpatched system using the target
3108 private key in a scenario with persistent DH parameters and a private
3109 key that is shared between multiple clients.
3111 This only affects processors that support the BMI1, BMI2 and ADX extensions
3112 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3114 This issue was reported to OpenSSL by the OSS-Fuzz project.
3119 * Malformed X.509 IPAddressFamily could cause OOB read
3121 If an X.509 certificate has a malformed IPAddressFamily extension,
3122 OpenSSL could do a one-byte buffer overread. The most likely result
3123 would be an erroneous display of the certificate in text format.
3125 This issue was reported to OpenSSL by the OSS-Fuzz project.
3130 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3132 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3133 platform rather than 'mingw'.
3137 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3138 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3139 which is the minimum version we support.
3143 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3145 * Encrypt-Then-Mac renegotiation crash
3147 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3148 negotiated where it was not in the original handshake (or vice-versa) then
3149 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3150 and servers are affected.
3152 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3157 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3159 * Truncated packet could crash via OOB read
3161 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3162 cipher is being used, then a truncated packet can cause that host to
3163 perform an out-of-bounds read, usually resulting in a crash.
3165 This issue was reported to OpenSSL by Robert Święcki of Google.
3170 * Bad (EC)DHE parameters cause a client crash
3172 If a malicious server supplies bad parameters for a DHE or ECDHE key
3173 exchange then this can result in the client attempting to dereference a
3174 NULL pointer leading to a client crash. This could be exploited in a Denial
3177 This issue was reported to OpenSSL by Guido Vranken.
3182 * BN_mod_exp may produce incorrect results on x86_64
3184 There is a carry propagating bug in the x86_64 Montgomery squaring
3185 procedure. No EC algorithms are affected. Analysis suggests that attacks
3186 against RSA and DSA as a result of this defect would be very difficult to
3187 perform and are not believed likely. Attacks against DH are considered just
3188 feasible (although very difficult) because most of the work necessary to
3189 deduce information about a private key may be performed offline. The amount
3190 of resources required for such an attack would be very significant and
3191 likely only accessible to a limited number of attackers. An attacker would
3192 additionally need online access to an unpatched system using the target
3193 private key in a scenario with persistent DH parameters and a private
3194 key that is shared between multiple clients. For example this can occur by
3195 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3196 similar to CVE-2015-3193 but must be treated as a separate problem.
3198 This issue was reported to OpenSSL by the OSS-Fuzz project.
3203 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3205 * ChaCha20/Poly1305 heap-buffer-overflow
3207 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3208 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3209 crash. This issue is not considered to be exploitable beyond a DoS.
3211 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3216 * CMS Null dereference
3218 Applications parsing invalid CMS structures can crash with a NULL pointer
3219 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3220 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3221 structure callback if an attempt is made to free certain invalid encodings.
3222 Only CHOICE structures using a callback which do not handle NULL value are
3225 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3230 * Montgomery multiplication may produce incorrect results
3232 There is a carry propagating bug in the Broadwell-specific Montgomery
3233 multiplication procedure that handles input lengths divisible by, but
3234 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3235 and DH private keys are impossible. This is because the subroutine in
3236 question is not used in operations with the private key itself and an input
3237 of the attacker's direct choice. Otherwise the bug can manifest itself as
3238 transient authentication and key negotiation failures or reproducible
3239 erroneous outcome of public-key operations with specially crafted input.
3240 Among EC algorithms only Brainpool P-512 curves are affected and one
3241 presumably can attack ECDH key negotiation. Impact was not analyzed in
3242 detail, because pre-requisites for attack are considered unlikely. Namely
3243 multiple clients have to choose the curve in question and the server has to
3244 share the private key among them, neither of which is default behaviour.
3245 Even then only clients that chose the curve will be affected.
3247 This issue was publicly reported as transient failures and was not
3248 initially recognized as a security issue. Thanks to Richard Morgan for
3249 providing reproducible case.
3254 * Removed automatic addition of RPATH in shared libraries and executables,
3255 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3259 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3261 * Fix Use After Free for large message sizes
3263 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3264 message larger than approx 16k is received then the underlying buffer to
3265 store the incoming message is reallocated and moved. Unfortunately a
3266 dangling pointer to the old location is left which results in an attempt to
3267 write to the previously freed location. This is likely to result in a
3268 crash, however it could potentially lead to execution of arbitrary code.
3270 This issue only affects OpenSSL 1.1.0a.
3272 This issue was reported to OpenSSL by Robert Święcki.
3277 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3279 * OCSP Status Request extension unbounded memory growth
3281 A malicious client can send an excessively large OCSP Status Request
3282 extension. If that client continually requests renegotiation, sending a
3283 large OCSP Status Request extension each time, then there will be unbounded
3284 memory growth on the server. This will eventually lead to a Denial Of
3285 Service attack through memory exhaustion. Servers with a default
3286 configuration are vulnerable even if they do not support OCSP. Builds using
3287 the "no-ocsp" build time option are not affected.
3289 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3294 * SSL_peek() hang on empty record
3296 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3297 sends an empty record. This could be exploited by a malicious peer in a
3298 Denial Of Service attack.
3300 This issue was reported to OpenSSL by Alex Gaynor.
3305 * Excessive allocation of memory in tls_get_message_header() and
3306 dtls1_preprocess_fragment()
3308 A (D)TLS message includes 3 bytes for its length in the header for the
3309 message. This would allow for messages up to 16Mb in length. Messages of
3310 this length are excessive and OpenSSL includes a check to ensure that a
3311 peer is sending reasonably sized messages in order to avoid too much memory
3312 being consumed to service a connection. A flaw in the logic of version
3313 1.1.0 means that memory for the message is allocated too early, prior to
3314 the excessive message length check. Due to way memory is allocated in
3315 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3316 to service a connection. This could lead to a Denial of Service through
3317 memory exhaustion. However, the excessive message length check still takes
3318 place, and this would cause the connection to immediately fail. Assuming
3319 that the application calls SSL_free() on the failed connection in a timely
3320 manner then the 21Mb of allocated memory will then be immediately freed
3321 again. Therefore the excessive memory allocation will be transitory in
3322 nature. This then means that there is only a security impact if:
3324 1) The application does not call SSL_free() in a timely manner in the event
3325 that the connection fails
3327 2) The application is working in a constrained environment where there is
3328 very little free memory
3330 3) The attacker initiates multiple connection attempts such that there are
3331 multiple connections in a state where memory has been allocated for the
3332 connection; SSL_free() has not yet been called; and there is insufficient
3333 memory to service the multiple requests.
3335 Except in the instance of (1) above any Denial Of Service is likely to be
3336 transitory because as soon as the connection fails the memory is
3337 subsequently freed again in the SSL_free() call. However there is an
3338 increased risk during this period of application crashes due to the lack of
3339 memory - which would then mean a more serious Denial of Service.
3341 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3342 (CVE-2016-6307 and CVE-2016-6308)
3346 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3347 had to be removed. Primary reason is that vendor assembler can't
3348 assemble our modules with -KPIC flag. As result it, assembly
3349 support, was not even available as option. But its lack means
3350 lack of side-channel resistant code, which is incompatible with
3351 security by todays standards. Fortunately gcc is readily available
3352 prepackaged option, which we firmly point at...
3356 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3358 * Windows command-line tool supports UTF-8 opt-in option for arguments
3359 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3360 (to any value) allows Windows user to access PKCS#12 file generated
3361 with Windows CryptoAPI and protected with non-ASCII password, as well
3362 as files generated under UTF-8 locale on Linux also protected with
3367 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3368 have been disabled by default and removed from DEFAULT, just like RC4.
3369 See the RC4 item below to re-enable both.
3373 * The method for finding the storage location for the Windows RAND seed file
3374 has changed. First we check %RANDFILE%. If that is not set then we check
3375 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3376 all else fails we fall back to C:\.
3380 * The EVP_EncryptUpdate() function has had its return type changed from void
3381 to int. A return of 0 indicates and error while a return of 1 indicates
3386 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3387 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3388 off the constant time implementation for RSA, DSA and DH have been made
3389 no-ops and deprecated.
3393 * Windows RAND implementation was simplified to only get entropy by
3394 calling CryptGenRandom(). Various other RAND-related tickets
3397 *Joseph Wylie Yandle, Rich Salz*
3399 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3400 and `OPENSSL_LH_`, respectively. The old names are available
3401 with API compatibility. They new names are now completely documented.
3405 * Unify TYPE_up_ref(obj) methods signature.
3406 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3407 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3408 int (instead of void) like all others TYPE_up_ref() methods.
3409 So now these methods also check the return value of CRYPTO_atomic_add(),
3410 and the validity of object reference counter.
3412 *fdasilvayy@gmail.com*
3414 * With Windows Visual Studio builds, the .pdb files are installed
3415 alongside the installed libraries and executables. For a static
3416 library installation, ossl_static.pdb is the associate compiler
3417 generated .pdb file to be used when linking programs.
3421 * Remove openssl.spec. Packaging files belong with the packagers.
3425 * Automatic Darwin/OSX configuration has had a refresh, it will now
3426 recognise x86_64 architectures automatically. You can still decide
3427 to build for a different bitness with the environment variable
3428 KERNEL_BITS (can be 32 or 64), for example:
3430 KERNEL_BITS=32 ./config
3434 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3435 256 bit AES and HMAC with SHA256.
3439 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3443 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3447 * To enable users to have their own config files and build file templates,
3448 Configure looks in the directory indicated by the environment variable
3449 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3450 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3451 name and is used as is.
3455 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3456 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3457 X509_CERT_FILE_CTX was removed.
3461 * "shared" builds are now the default. To create only static libraries use
3462 the "no-shared" Configure option.
3466 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3467 All of these option have not worked for some while and are fundamental
3472 * Make various cleanup routines no-ops and mark them as deprecated. Most
3473 global cleanup functions are no longer required because they are handled
3474 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3475 Explicitly de-initing can cause problems (e.g. where a library that uses
3476 OpenSSL de-inits, but an application is still using it). The affected
3477 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3478 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3479 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3480 COMP_zlib_cleanup().
3484 * --strict-warnings no longer enables runtime debugging options
3485 such as REF_DEBUG. Instead, debug options are automatically
3486 enabled with '--debug' builds.
3488 *Andy Polyakov, Emilia Käsper*
3490 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3491 have been moved out of the public header files. New functions for managing
3492 these have been added.
3496 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3497 objects have been moved out of the public header files. New
3498 functions for managing these have been added.
3502 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3503 have been moved out of the public header files. New functions for managing
3504 these have been added.
3508 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3509 moved out of the public header files. New functions for managing these
3514 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3518 * Removed the mk1mf build scripts.
3522 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3523 it is always safe to #include a header now.
3527 * Removed the aged BC-32 config and all its supporting scripts
3531 * Removed support for Ultrix, Netware, and OS/2.
3535 * Add support for HKDF.
3537 *Alessandro Ghedini*
3539 * Add support for blake2b and blake2s
3543 * Added support for "pipelining". Ciphers that have the
3544 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3545 encryptions/decryptions simultaneously. There are currently no built-in
3546 ciphers with this property but the expectation is that engines will be able
3547 to offer it to significantly improve throughput. Support has been extended
3548 into libssl so that multiple records for a single connection can be
3549 processed in one go (for >=TLS 1.1).
3553 * Added the AFALG engine. This is an async capable engine which is able to
3554 offload work to the Linux kernel. In this initial version it only supports
3555 AES128-CBC. The kernel must be version 4.1.0 or greater.
3559 * OpenSSL now uses a new threading API. It is no longer necessary to
3560 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3561 are two supported threading models: pthreads and windows threads. It is
3562 also possible to configure OpenSSL at compile time for "no-threads". The
3563 old threading API should no longer be used. The functions have been
3564 replaced with "no-op" compatibility macros.
3566 *Alessandro Ghedini, Matt Caswell*
3568 * Modify behavior of ALPN to invoke callback after SNI/servername
3569 callback, such that updates to the SSL_CTX affect ALPN.
3573 * Add SSL_CIPHER queries for authentication and key-exchange.
3577 * Changes to the DEFAULT cipherlist:
3578 - Prefer (EC)DHE handshakes over plain RSA.
3579 - Prefer AEAD ciphers over legacy ciphers.
3580 - Prefer ECDSA over RSA when both certificates are available.
3581 - Prefer TLSv1.2 ciphers/PRF.
3582 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3587 * Change the ECC default curve list to be this, in order: x25519,
3588 secp256r1, secp521r1, secp384r1.
3592 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3593 disabled by default. They can be re-enabled using the
3594 enable-weak-ssl-ciphers option to Configure.
3598 * If the server has ALPN configured, but supports no protocols that the
3599 client advertises, send a fatal "no_application_protocol" alert.
3600 This behaviour is SHALL in RFC 7301, though it isn't universally
3601 implemented by other servers.
3605 * Add X25519 support.
3606 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3607 for public and private key encoding using the format documented in
3608 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3609 key generation and key derivation.
3611 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3616 * Deprecate SRP_VBASE_get_by_user.
3617 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3618 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3619 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3620 seed, even if the seed is configured.
3622 Users should use SRP_VBASE_get1_by_user instead. Note that in
3623 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3624 also that even though configuring the SRP seed attempts to hide
3625 invalid usernames by continuing the handshake with fake
3626 credentials, this behaviour is not constant time and no strong
3627 guarantees are made that the handshake is indistinguishable from
3628 that of a valid user.
3632 * Configuration change; it's now possible to build dynamic engines
3633 without having to build shared libraries and vice versa. This
3634 only applies to the engines in `engines/`, those in `crypto/engine/`
3635 will always be built into libcrypto (i.e. "static").
3637 Building dynamic engines is enabled by default; to disable, use
3638 the configuration option "disable-dynamic-engine".
3640 The only requirements for building dynamic engines are the
3641 presence of the DSO module and building with position independent
3642 code, so they will also automatically be disabled if configuring
3643 with "disable-dso" or "disable-pic".
3645 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3646 are also taken away from openssl/opensslconf.h, as they are
3651 * Configuration change; if there is a known flag to compile
3652 position independent code, it will always be applied on the
3653 libcrypto and libssl object files, and never on the application
3654 object files. This means other libraries that use routines from
3655 libcrypto / libssl can be made into shared libraries regardless
3656 of how OpenSSL was configured.
3658 If this isn't desirable, the configuration options "disable-pic"
3659 or "no-pic" can be used to disable the use of PIC. This will
3660 also disable building shared libraries and dynamic engines.
3664 * Removed JPAKE code. It was experimental and has no wide use.
3668 * The INSTALL_PREFIX Makefile variable has been renamed to
3669 DESTDIR. That makes for less confusion on what this variable
3670 is for. Also, the configuration option --install_prefix is
3675 * Heartbeat for TLS has been removed and is disabled by default
3676 for DTLS; configure with enable-heartbeats. Code that uses the
3677 old #define's might need to be updated.
3679 *Emilia Käsper, Rich Salz*
3681 * Rename REF_CHECK to REF_DEBUG.
3685 * New "unified" build system
3687 The "unified" build system is aimed to be a common system for all
3688 platforms we support. With it comes new support for VMS.
3690 This system builds supports building in a different directory tree
3691 than the source tree. It produces one Makefile (for unix family
3692 or lookalikes), or one descrip.mms (for VMS).
3694 The source of information to make the Makefile / descrip.mms is
3695 small files called 'build.info', holding the necessary
3696 information for each directory with source to compile, and a
3697 template in Configurations, like unix-Makefile.tmpl or
3700 With this change, the library names were also renamed on Windows
3701 and on VMS. They now have names that are closer to the standard
3702 on Unix, and include the major version number, and in certain
3703 cases, the architecture they are built for. See "Notes on shared
3704 libraries" in INSTALL.
3706 We rely heavily on the perl module Text::Template.
3710 * Added support for auto-initialisation and de-initialisation of the library.
3711 OpenSSL no longer requires explicit init or deinit routines to be called,
3712 except in certain circumstances. See the OPENSSL_init_crypto() and
3713 OPENSSL_init_ssl() man pages for further information.
3717 * The arguments to the DTLSv1_listen function have changed. Specifically the
3718 "peer" argument is now expected to be a BIO_ADDR object.
3720 * Rewrite of BIO networking library. The BIO library lacked consistent
3721 support of IPv6, and adding it required some more extensive
3722 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3723 which hold all types of addresses and chains of address information.
3724 It also introduces a new API, with functions like BIO_socket,
3725 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3726 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3727 have been adapted accordingly.
3731 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3736 * CRIME protection: disable compression by default, even if OpenSSL is
3737 compiled with zlib enabled. Applications can still enable compression
3738 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3739 using the SSL_CONF library to configure compression.
3743 * The signature of the session callback configured with
3744 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3745 was explicitly marked as `const unsigned char*` instead of
3750 * Always DPURIFY. Remove the use of uninitialized memory in the
3751 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3755 * Removed many obsolete configuration items, including
3756 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3757 MD2_CHAR, MD2_INT, MD2_LONG
3759 IDEA_SHORT, IDEA_LONG
3760 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3762 *Rich Salz, with advice from Andy Polyakov*
3764 * Many BN internals have been moved to an internal header file.
3766 *Rich Salz with help from Andy Polyakov*
3768 * Configuration and writing out the results from it has changed.
3769 Files such as Makefile include/openssl/opensslconf.h and are now
3770 produced through general templates, such as Makefile.in and
3771 crypto/opensslconf.h.in and some help from the perl module
3774 Also, the center of configuration information is no longer
3775 Makefile. Instead, Configure produces a perl module in
3776 configdata.pm which holds most of the config data (in the hash
3777 table %config), the target data that comes from the target
3778 configuration in one of the `Configurations/*.conf` files (in
3783 * To clarify their intended purposes, the Configure options
3784 --prefix and --openssldir change their semantics, and become more
3785 straightforward and less interdependent.
3787 --prefix shall be used exclusively to give the location INSTALLTOP
3788 where programs, scripts, libraries, include files and manuals are
3789 going to be installed. The default is now /usr/local.
3791 --openssldir shall be used exclusively to give the default
3792 location OPENSSLDIR where certificates, private keys, CRLs are
3793 managed. This is also where the default openssl.cnf gets
3795 If the directory given with this option is a relative path, the
3796 values of both the --prefix value and the --openssldir value will
3797 be combined to become OPENSSLDIR.
3798 The default for --openssldir is INSTALLTOP/ssl.
3800 Anyone who uses --openssldir to specify where OpenSSL is to be
3801 installed MUST change to use --prefix instead.
3805 * The GOST engine was out of date and therefore it has been removed. An up
3806 to date GOST engine is now being maintained in an external repository.
3807 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3808 support for GOST ciphersuites (these are only activated if a GOST engine
3813 * EGD is no longer supported by default; use enable-egd when
3816 *Ben Kaduk and Rich Salz*
3818 * The distribution now has Makefile.in files, which are used to
3819 create Makefile's when Configure is run. *Configure must be run
3820 before trying to build now.*
3824 * The return value for SSL_CIPHER_description() for error conditions
3829 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3831 Obtaining and performing DNSSEC validation of TLSA records is
3832 the application's responsibility. The application provides
3833 the TLSA records of its choice to OpenSSL, and these are then
3834 used to authenticate the peer.
3836 The TLSA records need not even come from DNS. They can, for
3837 example, be used to implement local end-entity certificate or
3838 trust-anchor "pinning", where the "pin" data takes the form
3839 of TLSA records, which can augment or replace verification
3840 based on the usual WebPKI public certification authorities.
3844 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3845 continues to support deprecated interfaces in default builds.
3846 However, applications are strongly advised to compile their
3847 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3848 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3849 or the 1.1.0 releases.
3851 In environments in which all applications have been ported to
3852 not use any deprecated interfaces OpenSSL's Configure script
3853 should be used with the --api=1.1.0 option to entirely remove
3854 support for the deprecated features from the library and
3855 unconditionally disable them in the installed headers.
3856 Essentially the same effect can be achieved with the "no-deprecated"
3857 argument to Configure, except that this will always restrict
3858 the build to just the latest API, rather than a fixed API
3861 As applications are ported to future revisions of the API,
3862 they should update their compile-time OPENSSL_API_COMPAT define
3863 accordingly, but in most cases should be able to continue to
3864 compile with later releases.
3866 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3867 0x10000000L and 0x00908000L, respectively. However those
3868 versions did not support the OPENSSL_API_COMPAT feature, and
3869 so applications are not typically tested for explicit support
3870 of just the undeprecated features of either release.
3874 * Add support for setting the minimum and maximum supported protocol.
3875 It can bet set via the SSL_set_min_proto_version() and
3876 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3877 MaxProtocol. It's recommended to use the new APIs to disable
3878 protocols instead of disabling individual protocols using
3879 SSL_set_options() or SSL_CONF's Protocol. This change also
3880 removes support for disabling TLS 1.2 in the OpenSSL TLS
3881 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3885 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3889 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3890 and integrates ECDSA and ECDH functionality into EC. Implementations can
3891 now redirect key generation and no longer need to convert to or from
3894 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3895 include the ec.h header file instead.
3899 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3900 ciphers who are no longer supported and drops support the ephemeral RSA key
3901 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3905 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3906 opaque. For HMAC_CTX, the following constructors and destructors
3909 HMAC_CTX *HMAC_CTX_new(void);
3910 void HMAC_CTX_free(HMAC_CTX *ctx);
3912 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3913 destroy such methods has been added. See EVP_MD_meth_new(3) and
3914 EVP_CIPHER_meth_new(3) for documentation.
3917 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3918 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3919 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3920 an already created structure.
3921 2) For consistency with the majority of our object creators and
3922 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3923 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3924 for deprecated builds.
3928 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3929 cryptographic operations to be performed asynchronously as long as an
3930 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3931 further details. Libssl has also had this capability integrated with the
3932 introduction of the new mode SSL_MODE_ASYNC and associated error
3933 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3934 pages. This work was developed in partnership with Intel Corp.
3938 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3939 always enabled now. If you want to disable the support you should
3940 exclude it using the list of supported ciphers. This also means that the
3941 "-no_ecdhe" option has been removed from s_server.
3945 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3946 SSL_{CTX_}set1_curves() which can set a list.
3950 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3951 curve you want to support using SSL_{CTX_}set1_curves().
3955 * State machine rewrite. The state machine code has been significantly
3956 refactored in order to remove much duplication of code and solve issues
3957 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3958 further details). This change does have some associated API changes.
3959 Notably the SSL_state() function has been removed and replaced by
3960 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3961 SSL_set_state() has been removed altogether. The previous handshake states
3962 defined in ssl.h and ssl3.h have also been removed.
3966 * All instances of the string "ssleay" in the public API were replaced
3967 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3968 Some error codes related to internal RSA_eay API's were renamed.
3972 * The demo files in crypto/threads were moved to demo/threads.
3976 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3979 *Matt Caswell, Rich Salz*
3981 * New ASN.1 embed macro.
3983 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3984 structure is not allocated: it is part of the parent. That is instead of
3992 This reduces memory fragmentation and make it impossible to accidentally
3993 set a mandatory field to NULL.
3995 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3996 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3997 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4002 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4006 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4007 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4008 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4009 DES and RC4 ciphersuites.
4013 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4014 This changes the decoding behaviour for some invalid messages,
4015 though the change is mostly in the more lenient direction, and
4016 legacy behaviour is preserved as much as possible.
4020 * Fix no-stdio build.
4021 *David Woodhouse <David.Woodhouse@intel.com> and also*
4022 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4024 * New testing framework
4025 The testing framework has been largely rewritten and is now using
4026 perl and the perl modules Test::Harness and an extended variant of
4027 Test::More called OpenSSL::Test to do its work. All test scripts in
4028 test/ have been rewritten into test recipes, and all direct calls to
4029 executables in test/Makefile have become individual recipes using the
4030 simplified testing OpenSSL::Test::Simple.
4032 For documentation on our testing modules, do:
4034 perldoc test/testlib/OpenSSL/Test/Simple.pm
4035 perldoc test/testlib/OpenSSL/Test.pm
4039 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4040 are used; the latter aborts on memory leaks (usually checked on exit).
4041 Some undocumented "set malloc, etc., hooks" functions were removed
4042 and others were changed. All are now documented.
4046 * In DSA_generate_parameters_ex, if the provided seed is too short,
4049 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4051 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4052 from RFC4279, RFC4785, RFC5487, RFC5489.
4054 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4055 original RSA_PSK patch.
4059 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4060 era flag was never set throughout the codebase (only read). Also removed
4061 SSL3_FLAGS_POP_BUFFER which was only used if
4062 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4066 * Changed the default name options in the "ca", "crl", "req" and "x509"
4067 to be "oneline" instead of "compat".
4071 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4072 not aware of clients that still exhibit this bug, and the workaround
4073 hasn't been working properly for a while.
4077 * The return type of BIO_number_read() and BIO_number_written() as well as
4078 the corresponding num_read and num_write members in the BIO structure has
4079 changed from unsigned long to uint64_t. On platforms where an unsigned
4080 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4085 * Given the pervasive nature of TLS extensions it is inadvisable to run
4086 OpenSSL without support for them. It also means that maintaining
4087 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4088 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4092 * Removed support for the two export grade static DH ciphersuites
4093 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4094 were newly added (along with a number of other static DH ciphersuites) to
4095 1.0.2. However the two export ones have *never* worked since they were
4096 introduced. It seems strange in any case to be adding new export
4097 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4101 * Version negotiation has been rewritten. In particular SSLv23_method(),
4102 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4103 and turned into macros which simply call the new preferred function names
4104 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4105 should use the new names instead. Also as part of this change the ssl23.h
4106 header file has been removed.
4110 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4111 code and the associated standard is no longer considered fit-for-purpose.
4115 * RT2547 was closed. When generating a private key, try to make the
4116 output file readable only by the owner. This behavior change might
4117 be noticeable when interacting with other software.
4119 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4124 * Added HTTP GET support to the ocsp command.
4128 * Changed default digest for the dgst and enc commands from MD5 to
4133 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4137 * Added support for TLS extended master secret from
4138 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4139 initial patch which was a great help during development.
4143 * All libssl internal structures have been removed from the public header
4144 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4145 now redundant). Users should not attempt to access internal structures
4146 directly. Instead they should use the provided API functions.
4150 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4151 Access to deprecated functions can be re-enabled by running config with
4152 "enable-deprecated". In addition applications wishing to use deprecated
4153 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4154 will, by default, disable some transitive includes that previously existed
4155 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4159 * Added support for OCB mode. OpenSSL has been granted a patent license
4160 compatible with the OpenSSL license for use of OCB. Details are available
4161 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4162 for OCB can be removed by calling config with no-ocb.
4166 * SSLv2 support has been removed. It still supports receiving a SSLv2
4167 compatible client hello.
4171 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4172 done while fixing the error code for the key-too-small case.
4174 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4176 * CA.sh has been removed; use CA.pl instead.
4180 * Removed old DES API.
4184 * Remove various unsupported platforms:
4190 Sinix/ReliantUNIX RM400
4195 16-bit platforms such as WIN16
4199 * Clean up OPENSSL_NO_xxx #define's
4200 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4201 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4202 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4203 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4204 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4205 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4206 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4207 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4208 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4209 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4213 * Cleaned up dead code
4214 Remove all but one '#ifdef undef' which is to be looked at.
4218 * Clean up calling of xxx_free routines.
4219 Just like free(), fix most of the xxx_free routines to accept
4220 NULL. Remove the non-null checks from callers. Save much code.
4224 * Add secure heap for storage of private keys (when possible).
4225 Add BIO_s_secmem(), CBIGNUM, etc.
4226 Contributed by Akamai Technologies under our Corporate CLA.
4230 * Experimental support for a new, fast, unbiased prime candidate generator,
4231 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4233 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4235 * New output format NSS in the sess_id command line tool. This allows
4236 exporting the session id and the master key in NSS keylog format.
4238 *Martin Kaiser <martin@kaiser.cx>*
4240 * Harmonize version and its documentation. -f flag is used to display
4243 *mancha <mancha1@zoho.com>*
4245 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4246 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4248 *mancha <mancha1@zoho.com>*
4250 * Fix some double frees. These are not thought to be exploitable.
4252 *mancha <mancha1@zoho.com>*
4254 * A missing bounds check in the handling of the TLS heartbeat extension
4255 can be used to reveal up to 64k of memory to a connected client or
4258 Thanks for Neel Mehta of Google Security for discovering this bug and to
4259 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4260 preparing the fix ([CVE-2014-0160])
4262 *Adam Langley, Bodo Moeller*
4264 * Fix for the attack described in the paper "Recovering OpenSSL
4265 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4266 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4267 <http://eprint.iacr.org/2014/140>
4269 Thanks to Yuval Yarom and Naomi Benger for discovering this
4270 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4272 *Yuval Yarom and Naomi Benger*
4274 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4275 this fixes a limitation in previous versions of OpenSSL.
4279 * Experimental encrypt-then-mac support.
4281 Experimental support for encrypt then mac from
4282 draft-gutmann-tls-encrypt-then-mac-02.txt
4284 To enable it set the appropriate extension number (0x42 for the test
4285 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4287 For non-compliant peers (i.e. just about everything) this should have no
4290 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4294 * Add EVP support for key wrapping algorithms, to avoid problems with
4295 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4296 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4297 algorithms and include tests cases.
4301 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4306 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4307 MGF1 digest and OAEP label.
4311 * Make openssl verify return errors.
4313 *Chris Palmer <palmer@google.com> and Ben Laurie*
4315 * New function ASN1_TIME_diff to calculate the difference between two
4316 ASN1_TIME structures or one structure and the current time.
4320 * Update fips_test_suite to support multiple command line options. New
4321 test to induce all self test errors in sequence and check expected
4326 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4327 sign or verify all in one operation.
4331 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4332 test programs and fips_test_suite. Includes functionality to parse
4333 the minimal script output of fipsalgest.pl directly.
4337 * Add authorisation parameter to FIPS_module_mode_set().
4341 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4345 * Use separate DRBG fields for internal and external flags. New function
4346 FIPS_drbg_health_check() to perform on demand health checking. Add
4347 generation tests to fips_test_suite with reduced health check interval to
4348 demonstrate periodic health checking. Add "nodh" option to
4349 fips_test_suite to skip very slow DH test.
4353 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4358 * More extensive health check for DRBG checking many more failure modes.
4359 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4360 combination: call this in fips_test_suite.
4364 * Add support for canonical generation of DSA parameter 'g'. See
4367 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4368 POST to handle HMAC cases.
4372 * Add functions FIPS_module_version() and FIPS_module_version_text()
4373 to return numerical and string versions of the FIPS module number.
4377 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4378 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4379 outside the validated module in the FIPS capable OpenSSL.
4383 * Minor change to DRBG entropy callback semantics. In some cases
4384 there is no multiple of the block length between min_len and
4385 max_len. Allow the callback to return more than max_len bytes
4386 of entropy but discard any extra: it is the callback's responsibility
4387 to ensure that the extra data discarded does not impact the
4388 requested amount of entropy.
4392 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4393 information in FIPS186-3, SP800-57 and SP800-131A.
4397 * CCM support via EVP. Interface is very similar to GCM case except we
4398 must supply all data in one chunk (i.e. no update, final) and the
4399 message length must be supplied if AAD is used. Add algorithm test
4404 * Initial version of POST overhaul. Add POST callback to allow the status
4405 of POST to be monitored and/or failures induced. Modify fips_test_suite
4406 to use callback. Always run all selftests even if one fails.
4410 * XTS support including algorithm test driver in the fips_gcmtest program.
4411 Note: this does increase the maximum key length from 32 to 64 bytes but
4412 there should be no binary compatibility issues as existing applications
4413 will never use XTS mode.
4417 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4418 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4419 performs algorithm blocking for unapproved PRNG types. Also do not
4420 set PRNG type in FIPS_mode_set(): leave this to the application.
4421 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4422 the standard OpenSSL PRNG: set additional data to a date time vector.
4426 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4427 This shouldn't present any incompatibility problems because applications
4428 shouldn't be using these directly and any that are will need to rethink
4429 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4433 * Extensive self tests and health checking required by SP800-90 DRBG.
4434 Remove strength parameter from FIPS_drbg_instantiate and always
4435 instantiate at maximum supported strength.
4439 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4443 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4447 * New function DH_compute_key_padded() to compute a DH key and pad with
4448 leading zeroes if needed: this complies with SP800-56A et al.
4452 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4453 anything, incomplete, subject to change and largely untested at present.
4457 * Modify fipscanisteronly build option to only build the necessary object
4458 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4462 * Add experimental option FIPSSYMS to give all symbols in
4463 fipscanister.o and FIPS or fips prefix. This will avoid
4464 conflicts with future versions of OpenSSL. Add perl script
4465 util/fipsas.pl to preprocess assembly language source files
4466 and rename any affected symbols.
4470 * Add selftest checks and algorithm block of non-fips algorithms in
4471 FIPS mode. Remove DES2 from selftests.
4475 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4476 return internal method without any ENGINE dependencies. Add new
4477 tiny fips sign and verify functions.
4481 * New build option no-ec2m to disable characteristic 2 code.
4485 * New build option "fipscanisteronly". This only builds fipscanister.o
4486 and (currently) associated fips utilities. Uses the file Makefile.fips
4487 instead of Makefile.org as the prototype.
4491 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4492 Update fips_gcmtest to use IV generator.
4496 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4497 setting output buffer to NULL. The `*Final` function must be
4498 called although it will not retrieve any additional data. The tag
4499 can be set or retrieved with a ctrl. The IV length is by default 12
4500 bytes (96 bits) but can be set to an alternative value. If the IV
4501 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4506 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4507 underlying do_cipher function handles all cipher semantics itself
4508 including padding and finalisation. This is useful if (for example)
4509 an ENGINE cipher handles block padding itself. The behaviour of
4510 do_cipher is subtly changed if this flag is set: the return value
4511 is the number of characters written to the output buffer (zero is
4512 no longer an error code) or a negative error code. Also if the
4513 input buffer is NULL and length 0 finalisation should be performed.
4517 * If a candidate issuer certificate is already part of the constructed
4518 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4522 * Improve forward-security support: add functions
4524 void SSL_CTX_set_not_resumable_session_callback(
4525 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4526 void SSL_set_not_resumable_session_callback(
4527 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4529 for use by SSL/TLS servers; the callback function will be called whenever a
4530 new session is created, and gets to decide whether the session may be
4531 cached to make it resumable (return 0) or not (return 1). (As by the
4532 SSL/TLS protocol specifications, the session_id sent by the server will be
4533 empty to indicate that the session is not resumable; also, the server will
4534 not generate RFC 4507 (RFC 5077) session tickets.)
4536 A simple reasonable callback implementation is to return is_forward_secure.
4537 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4538 by the SSL/TLS server library, indicating whether it can provide forward
4541 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4543 * New -verify_name option in command line utilities to set verification
4548 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4549 Add CMAC pkey methods.
4553 * Experimental renegotiation in s_server -www mode. If the client
4554 browses /reneg connection is renegotiated. If /renegcert it is
4555 renegotiated requesting a certificate.
4559 * Add an "external" session cache for debugging purposes to s_server. This
4560 should help trace issues which normally are only apparent in deployed
4561 multi-process servers.
4565 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4566 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4567 BIO_set_cipher() and some obscure PEM functions were changed so they
4568 can now return an error. The RAND changes required a change to the
4569 RAND_METHOD structure.
4573 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4574 a gcc attribute to warn if the result of a function is ignored. This
4575 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4576 whose return value is often ignored.
4580 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4581 These allow SCTs (signed certificate timestamps) to be requested and
4582 validated when establishing a connection.
4584 *Rob Percival <robpercival@google.com>*
4589 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4591 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4592 used even when parsing explicit parameters, when loading a encoded key
4593 or calling `EC_GROUP_new_from_ecpkparameters()`/
4594 `EC_GROUP_new_from_ecparameters()`.
4595 This prevents bypass of security hardening and performance gains,
4596 especially for curves with specialized EC_METHODs.
4597 By default, if a key encoded with explicit parameters is loaded and later
4598 encoded, the output is still encoded with explicit parameters, even if
4599 internally a "named" EC_GROUP is used for computation.
4603 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4604 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4605 NULL. After this change, only the cofactor parameter can be NULL. It also
4606 does some minimal sanity checks on the passed order.
4611 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4612 An attack is simple, if the first CMS_recipientInfo is valid but the
4613 second CMS_recipientInfo is chosen ciphertext. If the second
4614 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4615 encryption key will be replaced by garbage, and the message cannot be
4616 decoded, but if the RSA decryption fails, the correct encryption key is
4617 used and the recipient will not notice the attack.
4618 As a work around for this potential attack the length of the decrypted
4619 key must be equal to the cipher default key length, in case the
4620 certifiate is not given and all recipientInfo are tried out.
4621 The old behaviour can be re-enabled in the CMS code by setting the
4622 CMS_DEBUG_DECRYPT flag.
4627 * Document issue with installation paths in diverse Windows builds
4629 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4630 binaries and run-time config file.
4635 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4637 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4638 This changes the size when using the `genpkey` command when no size is given.
4639 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4640 generation commands to use 2048 bits by default.
4644 * Add FIPS support for Android Arm 64-bit
4646 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4647 Module in Version 2.0.10. For some reason, the corresponding target
4648 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4649 built with FIPS support on Android Arm 64-bit. This omission has been
4652 *Matthias St. Pierre*
4654 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4656 * 0-byte record padding oracle
4658 If an application encounters a fatal protocol error and then calls
4659 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4660 then OpenSSL can respond differently to the calling application if a 0 byte
4661 record is received with invalid padding compared to if a 0 byte record is
4662 received with an invalid MAC. If the application then behaves differently
4663 based on that in a way that is detectable to the remote peer, then this
4664 amounts to a padding oracle that could be used to decrypt data.
4666 In order for this to be exploitable "non-stitched" ciphersuites must be in
4667 use. Stitched ciphersuites are optimised implementations of certain
4668 commonly used ciphersuites. Also the application must call SSL_shutdown()
4669 twice even if a protocol error has occurred (applications should not do
4670 this but some do anyway).
4672 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4673 Aviram, with additional investigation by Steven Collison and Andrew
4674 Hourselt. It was reported to OpenSSL on 10th December 2018.
4679 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4683 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4685 * Microarchitecture timing vulnerability in ECC scalar multiplication
4687 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4688 shown to be vulnerable to a microarchitecture timing side channel attack.
4689 An attacker with sufficient access to mount local timing attacks during
4690 ECDSA signature generation could recover the private key.
4692 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4693 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4699 * Timing vulnerability in DSA signature generation
4701 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4702 timing side channel attack. An attacker could use variations in the signing
4703 algorithm to recover the private key.
4705 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4710 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4711 Module, accidentally introduced while backporting security fixes from the
4712 development branch and hindering the use of ECC in FIPS mode.
4716 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4718 * Client DoS due to large DH parameter
4720 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4721 malicious server can send a very large prime value to the client. This will
4722 cause the client to spend an unreasonably long period of time generating a
4723 key for this prime resulting in a hang until the client has finished. This
4724 could be exploited in a Denial Of Service attack.
4726 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4731 * Cache timing vulnerability in RSA Key Generation
4733 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4734 a cache timing side channel attack. An attacker with sufficient access to
4735 mount cache timing attacks during the RSA key generation process could
4736 recover the private key.
4738 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4739 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4744 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4745 parameter is no longer accepted, as it leads to a corrupt table. NULL
4746 pem_str is reserved for alias entries only.
4750 * Revert blinding in ECDSA sign and instead make problematic addition
4751 length-invariant. Switch even to fixed-length Montgomery multiplication.
4755 * Change generating and checking of primes so that the error rate of not
4756 being prime depends on the intended use based on the size of the input.
4757 For larger primes this will result in more rounds of Miller-Rabin.
4758 The maximal error rate for primes with more than 1080 bits is lowered
4761 *Kurt Roeckx, Annie Yousar*
4763 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4767 * Add blinding to ECDSA and DSA signatures to protect against side channel
4768 attacks discovered by Keegan Ryan (NCC Group).
4772 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4773 now allow empty (zero character) pass phrases.
4777 * Certificate time validation (X509_cmp_time) enforces stricter
4778 compliance with RFC 5280. Fractional seconds and timezone offsets
4779 are no longer allowed.
4783 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4785 * Constructed ASN.1 types with a recursive definition could exceed the stack
4787 Constructed ASN.1 types with a recursive definition (such as can be found
4788 in PKCS7) could eventually exceed the stack given malicious input with
4789 excessive recursion. This could result in a Denial Of Service attack. There
4790 are no such structures used within SSL/TLS that come from untrusted sources
4791 so this is considered safe.
4793 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4799 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4801 * Read/write after SSL object in error state
4803 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4804 mechanism. The intent was that if a fatal error occurred during a handshake
4805 then OpenSSL would move into the error state and would immediately fail if
4806 you attempted to continue the handshake. This works as designed for the
4807 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4808 SSL_connect()), however due to a bug it does not work correctly if
4809 SSL_read() or SSL_write() is called directly. In that scenario, if the
4810 handshake fails then a fatal error will be returned in the initial function
4811 call. If SSL_read()/SSL_write() is subsequently called by the application
4812 for the same SSL object then it will succeed and the data is passed without
4813 being decrypted/encrypted directly from the SSL/TLS record layer.
4815 In order to exploit this issue an application bug would have to be present
4816 that resulted in a call to SSL_read()/SSL_write() being issued after having
4817 already received a fatal error.
4819 This issue was reported to OpenSSL by David Benjamin (Google).
4824 * rsaz_1024_mul_avx2 overflow bug on x86_64
4826 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4827 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4828 Analysis suggests that attacks against RSA and DSA as a result of this
4829 defect would be very difficult to perform and are not believed likely.
4830 Attacks against DH1024 are considered just feasible, because most of the
4831 work necessary to deduce information about a private key may be performed
4832 offline. The amount of resources required for such an attack would be
4833 significant. However, for an attack on TLS to be meaningful, the server
4834 would have to share the DH1024 private key among multiple clients, which is
4835 no longer an option since CVE-2016-0701.
4837 This only affects processors that support the AVX2 but not ADX extensions
4838 like Intel Haswell (4th generation).
4840 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4841 was originally found via the OSS-Fuzz project.
4846 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4848 * bn_sqrx8x_internal carry bug on x86_64
4850 There is a carry propagating bug in the x86_64 Montgomery squaring
4851 procedure. No EC algorithms are affected. Analysis suggests that attacks
4852 against RSA and DSA as a result of this defect would be very difficult to
4853 perform and are not believed likely. Attacks against DH are considered just
4854 feasible (although very difficult) because most of the work necessary to
4855 deduce information about a private key may be performed offline. The amount
4856 of resources required for such an attack would be very significant and
4857 likely only accessible to a limited number of attackers. An attacker would
4858 additionally need online access to an unpatched system using the target
4859 private key in a scenario with persistent DH parameters and a private
4860 key that is shared between multiple clients.
4862 This only affects processors that support the BMI1, BMI2 and ADX extensions
4863 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4865 This issue was reported to OpenSSL by the OSS-Fuzz project.
4870 * Malformed X.509 IPAddressFamily could cause OOB read
4872 If an X.509 certificate has a malformed IPAddressFamily extension,
4873 OpenSSL could do a one-byte buffer overread. The most likely result
4874 would be an erroneous display of the certificate in text format.
4876 This issue was reported to OpenSSL by the OSS-Fuzz project.
4880 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4882 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4883 platform rather than 'mingw'.
4887 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4889 * Truncated packet could crash via OOB read
4891 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4892 cipher is being used, then a truncated packet can cause that host to
4893 perform an out-of-bounds read, usually resulting in a crash.
4895 This issue was reported to OpenSSL by Robert Święcki of Google.
4900 * BN_mod_exp may produce incorrect results on x86_64
4902 There is a carry propagating bug in the x86_64 Montgomery squaring
4903 procedure. No EC algorithms are affected. Analysis suggests that attacks
4904 against RSA and DSA as a result of this defect would be very difficult to
4905 perform and are not believed likely. Attacks against DH are considered just
4906 feasible (although very difficult) because most of the work necessary to
4907 deduce information about a private key may be performed offline. The amount
4908 of resources required for such an attack would be very significant and
4909 likely only accessible to a limited number of attackers. An attacker would
4910 additionally need online access to an unpatched system using the target
4911 private key in a scenario with persistent DH parameters and a private
4912 key that is shared between multiple clients. For example this can occur by
4913 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4914 similar to CVE-2015-3193 but must be treated as a separate problem.
4916 This issue was reported to OpenSSL by the OSS-Fuzz project.
4921 * Montgomery multiplication may produce incorrect results
4923 There is a carry propagating bug in the Broadwell-specific Montgomery
4924 multiplication procedure that handles input lengths divisible by, but
4925 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4926 and DH private keys are impossible. This is because the subroutine in
4927 question is not used in operations with the private key itself and an input
4928 of the attacker's direct choice. Otherwise the bug can manifest itself as
4929 transient authentication and key negotiation failures or reproducible
4930 erroneous outcome of public-key operations with specially crafted input.
4931 Among EC algorithms only Brainpool P-512 curves are affected and one
4932 presumably can attack ECDH key negotiation. Impact was not analyzed in
4933 detail, because pre-requisites for attack are considered unlikely. Namely
4934 multiple clients have to choose the curve in question and the server has to
4935 share the private key among them, neither of which is default behaviour.
4936 Even then only clients that chose the curve will be affected.
4938 This issue was publicly reported as transient failures and was not
4939 initially recognized as a security issue. Thanks to Richard Morgan for
4940 providing reproducible case.
4945 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4946 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4947 prevent issues where no progress is being made and the peer continually
4948 sends unrecognised record types, using up resources processing them.
4952 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4954 * Missing CRL sanity check
4956 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4957 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4958 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4960 This issue only affects the OpenSSL 1.0.2i
4965 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4967 * OCSP Status Request extension unbounded memory growth
4969 A malicious client can send an excessively large OCSP Status Request
4970 extension. If that client continually requests renegotiation, sending a
4971 large OCSP Status Request extension each time, then there will be unbounded
4972 memory growth on the server. This will eventually lead to a Denial Of
4973 Service attack through memory exhaustion. Servers with a default
4974 configuration are vulnerable even if they do not support OCSP. Builds using
4975 the "no-ocsp" build time option are not affected.
4977 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4982 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4985 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4991 * OOB write in MDC2_Update()
4993 An overflow can occur in MDC2_Update() either if called directly or
4994 through the EVP_DigestUpdate() function using MDC2. If an attacker
4995 is able to supply very large amounts of input data after a previous
4996 call to EVP_EncryptUpdate() with a partial block then a length check
4997 can overflow resulting in a heap corruption.
4999 The amount of data needed is comparable to SIZE_MAX which is impractical
5002 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5007 * Malformed SHA512 ticket DoS
5009 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5010 DoS attack where a malformed ticket will result in an OOB read which will
5013 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5014 a custom server callback and ticket lookup mechanism.
5016 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5021 * OOB write in BN_bn2dec()
5023 The function BN_bn2dec() does not check the return value of BN_div_word().
5024 This can cause an OOB write if an application uses this function with an
5025 overly large BIGNUM. This could be a problem if an overly large certificate
5026 or CRL is printed out from an untrusted source. TLS is not affected because
5027 record limits will reject an oversized certificate before it is parsed.
5029 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5034 * OOB read in TS_OBJ_print_bio()
5036 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5037 the total length the OID text representation would use and not the amount
5038 of data written. This will result in OOB reads when large OIDs are
5041 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5046 * Pointer arithmetic undefined behaviour
5048 Avoid some undefined pointer arithmetic
5050 A common idiom in the codebase is to check limits in the following manner:
5053 Where "p" points to some malloc'd data of SIZE bytes and
5056 "len" here could be from some externally supplied data (e.g. from a TLS
5059 The rules of C pointer arithmetic are such that "p + len" is only well
5060 defined where len <= SIZE. Therefore the above idiom is actually
5061 undefined behaviour.
5063 For example this could cause problems if some malloc implementation
5064 provides an address for "p" such that "p + len" actually overflows for
5065 values of len that are too big and therefore p + len < limit.
5067 This issue was reported to OpenSSL by Guido Vranken
5072 * Constant time flag not preserved in DSA signing
5074 Operations in the DSA signing algorithm should run in constant time in
5075 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5076 implementation means that a non-constant time codepath is followed for
5077 certain operations. This has been demonstrated through a cache-timing
5078 attack to be sufficient for an attacker to recover the private DSA key.
5080 This issue was reported by César Pereida (Aalto University), Billy Brumley
5081 (Tampere University of Technology), and Yuval Yarom (The University of
5082 Adelaide and NICTA).
5087 * DTLS buffered message DoS
5089 In a DTLS connection where handshake messages are delivered out-of-order
5090 those messages that OpenSSL is not yet ready to process will be buffered
5091 for later use. Under certain circumstances, a flaw in the logic means that
5092 those messages do not get removed from the buffer even though the handshake
5093 has been completed. An attacker could force up to approx. 15 messages to
5094 remain in the buffer when they are no longer required. These messages will
5095 be cleared when the DTLS connection is closed. The default maximum size for
5096 a message is 100k. Therefore the attacker could force an additional 1500k
5097 to be consumed per connection. By opening many simulataneous connections an
5098 attacker could cause a DoS attack through memory exhaustion.
5100 This issue was reported to OpenSSL by Quan Luo.
5105 * DTLS replay protection DoS
5107 A flaw in the DTLS replay attack protection mechanism means that records
5108 that arrive for future epochs update the replay protection "window" before
5109 the MAC for the record has been validated. This could be exploited by an
5110 attacker by sending a record for the next epoch (which does not have to
5111 decrypt or have a valid MAC), with a very large sequence number. This means
5112 that all subsequent legitimate packets are dropped causing a denial of
5113 service for a specific DTLS connection.
5115 This issue was reported to OpenSSL by the OCAP audit team.
5120 * Certificate message OOB reads
5122 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5123 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5124 theoretical DoS risk but this has not been observed in practice on common
5127 The messages affected are client certificate, client certificate request
5128 and server certificate. As a result the attack can only be performed
5129 against a client or a server which enables client authentication.
5131 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5136 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5138 * Prevent padding oracle in AES-NI CBC MAC check
5140 A MITM attacker can use a padding oracle attack to decrypt traffic
5141 when the connection uses an AES CBC cipher and the server support
5144 This issue was introduced as part of the fix for Lucky 13 padding
5145 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5146 constant time by making sure that always the same bytes are read and
5147 compared against either the MAC or padding bytes. But it no longer
5148 checked that there was enough data to have both the MAC and padding
5151 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5155 * Fix EVP_EncodeUpdate overflow
5157 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5158 Base64 encoding of binary data. If an attacker is able to supply very large
5159 amounts of input data then a length check can overflow resulting in a heap
5162 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5163 the `PEM_write_bio*` family of functions. These are mainly used within the
5164 OpenSSL command line applications, so any application which processes data
5165 from an untrusted source and outputs it as a PEM file should be considered
5166 vulnerable to this issue. User applications that call these APIs directly
5167 with large amounts of untrusted data may also be vulnerable.
5169 This issue was reported by Guido Vranken.
5174 * Fix EVP_EncryptUpdate overflow
5176 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5177 is able to supply very large amounts of input data after a previous call to
5178 EVP_EncryptUpdate() with a partial block then a length check can overflow
5179 resulting in a heap corruption. Following an analysis of all OpenSSL
5180 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5181 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5182 the first called function after an EVP_EncryptInit(), and therefore that
5183 specific call must be safe. The second form is where the length passed to
5184 EVP_EncryptUpdate() can be seen from the code to be some small value and
5185 therefore there is no possibility of an overflow. Since all instances are
5186 one of these two forms, it is believed that there can be no overflows in
5187 internal code due to this problem. It should be noted that
5188 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5189 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5190 of these calls have also been analysed too and it is believed there are no
5191 instances in internal usage where an overflow could occur.
5193 This issue was reported by Guido Vranken.
5198 * Prevent ASN.1 BIO excessive memory allocation
5200 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5201 a short invalid encoding can cause allocation of large amounts of memory
5202 potentially consuming excessive resources or exhausting memory.
5204 Any application parsing untrusted data through d2i BIO functions is
5205 affected. The memory based functions such as d2i_X509() are *not* affected.
5206 Since the memory based functions are used by the TLS library, TLS
5207 applications are not affected.
5209 This issue was reported by Brian Carpenter.
5216 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5217 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5218 in arbitrary stack data being returned in the buffer.
5220 This issue was reported by Guido Vranken.
5225 * Modify behavior of ALPN to invoke callback after SNI/servername
5226 callback, such that updates to the SSL_CTX affect ALPN.
5230 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5235 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5236 methods are enabled and ssl2 is disabled the methods return NULL.
5240 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5242 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5243 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5244 provide any "EXPORT" or "LOW" strength ciphers.
5248 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5249 is by default disabled at build-time. Builds that are not configured with
5250 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5251 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5252 will need to explicitly call either of:
5254 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5256 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5258 as appropriate. Even if either of those is used, or the application
5259 explicitly uses the version-specific SSLv2_method() or its client and
5260 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5261 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5262 ciphers, and SSLv2 56-bit DES are no longer available.
5267 * Fix a double-free in DSA code
5269 A double free bug was discovered when OpenSSL parses malformed DSA private
5270 keys and could lead to a DoS attack or memory corruption for applications
5271 that receive DSA private keys from untrusted sources. This scenario is
5274 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5280 * Disable SRP fake user seed to address a server memory leak.
5282 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5284 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5285 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5286 was changed to ignore the "fake user" SRP seed, even if the seed
5289 Users should use SRP_VBASE_get1_by_user instead. Note that in
5290 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5291 also that even though configuring the SRP seed attempts to hide
5292 invalid usernames by continuing the handshake with fake
5293 credentials, this behaviour is not constant time and no strong
5294 guarantees are made that the handshake is indistinguishable from
5295 that of a valid user.
5300 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5302 In the BN_hex2bn function the number of hex digits is calculated using an
5303 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5304 large values of `i` this can result in `bn_expand` not allocating any
5305 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5306 field as NULL leading to a subsequent NULL ptr deref. For very large values
5307 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5308 In this case memory is allocated to the internal BIGNUM data field, but it
5309 is insufficiently sized leading to heap corruption. A similar issue exists
5310 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5311 is ever called by user applications with very large untrusted hex/dec data.
5312 This is anticipated to be a rare occurrence.
5314 All OpenSSL internal usage of these functions use data that is not expected
5315 to be untrusted, e.g. config file data or application command line
5316 arguments. If user developed applications generate config file data based
5317 on untrusted data then it is possible that this could also lead to security
5318 consequences. This is also anticipated to be rare.
5320 This issue was reported to OpenSSL by Guido Vranken.
5325 * Fix memory issues in `BIO_*printf` functions
5327 The internal `fmtstr` function used in processing a "%s" format string in
5328 the `BIO_*printf` functions could overflow while calculating the length of a
5329 string and cause an OOB read when printing very long strings.
5331 Additionally the internal `doapr_outch` function can attempt to write to an
5332 OOB memory location (at an offset from the NULL pointer) in the event of a
5333 memory allocation failure. In 1.0.2 and below this could be caused where
5334 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5335 could be in processing a very long "%s" format string. Memory leaks can
5338 The first issue may mask the second issue dependent on compiler behaviour.
5339 These problems could enable attacks where large amounts of untrusted data
5340 is passed to the `BIO_*printf` functions. If applications use these functions
5341 in this way then they could be vulnerable. OpenSSL itself uses these
5342 functions when printing out human-readable dumps of ASN.1 data. Therefore
5343 applications that print this data could be vulnerable if the data is from
5344 untrusted sources. OpenSSL command line applications could also be
5345 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5346 as command line arguments.
5348 Libssl is not considered directly vulnerable. Additionally certificates etc
5349 received via remote connections via libssl are also unlikely to be able to
5350 trigger these issues because of message size limits enforced within libssl.
5352 This issue was reported to OpenSSL Guido Vranken.
5357 * Side channel attack on modular exponentiation
5359 A side-channel attack was found which makes use of cache-bank conflicts on
5360 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5361 of RSA keys. The ability to exploit this issue is limited as it relies on
5362 an attacker who has control of code in a thread running on the same
5363 hyper-threaded core as the victim thread which is performing decryptions.
5365 This issue was reported to OpenSSL by Yuval Yarom, The University of
5366 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5367 Nadia Heninger, University of Pennsylvania with more information at
5368 <http://cachebleed.info>.
5373 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5374 if no keysize is specified with default_bits. This fixes an
5375 omission in an earlier change that changed all RSA/DSA key generation
5376 commands to use 2048 bits by default.
5380 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5382 * DH small subgroups
5384 Historically OpenSSL only ever generated DH parameters based on "safe"
5385 primes. More recently (in version 1.0.2) support was provided for
5386 generating X9.42 style parameter files such as those required for RFC 5114
5387 support. The primes used in such files may not be "safe". Where an
5388 application is using DH configured with parameters based on primes that are
5389 not "safe" then an attacker could use this fact to find a peer's private
5390 DH exponent. This attack requires that the attacker complete multiple
5391 handshakes in which the peer uses the same private DH exponent. For example
5392 this could be used to discover a TLS server's private DH exponent if it's
5393 reusing the private DH exponent or it's using a static DH ciphersuite.
5395 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5396 TLS. It is not on by default. If the option is not set then the server
5397 reuses the same private DH exponent for the life of the server process and
5398 would be vulnerable to this attack. It is believed that many popular
5399 applications do set this option and would therefore not be at risk.
5401 The fix for this issue adds an additional check where a "q" parameter is
5402 available (as is the case in X9.42 based parameters). This detects the
5403 only known attack, and is the only possible defense for static DH
5404 ciphersuites. This could have some performance impact.
5406 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5407 default and cannot be disabled. This could have some performance impact.
5409 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5414 * SSLv2 doesn't block disabled ciphers
5416 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5417 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5418 been disabled, provided that the SSLv2 protocol was not also disabled via
5421 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5422 and Sebastian Schinzel.
5427 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5429 * BN_mod_exp may produce incorrect results on x86_64
5431 There is a carry propagating bug in the x86_64 Montgomery squaring
5432 procedure. No EC algorithms are affected. Analysis suggests that attacks
5433 against RSA and DSA as a result of this defect would be very difficult to
5434 perform and are not believed likely. Attacks against DH are considered just
5435 feasible (although very difficult) because most of the work necessary to
5436 deduce information about a private key may be performed offline. The amount
5437 of resources required for such an attack would be very significant and
5438 likely only accessible to a limited number of attackers. An attacker would
5439 additionally need online access to an unpatched system using the target
5440 private key in a scenario with persistent DH parameters and a private
5441 key that is shared between multiple clients. For example this can occur by
5442 default in OpenSSL DHE based SSL/TLS ciphersuites.
5444 This issue was reported to OpenSSL by Hanno Böck.
5449 * Certificate verify crash with missing PSS parameter
5451 The signature verification routines will crash with a NULL pointer
5452 dereference if presented with an ASN.1 signature using the RSA PSS
5453 algorithm and absent mask generation function parameter. Since these
5454 routines are used to verify certificate signature algorithms this can be
5455 used to crash any certificate verification operation and exploited in a
5456 DoS attack. Any application which performs certificate verification is
5457 vulnerable including OpenSSL clients and servers which enable client
5460 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5465 * X509_ATTRIBUTE memory leak
5467 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5468 memory. This structure is used by the PKCS#7 and CMS routines so any
5469 application which reads PKCS#7 or CMS data from untrusted sources is
5470 affected. SSL/TLS is not affected.
5472 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5478 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5479 This changes the decoding behaviour for some invalid messages,
5480 though the change is mostly in the more lenient direction, and
5481 legacy behaviour is preserved as much as possible.
5485 * In DSA_generate_parameters_ex, if the provided seed is too short,
5488 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5490 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5492 * Alternate chains certificate forgery
5494 During certificate verification, OpenSSL will attempt to find an
5495 alternative certificate chain if the first attempt to build such a chain
5496 fails. An error in the implementation of this logic can mean that an
5497 attacker could cause certain checks on untrusted certificates to be
5498 bypassed, such as the CA flag, enabling them to use a valid leaf
5499 certificate to act as a CA and "issue" an invalid certificate.
5501 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5506 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5508 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5509 incompatibility in the handling of HMAC. The previous ABI has now been
5514 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5516 * Malformed ECParameters causes infinite loop
5518 When processing an ECParameters structure OpenSSL enters an infinite loop
5519 if the curve specified is over a specially malformed binary polynomial
5522 This can be used to perform denial of service against any
5523 system which processes public keys, certificate requests or
5524 certificates. This includes TLS clients and TLS servers with
5525 client authentication enabled.
5527 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5532 * Exploitable out-of-bounds read in X509_cmp_time
5534 X509_cmp_time does not properly check the length of the ASN1_TIME
5535 string and can read a few bytes out of bounds. In addition,
5536 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5539 An attacker can use this to craft malformed certificates and CRLs of
5540 various sizes and potentially cause a segmentation fault, resulting in
5541 a DoS on applications that verify certificates or CRLs. TLS clients
5542 that verify CRLs are affected. TLS clients and servers with client
5543 authentication enabled may be affected if they use custom verification
5546 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5547 independently by Hanno Böck.
5552 * PKCS7 crash with missing EnvelopedContent
5554 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5555 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5556 with missing content and trigger a NULL pointer dereference on parsing.
5558 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5559 structures from untrusted sources are affected. OpenSSL clients and
5560 servers are not affected.
5562 This issue was reported to OpenSSL by Michal Zalewski (Google).
5567 * CMS verify infinite loop with unknown hash function
5569 When verifying a signedData message the CMS code can enter an infinite loop
5570 if presented with an unknown hash function OID. This can be used to perform
5571 denial of service against any system which verifies signedData messages using
5573 This issue was reported to OpenSSL by Johannes Bauer.
5578 * Race condition handling NewSessionTicket
5580 If a NewSessionTicket is received by a multi-threaded client when attempting to
5581 reuse a previous ticket then a race condition can occur potentially leading to
5582 a double free of the ticket data.
5587 * Only support 256-bit or stronger elliptic curves with the
5588 'ecdh_auto' setting (server) or by default (client). Of supported
5589 curves, prefer P-256 (both).
5593 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5595 * ClientHello sigalgs DoS fix
5597 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5598 invalid signature algorithms extension a NULL pointer dereference will
5599 occur. This can be exploited in a DoS attack against the server.
5601 This issue was was reported to OpenSSL by David Ramos of Stanford
5605 *Stephen Henson and Matt Caswell*
5607 * Multiblock corrupted pointer fix
5609 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5610 feature only applies on 64 bit x86 architecture platforms that support AES
5611 NI instructions. A defect in the implementation of "multiblock" can cause
5612 OpenSSL's internal write buffer to become incorrectly set to NULL when
5613 using non-blocking IO. Typically, when the user application is using a
5614 socket BIO for writing, this will only result in a failed connection.
5615 However if some other BIO is used then it is likely that a segmentation
5616 fault will be triggered, thus enabling a potential DoS attack.
5618 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5623 * Segmentation fault in DTLSv1_listen fix
5625 The DTLSv1_listen function is intended to be stateless and processes the
5626 initial ClientHello from many peers. It is common for user code to loop
5627 over the call to DTLSv1_listen until a valid ClientHello is received with
5628 an associated cookie. A defect in the implementation of DTLSv1_listen means
5629 that state is preserved in the SSL object from one invocation to the next
5630 that can lead to a segmentation fault. Errors processing the initial
5631 ClientHello can trigger this scenario. An example of such an error could be
5632 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5635 This issue was reported to OpenSSL by Per Allansson.
5640 * Segmentation fault in ASN1_TYPE_cmp fix
5642 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5643 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5644 certificate signature algorithm consistency this can be used to crash any
5645 certificate verification operation and exploited in a DoS attack. Any
5646 application which performs certificate verification is vulnerable including
5647 OpenSSL clients and servers which enable client authentication.
5652 * Segmentation fault for invalid PSS parameters fix
5654 The signature verification routines will crash with a NULL pointer
5655 dereference if presented with an ASN.1 signature using the RSA PSS
5656 algorithm and invalid parameters. Since these routines are used to verify
5657 certificate signature algorithms this can be used to crash any
5658 certificate verification operation and exploited in a DoS attack. Any
5659 application which performs certificate verification is vulnerable including
5660 OpenSSL clients and servers which enable client authentication.
5662 This issue was was reported to OpenSSL by Brian Carpenter.
5667 * ASN.1 structure reuse memory corruption fix
5669 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5670 memory corruption via an invalid write. Such reuse is and has been
5671 strongly discouraged and is believed to be rare.
5673 Applications that parse structures containing CHOICE or ANY DEFINED BY
5674 components may be affected. Certificate parsing (d2i_X509 and related
5675 functions) are however not affected. OpenSSL clients and servers are
5681 * PKCS7 NULL pointer dereferences fix
5683 The PKCS#7 parsing code does not handle missing outer ContentInfo
5684 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5685 missing content and trigger a NULL pointer dereference on parsing.
5687 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5688 otherwise parse PKCS#7 structures from untrusted sources are
5689 affected. OpenSSL clients and servers are not affected.
5691 This issue was reported to OpenSSL by Michal Zalewski (Google).
5696 * DoS via reachable assert in SSLv2 servers fix
5698 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5699 servers that both support SSLv2 and enable export cipher suites by sending
5700 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5702 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5703 (OpenSSL development team).
5708 * Empty CKE with client auth and DHE fix
5710 If client auth is used then a server can seg fault in the event of a DHE
5711 ciphersuite being selected and a zero length ClientKeyExchange message
5712 being sent by the client. This could be exploited in a DoS attack.
5717 * Handshake with unseeded PRNG fix
5719 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5720 with an unseeded PRNG. The conditions are:
5721 - The client is on a platform where the PRNG has not been seeded
5722 automatically, and the user has not seeded manually
5723 - A protocol specific client method version has been used (i.e. not
5724 SSL_client_methodv23)
5725 - A ciphersuite is used that does not require additional random data from
5726 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5728 If the handshake succeeds then the client random that has been used will
5729 have been generated from a PRNG with insufficient entropy and therefore the
5730 output may be predictable.
5732 For example using the following command with an unseeded openssl will
5733 succeed on an unpatched platform:
5735 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5740 * Use After Free following d2i_ECPrivatekey error fix
5742 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5743 could cause a use after free condition. This, in turn, could cause a double
5744 free in several private key parsing functions (such as d2i_PrivateKey
5745 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5746 for applications that receive EC private keys from untrusted
5747 sources. This scenario is considered rare.
5749 This issue was discovered by the BoringSSL project and fixed in their
5755 * X509_to_X509_REQ NULL pointer deref fix
5757 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5758 the certificate key is invalid. This function is rarely used in practice.
5760 This issue was discovered by Brian Carpenter.
5765 * Removed the export ciphers from the DEFAULT ciphers
5769 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5771 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5772 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5773 So far those who have to target multiple platforms would compromise
5774 and argue that binary targeting say ARMv5 would still execute on
5775 ARMv8. "Universal" build resolves this compromise by providing
5776 near-optimal performance even on newer platforms.
5780 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5781 (other platforms pending).
5783 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5785 * Add support for the SignedCertificateTimestampList certificate and
5786 OCSP response extensions from RFC6962.
5790 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5791 for corner cases. (Certain input points at infinity could lead to
5792 bogus results, with non-infinity inputs mapped to infinity too.)
5796 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5797 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5798 common cases are optimized and there still is room for further
5799 improvements. Vector Permutation AES for Altivec is also added.
5803 * Add support for little-endian ppc64 Linux target.
5805 *Marcelo Cerri (IBM)*
5807 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5808 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5809 are optimized and there still is room for further improvements.
5810 Both 32- and 64-bit modes are supported.
5812 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5814 * Improved ARMv7 NEON support.
5818 * Support for SPARC Architecture 2011 crypto extensions, first
5819 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5820 SHA256/512, MD5, GHASH and modular exponentiation.
5822 *Andy Polyakov, David Miller*
5824 * Accelerated modular exponentiation for Intel processors, a.k.a.
5827 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5829 * Support for new and upcoming Intel processors, including AVX2,
5830 BMI and SHA ISA extensions. This includes additional "stitched"
5831 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5834 This work was sponsored by Intel Corp.
5838 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5839 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5840 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5844 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5845 this fixes a limitation in previous versions of OpenSSL.
5849 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5850 MGF1 digest and OAEP label.
5854 * Add EVP support for key wrapping algorithms, to avoid problems with
5855 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5856 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5857 algorithms and include tests cases.
5861 * Add functions to allocate and set the fields of an ECDSA_METHOD
5864 *Douglas E. Engert, Steve Henson*
5866 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5867 difference in days and seconds between two tm or ASN1_TIME structures.
5871 * Add -rev test option to s_server to just reverse order of characters
5872 received by client and send back to server. Also prints an abbreviated
5873 summary of the connection parameters.
5877 * New option -brief for s_client and s_server to print out a brief summary
5878 of connection parameters.
5882 * Add callbacks for arbitrary TLS extensions.
5884 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5886 * New option -crl_download in several openssl utilities to download CRLs
5887 from CRLDP extension in certificates.
5891 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5895 * New function X509_CRL_diff to generate a delta CRL from the difference
5896 of two full CRLs. Add support to "crl" utility.
5900 * New functions to set lookup_crls function and to retrieve
5901 X509_STORE from X509_STORE_CTX.
5905 * Print out deprecated issuer and subject unique ID fields in
5910 * Extend OCSP I/O functions so they can be used for simple general purpose
5911 HTTP as well as OCSP. New wrapper function which can be used to download
5912 CRLs using the OCSP API.
5916 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5920 * `SSL_CONF*` functions. These provide a common framework for application
5921 configuration using configuration files or command lines.
5925 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5926 message callback and prints the results. Needs compile time option
5927 "enable-ssl-trace". New options to s_client and s_server to enable
5932 * New ctrl and macro to retrieve supported points extensions.
5933 Print out extension in s_server and s_client.
5937 * New functions to retrieve certificate signature and signature
5942 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5947 * New Suite B modes for TLS code. These use and enforce the requirements
5948 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5949 only use Suite B curves. The Suite B modes can be set by using the
5950 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5954 * New chain verification flags for Suite B levels of security. Check
5955 algorithms are acceptable when flags are set in X509_verify_cert.
5959 * Make tls1_check_chain return a set of flags indicating checks passed
5960 by a certificate chain. Add additional tests to handle client
5961 certificates: checks for matching certificate type and issuer name
5966 * If an attempt is made to use a signature algorithm not in the peer
5967 preference list abort the handshake. If client has no suitable
5968 signature algorithms in response to a certificate request do not
5969 use the certificate.
5973 * If server EC tmp key is not in client preference list abort handshake.
5977 * Add support for certificate stores in CERT structure. This makes it
5978 possible to have different stores per SSL structure or one store in
5979 the parent SSL_CTX. Include distinct stores for certificate chain
5980 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5981 to build and store a certificate chain in CERT structure: returning
5982 an error if the chain cannot be built: this will allow applications
5983 to test if a chain is correctly configured.
5985 Note: if the CERT based stores are not set then the parent SSL_CTX
5986 store is used to retain compatibility with existing behaviour.
5990 * New function ssl_set_client_disabled to set a ciphersuite disabled
5991 mask based on the current session, check mask when sending client
5992 hello and checking the requested ciphersuite.
5996 * New ctrls to retrieve and set certificate types in a certificate
5997 request message. Print out received values in s_client. If certificate
5998 types is not set with custom values set sensible values based on
5999 supported signature algorithms.
6003 * Support for distinct client and server supported signature algorithms.
6007 * Add certificate callback. If set this is called whenever a certificate
6008 is required by client or server. An application can decide which
6009 certificate chain to present based on arbitrary criteria: for example
6010 supported signature algorithms. Add very simple example to s_server.
6011 This fixes many of the problems and restrictions of the existing client
6012 certificate callback: for example you can now clear an existing
6013 certificate and specify the whole chain.
6017 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6018 the certificate can be used for (if anything). Set valid_flags field
6019 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6020 to have similar checks in it.
6022 Add new "cert_flags" field to CERT structure and include a "strict mode".
6023 This enforces some TLS certificate requirements (such as only permitting
6024 certificate signature algorithms contained in the supported algorithms
6025 extension) which some implementations ignore: this option should be used
6026 with caution as it could cause interoperability issues.
6030 * Update and tidy signature algorithm extension processing. Work out
6031 shared signature algorithms based on preferences and peer algorithms
6032 and print them out in s_client and s_server. Abort handshake if no
6033 shared signature algorithms.
6037 * Add new functions to allow customised supported signature algorithms
6038 for SSL and SSL_CTX structures. Add options to s_client and s_server
6043 * New function SSL_certs_clear() to delete all references to certificates
6044 from an SSL structure. Before this once a certificate had been added
6045 it couldn't be removed.
6049 * Integrate hostname, email address and IP address checking with certificate
6050 verification. New verify options supporting checking in openssl utility.
6054 * Fixes and wildcard matching support to hostname and email checking
6055 functions. Add manual page.
6057 *Florian Weimer (Red Hat Product Security Team)*
6059 * New functions to check a hostname email or IP address against a
6060 certificate. Add options x509 utility to print results of checks against
6065 * Fix OCSP checking.
6067 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6069 * Initial experimental support for explicitly trusted non-root CAs.
6070 OpenSSL still tries to build a complete chain to a root but if an
6071 intermediate CA has a trust setting included that is used. The first
6072 setting is used: whether to trust (e.g., -addtrust option to the x509
6077 * Add -trusted_first option which attempts to find certificates in the
6078 trusted store even if an untrusted chain is also supplied.
6082 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6083 platform support for Linux and Android.
6087 * Support for linux-x32, ILP32 environment in x86_64 framework.
6091 * Experimental multi-implementation support for FIPS capable OpenSSL.
6092 When in FIPS mode the approved implementations are used as normal,
6093 when not in FIPS mode the internal unapproved versions are used instead.
6094 This means that the FIPS capable OpenSSL isn't forced to use the
6095 (often lower performance) FIPS implementations outside FIPS mode.
6099 * Transparently support X9.42 DH parameters when calling
6100 PEM_read_bio_DHparameters. This means existing applications can handle
6101 the new parameter format automatically.
6105 * Initial experimental support for X9.42 DH parameter format: mainly
6106 to support use of 'q' parameter for RFC5114 parameters.
6110 * Add DH parameters from RFC5114 including test data to dhtest.
6114 * Support for automatic EC temporary key parameter selection. If enabled
6115 the most preferred EC parameters are automatically used instead of
6116 hardcoded fixed parameters. Now a server just has to call:
6117 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6118 support ECDH and use the most appropriate parameters.
6122 * Enhance and tidy EC curve and point format TLS extension code. Use
6123 static structures instead of allocation if default values are used.
6124 New ctrls to set curves we wish to support and to retrieve shared curves.
6125 Print out shared curves in s_server. New options to s_server and s_client
6126 to set list of supported curves.
6130 * New ctrls to retrieve supported signature algorithms and
6131 supported curve values as an array of NIDs. Extend openssl utility
6132 to print out received values.
6136 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6137 between NIDs and the more common NIST names such as "P-256". Enhance
6138 ecparam utility and ECC method to recognise the NIST names for curves.
6142 * Enhance SSL/TLS certificate chain handling to support different
6143 chains for each certificate instead of one chain in the parent SSL_CTX.
6147 * Support for fixed DH ciphersuite client authentication: where both
6148 server and client use DH certificates with common parameters.
6152 * Support for fixed DH ciphersuites: those requiring DH server
6157 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6159 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6160 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6161 X509_CINF_get_signature were reverted post internal team review.
6166 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6168 * OCSP Status Request extension unbounded memory growth
6170 A malicious client can send an excessively large OCSP Status Request
6171 extension. If that client continually requests renegotiation, sending a
6172 large OCSP Status Request extension each time, then there will be unbounded
6173 memory growth on the server. This will eventually lead to a Denial Of
6174 Service attack through memory exhaustion. Servers with a default
6175 configuration are vulnerable even if they do not support OCSP. Builds using
6176 the "no-ocsp" build time option are not affected.
6178 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6183 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6186 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6192 * OOB write in MDC2_Update()
6194 An overflow can occur in MDC2_Update() either if called directly or
6195 through the EVP_DigestUpdate() function using MDC2. If an attacker
6196 is able to supply very large amounts of input data after a previous
6197 call to EVP_EncryptUpdate() with a partial block then a length check
6198 can overflow resulting in a heap corruption.
6200 The amount of data needed is comparable to SIZE_MAX which is impractical
6203 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6208 * Malformed SHA512 ticket DoS
6210 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6211 DoS attack where a malformed ticket will result in an OOB read which will
6214 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6215 a custom server callback and ticket lookup mechanism.
6217 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6222 * OOB write in BN_bn2dec()
6224 The function BN_bn2dec() does not check the return value of BN_div_word().
6225 This can cause an OOB write if an application uses this function with an
6226 overly large BIGNUM. This could be a problem if an overly large certificate
6227 or CRL is printed out from an untrusted source. TLS is not affected because
6228 record limits will reject an oversized certificate before it is parsed.
6230 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6235 * OOB read in TS_OBJ_print_bio()
6237 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6238 the total length the OID text representation would use and not the amount
6239 of data written. This will result in OOB reads when large OIDs are
6242 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6247 * Pointer arithmetic undefined behaviour
6249 Avoid some undefined pointer arithmetic
6251 A common idiom in the codebase is to check limits in the following manner:
6254 Where "p" points to some malloc'd data of SIZE bytes and
6257 "len" here could be from some externally supplied data (e.g. from a TLS
6260 The rules of C pointer arithmetic are such that "p + len" is only well
6261 defined where len <= SIZE. Therefore the above idiom is actually
6262 undefined behaviour.
6264 For example this could cause problems if some malloc implementation
6265 provides an address for "p" such that "p + len" actually overflows for
6266 values of len that are too big and therefore p + len < limit.
6268 This issue was reported to OpenSSL by Guido Vranken
6273 * Constant time flag not preserved in DSA signing
6275 Operations in the DSA signing algorithm should run in constant time in
6276 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6277 implementation means that a non-constant time codepath is followed for
6278 certain operations. This has been demonstrated through a cache-timing
6279 attack to be sufficient for an attacker to recover the private DSA key.
6281 This issue was reported by César Pereida (Aalto University), Billy Brumley
6282 (Tampere University of Technology), and Yuval Yarom (The University of
6283 Adelaide and NICTA).
6288 * DTLS buffered message DoS
6290 In a DTLS connection where handshake messages are delivered out-of-order
6291 those messages that OpenSSL is not yet ready to process will be buffered
6292 for later use. Under certain circumstances, a flaw in the logic means that
6293 those messages do not get removed from the buffer even though the handshake
6294 has been completed. An attacker could force up to approx. 15 messages to
6295 remain in the buffer when they are no longer required. These messages will
6296 be cleared when the DTLS connection is closed. The default maximum size for
6297 a message is 100k. Therefore the attacker could force an additional 1500k
6298 to be consumed per connection. By opening many simulataneous connections an
6299 attacker could cause a DoS attack through memory exhaustion.
6301 This issue was reported to OpenSSL by Quan Luo.
6306 * DTLS replay protection DoS
6308 A flaw in the DTLS replay attack protection mechanism means that records
6309 that arrive for future epochs update the replay protection "window" before
6310 the MAC for the record has been validated. This could be exploited by an
6311 attacker by sending a record for the next epoch (which does not have to
6312 decrypt or have a valid MAC), with a very large sequence number. This means
6313 that all subsequent legitimate packets are dropped causing a denial of
6314 service for a specific DTLS connection.
6316 This issue was reported to OpenSSL by the OCAP audit team.
6321 * Certificate message OOB reads
6323 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6324 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6325 theoretical DoS risk but this has not been observed in practice on common
6328 The messages affected are client certificate, client certificate request
6329 and server certificate. As a result the attack can only be performed
6330 against a client or a server which enables client authentication.
6332 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6337 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6339 * Prevent padding oracle in AES-NI CBC MAC check
6341 A MITM attacker can use a padding oracle attack to decrypt traffic
6342 when the connection uses an AES CBC cipher and the server support
6345 This issue was introduced as part of the fix for Lucky 13 padding
6346 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6347 constant time by making sure that always the same bytes are read and
6348 compared against either the MAC or padding bytes. But it no longer
6349 checked that there was enough data to have both the MAC and padding
6352 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6357 * Fix EVP_EncodeUpdate overflow
6359 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6360 Base64 encoding of binary data. If an attacker is able to supply very large
6361 amounts of input data then a length check can overflow resulting in a heap
6364 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6365 the `PEM_write_bio*` family of functions. These are mainly used within the
6366 OpenSSL command line applications, so any application which processes data
6367 from an untrusted source and outputs it as a PEM file should be considered
6368 vulnerable to this issue. User applications that call these APIs directly
6369 with large amounts of untrusted data may also be vulnerable.
6371 This issue was reported by Guido Vranken.
6376 * Fix EVP_EncryptUpdate overflow
6378 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6379 is able to supply very large amounts of input data after a previous call to
6380 EVP_EncryptUpdate() with a partial block then a length check can overflow
6381 resulting in a heap corruption. Following an analysis of all OpenSSL
6382 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6383 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6384 the first called function after an EVP_EncryptInit(), and therefore that
6385 specific call must be safe. The second form is where the length passed to
6386 EVP_EncryptUpdate() can be seen from the code to be some small value and
6387 therefore there is no possibility of an overflow. Since all instances are
6388 one of these two forms, it is believed that there can be no overflows in
6389 internal code due to this problem. It should be noted that
6390 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6391 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6392 of these calls have also been analysed too and it is believed there are no
6393 instances in internal usage where an overflow could occur.
6395 This issue was reported by Guido Vranken.
6400 * Prevent ASN.1 BIO excessive memory allocation
6402 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6403 a short invalid encoding can casuse allocation of large amounts of memory
6404 potentially consuming excessive resources or exhausting memory.
6406 Any application parsing untrusted data through d2i BIO functions is
6407 affected. The memory based functions such as d2i_X509() are *not* affected.
6408 Since the memory based functions are used by the TLS library, TLS
6409 applications are not affected.
6411 This issue was reported by Brian Carpenter.
6418 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6419 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6420 in arbitrary stack data being returned in the buffer.
6422 This issue was reported by Guido Vranken.
6427 * Modify behavior of ALPN to invoke callback after SNI/servername
6428 callback, such that updates to the SSL_CTX affect ALPN.
6432 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6437 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6438 methods are enabled and ssl2 is disabled the methods return NULL.
6442 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6444 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6445 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6446 provide any "EXPORT" or "LOW" strength ciphers.
6450 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6451 is by default disabled at build-time. Builds that are not configured with
6452 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6453 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6454 will need to explicitly call either of:
6456 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6458 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6460 as appropriate. Even if either of those is used, or the application
6461 explicitly uses the version-specific SSLv2_method() or its client and
6462 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6463 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6464 ciphers, and SSLv2 56-bit DES are no longer available.
6469 * Fix a double-free in DSA code
6471 A double free bug was discovered when OpenSSL parses malformed DSA private
6472 keys and could lead to a DoS attack or memory corruption for applications
6473 that receive DSA private keys from untrusted sources. This scenario is
6476 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6482 * Disable SRP fake user seed to address a server memory leak.
6484 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6486 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6487 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6488 was changed to ignore the "fake user" SRP seed, even if the seed
6491 Users should use SRP_VBASE_get1_by_user instead. Note that in
6492 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6493 also that even though configuring the SRP seed attempts to hide
6494 invalid usernames by continuing the handshake with fake
6495 credentials, this behaviour is not constant time and no strong
6496 guarantees are made that the handshake is indistinguishable from
6497 that of a valid user.
6502 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6504 In the BN_hex2bn function the number of hex digits is calculated using an
6505 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6506 large values of `i` this can result in `bn_expand` not allocating any
6507 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6508 field as NULL leading to a subsequent NULL ptr deref. For very large values
6509 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6510 In this case memory is allocated to the internal BIGNUM data field, but it
6511 is insufficiently sized leading to heap corruption. A similar issue exists
6512 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6513 is ever called by user applications with very large untrusted hex/dec data.
6514 This is anticipated to be a rare occurrence.
6516 All OpenSSL internal usage of these functions use data that is not expected
6517 to be untrusted, e.g. config file data or application command line
6518 arguments. If user developed applications generate config file data based
6519 on untrusted data then it is possible that this could also lead to security
6520 consequences. This is also anticipated to be rare.
6522 This issue was reported to OpenSSL by Guido Vranken.
6527 * Fix memory issues in `BIO_*printf` functions
6529 The internal `fmtstr` function used in processing a "%s" format string in
6530 the `BIO_*printf` functions could overflow while calculating the length of a
6531 string and cause an OOB read when printing very long strings.
6533 Additionally the internal `doapr_outch` function can attempt to write to an
6534 OOB memory location (at an offset from the NULL pointer) in the event of a
6535 memory allocation failure. In 1.0.2 and below this could be caused where
6536 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6537 could be in processing a very long "%s" format string. Memory leaks can
6540 The first issue may mask the second issue dependent on compiler behaviour.
6541 These problems could enable attacks where large amounts of untrusted data
6542 is passed to the `BIO_*printf` functions. If applications use these functions
6543 in this way then they could be vulnerable. OpenSSL itself uses these
6544 functions when printing out human-readable dumps of ASN.1 data. Therefore
6545 applications that print this data could be vulnerable if the data is from
6546 untrusted sources. OpenSSL command line applications could also be
6547 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6548 as command line arguments.
6550 Libssl is not considered directly vulnerable. Additionally certificates etc
6551 received via remote connections via libssl are also unlikely to be able to
6552 trigger these issues because of message size limits enforced within libssl.
6554 This issue was reported to OpenSSL Guido Vranken.
6559 * Side channel attack on modular exponentiation
6561 A side-channel attack was found which makes use of cache-bank conflicts on
6562 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6563 of RSA keys. The ability to exploit this issue is limited as it relies on
6564 an attacker who has control of code in a thread running on the same
6565 hyper-threaded core as the victim thread which is performing decryptions.
6567 This issue was reported to OpenSSL by Yuval Yarom, The University of
6568 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6569 Nadia Heninger, University of Pennsylvania with more information at
6570 <http://cachebleed.info>.
6575 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6576 if no keysize is specified with default_bits. This fixes an
6577 omission in an earlier change that changed all RSA/DSA key generation
6578 commands to use 2048 bits by default.
6582 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6584 * Protection for DH small subgroup attacks
6586 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6587 switched on by default and cannot be disabled. This could have some
6592 * SSLv2 doesn't block disabled ciphers
6594 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6595 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6596 been disabled, provided that the SSLv2 protocol was not also disabled via
6599 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6600 and Sebastian Schinzel.
6605 * Reject DH handshakes with parameters shorter than 1024 bits.
6609 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6611 * Certificate verify crash with missing PSS parameter
6613 The signature verification routines will crash with a NULL pointer
6614 dereference if presented with an ASN.1 signature using the RSA PSS
6615 algorithm and absent mask generation function parameter. Since these
6616 routines are used to verify certificate signature algorithms this can be
6617 used to crash any certificate verification operation and exploited in a
6618 DoS attack. Any application which performs certificate verification is
6619 vulnerable including OpenSSL clients and servers which enable client
6622 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6627 * X509_ATTRIBUTE memory leak
6629 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6630 memory. This structure is used by the PKCS#7 and CMS routines so any
6631 application which reads PKCS#7 or CMS data from untrusted sources is
6632 affected. SSL/TLS is not affected.
6634 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6640 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6641 This changes the decoding behaviour for some invalid messages,
6642 though the change is mostly in the more lenient direction, and
6643 legacy behaviour is preserved as much as possible.
6647 * In DSA_generate_parameters_ex, if the provided seed is too short,
6648 use a random seed, as already documented.
6650 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6652 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6654 * Alternate chains certificate forgery
6656 During certificate verfification, OpenSSL will attempt to find an
6657 alternative certificate chain if the first attempt to build such a chain
6658 fails. An error in the implementation of this logic can mean that an
6659 attacker could cause certain checks on untrusted certificates to be
6660 bypassed, such as the CA flag, enabling them to use a valid leaf
6661 certificate to act as a CA and "issue" an invalid certificate.
6663 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6669 * Race condition handling PSK identify hint
6671 If PSK identity hints are received by a multi-threaded client then
6672 the values are wrongly updated in the parent SSL_CTX structure. This can
6673 result in a race condition potentially leading to a double free of the
6679 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6681 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6682 incompatibility in the handling of HMAC. The previous ABI has now been
6685 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6687 * Malformed ECParameters causes infinite loop
6689 When processing an ECParameters structure OpenSSL enters an infinite loop
6690 if the curve specified is over a specially malformed binary polynomial
6693 This can be used to perform denial of service against any
6694 system which processes public keys, certificate requests or
6695 certificates. This includes TLS clients and TLS servers with
6696 client authentication enabled.
6698 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6703 * Exploitable out-of-bounds read in X509_cmp_time
6705 X509_cmp_time does not properly check the length of the ASN1_TIME
6706 string and can read a few bytes out of bounds. In addition,
6707 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6710 An attacker can use this to craft malformed certificates and CRLs of
6711 various sizes and potentially cause a segmentation fault, resulting in
6712 a DoS on applications that verify certificates or CRLs. TLS clients
6713 that verify CRLs are affected. TLS clients and servers with client
6714 authentication enabled may be affected if they use custom verification
6717 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6718 independently by Hanno Böck.
6723 * PKCS7 crash with missing EnvelopedContent
6725 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6726 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6727 with missing content and trigger a NULL pointer dereference on parsing.
6729 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6730 structures from untrusted sources are affected. OpenSSL clients and
6731 servers are not affected.
6733 This issue was reported to OpenSSL by Michal Zalewski (Google).
6738 * CMS verify infinite loop with unknown hash function
6740 When verifying a signedData message the CMS code can enter an infinite loop
6741 if presented with an unknown hash function OID. This can be used to perform
6742 denial of service against any system which verifies signedData messages using
6744 This issue was reported to OpenSSL by Johannes Bauer.
6749 * Race condition handling NewSessionTicket
6751 If a NewSessionTicket is received by a multi-threaded client when attempting to
6752 reuse a previous ticket then a race condition can occur potentially leading to
6753 a double free of the ticket data.
6758 * Reject DH handshakes with parameters shorter than 768 bits.
6760 *Kurt Roeckx and Emilia Kasper*
6762 * dhparam: generate 2048-bit parameters by default.
6764 *Kurt Roeckx and Emilia Kasper*
6766 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6768 * Segmentation fault in ASN1_TYPE_cmp fix
6770 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6771 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6772 certificate signature algorithm consistency this can be used to crash any
6773 certificate verification operation and exploited in a DoS attack. Any
6774 application which performs certificate verification is vulnerable including
6775 OpenSSL clients and servers which enable client authentication.
6780 * ASN.1 structure reuse memory corruption fix
6782 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6783 memory corruption via an invalid write. Such reuse is and has been
6784 strongly discouraged and is believed to be rare.
6786 Applications that parse structures containing CHOICE or ANY DEFINED BY
6787 components may be affected. Certificate parsing (d2i_X509 and related
6788 functions) are however not affected. OpenSSL clients and servers are
6794 * PKCS7 NULL pointer dereferences fix
6796 The PKCS#7 parsing code does not handle missing outer ContentInfo
6797 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6798 missing content and trigger a NULL pointer dereference on parsing.
6800 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6801 otherwise parse PKCS#7 structures from untrusted sources are
6802 affected. OpenSSL clients and servers are not affected.
6804 This issue was reported to OpenSSL by Michal Zalewski (Google).
6809 * DoS via reachable assert in SSLv2 servers fix
6811 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6812 servers that both support SSLv2 and enable export cipher suites by sending
6813 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6815 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6816 (OpenSSL development team).
6821 * Use After Free following d2i_ECPrivatekey error fix
6823 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6824 could cause a use after free condition. This, in turn, could cause a double
6825 free in several private key parsing functions (such as d2i_PrivateKey
6826 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6827 for applications that receive EC private keys from untrusted
6828 sources. This scenario is considered rare.
6830 This issue was discovered by the BoringSSL project and fixed in their
6836 * X509_to_X509_REQ NULL pointer deref fix
6838 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6839 the certificate key is invalid. This function is rarely used in practice.
6841 This issue was discovered by Brian Carpenter.
6846 * Removed the export ciphers from the DEFAULT ciphers
6850 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6852 * Build fixes for the Windows and OpenVMS platforms
6854 *Matt Caswell and Richard Levitte*
6856 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6858 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6859 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6860 dereference. This could lead to a Denial Of Service attack. Thanks to
6861 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6866 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6867 dtls1_buffer_record function under certain conditions. In particular this
6868 could occur if an attacker sent repeated DTLS records with the same
6869 sequence number but for the next epoch. The memory leak could be exploited
6870 by an attacker in a Denial of Service attack through memory exhaustion.
6871 Thanks to Chris Mueller for reporting this issue.
6876 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6877 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6878 method would be set to NULL which could later result in a NULL pointer
6879 dereference. Thanks to Frank Schmirler for reporting this issue.
6884 * Abort handshake if server key exchange message is omitted for ephemeral
6887 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6888 reporting this issue.
6893 * Remove non-export ephemeral RSA code on client and server. This code
6894 violated the TLS standard by allowing the use of temporary RSA keys in
6895 non-export ciphersuites and could be used by a server to effectively
6896 downgrade the RSA key length used to a value smaller than the server
6897 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6898 INRIA or reporting this issue.
6903 * Fixed issue where DH client certificates are accepted without verification.
6904 An OpenSSL server will accept a DH certificate for client authentication
6905 without the certificate verify message. This effectively allows a client to
6906 authenticate without the use of a private key. This only affects servers
6907 which trust a client certificate authority which issues certificates
6908 containing DH keys: these are extremely rare and hardly ever encountered.
6909 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6915 * Ensure that the session ID context of an SSL is updated when its
6916 SSL_CTX is updated via SSL_set_SSL_CTX.
6918 The session ID context is typically set from the parent SSL_CTX,
6919 and can vary with the CTX.
6923 * Fix various certificate fingerprint issues.
6925 By using non-DER or invalid encodings outside the signed portion of a
6926 certificate the fingerprint can be changed without breaking the signature.
6927 Although no details of the signed portion of the certificate can be changed
6928 this can cause problems with some applications: e.g. those using the
6929 certificate fingerprint for blacklists.
6931 1. Reject signatures with non zero unused bits.
6933 If the BIT STRING containing the signature has non zero unused bits reject
6934 the signature. All current signature algorithms require zero unused bits.
6936 2. Check certificate algorithm consistency.
6938 Check the AlgorithmIdentifier inside TBS matches the one in the
6939 certificate signature. NB: this will result in signature failure
6940 errors for some broken certificates.
6942 Thanks to Konrad Kraszewski from Google for reporting this issue.
6944 3. Check DSA/ECDSA signatures use DER.
6946 Re-encode DSA/ECDSA signatures and compare with the original received
6947 signature. Return an error if there is a mismatch.
6949 This will reject various cases including garbage after signature
6950 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6951 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6952 (negative or with leading zeroes).
6954 Further analysis was conducted and fixes were developed by Stephen Henson
6955 of the OpenSSL core team.
6961 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6962 results on some platforms, including x86_64. This bug occurs at random
6963 with a very low probability, and is not known to be exploitable in any
6964 way, though its exact impact is difficult to determine. Thanks to Pieter
6965 Wuille (Blockstream) who reported this issue and also suggested an initial
6966 fix. Further analysis was conducted by the OpenSSL development team and
6967 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6968 the OpenSSL core team.
6973 * Do not resume sessions on the server if the negotiated protocol
6974 version does not match the session's version. Resuming with a different
6975 version, while not strictly forbidden by the RFC, is of questionable
6976 sanity and breaks all known clients.
6978 *David Benjamin, Emilia Käsper*
6980 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6981 early CCS messages during renegotiation. (Note that because
6982 renegotiation is encrypted, this early CCS was not exploitable.)
6986 * Tighten client-side session ticket handling during renegotiation:
6987 ensure that the client only accepts a session ticket if the server sends
6988 the extension anew in the ServerHello. Previously, a TLS client would
6989 reuse the old extension state and thus accept a session ticket if one was
6990 announced in the initial ServerHello.
6992 Similarly, ensure that the client requires a session ticket if one
6993 was advertised in the ServerHello. Previously, a TLS client would
6994 ignore a missing NewSessionTicket message.
6998 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7002 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7003 sends a carefully crafted handshake message, to cause OpenSSL to fail
7004 to free up to 64k of memory causing a memory leak. This could be
7005 exploited in a Denial Of Service attack. This issue affects OpenSSL
7006 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7007 whether SRTP is used or configured. Implementations of OpenSSL that
7008 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7010 The fix was developed by the OpenSSL team.
7015 * Session Ticket Memory Leak.
7017 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7018 integrity of that ticket is first verified. In the event of a session
7019 ticket integrity check failing, OpenSSL will fail to free memory
7020 causing a memory leak. By sending a large number of invalid session
7021 tickets an attacker could exploit this issue in a Denial Of Service
7027 * Build option no-ssl3 is incomplete.
7029 When OpenSSL is configured with "no-ssl3" as a build option, servers
7030 could accept and complete a SSL 3.0 handshake, and clients could be
7031 configured to send them.
7034 *Akamai and the OpenSSL team*
7036 * Add support for TLS_FALLBACK_SCSV.
7037 Client applications doing fallback retries should call
7038 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7041 *Adam Langley, Bodo Moeller*
7043 * Add additional DigestInfo checks.
7045 Re-encode DigestInto in DER and check against the original when
7046 verifying RSA signature: this will reject any improperly encoded
7047 DigestInfo structures.
7049 Note: this is a precautionary measure and no attacks are currently known.
7053 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7055 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7056 SRP code can be overrun an internal buffer. Add sanity check that
7057 g, A, B < N to SRP code.
7059 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7060 Group for discovering this issue.
7065 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7066 TLS 1.0 instead of higher protocol versions when the ClientHello message
7067 is badly fragmented. This allows a man-in-the-middle attacker to force a
7068 downgrade to TLS 1.0 even if both the server and the client support a
7069 higher protocol version, by modifying the client's TLS records.
7071 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7072 researching this issue.
7077 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7078 to a denial of service attack. A malicious server can crash the client
7079 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7080 ciphersuite and sending carefully crafted handshake messages.
7082 Thanks to Felix Gröbert (Google) for discovering and researching this
7088 * By sending carefully crafted DTLS packets an attacker could cause openssl
7089 to leak memory. This can be exploited through a Denial of Service attack.
7090 Thanks to Adam Langley for discovering and researching this issue.
7095 * An attacker can force openssl to consume large amounts of memory whilst
7096 processing DTLS handshake messages. This can be exploited through a
7097 Denial of Service attack.
7098 Thanks to Adam Langley for discovering and researching this issue.
7103 * An attacker can force an error condition which causes openssl to crash
7104 whilst processing DTLS packets due to memory being freed twice. This
7105 can be exploited through a Denial of Service attack.
7106 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7112 * If a multithreaded client connects to a malicious server using a resumed
7113 session and the server sends an ec point format extension it could write
7114 up to 255 bytes to freed memory.
7116 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7122 * A malicious server can crash an OpenSSL client with a null pointer
7123 dereference (read) by specifying an SRP ciphersuite even though it was not
7124 properly negotiated with the client. This can be exploited through a
7125 Denial of Service attack.
7127 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7128 discovering and researching this issue.
7133 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7134 X509_name_oneline, X509_name_print_ex et al. to leak some information
7135 from the stack. Applications may be affected if they echo pretty printing
7136 output to the attacker.
7138 Thanks to Ivan Fratric (Google) for discovering this issue.
7141 *Emilia Käsper, and Steve Henson*
7143 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7144 for corner cases. (Certain input points at infinity could lead to
7145 bogus results, with non-infinity inputs mapped to infinity too.)
7149 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7151 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7152 handshake can force the use of weak keying material in OpenSSL
7153 SSL/TLS clients and servers.
7155 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7156 researching this issue. ([CVE-2014-0224])
7158 *KIKUCHI Masashi, Steve Henson*
7160 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7161 OpenSSL DTLS client the code can be made to recurse eventually crashing
7164 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7167 *Imre Rad, Steve Henson*
7169 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7170 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7171 client or server. This is potentially exploitable to run arbitrary
7172 code on a vulnerable client or server.
7174 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7176 *Jüri Aedla, Steve Henson*
7178 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7179 are subject to a denial of service attack.
7181 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7182 this issue. ([CVE-2014-3470])
7184 *Felix Gröbert, Ivan Fratric, Steve Henson*
7186 * Harmonize version and its documentation. -f flag is used to display
7189 *mancha <mancha1@zoho.com>*
7191 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7192 in i2d_ECPrivateKey.
7194 *mancha <mancha1@zoho.com>*
7196 * Fix some double frees. These are not thought to be exploitable.
7198 *mancha <mancha1@zoho.com>*
7200 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7202 * A missing bounds check in the handling of the TLS heartbeat extension
7203 can be used to reveal up to 64k of memory to a connected client or
7206 Thanks for Neel Mehta of Google Security for discovering this bug and to
7207 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7208 preparing the fix ([CVE-2014-0160])
7210 *Adam Langley, Bodo Moeller*
7212 * Fix for the attack described in the paper "Recovering OpenSSL
7213 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7214 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7215 <http://eprint.iacr.org/2014/140>
7217 Thanks to Yuval Yarom and Naomi Benger for discovering this
7218 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7220 *Yuval Yarom and Naomi Benger*
7222 * TLS pad extension: draft-agl-tls-padding-03
7224 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7225 TLS client Hello record length value would otherwise be > 255 and
7226 less that 512 pad with a dummy extension containing zeroes so it
7227 is at least 512 bytes long.
7229 *Adam Langley, Steve Henson*
7231 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7233 * Fix for TLS record tampering bug. A carefully crafted invalid
7234 handshake could crash OpenSSL with a NULL pointer exception.
7235 Thanks to Anton Johansson for reporting this issues.
7238 * Keep original DTLS digest and encryption contexts in retransmission
7239 structures so we can use the previous session parameters if they need
7240 to be resent. ([CVE-2013-6450])
7244 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7245 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7246 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7247 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7248 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7249 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7251 *Rob Stradling, Adam Langley*
7253 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7255 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7256 supporting platforms or when small records were transferred.
7258 *Andy Polyakov, Steve Henson*
7260 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7262 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7264 This addresses the flaw in CBC record processing discovered by
7265 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7266 at: <http://www.isg.rhul.ac.uk/tls/>
7268 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7269 Security Group at Royal Holloway, University of London
7270 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7271 Emilia Käsper for the initial patch.
7274 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7276 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7277 ciphersuites which can be exploited in a denial of service attack.
7278 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7279 and detecting this bug and to Wolfgang Ettlinger
7280 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7285 * Return an error when checking OCSP signatures when key is NULL.
7286 This fixes a DoS attack. ([CVE-2013-0166])
7290 * Make openssl verify return errors.
7292 *Chris Palmer <palmer@google.com> and Ben Laurie*
7294 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7295 the right response is stapled. Also change SSL_get_certificate()
7296 so it returns the certificate actually sent.
7297 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7299 *Rob Stradling <rob.stradling@comodo.com>*
7301 * Fix possible deadlock when decoding public keys.
7305 * Don't use TLS 1.0 record version number in initial client hello
7310 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7312 * Sanity check record length before skipping explicit IV in TLS
7313 1.2, 1.1 and DTLS to fix DoS attack.
7315 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7316 fuzzing as a service testing platform.
7321 * Initialise tkeylen properly when encrypting CMS messages.
7322 Thanks to Solar Designer of Openwall for reporting this issue.
7326 * In FIPS mode don't try to use composite ciphers as they are not
7331 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7333 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7334 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7335 mean any application compiled against OpenSSL 1.0.0 headers setting
7336 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7337 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7338 0x10000000L Any application which was previously compiled against
7339 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7340 will need to be recompiled as a result. Letting be results in
7341 inability to disable specifically TLS 1.1 and in client context,
7342 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7346 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7347 disable just protocol X, but all protocols above X *if* there are
7348 protocols *below* X still enabled. In more practical terms it means
7349 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7350 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7351 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7356 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7358 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7359 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7360 in CRYPTO_realloc_clean.
7362 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7363 issue and to Adam Langley <agl@chromium.org> for fixing it.
7366 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7368 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7372 * Workarounds for some broken servers that "hang" if a client hello
7373 record length exceeds 255 bytes.
7375 1. Do not use record version number > TLS 1.0 in initial client
7376 hello: some (but not all) hanging servers will now work.
7377 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7378 the number of ciphers sent in the client hello. This should be
7379 set to an even number, such as 50, for example by passing:
7380 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7381 Most broken servers should now work.
7382 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7383 TLS 1.2 client support entirely.
7387 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7391 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7393 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7394 STRING form instead of a DigestInfo.
7398 * The format used for MDC2 RSA signatures is inconsistent between EVP
7399 and the RSA_sign/RSA_verify functions. This was made more apparent when
7400 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7401 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7402 the correct format in RSA_verify so both forms transparently work.
7406 * Some servers which support TLS 1.0 can choke if we initially indicate
7407 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7408 encrypted premaster secret. As a workaround use the maximum permitted
7409 client version in client hello, this should keep such servers happy
7410 and still work with previous versions of OpenSSL.
7414 * Add support for TLS/DTLS heartbeats.
7416 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7418 * Add support for SCTP.
7420 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7422 * Improved PRNG seeding for VOS.
7424 *Paul Green <Paul.Green@stratus.com>*
7426 * Extensive assembler packs updates, most notably:
7428 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7429 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7430 - x86_64: bit-sliced AES implementation;
7431 - ARM: NEON support, contemporary platforms optimizations;
7432 - s390x: z196 support;
7433 - `*`: GHASH and GF(2^m) multiplication implementations;
7437 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7438 (removal of unnecessary code)
7440 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7442 * Add TLS key material exporter from RFC 5705.
7446 * Add DTLS-SRTP negotiation from RFC 5764.
7450 * Add Next Protocol Negotiation,
7451 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7452 disabled with a no-npn flag to config or Configure. Code donated
7455 *Adam Langley <agl@google.com> and Ben Laurie*
7457 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7458 NIST-P256, NIST-P521, with constant-time single point multiplication on
7459 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7460 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7461 Code made available under Apache License version 2.0.
7463 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7464 line to include this in your build of OpenSSL, and run "make depend" (or
7465 "make update"). This enables the following EC_METHODs:
7467 EC_GFp_nistp224_method()
7468 EC_GFp_nistp256_method()
7469 EC_GFp_nistp521_method()
7471 EC_GROUP_new_by_curve_name() will automatically use these (while
7472 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7475 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7477 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7478 all platforms. Move ssize_t definition from e_os.h to the public
7479 header file e_os2.h as it now appears in public header file cms.h
7483 * New -sigopt option to the ca, req and x509 utilities. Additional
7484 signature parameters can be passed using this option and in
7489 * Add RSA PSS signing function. This will generate and set the
7490 appropriate AlgorithmIdentifiers for PSS based on those in the
7491 corresponding EVP_MD_CTX structure. No application support yet.
7495 * Support for companion algorithm specific ASN1 signing routines.
7496 New function ASN1_item_sign_ctx() signs a pre-initialised
7497 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7498 the appropriate parameters.
7502 * Add new algorithm specific ASN1 verification initialisation function
7503 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7504 handling will be the same no matter what EVP_PKEY_METHOD is used.
7505 Add a PSS handler to support verification of PSS signatures: checked
7506 against a number of sample certificates.
7510 * Add signature printing for PSS. Add PSS OIDs.
7512 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7514 * Add algorithm specific signature printing. An individual ASN1 method
7515 can now print out signatures instead of the standard hex dump.
7517 More complex signatures (e.g. PSS) can print out more meaningful
7518 information. Include DSA version that prints out the signature
7523 * Password based recipient info support for CMS library: implementing
7528 * Split password based encryption into PBES2 and PBKDF2 functions. This
7529 neatly separates the code into cipher and PBE sections and is required
7530 for some algorithms that split PBES2 into separate pieces (such as
7531 password based CMS).
7535 * Session-handling fixes:
7536 - Fix handling of connections that are resuming with a session ID,
7537 but also support Session Tickets.
7538 - Fix a bug that suppressed issuing of a new ticket if the client
7539 presented a ticket with an expired session.
7540 - Try to set the ticket lifetime hint to something reasonable.
7541 - Make tickets shorter by excluding irrelevant information.
7542 - On the client side, don't ignore renewed tickets.
7544 *Adam Langley, Bodo Moeller (Google)*
7546 * Fix PSK session representation.
7550 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7552 This work was sponsored by Intel.
7556 * Add GCM support to TLS library. Some custom code is needed to split
7557 the IV between the fixed (from PRF) and explicit (from TLS record)
7558 portions. This adds all GCM ciphersuites supported by RFC5288 and
7559 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7560 add a special AESGCM string for GCM only.
7564 * Expand range of ctrls for AES GCM. Permit setting invocation
7565 field on decrypt and retrieval of invocation field only on encrypt.
7569 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7570 As required by RFC5289 these ciphersuites cannot be used if for
7571 versions of TLS earlier than 1.2.
7575 * For FIPS capable OpenSSL interpret a NULL default public key method
7576 as unset and return the appropriate default but do *not* set the default.
7577 This means we can return the appropriate method in applications that
7578 switch between FIPS and non-FIPS modes.
7582 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7583 ENGINE is used then we cannot handle that in the FIPS module so we
7584 keep original code iff non-FIPS operations are allowed.
7588 * Add -attime option to openssl utilities.
7590 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7592 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7596 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7597 FIPS EC methods unconditionally for now.
7601 * New build option no-ec2m to disable characteristic 2 code.
7605 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7606 all cases can be covered as some introduce binary incompatibilities.
7610 * Redirect RSA operations to FIPS module including keygen,
7611 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7615 * Add similar low-level API blocking to ciphers.
7619 * low-level digest APIs are not approved in FIPS mode: any attempt
7620 to use these will cause a fatal error. Applications that *really* want
7621 to use them can use the `private_*` version instead.
7625 * Redirect cipher operations to FIPS module for FIPS builds.
7629 * Redirect digest operations to FIPS module for FIPS builds.
7633 * Update build system to add "fips" flag which will link in fipscanister.o
7634 for static and shared library builds embedding a signature if needed.
7638 * Output TLS supported curves in preference order instead of numerical
7639 order. This is currently hardcoded for the highest order curves first.
7640 This should be configurable so applications can judge speed vs strength.
7644 * Add TLS v1.2 server support for client authentication.
7648 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7653 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7654 FIPS modules versions.
7658 * Add TLS v1.2 client side support for client authentication. Keep cache
7659 of handshake records longer as we don't know the hash algorithm to use
7660 until after the certificate request message is received.
7664 * Initial TLS v1.2 client support. Add a default signature algorithms
7665 extension including all the algorithms we support. Parse new signature
7666 format in client key exchange. Relax some ECC signing restrictions for
7667 TLS v1.2 as indicated in RFC5246.
7671 * Add server support for TLS v1.2 signature algorithms extension. Switch
7672 to new signature format when needed using client digest preference.
7673 All server ciphersuites should now work correctly in TLS v1.2. No client
7674 support yet and no support for client certificates.
7678 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7679 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7680 ciphersuites. At present only RSA key exchange ciphersuites work with
7681 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7682 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7683 and version checking.
7687 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7688 with this defined it will not be affected by any changes to ssl internal
7689 structures. Add several utility functions to allow openssl application
7690 to work with OPENSSL_NO_SSL_INTERN defined.
7694 * A long standing patch to add support for SRP from EdelWeb (Peter
7695 Sylvester and Christophe Renou) was integrated.
7696 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7697 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7700 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7704 * Permit abbreviated handshakes when renegotiating using the function
7705 SSL_renegotiate_abbreviated().
7707 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7709 * Add call to ENGINE_register_all_complete() to
7710 ENGINE_load_builtin_engines(), so some implementations get used
7711 automatically instead of needing explicit application support.
7715 * Add support for TLS key exporter as described in RFC5705.
7717 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7719 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7720 a few changes are required:
7722 Add SSL_OP_NO_TLSv1_1 flag.
7723 Add TLSv1_1 methods.
7724 Update version checking logic to handle version 1.1.
7725 Add explicit IV handling (ported from DTLS code).
7726 Add command line options to s_client/s_server.
7733 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7735 * X509_ATTRIBUTE memory leak
7737 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7738 memory. This structure is used by the PKCS#7 and CMS routines so any
7739 application which reads PKCS#7 or CMS data from untrusted sources is
7740 affected. SSL/TLS is not affected.
7742 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7748 * Race condition handling PSK identify hint
7750 If PSK identity hints are received by a multi-threaded client then
7751 the values are wrongly updated in the parent SSL_CTX structure. This can
7752 result in a race condition potentially leading to a double free of the
7758 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7760 * Malformed ECParameters causes infinite loop
7762 When processing an ECParameters structure OpenSSL enters an infinite loop
7763 if the curve specified is over a specially malformed binary polynomial
7766 This can be used to perform denial of service against any
7767 system which processes public keys, certificate requests or
7768 certificates. This includes TLS clients and TLS servers with
7769 client authentication enabled.
7771 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7776 * Exploitable out-of-bounds read in X509_cmp_time
7778 X509_cmp_time does not properly check the length of the ASN1_TIME
7779 string and can read a few bytes out of bounds. In addition,
7780 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7783 An attacker can use this to craft malformed certificates and CRLs of
7784 various sizes and potentially cause a segmentation fault, resulting in
7785 a DoS on applications that verify certificates or CRLs. TLS clients
7786 that verify CRLs are affected. TLS clients and servers with client
7787 authentication enabled may be affected if they use custom verification
7790 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7791 independently by Hanno Böck.
7796 * PKCS7 crash with missing EnvelopedContent
7798 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7799 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7800 with missing content and trigger a NULL pointer dereference on parsing.
7802 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7803 structures from untrusted sources are affected. OpenSSL clients and
7804 servers are not affected.
7806 This issue was reported to OpenSSL by Michal Zalewski (Google).
7811 * CMS verify infinite loop with unknown hash function
7813 When verifying a signedData message the CMS code can enter an infinite loop
7814 if presented with an unknown hash function OID. This can be used to perform
7815 denial of service against any system which verifies signedData messages using
7817 This issue was reported to OpenSSL by Johannes Bauer.
7822 * Race condition handling NewSessionTicket
7824 If a NewSessionTicket is received by a multi-threaded client when attempting to
7825 reuse a previous ticket then a race condition can occur potentially leading to
7826 a double free of the ticket data.
7831 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7833 * Segmentation fault in ASN1_TYPE_cmp fix
7835 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7836 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7837 certificate signature algorithm consistency this can be used to crash any
7838 certificate verification operation and exploited in a DoS attack. Any
7839 application which performs certificate verification is vulnerable including
7840 OpenSSL clients and servers which enable client authentication.
7845 * ASN.1 structure reuse memory corruption fix
7847 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7848 memory corruption via an invalid write. Such reuse is and has been
7849 strongly discouraged and is believed to be rare.
7851 Applications that parse structures containing CHOICE or ANY DEFINED BY
7852 components may be affected. Certificate parsing (d2i_X509 and related
7853 functions) are however not affected. OpenSSL clients and servers are
7859 * PKCS7 NULL pointer dereferences fix
7861 The PKCS#7 parsing code does not handle missing outer ContentInfo
7862 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7863 missing content and trigger a NULL pointer dereference on parsing.
7865 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7866 otherwise parse PKCS#7 structures from untrusted sources are
7867 affected. OpenSSL clients and servers are not affected.
7869 This issue was reported to OpenSSL by Michal Zalewski (Google).
7874 * DoS via reachable assert in SSLv2 servers fix
7876 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7877 servers that both support SSLv2 and enable export cipher suites by sending
7878 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7880 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7881 (OpenSSL development team).
7886 * Use After Free following d2i_ECPrivatekey error fix
7888 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7889 could cause a use after free condition. This, in turn, could cause a double
7890 free in several private key parsing functions (such as d2i_PrivateKey
7891 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7892 for applications that receive EC private keys from untrusted
7893 sources. This scenario is considered rare.
7895 This issue was discovered by the BoringSSL project and fixed in their
7901 * X509_to_X509_REQ NULL pointer deref fix
7903 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7904 the certificate key is invalid. This function is rarely used in practice.
7906 This issue was discovered by Brian Carpenter.
7911 * Removed the export ciphers from the DEFAULT ciphers
7915 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7917 * Build fixes for the Windows and OpenVMS platforms
7919 *Matt Caswell and Richard Levitte*
7921 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7923 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7924 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7925 dereference. This could lead to a Denial Of Service attack. Thanks to
7926 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7931 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7932 dtls1_buffer_record function under certain conditions. In particular this
7933 could occur if an attacker sent repeated DTLS records with the same
7934 sequence number but for the next epoch. The memory leak could be exploited
7935 by an attacker in a Denial of Service attack through memory exhaustion.
7936 Thanks to Chris Mueller for reporting this issue.
7941 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7942 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7943 method would be set to NULL which could later result in a NULL pointer
7944 dereference. Thanks to Frank Schmirler for reporting this issue.
7949 * Abort handshake if server key exchange message is omitted for ephemeral
7952 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7953 reporting this issue.
7958 * Remove non-export ephemeral RSA code on client and server. This code
7959 violated the TLS standard by allowing the use of temporary RSA keys in
7960 non-export ciphersuites and could be used by a server to effectively
7961 downgrade the RSA key length used to a value smaller than the server
7962 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7963 INRIA or reporting this issue.
7968 * Fixed issue where DH client certificates are accepted without verification.
7969 An OpenSSL server will accept a DH certificate for client authentication
7970 without the certificate verify message. This effectively allows a client to
7971 authenticate without the use of a private key. This only affects servers
7972 which trust a client certificate authority which issues certificates
7973 containing DH keys: these are extremely rare and hardly ever encountered.
7974 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7980 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7981 results on some platforms, including x86_64. This bug occurs at random
7982 with a very low probability, and is not known to be exploitable in any
7983 way, though its exact impact is difficult to determine. Thanks to Pieter
7984 Wuille (Blockstream) who reported this issue and also suggested an initial
7985 fix. Further analysis was conducted by the OpenSSL development team and
7986 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7987 the OpenSSL core team.
7992 * Fix various certificate fingerprint issues.
7994 By using non-DER or invalid encodings outside the signed portion of a
7995 certificate the fingerprint can be changed without breaking the signature.
7996 Although no details of the signed portion of the certificate can be changed
7997 this can cause problems with some applications: e.g. those using the
7998 certificate fingerprint for blacklists.
8000 1. Reject signatures with non zero unused bits.
8002 If the BIT STRING containing the signature has non zero unused bits reject
8003 the signature. All current signature algorithms require zero unused bits.
8005 2. Check certificate algorithm consistency.
8007 Check the AlgorithmIdentifier inside TBS matches the one in the
8008 certificate signature. NB: this will result in signature failure
8009 errors for some broken certificates.
8011 Thanks to Konrad Kraszewski from Google for reporting this issue.
8013 3. Check DSA/ECDSA signatures use DER.
8015 Reencode DSA/ECDSA signatures and compare with the original received
8016 signature. Return an error if there is a mismatch.
8018 This will reject various cases including garbage after signature
8019 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8020 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8021 (negative or with leading zeroes).
8023 Further analysis was conducted and fixes were developed by Stephen Henson
8024 of the OpenSSL core team.
8030 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8032 * Session Ticket Memory Leak.
8034 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8035 integrity of that ticket is first verified. In the event of a session
8036 ticket integrity check failing, OpenSSL will fail to free memory
8037 causing a memory leak. By sending a large number of invalid session
8038 tickets an attacker could exploit this issue in a Denial Of Service
8044 * Build option no-ssl3 is incomplete.
8046 When OpenSSL is configured with "no-ssl3" as a build option, servers
8047 could accept and complete a SSL 3.0 handshake, and clients could be
8048 configured to send them.
8051 *Akamai and the OpenSSL team*
8053 * Add support for TLS_FALLBACK_SCSV.
8054 Client applications doing fallback retries should call
8055 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8058 *Adam Langley, Bodo Moeller*
8060 * Add additional DigestInfo checks.
8062 Reencode DigestInto in DER and check against the original when
8063 verifying RSA signature: this will reject any improperly encoded
8064 DigestInfo structures.
8066 Note: this is a precautionary measure and no attacks are currently known.
8070 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8072 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8073 to a denial of service attack. A malicious server can crash the client
8074 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8075 ciphersuite and sending carefully crafted handshake messages.
8077 Thanks to Felix Gröbert (Google) for discovering and researching this
8083 * By sending carefully crafted DTLS packets an attacker could cause openssl
8084 to leak memory. This can be exploited through a Denial of Service attack.
8085 Thanks to Adam Langley for discovering and researching this issue.
8090 * An attacker can force openssl to consume large amounts of memory whilst
8091 processing DTLS handshake messages. This can be exploited through a
8092 Denial of Service attack.
8093 Thanks to Adam Langley for discovering and researching this issue.
8098 * An attacker can force an error condition which causes openssl to crash
8099 whilst processing DTLS packets due to memory being freed twice. This
8100 can be exploited through a Denial of Service attack.
8101 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8107 * If a multithreaded client connects to a malicious server using a resumed
8108 session and the server sends an ec point format extension it could write
8109 up to 255 bytes to freed memory.
8111 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8117 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8118 X509_name_oneline, X509_name_print_ex et al. to leak some information
8119 from the stack. Applications may be affected if they echo pretty printing
8120 output to the attacker.
8122 Thanks to Ivan Fratric (Google) for discovering this issue.
8125 *Emilia Käsper, and Steve Henson*
8127 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8128 for corner cases. (Certain input points at infinity could lead to
8129 bogus results, with non-infinity inputs mapped to infinity too.)
8133 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8135 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8136 handshake can force the use of weak keying material in OpenSSL
8137 SSL/TLS clients and servers.
8139 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8140 researching this issue. ([CVE-2014-0224])
8142 *KIKUCHI Masashi, Steve Henson*
8144 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8145 OpenSSL DTLS client the code can be made to recurse eventually crashing
8148 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8151 *Imre Rad, Steve Henson*
8153 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8154 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8155 client or server. This is potentially exploitable to run arbitrary
8156 code on a vulnerable client or server.
8158 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8160 *Jüri Aedla, Steve Henson*
8162 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8163 are subject to a denial of service attack.
8165 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8166 this issue. ([CVE-2014-3470])
8168 *Felix Gröbert, Ivan Fratric, Steve Henson*
8170 * Harmonize version and its documentation. -f flag is used to display
8173 *mancha <mancha1@zoho.com>*
8175 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8176 in i2d_ECPrivateKey.
8178 *mancha <mancha1@zoho.com>*
8180 * Fix some double frees. These are not thought to be exploitable.
8182 *mancha <mancha1@zoho.com>*
8184 * Fix for the attack described in the paper "Recovering OpenSSL
8185 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8186 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8187 <http://eprint.iacr.org/2014/140>
8189 Thanks to Yuval Yarom and Naomi Benger for discovering this
8190 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8192 *Yuval Yarom and Naomi Benger*
8194 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8196 * Keep original DTLS digest and encryption contexts in retransmission
8197 structures so we can use the previous session parameters if they need
8198 to be resent. ([CVE-2013-6450])
8202 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8203 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8204 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8205 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8206 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8207 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8209 *Rob Stradling, Adam Langley*
8211 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8213 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8215 This addresses the flaw in CBC record processing discovered by
8216 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8217 at: <http://www.isg.rhul.ac.uk/tls/>
8219 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8220 Security Group at Royal Holloway, University of London
8221 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8222 Emilia Käsper for the initial patch.
8225 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8227 * Return an error when checking OCSP signatures when key is NULL.
8228 This fixes a DoS attack. ([CVE-2013-0166])
8232 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8233 the right response is stapled. Also change SSL_get_certificate()
8234 so it returns the certificate actually sent.
8235 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8236 (This is a backport)
8238 *Rob Stradling <rob.stradling@comodo.com>*
8240 * Fix possible deadlock when decoding public keys.
8244 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8246 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8249 * Sanity check record length before skipping explicit IV in DTLS
8252 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8253 fuzzing as a service testing platform.
8258 * Initialise tkeylen properly when encrypting CMS messages.
8259 Thanks to Solar Designer of Openwall for reporting this issue.
8263 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8265 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8266 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8267 in CRYPTO_realloc_clean.
8269 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8270 issue and to Adam Langley <agl@chromium.org> for fixing it.
8273 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8275 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8277 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8278 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8279 content decryption and always return the same error. Note: this attack
8280 needs on average 2^20 messages so it only affects automated senders. The
8281 old behaviour can be re-enabled in the CMS code by setting the
8282 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8283 an MMA defence is not necessary.
8284 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8285 this issue. ([CVE-2012-0884])
8289 * Fix CVE-2011-4619: make sure we really are receiving a
8290 client hello before rejecting multiple SGC restarts. Thanks to
8291 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8295 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8297 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8298 Thanks to Antonio Martin, Enterprise Secure Access Research and
8299 Development, Cisco Systems, Inc. for discovering this bug and
8300 preparing a fix. ([CVE-2012-0050])
8304 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8306 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8307 of the Vaudenay padding oracle attack on CBC mode encryption
8308 which enables an efficient plaintext recovery attack against
8309 the OpenSSL implementation of DTLS. Their attack exploits timing
8310 differences arising during decryption processing. A research
8311 paper describing this attack can be found at:
8312 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8313 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8314 Security Group at Royal Holloway, University of London
8315 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8316 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8317 for preparing the fix. ([CVE-2011-4108])
8319 *Robin Seggelmann, Michael Tuexen*
8321 * Clear bytes used for block padding of SSL 3.0 records.
8324 *Adam Langley (Google)*
8326 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8327 Kadianakis <desnacked@gmail.com> for discovering this issue and
8328 Adam Langley for preparing the fix. ([CVE-2011-4619])
8330 *Adam Langley (Google)*
8332 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8334 *Andrey Kulikov <amdeich@gmail.com>*
8336 * Prevent malformed RFC3779 data triggering an assertion failure.
8337 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8338 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8340 *Rob Austein <sra@hactrn.net>*
8342 * Improved PRNG seeding for VOS.
8344 *Paul Green <Paul.Green@stratus.com>*
8346 * Fix ssl_ciph.c set-up race.
8348 *Adam Langley (Google)*
8350 * Fix spurious failures in ecdsatest.c.
8352 *Emilia Käsper (Google)*
8354 * Fix the BIO_f_buffer() implementation (which was mixing different
8355 interpretations of the `..._len` fields).
8357 *Adam Langley (Google)*
8359 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8360 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8361 threads won't reuse the same blinding coefficients.
8363 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8364 lock to call BN_BLINDING_invert_ex, and avoids one use of
8365 BN_BLINDING_update for each BN_BLINDING structure (previously,
8366 the last update always remained unused).
8368 *Emilia Käsper (Google)*
8370 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8372 *Bob Buckholz (Google)*
8374 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8376 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8377 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8379 *Kaspar Brand <ossl@velox.ch>*
8381 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8382 for multi-threaded use of ECDH. ([CVE-2011-3210])
8384 *Adam Langley (Google)*
8386 * Fix x509_name_ex_d2i memory leak on bad inputs.
8390 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8391 signature public key algorithm by using OID xref utilities instead.
8392 Before this you could only use some ECC ciphersuites with SHA1 only.
8396 * Add protection against ECDSA timing attacks as mentioned in the paper
8397 by Billy Bob Brumley and Nicola Tuveri, see:
8398 <http://eprint.iacr.org/2011/232.pdf>
8400 *Billy Bob Brumley and Nicola Tuveri*
8402 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8404 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8406 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8408 * Fix bug in string printing code: if *any* escaping is enabled we must
8409 escape the escape character (backslash) or the resulting string is
8414 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8416 * Disable code workaround for ancient and obsolete Netscape browsers
8417 and servers: an attacker can use it in a ciphersuite downgrade attack.
8418 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8422 * Fixed J-PAKE implementation error, originally discovered by
8423 Sebastien Martini, further info and confirmation from Stefan
8424 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8428 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8430 * Fix extension code to avoid race conditions which can result in a buffer
8431 overrun vulnerability: resumed sessions must not be modified as they can
8432 be shared by multiple threads. CVE-2010-3864
8436 * Fix WIN32 build system to correctly link an ENGINE directory into
8441 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8443 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8446 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8448 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8450 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8451 context. The operation can be customised via the ctrl mechanism in
8452 case ENGINEs want to include additional functionality.
8456 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8460 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8461 output hashes compatible with older versions of OpenSSL.
8463 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8465 * Fix compression algorithm handling: if resuming a session use the
8466 compression algorithm of the resumed session instead of determining
8467 it from client hello again. Don't allow server to change algorithm.
8471 * Add load_crls() function to commands tidying load_certs() too. Add option
8472 to verify utility to allow additional CRLs to be included.
8476 * Update OCSP request code to permit adding custom headers to the request:
8477 some responders need this.
8481 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8484 *Julia Lawall <julia@diku.dk>*
8486 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8487 needlessly dereferenced structures, used obsolete functions and
8488 didn't handle all updated verify codes correctly.
8492 * Disable MD2 in the default configuration.
8496 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8497 indicate the initial BIO being pushed or popped. This makes it possible
8498 to determine whether the BIO is the one explicitly called or as a result
8499 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8500 it handles reference counts correctly and doesn't zero out the I/O bio
8501 when it is not being explicitly popped. WARNING: applications which
8502 included workarounds for the old buggy behaviour will need to be modified
8503 or they could free up already freed BIOs.
8507 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8508 renaming to all platforms (within the 0.9.8 branch, this was
8509 done conditionally on Netware platforms to avoid a name clash).
8511 *Guenter <lists@gknw.net>*
8513 * Add ECDHE and PSK support to DTLS.
8515 *Michael Tuexen <tuexen@fh-muenster.de>*
8517 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8522 * Add "missing" function EVP_MD_flags() (without this the only way to
8523 retrieve a digest flags is by accessing the structure directly. Update
8524 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8525 or cipher is registered as in the "from" argument. Print out all
8526 registered digests in the dgst usage message instead of manually
8527 attempting to work them out.
8531 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8532 this allows the use of compression and extensions. Change default cipher
8533 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8534 by default unless an application cipher string requests it.
8538 * Alter match criteria in PKCS12_parse(). It used to try to use local
8539 key ids to find matching certificates and keys but some PKCS#12 files
8540 don't follow the (somewhat unwritten) rules and this strategy fails.
8541 Now just gather all certificates together and the first private key
8542 then look for the first certificate that matches the key.
8546 * Support use of registered digest and cipher names for dgst and cipher
8547 commands instead of having to add each one as a special case. So now
8554 openssl dgst -sha256 foo
8556 and this works for ENGINE based algorithms too.
8560 * Update Gost ENGINE to support parameter files.
8562 *Victor B. Wagner <vitus@cryptocom.ru>*
8564 * Support GeneralizedTime in ca utility.
8566 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8568 * Enhance the hash format used for certificate directory links. The new
8569 form uses the canonical encoding (meaning equivalent names will work
8570 even if they aren't identical) and uses SHA1 instead of MD5. This form
8571 is incompatible with the older format and as a result c_rehash should
8572 be used to rebuild symbolic links.
8576 * Make PKCS#8 the default write format for private keys, replacing the
8577 traditional format. This form is standardised, more secure and doesn't
8578 include an implicit MD5 dependency.
8582 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8583 committed to OpenSSL should pass this lot as a minimum.
8587 * Add session ticket override functionality for use by EAP-FAST.
8589 *Jouni Malinen <j@w1.fi>*
8591 * Modify HMAC functions to return a value. Since these can be implemented
8592 in an ENGINE errors can occur.
8596 * Type-checked OBJ_bsearch_ex.
8600 * Type-checked OBJ_bsearch. Also some constification necessitated
8601 by type-checking. Still to come: TXT_DB, bsearch(?),
8602 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8607 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8608 seconds to a tm structure directly, instead of going through OS
8609 specific date routines. This avoids any issues with OS routines such
8610 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8611 and X509_time_adj_ex() to cover the extended range. The existing
8612 X509_time_adj() is still usable and will no longer have any date issues.
8616 * Delta CRL support. New use deltas option which will attempt to locate
8617 and search any appropriate delta CRLs available.
8619 This work was sponsored by Google.
8623 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8624 code and add additional score elements. Validate alternate CRL paths
8625 as part of the CRL checking and indicate a new error "CRL path validation
8626 error" in this case. Applications wanting additional details can use
8627 the verify callback and check the new "parent" field. If this is not
8628 NULL CRL path validation is taking place. Existing applications won't
8629 see this because it requires extended CRL support which is off by
8632 This work was sponsored by Google.
8636 * Support for freshest CRL extension.
8638 This work was sponsored by Google.
8642 * Initial indirect CRL support. Currently only supported in the CRLs
8643 passed directly and not via lookup. Process certificate issuer
8644 CRL entry extension and lookup CRL entries by bother issuer name
8645 and serial number. Check and process CRL issuer entry in IDP extension.
8647 This work was sponsored by Google.
8651 * Add support for distinct certificate and CRL paths. The CRL issuer
8652 certificate is validated separately in this case. Only enabled if
8653 an extended CRL support flag is set: this flag will enable additional
8654 CRL functionality in future.
8656 This work was sponsored by Google.
8660 * Add support for policy mappings extension.
8662 This work was sponsored by Google.
8666 * Fixes to pathlength constraint, self issued certificate handling,
8667 policy processing to align with RFC3280 and PKITS tests.
8669 This work was sponsored by Google.
8673 * Support for name constraints certificate extension. DN, email, DNS
8674 and URI types are currently supported.
8676 This work was sponsored by Google.
8680 * To cater for systems that provide a pointer-based thread ID rather
8681 than numeric, deprecate the current numeric thread ID mechanism and
8682 replace it with a structure and associated callback type. This
8683 mechanism allows a numeric "hash" to be extracted from a thread ID in
8684 either case, and on platforms where pointers are larger than 'long',
8685 mixing is done to help ensure the numeric 'hash' is usable even if it
8686 can't be guaranteed unique. The default mechanism is to use "&errno"
8687 as a pointer-based thread ID to distinguish between threads.
8689 Applications that want to provide their own thread IDs should now use
8690 CRYPTO_THREADID_set_callback() to register a callback that will call
8691 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8693 Note that ERR_remove_state() is now deprecated, because it is tied
8694 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8695 to free the current thread's error state should be replaced by
8696 ERR_remove_thread_state(NULL).
8698 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8699 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8700 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8701 application was previously providing a numeric thread callback that
8702 was inappropriate for distinguishing threads, then uniqueness might
8703 have been obtained with &errno that happened immediately in the
8704 intermediate development versions of OpenSSL; this is no longer the
8705 case, the numeric thread callback will now override the automatic use
8708 *Geoff Thorpe, with help from Bodo Moeller*
8710 * Initial support for different CRL issuing certificates. This covers a
8711 simple case where the self issued certificates in the chain exist and
8712 the real CRL issuer is higher in the existing chain.
8714 This work was sponsored by Google.
8718 * Removed effectively defunct crypto/store from the build.
8722 * Revamp of STACK to provide stronger type-checking. Still to come:
8723 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8724 ASN1_STRING, CONF_VALUE.
8728 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8729 RAM on SSL connections. This option can save about 34k per idle SSL.
8733 * Revamp of LHASH to provide stronger type-checking. Still to come:
8734 STACK, TXT_DB, bsearch, qsort.
8738 * Initial support for Cryptographic Message Syntax (aka CMS) based
8739 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8740 support for data, signedData, compressedData, digestedData and
8741 encryptedData, envelopedData types included. Scripts to check against
8742 RFC4134 examples draft and interop and consistency checks of many
8743 content types and variants.
8747 * Add options to enc utility to support use of zlib compression BIO.
8751 * Extend mk1mf to support importing of options and assembly language
8752 files from Configure script, currently only included in VC-WIN32.
8753 The assembly language rules can now optionally generate the source
8754 files from the associated perl scripts.
8758 * Implement remaining functionality needed to support GOST ciphersuites.
8759 Interop testing has been performed using CryptoPro implementations.
8761 *Victor B. Wagner <vitus@cryptocom.ru>*
8763 * s390x assembler pack.
8767 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8772 * Implement Opaque PRF Input TLS extension as specified in
8773 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8774 official specification yet and no extension type assignment by
8775 IANA exists, this extension (for now) will have to be explicitly
8776 enabled when building OpenSSL by providing the extension number
8777 to use. For example, specify an option
8779 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8781 to the "config" or "Configure" script to enable the extension,
8782 assuming extension number 0x9527 (which is a completely arbitrary
8783 and unofficial assignment based on the MD5 hash of the Internet
8784 Draft). Note that by doing so, you potentially lose
8785 interoperability with other TLS implementations since these might
8786 be using the same extension number for other purposes.
8788 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8789 opaque PRF input value to use in the handshake. This will create
8790 an internal copy of the length-'len' string at 'src', and will
8791 return non-zero for success.
8793 To get more control and flexibility, provide a callback function
8796 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8797 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8801 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8804 Callback function 'cb' will be called in handshakes, and is
8805 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8806 Argument 'arg' is for application purposes (the value as given to
8807 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8808 be provided to the callback function). The callback function
8809 has to return non-zero to report success: usually 1 to use opaque
8810 PRF input just if possible, or 2 to enforce use of the opaque PRF
8811 input. In the latter case, the library will abort the handshake
8812 if opaque PRF input is not successfully negotiated.
8814 Arguments 'peerinput' and 'len' given to the callback function
8815 will always be NULL and 0 in the case of a client. A server will
8816 see the client's opaque PRF input through these variables if
8817 available (NULL and 0 otherwise). Note that if the server
8818 provides an opaque PRF input, the length must be the same as the
8819 length of the client's opaque PRF input.
8821 Note that the callback function will only be called when creating
8822 a new session (session resumption can resume whatever was
8823 previously negotiated), and will not be called in SSL 2.0
8824 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8825 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8826 for applications that need to enforce opaque PRF input.
8830 * Update ssl code to support digests other than SHA1+MD5 for handshake
8833 *Victor B. Wagner <vitus@cryptocom.ru>*
8835 * Add RFC4507 support to OpenSSL. This includes the corrections in
8836 RFC4507bis. The encrypted ticket format is an encrypted encoded
8837 SSL_SESSION structure, that way new session features are automatically
8840 If a client application caches session in an SSL_SESSION structure
8841 support is transparent because tickets are now stored in the encoded
8844 The SSL_CTX structure automatically generates keys for ticket
8845 protection in servers so again support should be possible
8846 with no application modification.
8848 If a client or server wishes to disable RFC4507 support then the option
8849 SSL_OP_NO_TICKET can be set.
8851 Add a TLS extension debugging callback to allow the contents of any client
8852 or server extensions to be examined.
8854 This work was sponsored by Google.
8858 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8859 OpenSSL should now compile cleanly on gcc 4.2
8861 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8863 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8864 support including streaming MAC support: this is required for GOST
8865 ciphersuite support.
8867 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8869 * Add option -stream to use PKCS#7 streaming in smime utility. New
8870 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8871 to output in BER and PEM format.
8875 * Experimental support for use of HMAC via EVP_PKEY interface. This
8876 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8877 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8878 ENGINE support for HMAC keys which are unextractable. New -mac and
8879 -macopt options to dgst utility.
8883 * New option -sigopt to dgst utility. Update dgst to use
8884 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8885 alternative signing parameters such as X9.31 or PSS in the dgst
8890 * Change ssl_cipher_apply_rule(), the internal function that does
8891 the work each time a ciphersuite string requests enabling
8892 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8893 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8894 the order of disabled ciphersuites such that those ciphersuites
8895 that most recently went from enabled to disabled not only stay
8896 in order with respect to each other, but also have higher priority
8897 than other disabled ciphersuites the next time ciphersuites are
8900 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8901 the same ciphersuites as with "HIGH" alone, but in a specific
8902 order where the PSK ciphersuites come first (since they are the
8903 most recently disabled ciphersuites when "HIGH" is parsed).
8905 Also, change ssl_create_cipher_list() (using this new
8906 functionality) such that between otherwise identical
8907 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8912 * Change ssl_create_cipher_list() so that it automatically
8913 arranges the ciphersuites in reasonable order before starting
8914 to process the rule string. Thus, the definition for "DEFAULT"
8915 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8916 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8917 This makes it much easier to arrive at a reasonable default order
8918 in applications for which anonymous ciphers are OK (meaning
8919 that you can't actually use DEFAULT).
8921 *Bodo Moeller; suggested by Victor Duchovni*
8923 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8924 processing) into multiple integers instead of setting
8925 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8926 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8927 (These masks as well as the individual bit definitions are hidden
8928 away into the non-exported interface ssl/ssl_locl.h, so this
8929 change to the definition of the SSL_CIPHER structure shouldn't
8930 affect applications.) This give us more bits for each of these
8931 categories, so there is no longer a need to coagulate AES128 and
8932 AES256 into a single algorithm bit, and to coagulate Camellia128
8933 and Camellia256 into a single algorithm bit, which has led to all
8936 Thus, among other things, the kludge introduced in 0.9.7m and
8937 0.9.8e for masking out AES256 independently of AES128 or masking
8938 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8940 With the change, we also introduce new ciphersuite aliases that
8941 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8946 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8947 Use the leftmost N bytes of the signature input if the input is
8948 larger than the prime q (with N being the size in bytes of q).
8952 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8953 it yet and it is largely untested.
8957 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8961 * Initial incomplete changes to avoid need for function casts in OpenSSL
8962 some compilers (gcc 4.2 and later) reject their use. Safestack is
8963 reimplemented. Update ASN1 to avoid use of legacy functions.
8967 * Win32/64 targets are linked with Winsock2.
8971 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8972 to external functions. This can be used to increase CRL handling
8973 efficiency especially when CRLs are very large by (for example) storing
8974 the CRL revoked certificates in a database.
8978 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8979 new CRLs added to a directory can be used. New command line option
8980 -verify_return_error to s_client and s_server. This causes real errors
8981 to be returned by the verify callback instead of carrying on no matter
8982 what. This reflects the way a "real world" verify callback would behave.
8986 * GOST engine, supporting several GOST algorithms and public key formats.
8987 Kindly donated by Cryptocom.
8991 * Partial support for Issuing Distribution Point CRL extension. CRLs
8992 partitioned by DP are handled but no indirect CRL or reason partitioning
8993 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8994 selected via a scoring technique which handles IDP and AKID in CRLs.
8998 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8999 will ultimately be used for all verify operations: this will remove the
9000 X509_STORE dependency on certificate verification and allow alternative
9001 lookup methods. X509_STORE based implementations of these two callbacks.
9005 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9006 Modify get_crl() to find a valid (unexpired) CRL if possible.
9010 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9011 this would be called X509_CRL_cmp() but that name is already used by
9012 a function that just compares CRL issuer names. Cache several CRL
9013 extensions in X509_CRL structure and cache CRLDP in X509.
9017 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9018 this maps equivalent X509_NAME structures into a consistent structure.
9019 Name comparison can then be performed rapidly using memcmp().
9023 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9028 * Allow digests to supply their own micalg string for S/MIME type using
9029 the ctrl EVP_MD_CTRL_MICALG.
9033 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9034 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9035 ctrl. It can then customise the structure before and/or after signing
9040 * New function OBJ_add_sigid() to allow application defined signature OIDs
9041 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9042 to free up any added signature OIDs.
9046 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9047 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9048 digest and cipher tables. New options added to openssl utility:
9049 list-message-digest-algorithms and list-cipher-algorithms.
9053 * Change the array representation of binary polynomials: the list
9054 of degrees of non-zero coefficients is now terminated with -1.
9055 Previously it was terminated with 0, which was also part of the
9056 value; thus, the array representation was not applicable to
9057 polynomials where t^0 has coefficient zero. This change makes
9058 the array representation useful in a more general context.
9062 * Various modifications and fixes to SSL/TLS cipher string
9063 handling. For ECC, the code now distinguishes between fixed ECDH
9064 with RSA certificates on the one hand and with ECDSA certificates
9065 on the other hand, since these are separate ciphersuites. The
9066 unused code for Fortezza ciphersuites has been removed.
9068 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9069 (not "ECDHE"). For consistency with the code for DH
9070 certificates, use of ECDH certificates is now considered ECDH
9071 authentication, not RSA or ECDSA authentication (the latter is
9072 merely the CA's signing algorithm and not actively used in the
9075 The temporary ciphersuite alias "ECCdraft" is no longer
9076 available, and ECC ciphersuites are no longer excluded from "ALL"
9077 and "DEFAULT". The following aliases now exist for RFC 4492
9078 ciphersuites, most of these by analogy with the DH case:
9080 kECDHr - ECDH cert, signed with RSA
9081 kECDHe - ECDH cert, signed with ECDSA
9082 kECDH - ECDH cert (signed with either RSA or ECDSA)
9083 kEECDH - ephemeral ECDH
9084 ECDH - ECDH cert or ephemeral ECDH
9090 AECDH - anonymous ECDH
9091 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9095 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9096 Use correct micalg parameters depending on digest(s) in signed message.
9100 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9101 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9105 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9106 an engine to register a method. Add ENGINE lookups for methods and
9107 functional reference processing.
9111 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9112 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9117 * New -resign option to smime utility. This adds one or more signers
9118 to an existing PKCS#7 signedData structure. Also -md option to use an
9119 alternative message digest algorithm for signing.
9123 * Tidy up PKCS#7 routines and add new functions to make it easier to
9124 create PKCS7 structures containing multiple signers. Update smime
9125 application to support multiple signers.
9129 * New -macalg option to pkcs12 utility to allow setting of an alternative
9134 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9135 Reorganize PBE internals to lookup from a static table using NIDs,
9136 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9137 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9138 PRF which will be automatically used with PBES2.
9142 * Replace the algorithm specific calls to generate keys in "req" with the
9147 * Update PKCS#7 enveloped data routines to use new API. This is now
9148 supported by any public key method supporting the encrypt operation. A
9149 ctrl is added to allow the public key algorithm to examine or modify
9150 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9155 * Add a ctrl to asn1 method to allow a public key algorithm to express
9156 a default digest type to use. In most cases this will be SHA1 but some
9157 algorithms (such as GOST) need to specify an alternative digest. The
9158 return value indicates how strong the preference is 1 means optional and
9159 2 is mandatory (that is it is the only supported type). Modify
9160 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9161 use the default md. Update openssl utilities to use the default digest
9162 type for signing if it is not explicitly indicated.
9166 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9167 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9168 signing method from the key type. This effectively removes the link
9169 between digests and public key types.
9173 * Add an OID cross reference table and utility functions. Its purpose is to
9174 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9175 rsaEncryption. This will allow some of the algorithm specific hackery
9176 needed to use the correct OID to be removed.
9180 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9181 structures for PKCS7_sign(). They are now set up by the relevant public
9186 * Add provisional EC pkey method with support for ECDSA and ECDH.
9190 * Add support for key derivation (agreement) in the API, DH method and
9195 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9196 public and private key formats. As a side effect these add additional
9197 command line functionality not previously available: DSA signatures can be
9198 generated and verified using pkeyutl and DH key support and generation in
9205 *Oliver Tappe <zooey@hirschkaefer.de>*
9207 * New make target "install_html_docs" installs HTML renditions of the
9210 *Oliver Tappe <zooey@hirschkaefer.de>*
9212 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9213 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9214 support key and parameter generation and add initial key generation
9215 functionality for RSA.
9219 * Add functions for main EVP_PKEY_method operations. The undocumented
9220 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9221 `EVP_PKEY_{encrypt,decrypt}_old`.
9225 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9226 key API, doesn't do much yet.
9230 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9231 public key algorithms. New option to openssl utility:
9232 "list-public-key-algorithms" to print out info.
9236 * Implement the Supported Elliptic Curves Extension for
9237 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9241 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9242 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9246 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9247 utilities such as rsa, dsa, dsaparam etc except they process any key
9252 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9253 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9254 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9259 * Initial support for pluggable public key ASN1.
9260 De-spaghettify the public key ASN1 handling. Move public and private
9261 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9262 algorithm specific handling to a single module within the relevant
9263 algorithm directory. Add functions to allow (near) opaque processing
9264 of public and private key structures.
9268 * Implement the Supported Point Formats Extension for
9269 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9273 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9274 for the psk identity [hint] and the psk callback functions to the
9275 SSL_SESSION, SSL and SSL_CTX structure.
9278 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9282 SSL_CTX_use_psk_identity_hint
9283 SSL_get_psk_identity_hint
9284 SSL_get_psk_identity
9285 SSL_use_psk_identity_hint
9287 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9289 * Add RFC 3161 compliant time stamp request creation, response generation
9290 and response verification functionality.
9292 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9294 * Add initial support for TLS extensions, specifically for the server_name
9295 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9296 have new members for a host name. The SSL data structure has an
9297 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9298 stored in that context to allow for session resumption, even after the
9299 SSL has been switched to a new SSL_CTX in reaction to a client's
9300 server_name extension.
9302 New functions (subject to change):
9304 SSL_get_servername()
9305 SSL_get_servername_type()
9308 New CTRL codes and macros (subject to change):
9310 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9311 - SSL_CTX_set_tlsext_servername_callback()
9312 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9313 - SSL_CTX_set_tlsext_servername_arg()
9314 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9316 openssl s_client has a new '-servername ...' option.
9318 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9319 '-key2 ...', '-servername_fatal' (subject to change). This allows
9320 testing the HostName extension for a specific single host name ('-cert'
9321 and '-key' remain fallbacks for handshakes without HostName
9322 negotiation). If the unrecognized_name alert has to be sent, this by
9323 default is a warning; it becomes fatal with the '-servername_fatal'
9326 *Peter Sylvester, Remy Allais, Christophe Renou*
9328 * Whirlpool hash implementation is added.
9332 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9333 bn(64,32). Because of instruction set limitations it doesn't have
9334 any negative impact on performance. This was done mostly in order
9335 to make it possible to share assembler modules, such as bn_mul_mont
9336 implementations, between 32- and 64-bit builds without hassle.
9340 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9341 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9346 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9347 dedicated Montgomery multiplication procedure, is introduced.
9348 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9349 "64-bit" performance on certain 32-bit targets.
9353 * New option SSL_OP_NO_COMP to disable use of compression selectively
9354 in SSL structures. New SSL ctrl to set maximum send fragment size.
9355 Save memory by setting the I/O buffer sizes dynamically instead of
9356 using the maximum available value.
9360 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9361 in addition to the text details.
9365 * Very, very preliminary EXPERIMENTAL support for printing of general
9366 ASN1 structures. This currently produces rather ugly output and doesn't
9367 handle several customised structures at all.
9371 * Integrated support for PVK file format and some related formats such
9372 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9373 these in the 'rsa' and 'dsa' utilities.
9377 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9381 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9382 place for the (very old) "NETSCAPE" format certificates which are now
9383 handled using new ASN1 code equivalents.
9387 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9388 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9389 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9393 * Modify CRL distribution points extension code to print out previously
9394 unsupported fields. Enhance extension setting code to allow setting of
9399 * Add print and set support for Issuing Distribution Point CRL extension.
9403 * Change 'Configure' script to enable Camellia by default.
9410 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9412 * When rejecting SSL/TLS records due to an incorrect version number, never
9413 update s->server with a new major version number. As of
9414 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9415 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9416 the previous behavior could result in a read attempt at NULL when
9417 receiving specific incorrect SSL/TLS records once record payload
9418 protection is active. ([CVE-2010-0740])
9420 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9422 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9423 could be crashed if the relevant tables were not present (e.g. chrooted).
9425 *Tomas Hoger <thoger@redhat.com>*
9427 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9429 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9431 *Martin Olsson, Neel Mehta*
9433 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9434 accommodate for stack sorting, always a write lock!).
9438 * On some versions of WIN32 Heap32Next is very slow. This can cause
9439 excessive delays in the RAND_poll(): over a minute. As a workaround
9440 include a time check in the inner Heap32Next loop too.
9444 * The code that handled flushing of data in SSL/TLS originally used the
9445 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9446 the problem outlined in PR#1949. The fix suggested there however can
9447 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9448 of Apache). So instead simplify the code to flush unconditionally.
9449 This should be fine since flushing with no data to flush is a no op.
9453 * Handle TLS versions 2.0 and later properly and correctly use the
9454 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9455 off ancient servers have a habit of sticking around for a while...
9459 * Modify compression code so it frees up structures without using the
9460 ex_data callbacks. This works around a problem where some applications
9461 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9462 restarting) then use compression (e.g. SSL with compression) later.
9463 This results in significant per-connection memory leaks and
9464 has caused some security issues including CVE-2008-1678 and
9469 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9470 change when encrypting or decrypting.
9474 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9475 connect and renegotiate with servers which do not support RI.
9476 Until RI is more widely deployed this option is enabled by default.
9480 * Add "missing" ssl ctrls to clear options and mode.
9484 * If client attempts to renegotiate and doesn't support RI respond with
9485 a no_renegotiation alert as required by RFC5746. Some renegotiating
9486 TLS clients will continue a connection gracefully when they receive
9487 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9488 waiting for a server hello which it will never receive. Now we treat a
9489 received no_renegotiation alert as a fatal error. This is because
9490 applications requesting a renegotiation might well expect it to succeed
9491 and would have no code in place to handle the server denying it so the
9492 only safe thing to do is to terminate the connection.
9496 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9497 peer supports secure renegotiation and 0 otherwise. Print out peer
9498 renegotiation support in s_client/s_server.
9502 * Replace the highly broken and deprecated SPKAC certification method with
9503 the updated NID creation version. This should correctly handle UTF8.
9507 * Implement RFC5746. Re-enable renegotiation but require the extension
9508 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9509 turns out to be a bad idea. It has been replaced by
9510 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9511 SSL_CTX_set_options(). This is really not recommended unless you
9512 know what you are doing.
9514 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9516 * Fixes to stateless session resumption handling. Use initial_ctx when
9517 issuing and attempting to decrypt tickets in case it has changed during
9518 servername handling. Use a non-zero length session ID when attempting
9519 stateless session resumption: this makes it possible to determine if
9520 a resumption has occurred immediately after receiving server hello
9521 (several places in OpenSSL subtly assume this) instead of later in
9526 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9527 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9528 fixes for a few places where the return code is not checked
9531 *Julia Lawall <julia@diku.dk>*
9533 * Add --strict-warnings option to Configure script to include devteam
9534 warnings in other configurations.
9538 * Add support for --libdir option and LIBDIR variable in makefiles. This
9539 makes it possible to install openssl libraries in locations which
9540 have names other than "lib", for example "/usr/lib64" which some
9543 *Steve Henson, based on patch from Jeremy Utley*
9545 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9546 X690 8.9.12 and can produce some misleading textual output of OIDs.
9548 *Steve Henson, reported by Dan Kaminsky*
9550 * Delete MD2 from algorithm tables. This follows the recommendation in
9551 several standards that it is not used in new applications due to
9552 several cryptographic weaknesses. For binary compatibility reasons
9553 the MD2 API is still compiled in by default.
9557 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9562 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9563 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9566 *Guenter <lists@gknw.net>*
9568 * Fix the server certificate chain building code to use X509_verify_cert(),
9569 it used to have an ad-hoc builder which was unable to cope with anything
9570 other than a simple chain.
9572 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9574 * Don't check self signed certificate signatures in X509_verify_cert()
9575 by default (a flag can override this): it just wastes time without
9576 adding any security. As a useful side effect self signed root CAs
9577 with non-FIPS digests are now usable in FIPS mode.
9581 * In dtls1_process_out_of_seq_message() the check if the current message
9582 is already buffered was missing. For every new message was memory
9583 allocated, allowing an attacker to perform an denial of service attack
9584 with sending out of seq handshake messages until there is no memory
9585 left. Additionally every future message was buffered, even if the
9586 sequence number made no sense and would be part of another handshake.
9587 So only messages with sequence numbers less than 10 in advance will be
9588 buffered. ([CVE-2009-1378])
9590 *Robin Seggelmann, discovered by Daniel Mentz*
9592 * Records are buffered if they arrive with a future epoch to be
9593 processed after finishing the corresponding handshake. There is
9594 currently no limitation to this buffer allowing an attacker to perform
9595 a DOS attack with sending records with future epochs until there is no
9596 memory left. This patch adds the pqueue_size() function to determine
9597 the size of a buffer and limits the record buffer to 100 entries.
9600 *Robin Seggelmann, discovered by Daniel Mentz*
9602 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9603 parent structure is freed. ([CVE-2009-1379])
9607 * Handle non-blocking I/O properly in SSL_shutdown() call.
9609 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9611 * Add `2.5.4.*` OIDs
9613 *Ilya O. <vrghost@gmail.com>*
9615 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9617 * Disable renegotiation completely - this fixes a severe security
9618 problem ([CVE-2009-3555]) at the cost of breaking all
9619 renegotiation. Renegotiation can be re-enabled by setting
9620 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9621 run-time. This is really not recommended unless you know what
9626 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9628 * Don't set val to NULL when freeing up structures, it is freed up by
9629 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9630 zeroing past the valid field. ([CVE-2009-0789])
9632 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9634 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9635 checked correctly. This would allow some invalid signed attributes to
9636 appear to verify correctly. ([CVE-2009-0591])
9638 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9640 * Reject UniversalString and BMPString types with invalid lengths. This
9641 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9642 a legal length. ([CVE-2009-0590])
9646 * Set S/MIME signing as the default purpose rather than setting it
9647 unconditionally. This allows applications to override it at the store
9652 * Permit restricted recursion of ASN1 strings. This is needed in practice
9653 to handle some structures.
9657 * Improve efficiency of mem_gets: don't search whole buffer each time
9660 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9662 * New -hex option for openssl rand.
9666 * Print out UTF8String and NumericString when parsing ASN1.
9670 * Support NumericString type for name components.
9674 * Allow CC in the environment to override the automatically chosen
9675 compiler. Note that nothing is done to ensure flags work with the
9680 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9682 * Properly check EVP_VerifyFinal() and similar return values
9685 *Ben Laurie, Bodo Moeller, Google Security Team*
9687 * Enable TLS extensions by default.
9691 * Allow the CHIL engine to be loaded, whether the application is
9692 multithreaded or not. (This does not release the developer from the
9693 obligation to set up the dynamic locking callbacks.)
9695 *Sander Temme <sander@temme.net>*
9697 * Use correct exit code if there is an error in dgst command.
9699 *Steve Henson; problem pointed out by Roland Dirlewanger*
9701 * Tweak Configure so that you need to say "experimental-jpake" to enable
9702 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9706 * Add experimental JPAKE support, including demo authentication in
9707 s_client and s_server.
9711 * Set the comparison function in v3_addr_canonize().
9713 *Rob Austein <sra@hactrn.net>*
9715 * Add support for XMPP STARTTLS in s_client.
9717 *Philip Paeps <philip@freebsd.org>*
9719 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9720 to ensure that even with this option, only ciphersuites in the
9721 server's preference list will be accepted. (Note that the option
9722 applies only when resuming a session, so the earlier behavior was
9723 just about the algorithm choice for symmetric cryptography.)
9727 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9729 * Fix NULL pointer dereference if a DTLS server received
9730 ChangeCipherSpec as first record ([CVE-2009-1386]).
9734 * Fix a state transition in s3_srvr.c and d1_srvr.c
9735 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9739 * The fix in 0.9.8c that supposedly got rid of unsafe
9740 double-checked locking was incomplete for RSA blinding,
9741 addressing just one layer of what turns out to have been
9742 doubly unsafe triple-checked locking.
9744 So now fix this for real by retiring the MONT_HELPER macro
9745 in crypto/rsa/rsa_eay.c.
9747 *Bodo Moeller; problem pointed out by Marius Schilder*
9749 * Various precautionary measures:
9751 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9753 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9754 (NB: This would require knowledge of the secret session ticket key
9755 to exploit, in which case you'd be SOL either way.)
9757 - Change bn_nist.c so that it will properly handle input BIGNUMs
9758 outside the expected range.
9760 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9763 *Neel Mehta, Bodo Moeller*
9765 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9766 the load fails. Useful for distros.
9768 *Ben Laurie and the FreeBSD team*
9770 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9774 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9778 * Expand ENGINE to support engine supplied SSL client certificate functions.
9780 This work was sponsored by Logica.
9784 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9785 keystores. Support for SSL/TLS client authentication too.
9786 Not compiled unless enable-capieng specified to Configure.
9788 This work was sponsored by Logica.
9792 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9793 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9794 attribute creation routines such as certificate requests and PKCS#12
9799 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9801 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9802 handshake which could lead to a client crash as found using the
9803 Codenomicon TLS test suite ([CVE-2008-1672])
9805 *Steve Henson, Mark Cox*
9807 * Fix double free in TLS server name extensions which could lead to
9808 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9812 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9814 Clear the error queue to ensure that error entries left from
9815 older function calls do not interfere with the correct operation.
9817 *Lutz Jaenicke, Erik de Castro Lopo*
9819 * Remove root CA certificates of commercial CAs:
9821 The OpenSSL project does not recommend any specific CA and does not
9822 have any policy with respect to including or excluding any CA.
9823 Therefore it does not make any sense to ship an arbitrary selection
9824 of root CA certificates with the OpenSSL software.
9828 * RSA OAEP patches to fix two separate invalid memory reads.
9829 The first one involves inputs when 'lzero' is greater than
9830 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9831 before the beginning of from). The second one involves inputs where
9832 the 'db' section contains nothing but zeroes (there is a one-byte
9833 invalid read after the end of 'db').
9835 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9837 * Partial backport from 0.9.9-dev:
9839 Introduce bn_mul_mont (dedicated Montgomery multiplication
9840 procedure) as a candidate for BIGNUM assembler implementation.
9841 While 0.9.9-dev uses assembler for various architectures, only
9842 x86_64 is available by default here in the 0.9.8 branch, and
9843 32-bit x86 is available through a compile-time setting.
9845 To try the 32-bit x86 assembler implementation, use Configure
9846 option "enable-montasm" (which exists only for this backport).
9848 As "enable-montasm" for 32-bit x86 disclaims code stability
9849 anyway, in this constellation we activate additional code
9850 backported from 0.9.9-dev for further performance improvements,
9851 namely BN_from_montgomery_word. (To enable this otherwise,
9852 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9854 *Andy Polyakov (backport partially by Bodo Moeller)*
9856 * Add TLS session ticket callback. This allows an application to set
9857 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9858 values. This is useful for key rollover for example where several key
9859 sets may exist with different names.
9863 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9864 This was broken until now in 0.9.8 releases, such that the only way
9865 a registered ENGINE could be used (assuming it initialises
9866 successfully on the host) was to explicitly set it as the default
9867 for the relevant algorithms. This is in contradiction with 0.9.7
9868 behaviour and the documentation. With this fix, when an ENGINE is
9869 registered into a given algorithm's table of implementations, the
9870 'uptodate' flag is reset so that auto-discovery will be used next
9871 time a new context for that algorithm attempts to select an
9874 *Ian Lister (tweaked by Geoff Thorpe)*
9876 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9877 implementation in the following ways:
9879 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9882 Lack of BER streaming support means one pass streaming processing is
9883 only supported if data is detached: setting the streaming flag is
9884 ignored for embedded content.
9886 CMS support is disabled by default and must be explicitly enabled
9887 with the enable-cms configuration option.
9891 * Update the GMP engine glue to do direct copies between BIGNUM and
9892 mpz_t when openssl and GMP use the same limb size. Otherwise the
9893 existing "conversion via a text string export" trick is still used.
9895 *Paul Sheer <paulsheer@gmail.com>*
9897 * Zlib compression BIO. This is a filter BIO which compressed and
9898 uncompresses any data passed through it.
9902 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9903 RFC3394 compatible AES key wrapping.
9907 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9908 sets string data without copying. X509_ALGOR_set0() and
9909 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9910 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9911 from an X509_ATTRIBUTE structure optionally checking it occurs only
9912 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9917 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9918 to get the expected BN_FLG_CONSTTIME behavior.
9920 *Bodo Moeller (Google)*
9924 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9925 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9926 - added some more tests to do_tests.pl
9927 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9928 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9929 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9930 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9931 - various changes to netware.pl to enable gcc-cross builds on Win32
9933 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9934 - various changes to fix missing prototype warnings
9935 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9936 - added AES, WHIRLPOOL and CPUID assembler code to build files
9937 - added missing AES assembler make rules to mk1mf.pl
9938 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9940 *Guenter Knauf <eflash@gmx.net>*
9942 * Implement certificate status request TLS extension defined in RFC3546.
9943 A client can set the appropriate parameters and receive the encoded
9944 OCSP response via a callback. A server can query the supplied parameters
9945 and set the encoded OCSP response in the callback. Add simplified examples
9946 to s_client and s_server.
9950 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9953 + Binary incompatibility of ssl_ctx_st structure
9954 + DTLS interoperation with non-compliant servers
9955 + Don't call get_session_cb() without proposed session
9956 + Fix ia64 assembler code
9958 *Andy Polyakov, Steve Henson*
9960 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9962 * DTLS Handshake overhaul. There were longstanding issues with
9963 OpenSSL DTLS implementation, which were making it impossible for
9964 RFC 4347 compliant client to communicate with OpenSSL server.
9965 Unfortunately just fixing these incompatibilities would "cut off"
9966 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9967 server keeps tolerating non RFC compliant syntax. The opposite is
9968 not true, 0.9.8f client can not communicate with earlier server.
9969 This update even addresses CVE-2007-4995.
9973 * Changes to avoid need for function casts in OpenSSL: some compilers
9974 (gcc 4.2 and later) reject their use.
9975 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9978 * Add RFC4507 support to OpenSSL. This includes the corrections in
9979 RFC4507bis. The encrypted ticket format is an encrypted encoded
9980 SSL_SESSION structure, that way new session features are automatically
9983 If a client application caches session in an SSL_SESSION structure
9984 support is transparent because tickets are now stored in the encoded
9987 The SSL_CTX structure automatically generates keys for ticket
9988 protection in servers so again support should be possible
9989 with no application modification.
9991 If a client or server wishes to disable RFC4507 support then the option
9992 SSL_OP_NO_TICKET can be set.
9994 Add a TLS extension debugging callback to allow the contents of any client
9995 or server extensions to be examined.
9997 This work was sponsored by Google.
10001 * Add initial support for TLS extensions, specifically for the server_name
10002 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10003 have new members for a host name. The SSL data structure has an
10004 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10005 stored in that context to allow for session resumption, even after the
10006 SSL has been switched to a new SSL_CTX in reaction to a client's
10007 server_name extension.
10009 New functions (subject to change):
10011 SSL_get_servername()
10012 SSL_get_servername_type()
10015 New CTRL codes and macros (subject to change):
10017 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10018 - SSL_CTX_set_tlsext_servername_callback()
10019 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10020 - SSL_CTX_set_tlsext_servername_arg()
10021 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10023 openssl s_client has a new '-servername ...' option.
10025 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10026 '-key2 ...', '-servername_fatal' (subject to change). This allows
10027 testing the HostName extension for a specific single host name ('-cert'
10028 and '-key' remain fallbacks for handshakes without HostName
10029 negotiation). If the unrecognized_name alert has to be sent, this by
10030 default is a warning; it becomes fatal with the '-servername_fatal'
10033 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10035 * Add AES and SSE2 assembly language support to VC++ build.
10039 * Mitigate attack on final subtraction in Montgomery reduction.
10043 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10044 (which previously caused an internal error).
10048 * Squeeze another 10% out of IGE mode when in != out.
10052 * AES IGE mode speedup.
10054 *Dean Gaudet (Google)*
10056 * Add the Korean symmetric 128-bit cipher SEED (see
10057 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10058 add SEED ciphersuites from RFC 4162:
10060 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10061 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10062 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10063 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10065 To minimize changes between patchlevels in the OpenSSL 0.9.8
10066 series, SEED remains excluded from compilation unless OpenSSL
10067 is configured with 'enable-seed'.
10069 *KISA, Bodo Moeller*
10071 * Mitigate branch prediction attacks, which can be practical if a
10072 single processor is shared, allowing a spy process to extract
10073 information. For detailed background information, see
10074 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10075 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10076 and Necessary Software Countermeasures"). The core of the change
10077 are new versions BN_div_no_branch() and
10078 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10079 respectively, which are slower, but avoid the security-relevant
10080 conditional branches. These are automatically called by BN_div()
10081 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10082 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10083 remove a conditional branch.
10085 BN_FLG_CONSTTIME is the new name for the previous
10086 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10087 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10088 in the exponent causes BN_mod_exp_mont() to use the alternative
10089 implementation in BN_mod_exp_mont_consttime().) The old name
10090 remains as a deprecated alias.
10092 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10093 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10094 constant-time implementations for more than just exponentiation.
10095 Here too the old name is kept as a deprecated alias.
10097 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10098 the BN_BLINDING structure gets an independent copy of the
10099 modulus. This means that the previous `BIGNUM *m` argument to
10100 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10101 essentially becomes `const BIGNUM *m`, although we can't actually
10102 change this in the header file before 0.9.9. It allows
10103 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10104 enable BN_FLG_CONSTTIME.
10106 *Matthew D Wood (Intel Corp)*
10108 * In the SSL/TLS server implementation, be strict about session ID
10109 context matching (which matters if an application uses a single
10110 external cache for different purposes). Previously,
10111 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10112 set. This did ensure strict client verification, but meant that,
10113 with applications using a single external cache for quite
10114 different requirements, clients could circumvent ciphersuite
10115 restrictions for a given session ID context by starting a session
10116 in a different context.
10120 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10121 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10122 authentication-only ciphersuites.
10126 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10127 not complete and could lead to a possible single byte overflow
10128 ([CVE-2007-5135]) [Ben Laurie]
10130 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10132 * Since AES128 and AES256 (and similarly Camellia128 and
10133 Camellia256) share a single mask bit in the logic of
10134 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10135 kludge to work properly if AES128 is available and AES256 isn't
10136 (or if Camellia128 is available and Camellia256 isn't).
10140 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10141 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10142 When a point or a seed is encoded in a BIT STRING, we need to
10143 prevent the removal of trailing zero bits to get the proper DER
10144 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10145 of a NamedBitList, for which trailing 0 bits need to be removed.)
10149 * Have SSL/TLS server implementation tolerate "mismatched" record
10150 protocol version while receiving ClientHello even if the
10151 ClientHello is fragmented. (The server can't insist on the
10152 particular protocol version it has chosen before the ServerHello
10153 message has informed the client about his choice.)
10157 * Add RFC 3779 support.
10159 *Rob Austein for ARIN, Ben Laurie*
10161 * Load error codes if they are not already present instead of using a
10162 static variable. This allows them to be cleanly unloaded and reloaded.
10163 Improve header file function name parsing.
10167 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10168 or CAPABILITY handshake as required by RFCs.
10170 *Goetz Babin-Ebell*
10172 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10174 * Introduce limits to prevent malicious keys being able to
10175 cause a denial of service. ([CVE-2006-2940])
10177 *Steve Henson, Bodo Moeller*
10179 * Fix ASN.1 parsing of certain invalid structures that can result
10180 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10182 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10183 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10185 * Fix SSL client code which could crash if connecting to a
10186 malicious SSLv2 server. ([CVE-2006-4343])
10188 *Tavis Ormandy and Will Drewry, Google Security Team*
10190 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10191 match only those. Before that, "AES256-SHA" would be interpreted
10192 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10193 the same strength classification in 0.9.7h) as we currently only
10194 have a single AES bit in the ciphersuite description bitmap.
10195 That change, however, also applied to ciphersuite strings such as
10196 "RC4-MD5" that intentionally matched multiple ciphersuites --
10197 namely, SSL 2.0 ciphersuites in addition to the more common ones
10198 from SSL 3.0/TLS 1.0.
10200 So we change the selection algorithm again: Naming an explicit
10201 ciphersuite selects this one ciphersuite, and any other similar
10202 ciphersuite (same bitmap) from *other* protocol versions.
10203 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10204 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10206 Since SSL 2.0 does not have any ciphersuites for which the
10207 128/256 bit distinction would be relevant, this works for now.
10208 The proper fix will be to use different bits for AES128 and
10209 AES256, which would have avoided the problems from the beginning;
10210 however, bits are scarce, so we can only do this in a new release
10211 (not just a patchlevel) when we can change the SSL_CIPHER
10212 definition to split the single 'unsigned long mask' bitmap into
10213 multiple values to extend the available space.
10217 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10219 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10220 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10222 * Add AES IGE and biIGE modes.
10226 * Change the Unix randomness entropy gathering to use poll() when
10227 possible instead of select(), since the latter has some
10228 undesirable limitations.
10230 *Darryl Miles via Richard Levitte and Bodo Moeller*
10232 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10233 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10234 cannot be implicitly activated as part of, e.g., the "AES" alias.
10235 However, please upgrade to OpenSSL 0.9.9[-dev] for
10236 non-experimental use of the ECC ciphersuites to get TLS extension
10237 support, which is required for curve and point format negotiation
10238 to avoid potential handshake problems.
10242 * Disable rogue ciphersuites:
10244 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10245 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10246 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10248 The latter two were purportedly from
10249 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10252 Also deactivate the remaining ciphersuites from
10253 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10254 unofficial, and the ID has long expired.
10258 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10259 dual-core machines) and other potential thread-safety issues.
10263 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10264 versions), which is now available for royalty-free use
10265 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10266 Also, add Camellia TLS ciphersuites from RFC 4132.
10268 To minimize changes between patchlevels in the OpenSSL 0.9.8
10269 series, Camellia remains excluded from compilation unless OpenSSL
10270 is configured with 'enable-camellia'.
10274 * Disable the padding bug check when compression is in use. The padding
10275 bug check assumes the first packet is of even length, this is not
10276 necessarily true if compression is enabled and can result in false
10277 positives causing handshake failure. The actual bug test is ancient
10278 code so it is hoped that implementations will either have fixed it by
10279 now or any which still have the bug do not support compression.
10283 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10285 * When applying a cipher rule check to see if string match is an explicit
10286 cipher suite and only match that one cipher suite if it is.
10290 * Link in manifests for VC++ if needed.
10292 *Austin Ziegler <halostatue@gmail.com>*
10294 * Update support for ECC-based TLS ciphersuites according to
10295 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10296 TLS extensions, which are supported starting with the 0.9.9
10297 branch, not in the OpenSSL 0.9.8 branch).
10301 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10302 opaque EVP_CIPHER_CTX handling.
10306 * Fixes and enhancements to zlib compression code. We now only use
10307 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10308 to conform with the standards mentioned here:
10309 <http://www.zlib.net/DLL_FAQ.txt>
10310 Static zlib linking now works on Windows and the new --with-zlib-include
10311 --with-zlib-lib options to Configure can be used to supply the location
10312 of the headers and library. Gracefully handle case where zlib library
10317 * Several fixes and enhancements to the OID generation code. The old code
10318 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10319 handle numbers larger than ULONG_MAX, truncated printing and had a
10320 non standard OBJ_obj2txt() behaviour.
10324 * Add support for building of engines under engine/ as shared libraries
10325 under VC++ build system.
10329 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10330 Hopefully, we will not see any false combination of paths any more.
10334 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10336 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10337 (part of SSL_OP_ALL). This option used to disable the
10338 countermeasure against man-in-the-middle protocol-version
10339 rollback in the SSL 2.0 server implementation, which is a bad
10340 idea. ([CVE-2005-2969])
10342 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10343 for Information Security, National Institute of Advanced Industrial
10344 Science and Technology [AIST], Japan)*
10346 * Add two function to clear and return the verify parameter flags.
10350 * Keep cipherlists sorted in the source instead of sorting them at
10351 runtime, thus removing the need for a lock.
10355 * Avoid some small subgroup attacks in Diffie-Hellman.
10357 *Nick Mathewson and Ben Laurie*
10359 * Add functions for well-known primes.
10363 * Extended Windows CE support.
10365 *Satoshi Nakamura and Andy Polyakov*
10367 * Initialize SSL_METHOD structures at compile time instead of during
10368 runtime, thus removing the need for a lock.
10372 * Make PKCS7_decrypt() work even if no certificate is supplied by
10373 attempting to decrypt each encrypted key in turn. Add support to
10378 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10380 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10383 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10387 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10388 key into the same file any more.
10392 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10396 * Add -utf8 command line and config file option to 'ca'.
10398 *Stefan <stf@udoma.org*
10400 * Removed the macro des_crypt(), as it seems to conflict with some
10401 libraries. Use DES_crypt().
10405 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10406 involves renaming the source and generated shared-libs for
10407 both. The engines will accept the corrected or legacy ids
10408 ('ncipher' and '4758_cca' respectively) when binding. NB,
10409 this only applies when building 'shared'.
10411 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10413 * Add attribute functions to EVP_PKEY structure. Modify
10414 PKCS12_create() to recognize a CSP name attribute and
10415 use it. Make -CSP option work again in pkcs12 utility.
10419 * Add new functionality to the bn blinding code:
10420 - automatic re-creation of the BN_BLINDING parameters after
10421 a fixed number of uses (currently 32)
10422 - add new function for parameter creation
10423 - introduce flags to control the update behaviour of the
10424 BN_BLINDING parameters
10425 - hide BN_BLINDING structure
10426 Add a second BN_BLINDING slot to the RSA structure to improve
10427 performance when a single RSA object is shared among several
10432 * Add support for DTLS.
10434 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10436 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10437 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10441 * Remove buggy and incomplete DH cert support from
10442 ssl/ssl_rsa.c and ssl/s3_both.c
10446 * Use SHA-1 instead of MD5 as the default digest algorithm for
10447 the `apps/openssl` commands.
10451 * Compile clean with "-Wall -Wmissing-prototypes
10452 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10453 DEBUG_SAFESTACK must also be set.
10457 * Change ./Configure so that certain algorithms can be disabled by default.
10458 The new counterpiece to "no-xxx" is "enable-xxx".
10460 The patented RC5 and MDC2 algorithms will now be disabled unless
10461 "enable-rc5" and "enable-mdc2", respectively, are specified.
10463 (IDEA remains enabled despite being patented. This is because IDEA
10464 is frequently required for interoperability, and there is no license
10465 fee for non-commercial use. As before, "no-idea" can be used to
10466 avoid this algorithm.)
10470 * Add processing of proxy certificates (see RFC 3820). This work was
10471 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10472 EGEE (Enabling Grids for E-science in Europe).
10476 * RC4 performance overhaul on modern architectures/implementations, such
10477 as Intel P4, IA-64 and AMD64.
10481 * New utility extract-section.pl. This can be used specify an alternative
10482 section number in a pod file instead of having to treat each file as
10483 a separate case in Makefile. This can be done by adding two lines to the
10486 =for comment openssl_section:XXX
10488 The blank line is mandatory.
10492 * New arguments -certform, -keyform and -pass for s_client and s_server
10493 to allow alternative format key and certificate files and passphrase
10498 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10499 update associated structures and add various utility functions.
10501 Add new policy related verify parameters, include policy checking in
10502 standard verify code. Enhance 'smime' application with extra parameters
10503 to support policy checking and print out.
10507 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10508 Nehemiah processors. These extensions support AES encryption in hardware
10509 as well as RNG (though RNG support is currently disabled).
10511 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10513 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10517 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10519 *Andy Polyakov and a number of other people*
10521 * Improved PowerPC platform support. Most notably BIGNUM assembler
10522 implementation contributed by IBM.
10524 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10526 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10527 exponent rather than 'unsigned long'. There is a corresponding change to
10528 the new 'rsa_keygen' element of the RSA_METHOD structure.
10530 *Jelte Jansen, Geoff Thorpe*
10532 * Functionality for creating the initial serial number file is now
10533 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10535 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10536 number file to 1, which is bound to cause problems. To avoid
10537 the problems while respecting compatibility between different 0.9.7
10538 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10539 CA.pl for serial number initialization. With the new release 0.9.8,
10540 we can fix the problem directly in the 'ca' utility.)
10544 * Reduced header interdependencies by declaring more opaque objects in
10545 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10546 give fewer recursive includes, which could break lazy source code - so
10547 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10548 developers should define this symbol when building and using openssl to
10549 ensure they track the recommended behaviour, interfaces, [etc], but
10550 backwards-compatible behaviour prevails when this isn't defined.
10554 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10558 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10559 This will generate a random key of the appropriate length based on the
10560 cipher context. The EVP_CIPHER can provide its own random key generation
10561 routine to support keys of a specific form. This is used in the des and
10562 3des routines to generate a key of the correct parity. Update S/MIME
10563 code to use new functions and hence generate correct parity DES keys.
10564 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10565 valid (weak or incorrect parity).
10569 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10570 as looking them up. This is useful when the verified structure may contain
10571 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10572 present unless the new PKCS7_NO_CRL flag is asserted.
10576 * Extend ASN1 oid configuration module. It now additionally accepts the
10579 shortName = some long name, 1.2.3.4
10583 * Reimplemented the BN_CTX implementation. There is now no more static
10584 limitation on the number of variables it can handle nor the depth of the
10585 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10586 information can now expand as required, and rather than having a single
10587 static array of bignums, BN_CTX now uses a linked-list of such arrays
10588 allowing it to expand on demand whilst maintaining the usefulness of
10589 BN_CTX's "bundling".
10593 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10594 to allow all RSA operations to function using a single BN_CTX.
10598 * Preliminary support for certificate policy evaluation and checking. This
10599 is initially intended to pass the tests outlined in "Conformance Testing
10600 of Relying Party Client Certificate Path Processing Logic" v1.07.
10604 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10605 remained unused and not that useful. A variety of other little bignum
10606 tweaks and fixes have also been made continuing on from the audit (see
10611 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10612 associated ASN1, EVP and SSL functions and old ASN1 macros.
10616 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10617 and this should never fail. So the return value from the use of
10618 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10619 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10623 * BN_CTX_get() should return zero-valued bignums, providing the same
10624 initialised value as BN_new().
10626 *Geoff Thorpe, suggested by Ulf Möller*
10628 * Support for inhibitAnyPolicy certificate extension.
10632 * An audit of the BIGNUM code is underway, for which debugging code is
10633 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10634 is considered valid when processing BIGNUMs, and causes execution to
10635 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10636 further steps are taken to deliberately pollute unused data in BIGNUM
10637 structures to try and expose faulty code further on. For now, openssl will
10638 (in its default mode of operation) continue to tolerate the inconsistent
10639 forms that it has tolerated in the past, but authors and packagers should
10640 consider trying openssl and their own applications when compiled with
10641 these debugging symbols defined. It will help highlight potential bugs in
10642 their own code, and will improve the test coverage for OpenSSL itself. At
10643 some point, these tighter rules will become openssl's default to improve
10644 maintainability, though the assert()s and other overheads will remain only
10645 in debugging configurations. See bn.h for more details.
10647 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10649 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10650 that can only be obtained through BN_CTX_new() (which implicitly
10651 initialises it). The presence of this function only made it possible
10652 to overwrite an existing structure (and cause memory leaks).
10656 * Because of the callback-based approach for implementing LHASH as a
10657 template type, lh_insert() adds opaque objects to hash-tables and
10658 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10659 to clean up those corresponding objects before destroying the hash table
10660 (and losing the object pointers). So some over-zealous constifications in
10661 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10662 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10663 prototyped to have "const" restrictions on the object pointers they are
10664 given (and so aren't required to cast them away any more).
10668 * The tmdiff.h API was so ugly and minimal that our own timing utility
10669 (speed) prefers to use its own implementation. The two implementations
10670 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10671 its object type properly exposed (MS_TM) instead of casting to/from
10672 `char *`. This may still change yet if someone realises MS_TM and
10674 aren't necessarily the greatest nomenclatures - but this is what was used
10675 internally to the implementation so I've used that for now.
10679 * Ensure that deprecated functions do not get compiled when
10680 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10681 the self-tests were still using deprecated key-generation functions so
10682 these have been updated also.
10686 * Reorganise PKCS#7 code to separate the digest location functionality
10687 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10688 New function PKCS7_set_digest() to set the digest type for PKCS#7
10689 digestedData type. Add additional code to correctly generate the
10690 digestedData type and add support for this type in PKCS7 initialization
10695 * New function PKCS7_set0_type_other() this initializes a PKCS7
10696 structure of type "other".
10700 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10701 sure the loop does correctly stop and breaking ("division by zero")
10702 modulus operations are not performed. The (pre-generated) prime
10703 table crypto/bn/bn_prime.h was already correct, but it could not be
10704 re-generated on some platforms because of the "division by zero"
10705 situation in the script.
10707 *Ralf S. Engelschall*
10709 * Update support for ECC-based TLS ciphersuites according to
10710 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10711 SHA-1 now is only used for "small" curves (where the
10712 representation of a field element takes up to 24 bytes); for
10713 larger curves, the field element resulting from ECDH is directly
10714 used as premaster secret.
10716 *Douglas Stebila (Sun Microsystems Laboratories)*
10718 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10719 curve secp160r1 to the tests.
10721 *Douglas Stebila (Sun Microsystems Laboratories)*
10723 * Add the possibility to load symbols globally with DSO.
10725 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10727 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10728 control of the error stack.
10732 * Add support for STORE in ENGINE.
10736 * Add the STORE type. The intention is to provide a common interface
10737 to certificate and key stores, be they simple file-based stores, or
10738 HSM-type store, or LDAP stores, or...
10739 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10743 * Add a generic structure called OPENSSL_ITEM. This can be used to
10744 pass a list of arguments to any function as well as provide a way
10745 for a function to pass data back to the caller.
10749 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10750 works like BUF_strdup() but can be used to duplicate a portion of
10751 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10756 * Add the function sk_find_ex() which works like sk_find(), but will
10757 return an index to an element even if an exact match couldn't be
10758 found. The index is guaranteed to point at the element where the
10759 searched-for key would be inserted to preserve sorting order.
10763 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10764 takes an extra flags argument for optional functionality. Currently,
10765 the following flags are defined:
10767 OBJ_BSEARCH_VALUE_ON_NOMATCH
10768 This one gets OBJ_bsearch_ex() to return a pointer to the first
10769 element where the comparing function returns a negative or zero
10772 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10773 This one gets OBJ_bsearch_ex() to return a pointer to the first
10774 element where the comparing function returns zero. This is useful
10775 if there are more than one element where the comparing function
10780 * Make it possible to create self-signed certificates with 'openssl ca'
10781 in such a way that the self-signed certificate becomes part of the
10782 CA database and uses the same mechanisms for serial number generation
10783 as all other certificate signing. The new flag '-selfsign' enables
10784 this functionality. Adapt CA.sh and CA.pl.in.
10788 * Add functionality to check the public key of a certificate request
10789 against a given private. This is useful to check that a certificate
10790 request can be signed by that key (self-signing).
10794 * Make it possible to have multiple active certificates with the same
10795 subject in the CA index file. This is done only if the keyword
10796 'unique_subject' is set to 'no' in the main CA section (default
10797 if 'CA_default') of the configuration file. The value is saved
10798 with the database itself in a separate index attribute file,
10799 named like the index file with '.attr' appended to the name.
10803 * Generate multi-valued AVAs using '+' notation in config files for
10808 * Support for nameConstraints certificate extension.
10812 * Support for policyConstraints certificate extension.
10816 * Support for policyMappings certificate extension.
10820 * Make sure the default DSA_METHOD implementation only uses its
10821 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10822 and change its own handlers to be NULL so as to remove unnecessary
10823 indirection. This lets alternative implementations fallback to the
10824 default implementation more easily.
10828 * Support for directoryName in GeneralName related extensions
10833 * Make it possible to link applications using Makefile.shared.
10834 Make that possible even when linking against static libraries!
10838 * Support for single pass processing for S/MIME signing. This now
10839 means that S/MIME signing can be done from a pipe, in addition
10840 cleartext signing (multipart/signed type) is effectively streaming
10841 and the signed data does not need to be all held in memory.
10843 This is done with a new flag PKCS7_STREAM. When this flag is set
10844 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10845 is done after the data is output (and digests calculated) in
10846 SMIME_write_PKCS7().
10850 * Add full support for -rpath/-R, both in shared libraries and
10851 applications, at least on the platforms where it's known how
10856 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10857 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10858 will now compute a table of multiples of the generator that
10859 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10860 faster (notably in the case of a single point multiplication,
10861 scalar * generator).
10863 *Nils Larsch, Bodo Moeller*
10865 * IPv6 support for certificate extensions. The various extensions
10866 which use the IP:a.b.c.d can now take IPv6 addresses using the
10867 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10872 * Added an ENGINE that implements RSA by performing private key
10873 exponentiations with the GMP library. The conversions to and from
10874 GMP's mpz_t format aren't optimised nor are any montgomery forms
10875 cached, and on x86 it appears OpenSSL's own performance has caught up.
10876 However there are likely to be other architectures where GMP could
10877 provide a boost. This ENGINE is not built in by default, but it can be
10878 specified at Configure time and should be accompanied by the necessary
10879 linker additions, eg;
10880 ./config -DOPENSSL_USE_GMP -lgmp
10884 * "openssl engine" will not display ENGINE/DSO load failure errors when
10885 testing availability of engines with "-t" - the old behaviour is
10886 produced by increasing the feature's verbosity with "-tt".
10890 * ECDSA routines: under certain error conditions uninitialized BN objects
10891 could be freed. Solution: make sure initialization is performed early
10892 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10897 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10898 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10899 software implementations. For DSA and DH, parameter generation can
10900 also be overridden by providing the appropriate method callbacks.
10904 * Change the "progress" mechanism used in key-generation and
10905 primality testing to functions that take a new BN_GENCB pointer in
10906 place of callback/argument pairs. The new API functions have `_ex`
10907 postfixes and the older functions are reimplemented as wrappers for
10908 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10909 declarations of the old functions to help (graceful) attempts to
10910 migrate to the new functions. Also, the new key-generation API
10911 functions operate on a caller-supplied key-structure and return
10912 success/failure rather than returning a key or NULL - this is to
10913 help make "keygen" another member function of RSA_METHOD etc.
10915 Example for using the new callback interface:
10917 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10918 void *my_arg = ...;
10921 BN_GENCB_set(&my_cb, my_callback, my_arg);
10923 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10924 /* For the meaning of a, b in calls to my_callback(), see the
10925 * documentation of the function that calls the callback.
10926 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10927 * my_callback should return 1 if it wants BN_is_prime_ex()
10928 * to continue, or 0 to stop.
10933 * Change the ZLIB compression method to be stateful, and make it
10934 available to TLS with the number defined in
10935 draft-ietf-tls-compression-04.txt.
10939 * Add the ASN.1 structures and functions for CertificatePair, which
10940 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10942 CertificatePair ::= SEQUENCE {
10943 forward [0] Certificate OPTIONAL,
10944 reverse [1] Certificate OPTIONAL,
10945 -- at least one of the pair shall be present -- }
10947 Also implement the PEM functions to read and write certificate
10948 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10950 This needed to be defined, mostly for the sake of the LDAP
10951 attribute crossCertificatePair, but may prove useful elsewhere as
10956 * Make it possible to inhibit symlinking of shared libraries in
10957 Makefile.shared, for Cygwin's sake.
10961 * Extend the BIGNUM API by creating a function
10962 void BN_set_negative(BIGNUM *a, int neg);
10963 and a macro that behave like
10964 int BN_is_negative(const BIGNUM *a);
10966 to avoid the need to access 'a->neg' directly in applications.
10970 * Implement fast modular reduction for pseudo-Mersenne primes
10971 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10972 EC_GROUP_new_curve_GFp() will now automatically use this
10975 *Nils Larsch <nla@trustcenter.de>*
10977 * Add new lock type (CRYPTO_LOCK_BN).
10981 * Change the ENGINE framework to automatically load engines
10982 dynamically from specific directories unless they could be
10983 found to already be built in or loaded. Move all the
10984 current engines except for the cryptodev one to a new
10985 directory engines/.
10986 The engines in engines/ are built as shared libraries if
10987 the "shared" options was given to ./Configure or ./config.
10988 Otherwise, they are inserted in libcrypto.a.
10989 /usr/local/ssl/engines is the default directory for dynamic
10990 engines, but that can be overridden at configure time through
10991 the usual use of --prefix and/or --openssldir, and at run
10992 time with the environment variable OPENSSL_ENGINES.
10994 *Geoff Thorpe and Richard Levitte*
10996 * Add Makefile.shared, a helper makefile to build shared
10997 libraries. Adapt Makefile.org.
11001 * Add version info to Win32 DLLs.
11003 *Peter 'Luna' Runestig" <peter@runestig.com>*
11005 * Add new 'medium level' PKCS#12 API. Certificates and keys
11006 can be added using this API to created arbitrary PKCS#12
11007 files while avoiding the low-level API.
11009 New options to PKCS12_create(), key or cert can be NULL and
11010 will then be omitted from the output file. The encryption
11011 algorithm NIDs can be set to -1 for no encryption, the mac
11012 iteration count can be set to 0 to omit the mac.
11014 Enhance pkcs12 utility by making the -nokeys and -nocerts
11015 options work when creating a PKCS#12 file. New option -nomac
11016 to omit the mac, NONE can be set for an encryption algorithm.
11017 New code is modified to use the enhanced PKCS12_create()
11018 instead of the low-level API.
11022 * Extend ASN1 encoder to support indefinite length constructed
11023 encoding. This can output sequences tags and octet strings in
11024 this form. Modify pk7_asn1.c to support indefinite length
11025 encoding. This is experimental and needs additional code to
11026 be useful, such as an ASN1 bio and some enhanced streaming
11029 Extend template encode functionality so that tagging is passed
11030 down to the template encoder.
11034 * Let 'openssl req' fail if an argument to '-newkey' is not
11035 recognized instead of using RSA as a default.
11039 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11040 As these are not official, they are not included in "ALL";
11041 the "ECCdraft" ciphersuite group alias can be used to select them.
11043 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11045 * Add ECDH engine support.
11047 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11049 * Add ECDH in new directory crypto/ecdh/.
11051 *Douglas Stebila (Sun Microsystems Laboratories)*
11053 * Let BN_rand_range() abort with an error after 100 iterations
11054 without success (which indicates a broken PRNG).
11058 * Change BN_mod_sqrt() so that it verifies that the input value
11059 is really the square of the return value. (Previously,
11060 BN_mod_sqrt would show GIGO behaviour.)
11064 * Add named elliptic curves over binary fields from X9.62, SECG,
11065 and WAP/WTLS; add OIDs that were still missing.
11067 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11069 * Extend the EC library for elliptic curves over binary fields
11070 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11073 EC_GF2m_simple_method
11077 EC_GROUP_new_curve_GF2m
11078 EC_GROUP_set_curve_GF2m
11079 EC_GROUP_get_curve_GF2m
11080 EC_POINT_set_affine_coordinates_GF2m
11081 EC_POINT_get_affine_coordinates_GF2m
11082 EC_POINT_set_compressed_coordinates_GF2m
11084 Point compression for binary fields is disabled by default for
11085 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11088 As binary polynomials are represented as BIGNUMs, various members
11089 of the EC_GROUP and EC_POINT data structures can be shared
11090 between the implementations for prime fields and binary fields;
11091 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11092 are essentially identical to their `..._GFp` counterparts.
11093 (For simplicity, the `..._GFp` prefix has been dropped from
11094 various internal method names.)
11096 An internal 'field_div' method (similar to 'field_mul' and
11097 'field_sqr') has been added; this is used only for binary fields.
11099 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11101 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11102 through methods ('mul', 'precompute_mult').
11104 The generic implementations (now internally called 'ec_wNAF_mul'
11105 and 'ec_wNAF_precomputed_mult') remain the default if these
11106 methods are undefined.
11108 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11110 * New function EC_GROUP_get_degree, which is defined through
11111 EC_METHOD. For curves over prime fields, this returns the bit
11112 length of the modulus.
11114 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11116 * New functions EC_GROUP_dup, EC_POINT_dup.
11117 (These simply call ..._new and ..._copy).
11119 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11121 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11122 Polynomials are represented as BIGNUMs (where the sign bit is not
11123 used) in the following functions [macros]:
11126 BN_GF2m_sub [= BN_GF2m_add]
11127 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11128 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11129 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11131 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11132 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11133 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11134 BN_GF2m_cmp [= BN_ucmp]
11136 (Note that only the 'mod' functions are actually for fields GF(2^m).
11137 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11139 For some functions, an the irreducible polynomial defining a
11140 field can be given as an 'unsigned int[]' with strictly
11141 decreasing elements giving the indices of those bits that are set;
11142 i.e., p[] represents the polynomial
11143 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11145 p[0] > p[1] > ... > p[k] = 0.
11146 This applies to the following functions:
11149 BN_GF2m_mod_mul_arr
11150 BN_GF2m_mod_sqr_arr
11151 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11152 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11153 BN_GF2m_mod_exp_arr
11154 BN_GF2m_mod_sqrt_arr
11155 BN_GF2m_mod_solve_quad_arr
11159 Conversion can be performed by the following functions:
11164 bntest.c has additional tests for binary polynomial arithmetic.
11166 Two implementations for BN_GF2m_mod_div() are available.
11167 The default algorithm simply uses BN_GF2m_mod_inv() and
11168 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11169 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11170 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11172 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11174 * Add new error code 'ERR_R_DISABLED' that can be used when some
11175 functionality is disabled at compile-time.
11177 *Douglas Stebila <douglas.stebila@sun.com>*
11179 * Change default behaviour of 'openssl asn1parse' so that more
11180 information is visible when viewing, e.g., a certificate:
11182 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11183 mode the content of non-printable OCTET STRINGs is output in a
11184 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11185 avoid the appearance of a printable string.
11187 *Nils Larsch <nla@trustcenter.de>*
11189 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11191 EC_GROUP_set_asn1_flag()
11192 EC_GROUP_get_asn1_flag()
11193 EC_GROUP_set_point_conversion_form()
11194 EC_GROUP_get_point_conversion_form()
11195 These control ASN1 encoding details:
11196 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11197 has been set to OPENSSL_EC_NAMED_CURVE.
11198 - Points are encoded in uncompressed form by default; options for
11199 asn1_for are as for point2oct, namely
11200 POINT_CONVERSION_COMPRESSED
11201 POINT_CONVERSION_UNCOMPRESSED
11202 POINT_CONVERSION_HYBRID
11204 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11206 EC_GROUP_set_seed()
11207 EC_GROUP_get0_seed()
11208 EC_GROUP_get_seed_len()
11209 This is used only for ASN1 purposes (so far).
11211 *Nils Larsch <nla@trustcenter.de>*
11213 * Add 'field_type' member to EC_METHOD, which holds the NID
11214 of the appropriate field type OID. The new function
11215 EC_METHOD_get_field_type() returns this value.
11217 *Nils Larsch <nla@trustcenter.de>*
11220 EC_POINT_point2bn()
11221 EC_POINT_bn2point()
11222 EC_POINT_point2hex()
11223 EC_POINT_hex2point()
11224 providing useful interfaces to EC_POINT_point2oct() and
11225 EC_POINT_oct2point().
11227 *Nils Larsch <nla@trustcenter.de>*
11229 * Change internals of the EC library so that the functions
11230 EC_GROUP_set_generator()
11231 EC_GROUP_get_generator()
11232 EC_GROUP_get_order()
11233 EC_GROUP_get_cofactor()
11234 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11235 to methods, which would lead to unnecessary code duplication when
11236 adding different types of curves.
11238 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11240 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11241 arithmetic, and such that modified wNAFs are generated
11242 (which avoid length expansion in many cases).
11246 * Add a function EC_GROUP_check_discriminant() (defined via
11247 EC_METHOD) that verifies that the curve discriminant is non-zero.
11249 Add a function EC_GROUP_check() that makes some sanity tests
11250 on a EC_GROUP, its generator and order. This includes
11251 EC_GROUP_check_discriminant().
11253 *Nils Larsch <nla@trustcenter.de>*
11255 * Add ECDSA in new directory crypto/ecdsa/.
11257 Add applications 'openssl ecparam' and 'openssl ecdsa'
11258 (these are based on 'openssl dsaparam' and 'openssl dsa').
11260 ECDSA support is also included in various other files across the
11261 library. Most notably,
11262 - 'openssl req' now has a '-newkey ecdsa:file' option;
11263 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11264 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11265 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11266 them suitable for ECDSA where domain parameters must be
11267 extracted before the specific public key;
11268 - ECDSA engine support has been added.
11270 *Nils Larsch <nla@trustcenter.de>*
11272 * Include some named elliptic curves, and add OIDs from X9.62,
11273 SECG, and WAP/WTLS. Each curve can be obtained from the new
11275 EC_GROUP_new_by_curve_name(),
11276 and the list of available named curves can be obtained with
11277 EC_get_builtin_curves().
11278 Also add a 'curve_name' member to EC_GROUP objects, which can be
11280 EC_GROUP_set_curve_name()
11281 EC_GROUP_get_curve_name()
11283 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11285 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11286 was actually never needed) and in BN_mul(). The removal in BN_mul()
11287 required a small change in bn_mul_part_recursive() and the addition
11288 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11289 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11290 bn_sub_words() and bn_add_words() except they take arrays with
11295 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11297 * Cleanse PEM buffers before freeing them since they may contain
11300 *Benjamin Bennett <ben@psc.edu>*
11302 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11303 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11304 authentication-only ciphersuites.
11308 * Since AES128 and AES256 share a single mask bit in the logic of
11309 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11310 kludge to work properly if AES128 is available and AES256 isn't.
11314 * Expand security boundary to match 1.1.1 module.
11318 * Remove redundant features: hash file source, editing of test vectors
11319 modify fipsld to use external fips_premain.c signature.
11323 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11324 run algorithm test programs.
11328 * Make algorithm test programs more tolerant of whitespace.
11332 * Have SSL/TLS server implementation tolerate "mismatched" record
11333 protocol version while receiving ClientHello even if the
11334 ClientHello is fragmented. (The server can't insist on the
11335 particular protocol version it has chosen before the ServerHello
11336 message has informed the client about his choice.)
11340 * Load error codes if they are not already present instead of using a
11341 static variable. This allows them to be cleanly unloaded and reloaded.
11345 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11347 * Introduce limits to prevent malicious keys being able to
11348 cause a denial of service. ([CVE-2006-2940])
11350 *Steve Henson, Bodo Moeller*
11352 * Fix ASN.1 parsing of certain invalid structures that can result
11353 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11355 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11356 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11358 * Fix SSL client code which could crash if connecting to a
11359 malicious SSLv2 server. ([CVE-2006-4343])
11361 *Tavis Ormandy and Will Drewry, Google Security Team*
11363 * Change ciphersuite string processing so that an explicit
11364 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11365 will no longer include "AES128-SHA"), and any other similar
11366 ciphersuite (same bitmap) from *other* protocol versions (so that
11367 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11368 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11369 changes from 0.9.8b and 0.9.8d.
11373 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11375 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11376 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11378 * Change the Unix randomness entropy gathering to use poll() when
11379 possible instead of select(), since the latter has some
11380 undesirable limitations.
11382 *Darryl Miles via Richard Levitte and Bodo Moeller*
11384 * Disable rogue ciphersuites:
11386 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11387 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11388 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11390 The latter two were purportedly from
11391 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11394 Also deactivate the remaining ciphersuites from
11395 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11396 unofficial, and the ID has long expired.
11400 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11401 dual-core machines) and other potential thread-safety issues.
11405 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11407 * Adapt fipsld and the build system to link against the validated FIPS
11408 module in FIPS mode.
11412 * Fixes for VC++ 2005 build under Windows.
11416 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11417 from a Windows bash shell such as MSYS. It is autodetected from the
11418 "config" script when run from a VC++ environment. Modify standard VC++
11419 build to use fipscanister.o from the GNU make build.
11423 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11425 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11426 The value now differs depending on if you build for FIPS or not.
11427 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11428 safely run with a non-FIPSed libcrypto, as it may crash because of
11429 the difference induced by this change.
11433 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11435 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11436 (part of SSL_OP_ALL). This option used to disable the
11437 countermeasure against man-in-the-middle protocol-version
11438 rollback in the SSL 2.0 server implementation, which is a bad
11439 idea. ([CVE-2005-2969])
11441 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11442 for Information Security, National Institute of Advanced Industrial
11443 Science and Technology [AIST, Japan)]*
11445 * Minimal support for X9.31 signatures and PSS padding modes. This is
11446 mainly for FIPS compliance and not fully integrated at this stage.
11450 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11451 the exponentiation using a fixed-length exponent. (Otherwise,
11452 the information leaked through timing could expose the secret key
11453 after many signatures; cf. Bleichenbacher's attack on DSA with
11458 * Make a new fixed-window mod_exp implementation the default for
11459 RSA, DSA, and DH private-key operations so that the sequence of
11460 squares and multiplies and the memory access pattern are
11461 independent of the particular secret key. This will mitigate
11462 cache-timing and potential related attacks.
11464 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11465 and this is automatically used by BN_mod_exp_mont() if the new flag
11466 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11467 will use this BN flag for private exponents unless the flag
11468 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11469 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11471 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11473 * Change the client implementation for SSLv23_method() and
11474 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11475 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11476 (Previously, the SSL 2.0 backwards compatible Client Hello
11477 message format would be used even with SSL_OP_NO_SSLv2.)
11481 * Add support for smime-type MIME parameter in S/MIME messages which some
11486 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11487 a threadsafe manner. Modify rsa code to use new function and add calls
11488 to dsa and dh code (which had race conditions before).
11492 * Include the fixed error library code in the C error file definitions
11493 instead of fixing them up at runtime. This keeps the error code
11494 structures constant.
11498 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11500 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11503 * Fixes for newer kerberos headers. NB: the casts are needed because
11504 the 'length' field is signed on one version and unsigned on another
11505 with no (?) obvious way to tell the difference, without these VC++
11506 complains. Also the "definition" of FAR (blank) is no longer included
11507 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11508 some needed definitions.
11512 * Undo Cygwin change.
11516 * Added support for proxy certificates according to RFC 3820.
11517 Because they may be a security thread to unaware applications,
11518 they must be explicitly allowed in run-time. See
11519 docs/HOWTO/proxy_certificates.txt for further information.
11523 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11525 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11526 server and client random values. Previously
11527 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11528 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11530 This change has negligible security impact because:
11532 1. Server and client random values still have 24 bytes of pseudo random
11535 2. Server and client random values are sent in the clear in the initial
11538 3. The master secret is derived using the premaster secret (48 bytes in
11539 size for static RSA ciphersuites) as well as client server and random
11542 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11545 *Stephen Henson, reported by UK NISCC*
11547 * Use Windows randomness collection on Cygwin.
11551 * Fix hang in EGD/PRNGD query when communication socket is closed
11552 prematurely by EGD/PRNGD.
11554 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11556 * Prompt for pass phrases when appropriate for PKCS12 input format.
11560 * Back-port of selected performance improvements from development
11561 branch, as well as improved support for PowerPC platforms.
11565 * Add lots of checks for memory allocation failure, error codes to indicate
11566 failure and freeing up memory if a failure occurs.
11568 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11570 * Add new -passin argument to dgst.
11574 * Perform some character comparisons of different types in X509_NAME_cmp:
11575 this is needed for some certificates that re-encode DNs into UTF8Strings
11576 (in violation of RFC3280) and can't or won't issue name rollover
11581 * Make an explicit check during certificate validation to see that
11582 the CA setting in each certificate on the chain is correct. As a
11583 side effect always do the following basic checks on extensions,
11584 not just when there's an associated purpose to the check:
11586 - if there is an unhandled critical extension (unless the user
11587 has chosen to ignore this fault)
11588 - if the path length has been exceeded (if one is set at all)
11589 - that certain extensions fit the associated purpose (if one has
11594 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11596 * Avoid a race condition when CRLs are checked in a multi threaded
11597 environment. This would happen due to the reordering of the revoked
11598 entries during signature checking and serial number lookup. Now the
11599 encoding is cached and the serial number sort performed under a lock.
11600 Add new STACK function sk_is_sorted().
11604 * Add Delta CRL to the extension code.
11608 * Various fixes to s3_pkt.c so alerts are sent properly.
11610 *David Holmes <d.holmes@f5.com>*
11612 * Reduce the chances of duplicate issuer name and serial numbers (in
11613 violation of RFC3280) using the OpenSSL certificate creation utilities.
11614 This is done by creating a random 64 bit value for the initial serial
11615 number when a serial number file is created or when a self signed
11616 certificate is created using 'openssl req -x509'. The initial serial
11617 number file is created using 'openssl x509 -next_serial' in CA.pl
11618 rather than being initialized to 1.
11622 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11624 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11625 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11627 *Joe Orton, Steve Henson*
11629 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11632 *Joe Orton, Steve Henson*
11634 * Make it possible to have multiple active certificates with the same
11635 subject in the CA index file. This is done only if the keyword
11636 'unique_subject' is set to 'no' in the main CA section (default
11637 if 'CA_default') of the configuration file. The value is saved
11638 with the database itself in a separate index attribute file,
11639 named like the index file with '.attr' appended to the name.
11643 * X509 verify fixes. Disable broken certificate workarounds when
11644 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11645 keyUsage extension present. Don't accept CRLs with unhandled critical
11646 extensions: since verify currently doesn't process CRL extensions this
11647 rejects a CRL with *any* critical extensions. Add new verify error codes
11652 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11653 A clarification of RFC2560 will require the use of OCTET STRINGs and
11654 some implementations cannot handle the current raw format. Since OpenSSL
11655 copies and compares OCSP nonces as opaque blobs without any attempt at
11656 parsing them this should not create any compatibility issues.
11660 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11661 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11662 this HMAC (and other) operations are several times slower than OpenSSL
11667 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11669 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11671 * Use the correct content when signing type "other".
11675 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11677 * Fix various bugs revealed by running the NISCC test suite:
11679 Stop out of bounds reads in the ASN1 code when presented with
11680 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11682 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11684 If verify callback ignores invalid public key errors don't try to check
11685 certificate signature with the NULL public key.
11689 * New -ignore_err option in ocsp application to stop the server
11690 exiting on the first error in a request.
11694 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11695 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11700 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11701 extra data after the compression methods not only for TLS 1.0
11702 but also for SSL 3.0 (as required by the specification).
11704 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11706 * Change X509_certificate_type() to mark the key as exported/exportable
11707 when it's 512 *bits* long, not 512 bytes.
11711 * Change AES_cbc_encrypt() so it outputs exact multiple of
11712 blocks during encryption.
11716 * Various fixes to base64 BIO and non blocking I/O. On write
11717 flushes were not handled properly if the BIO retried. On read
11718 data was not being buffered properly and had various logic bugs.
11719 This also affects blocking I/O when the data being decoded is a
11724 * Various S/MIME bugfixes and compatibility changes:
11725 output correct application/pkcs7 MIME type if
11726 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11727 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11728 of files as .eml work). Correctly handle very long lines in MIME
11733 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11735 * Countermeasure against the Klima-Pokorny-Rosa extension of
11736 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11737 a protocol version number mismatch like a decryption error
11738 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11742 * Turn on RSA blinding by default in the default implementation
11743 to avoid a timing attack. Applications that don't want it can call
11744 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11745 They would be ill-advised to do so in most cases.
11747 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11749 * Change RSA blinding code so that it works when the PRNG is not
11750 seeded (in this case, the secret RSA exponent is abused as
11751 an unpredictable seed -- if it is not unpredictable, there
11752 is no point in blinding anyway). Make RSA blinding thread-safe
11753 by remembering the creator's thread ID in rsa->blinding and
11754 having all other threads use local one-time blinding factors
11755 (this requires more computation than sharing rsa->blinding, but
11756 avoids excessive locking; and if an RSA object is not shared
11757 between threads, blinding will still be very fast).
11761 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11762 ENGINE as defaults for all supported algorithms irrespective of
11763 the 'flags' parameter. 'flags' is now honoured, so applications
11764 should make sure they are passing it correctly.
11768 * Target "mingw" now allows native Windows code to be generated in
11769 the Cygwin environment as well as with the MinGW compiler.
11773 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11775 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11776 via timing by performing a MAC computation even if incorrect
11777 block cipher padding has been found. This is a countermeasure
11778 against active attacks where the attacker has to distinguish
11779 between bad padding and a MAC verification error. ([CVE-2003-0078])
11781 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11782 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11783 Martin Vuagnoux (EPFL, Ilion)*
11785 * Make the no-err option work as intended. The intention with no-err
11786 is not to have the whole error stack handling routines removed from
11787 libcrypto, it's only intended to remove all the function name and
11788 reason texts, thereby removing some of the footprint that may not
11789 be interesting if those errors aren't displayed anyway.
11791 NOTE: it's still possible for any application or module to have its
11792 own set of error texts inserted. The routines are there, just not
11793 used by default when no-err is given.
11797 * Add support for FreeBSD on IA64.
11799 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11801 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11802 Kerberos function mit_des_cbc_cksum(). Before this change,
11803 the value returned by DES_cbc_cksum() was like the one from
11804 mit_des_cbc_cksum(), except the bytes were swapped.
11806 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11808 * Allow an application to disable the automatic SSL chain building.
11809 Before this a rather primitive chain build was always performed in
11810 ssl3_output_cert_chain(): an application had no way to send the
11811 correct chain if the automatic operation produced an incorrect result.
11813 Now the chain builder is disabled if either:
11815 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11817 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11819 The reasoning behind this is that an application would not want the
11820 auto chain building to take place if extra chain certificates are
11821 present and it might also want a means of sending no additional
11822 certificates (for example the chain has two certificates and the
11827 * Add the possibility to build without the ENGINE framework.
11829 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11831 * Under Win32 gmtime() can return NULL: check return value in
11832 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11836 * DSA routines: under certain error conditions uninitialized BN objects
11837 could be freed. Solution: make sure initialization is performed early
11838 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11839 Nils Larsch <nla@trustcenter.de> via PR#459)
11843 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11844 checked on reconnect on the client side, therefore session resumption
11845 could still fail with a "ssl session id is different" error. This
11846 behaviour is masked when SSL_OP_ALL is used due to
11847 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11848 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11849 followup to PR #377.
11853 * IA-32 assembler support enhancements: unified ELF targets, support
11854 for SCO/Caldera platforms, fix for Cygwin shared build.
11858 * Add support for FreeBSD on sparc64. As a consequence, support for
11859 FreeBSD on non-x86 processors is separate from x86 processors on
11860 the config script, much like the NetBSD support.
11862 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11864 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11866 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11869 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11870 code (06) was taken as the first octet of the session ID and the last
11871 octet was ignored consequently. As a result SSLv2 client side session
11872 caching could not have worked due to the session ID mismatch between
11874 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11879 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11880 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11885 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11886 seems that in spite of existing for more than a year, many application
11887 author have done nothing to provide the necessary callbacks, which
11888 means that this particular engine will not work properly anywhere.
11889 This is a very unfortunate situation which forces us, in the name
11890 of usability, to give the hw_ncipher.c a static lock, which is part
11892 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11893 appear in 0.9.8 or later. We EXPECT application authors to have
11894 dealt properly with this when 0.9.8 is released (unless we actually
11895 make such changes in the libcrypto locking code that changes will
11896 have to be made anyway).
11900 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11901 octets have been read, EOF or an error occurs. Without this change
11902 some truncated ASN1 structures will not produce an error.
11906 * Disable Heimdal support, since it hasn't been fully implemented.
11907 Still give the possibility to force the use of Heimdal, but with
11908 warnings and a request that patches get sent to openssl-dev.
11912 * Add the VC-CE target, introduce the WINCE sysname, and add
11913 INSTALL.WCE and appropriate conditionals to make it build.
11915 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11917 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11918 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11919 edit numbers of the version.
11921 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11923 * Introduce safe string copy and catenation functions
11924 (BUF_strlcpy() and BUF_strlcat()).
11926 *Ben Laurie (CHATS) and Richard Levitte*
11928 * Avoid using fixed-size buffers for one-line DNs.
11930 *Ben Laurie (CHATS)*
11932 * Add BUF_MEM_grow_clean() to avoid information leakage when
11933 resizing buffers containing secrets, and use where appropriate.
11935 *Ben Laurie (CHATS)*
11937 * Avoid using fixed size buffers for configuration file location.
11939 *Ben Laurie (CHATS)*
11941 * Avoid filename truncation for various CA files.
11943 *Ben Laurie (CHATS)*
11945 * Use sizeof in preference to magic numbers.
11947 *Ben Laurie (CHATS)*
11949 * Avoid filename truncation in cert requests.
11951 *Ben Laurie (CHATS)*
11953 * Add assertions to check for (supposedly impossible) buffer
11956 *Ben Laurie (CHATS)*
11958 * Don't cache truncated DNS entries in the local cache (this could
11959 potentially lead to a spoofing attack).
11961 *Ben Laurie (CHATS)*
11963 * Fix various buffers to be large enough for hex/decimal
11964 representations in a platform independent manner.
11966 *Ben Laurie (CHATS)*
11968 * Add CRYPTO_realloc_clean() to avoid information leakage when
11969 resizing buffers containing secrets, and use where appropriate.
11971 *Ben Laurie (CHATS)*
11973 * Add BIO_indent() to avoid much slightly worrying code to do
11976 *Ben Laurie (CHATS)*
11978 * Convert sprintf()/BIO_puts() to BIO_printf().
11980 *Ben Laurie (CHATS)*
11982 * buffer_gets() could terminate with the buffer only half
11985 *Ben Laurie (CHATS)*
11987 * Add assertions to prevent user-supplied crypto functions from
11988 overflowing internal buffers by having large block sizes, etc.
11990 *Ben Laurie (CHATS)*
11992 * New OPENSSL_assert() macro (similar to assert(), but enabled
11995 *Ben Laurie (CHATS)*
11997 * Eliminate unused copy of key in RC4.
11999 *Ben Laurie (CHATS)*
12001 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12003 *Ben Laurie (CHATS)*
12005 * Fix off-by-one error in EGD path.
12007 *Ben Laurie (CHATS)*
12009 * If RANDFILE path is too long, ignore instead of truncating.
12011 *Ben Laurie (CHATS)*
12013 * Eliminate unused and incorrectly sized X.509 structure
12016 *Ben Laurie (CHATS)*
12018 * Eliminate unused and dangerous function knumber().
12020 *Ben Laurie (CHATS)*
12022 * Eliminate unused and dangerous structure, KSSL_ERR.
12024 *Ben Laurie (CHATS)*
12026 * Protect against overlong session ID context length in an encoded
12027 session object. Since these are local, this does not appear to be
12030 *Ben Laurie (CHATS)*
12032 * Change from security patch (see 0.9.6e below) that did not affect
12033 the 0.9.6 release series:
12035 Remote buffer overflow in SSL3 protocol - an attacker could
12036 supply an oversized master key in Kerberos-enabled versions.
12039 *Ben Laurie (CHATS)*
12041 * Change the SSL kerb5 codes to match RFC 2712.
12045 * Make -nameopt work fully for req and add -reqopt switch.
12047 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12049 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12051 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12053 * Make sure tests can be performed even if the corresponding algorithms
12054 have been removed entirely. This was also the last step to make
12055 OpenSSL compilable with DJGPP under all reasonable conditions.
12057 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12059 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12060 to allow version independent disabling of normally unselected ciphers,
12061 which may be activated as a side-effect of selecting a single cipher.
12063 (E.g., cipher list string "RSA" enables ciphersuites that are left
12064 out of "ALL" because they do not provide symmetric encryption.
12065 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12067 *Lutz Jaenicke, Bodo Moeller*
12069 * Add appropriate support for separate platform-dependent build
12070 directories. The recommended way to make a platform-dependent
12071 build directory is the following (tested on Linux), maybe with
12074 # Place yourself outside of the OpenSSL source tree. In
12075 # this example, the environment variable OPENSSL_SOURCE
12076 # is assumed to contain the absolute OpenSSL source directory.
12077 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12078 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12079 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12080 mkdir -p `dirname $F`
12081 ln -s $OPENSSL_SOURCE/$F $F
12084 To be absolutely sure not to disturb the source tree, a "make clean"
12085 is a good thing. If it isn't successful, don't worry about it,
12086 it probably means the source directory is very clean.
12090 * Make sure any ENGINE control commands make local copies of string
12091 pointers passed to them whenever necessary. Otherwise it is possible
12092 the caller may have overwritten (or deallocated) the original string
12093 data when a later ENGINE operation tries to use the stored values.
12095 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12097 * Improve diagnostics in file reading and command-line digests.
12099 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12101 * Add AES modes CFB and OFB to the object database. Correct an
12102 error in AES-CFB decryption.
12106 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12107 allows existing EVP_CIPHER_CTX structures to be reused after
12108 calling `EVP_*Final()`. This behaviour is used by encryption
12109 BIOs and some applications. This has the side effect that
12110 applications must explicitly clean up cipher contexts with
12111 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12115 * Check the values of dna and dnb in bn_mul_recursive before calling
12116 bn_mul_comba (a non zero value means the a or b arrays do not contain
12117 n2 elements) and fallback to bn_mul_normal if either is not zero.
12121 * Fix escaping of non-ASCII characters when using the -subj option
12122 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12126 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12127 form for "surname", serialNumber has no short form.
12128 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12129 therefore remove "mail" short name for "internet 7".
12130 The OID for unique identifiers in X509 certificates is
12131 x500UniqueIdentifier, not uniqueIdentifier.
12132 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12136 * Add an "init" command to the ENGINE config module and auto initialize
12137 ENGINEs. Without any "init" command the ENGINE will be initialized
12138 after all ctrl commands have been executed on it. If init=1 the
12139 ENGINE is initialized at that point (ctrls before that point are run
12140 on the uninitialized ENGINE and after on the initialized one). If
12141 init=0 then the ENGINE will not be initialized at all.
12145 * Fix the 'app_verify_callback' interface so that the user-defined
12146 argument is actually passed to the callback: In the
12147 SSL_CTX_set_cert_verify_callback() prototype, the callback
12148 declaration has been changed from
12151 int (*cb)(X509_STORE_CTX *,void *);
12152 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12153 i=s->ctx->app_verify_callback(&ctx)
12154 has been changed into
12155 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12157 To update applications using SSL_CTX_set_cert_verify_callback(),
12158 a dummy argument can be added to their callback functions.
12160 *D. K. Smetters <smetters@parc.xerox.com>*
12162 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12164 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12166 * Add and OPENSSL_LOAD_CONF define which will cause
12167 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12168 This allows older applications to transparently support certain
12169 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12170 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12171 load the config file and OPENSSL_add_all_algorithms_conf() which will
12172 always load it have also been added.
12176 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12177 Adjust NIDs and EVP layer.
12179 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12181 * Config modules support in openssl utility.
12183 Most commands now load modules from the config file,
12184 though in a few (such as version) this isn't done
12185 because it couldn't be used for anything.
12187 In the case of ca and req the config file used is
12188 the same as the utility itself: that is the -config
12189 command line option can be used to specify an
12194 * Move default behaviour from OPENSSL_config(). If appname is NULL
12195 use "openssl_conf" if filename is NULL use default openssl config file.
12199 * Add an argument to OPENSSL_config() to allow the use of an alternative
12200 config section name. Add a new flag to tolerate a missing config file
12201 and move code to CONF_modules_load_file().
12205 * Support for crypto accelerator cards from Accelerated Encryption
12206 Processing, www.aep.ie. (Use engine 'aep')
12207 The support was copied from 0.9.6c [engine] and adapted/corrected
12208 to work with the new engine framework.
12210 *AEP Inc. and Richard Levitte*
12212 * Support for SureWare crypto accelerator cards from Baltimore
12213 Technologies. (Use engine 'sureware')
12214 The support was copied from 0.9.6c [engine] and adapted
12215 to work with the new engine framework.
12219 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12220 make the newer ENGINE framework commands for the CHIL engine work.
12222 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12224 * Make it possible to produce shared libraries on ReliantUNIX.
12226 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12228 * Add the configuration target debug-linux-ppro.
12229 Make 'openssl rsa' use the general key loading routines
12230 implemented in `apps.c`, and make those routines able to
12231 handle the key format FORMAT_NETSCAPE and the variant
12234 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12236 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12238 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12240 * Add -keyform to rsautl, and document -engine.
12242 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12244 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12245 BIO_R_NO_SUCH_FILE error code rather than the generic
12246 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12250 * Add new functions
12251 ERR_peek_last_error
12252 ERR_peek_last_error_line
12253 ERR_peek_last_error_line_data.
12254 These are similar to
12256 ERR_peek_error_line
12257 ERR_peek_error_line_data,
12258 but report on the latest error recorded rather than the first one
12259 still in the error queue.
12261 *Ben Laurie, Bodo Moeller*
12263 * default_algorithms option in ENGINE config module. This allows things
12265 default_algorithms = ALL
12266 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12270 * Preliminary ENGINE config module.
12274 * New experimental application configuration code.
12278 * Change the AES code to follow the same name structure as all other
12279 symmetric ciphers, and behave the same way. Move everything to
12280 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12282 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12284 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12286 *Ben Laurie and Theo de Raadt*
12288 * Add option to output public keys in req command.
12290 *Massimiliano Pala madwolf@openca.org*
12292 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12293 (up to about 10% better than before for P-192 and P-224).
12297 * New functions/macros
12299 SSL_CTX_set_msg_callback(ctx, cb)
12300 SSL_CTX_set_msg_callback_arg(ctx, arg)
12301 SSL_set_msg_callback(ssl, cb)
12302 SSL_set_msg_callback_arg(ssl, arg)
12304 to request calling a callback function
12306 void cb(int write_p, int version, int content_type,
12307 const void *buf, size_t len, SSL *ssl, void *arg)
12309 whenever a protocol message has been completely received
12310 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12311 protocol version according to which the SSL library interprets
12312 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12313 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12314 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12315 specification (change_cipher_spec(20), alert(21), handshake(22)).
12316 'buf' and 'len' point to the actual message, 'ssl' to the
12317 SSL object, and 'arg' is the application-defined value set by
12318 SSL[_CTX]_set_msg_callback_arg().
12320 'openssl s_client' and 'openssl s_server' have new '-msg' options
12321 to enable a callback that displays all protocol messages.
12325 * Change the shared library support so shared libraries are built as
12326 soon as the corresponding static library is finished, and thereby get
12327 openssl and the test programs linked against the shared library.
12328 This still only happens when the keyword "shard" has been given to
12329 the configuration scripts.
12331 NOTE: shared library support is still an experimental thing, and
12332 backward binary compatibility is still not guaranteed.
12334 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12336 * Add support for Subject Information Access extension.
12338 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12340 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12341 additional bytes when new memory had to be allocated, not just
12342 when reusing an existing buffer.
12346 * New command line and configuration option 'utf8' for the req command.
12347 This allows field values to be specified as UTF8 strings.
12351 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12352 runs for the former and machine-readable output for the latter.
12356 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12357 of the e-mail address in the DN (i.e., it will go into a certificate
12358 extension only). The new configuration file option 'email_in_dn = no'
12359 has the same effect.
12361 *Massimiliano Pala madwolf@openca.org*
12363 * Change all functions with names starting with `des_` to be starting
12364 with `DES_` instead. Add wrappers that are compatible with libdes,
12365 but are named `_ossl_old_des_*`. Finally, add macros that map the
12366 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12367 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12368 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12371 Since we provide two compatibility mappings, the user needs to
12372 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12373 compatibility is desired. The default (i.e., when that macro
12374 isn't defined) is OpenSSL 0.9.6c compatibility.
12376 There are also macros that enable and disable the support of old
12377 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12378 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12379 are defined, the default will apply: to support the old des routines.
12381 In either case, one must include openssl/des.h to get the correct
12382 definitions. Do not try to just include openssl/des_old.h, that
12385 NOTE: This is a major break of an old API into a new one. Software
12386 authors are encouraged to switch to the `DES_` style functions. Some
12387 time in the future, des_old.h and the libdes compatibility functions
12388 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12389 default), and then completely removed.
12393 * Test for certificates which contain unsupported critical extensions.
12394 If such a certificate is found during a verify operation it is
12395 rejected by default: this behaviour can be overridden by either
12396 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12397 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12398 X509_supported_extension() has also been added which returns 1 if a
12399 particular extension is supported.
12403 * Modify the behaviour of EVP cipher functions in similar way to digests
12404 to retain compatibility with existing code.
12408 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12409 compatibility with existing code. In particular the 'ctx' parameter does
12410 not have to be to be initialized before the call to EVP_DigestInit() and
12411 it is tidied up after a call to EVP_DigestFinal(). New function
12412 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12413 EVP_MD_CTX_copy() changed to not require the destination to be
12414 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12415 requires the destination to be valid.
12417 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12418 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12422 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12423 so that complete 'Handshake' protocol structures are kept in memory
12424 instead of overwriting 'msg_type' and 'length' with 'body' data.
12428 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12430 *Massimo Santin via Richard Levitte*
12432 * Major restructuring to the underlying ENGINE code. This includes
12433 reduction of linker bloat, separation of pure "ENGINE" manipulation
12434 (initialisation, etc) from functionality dealing with implementations
12435 of specific crypto interfaces. This change also introduces integrated
12436 support for symmetric ciphers and digest implementations - so ENGINEs
12437 can now accelerate these by providing EVP_CIPHER and EVP_MD
12438 implementations of their own. This is detailed in
12439 [crypto/engine/README.md](crypto/engine/README.md)
12440 as it couldn't be adequately described here. However, there are a few
12441 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12442 were changed in the original introduction of ENGINE code have now
12443 reverted back - the hooking from this code to ENGINE is now a good
12444 deal more passive and at run-time, operations deal directly with
12445 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12446 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12447 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12448 they were not being used by the framework as there is no concept of a
12449 BIGNUM_METHOD and they could not be generalised to the new
12450 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12451 ENGINE_cpy() has been removed as it cannot be consistently defined in
12456 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12460 * Change mkdef.pl to sort symbols that get the same entry number,
12461 and make sure the automatically generated functions `ERR_load_*`
12462 become part of libeay.num as well.
12466 * New function SSL_renegotiate_pending(). This returns true once
12467 renegotiation has been requested (either SSL_renegotiate() call
12468 or HelloRequest/ClientHello received from the peer) and becomes
12469 false once a handshake has been completed.
12470 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12471 sends a HelloRequest, but does not ensure that a handshake takes
12472 place. SSL_renegotiate_pending() is useful for checking if the
12473 client has followed the request.)
12477 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12478 By default, clients may request session resumption even during
12479 renegotiation (if session ID contexts permit); with this option,
12480 session resumption is possible only in the first handshake.
12482 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12483 more bits available for options that should not be part of
12484 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12488 * Add some demos for certificate and certificate request creation.
12492 * Make maximum certificate chain size accepted from the peer application
12493 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12494 "Douglas E. Engert" <deengert@anl.gov>.
12498 * Add support for shared libraries for Unixware-7
12499 (Boyd Lynn Gerber <gerberb@zenez.com>).
12503 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12504 be done prior to destruction. Use this to unload error strings from
12505 ENGINEs that load their own error strings. NB: This adds two new API
12506 functions to "get" and "set" this destroy handler in an ENGINE.
12510 * Alter all existing ENGINE implementations (except "openssl" and
12511 "openbsd") to dynamically instantiate their own error strings. This
12512 makes them more flexible to be built both as statically-linked ENGINEs
12513 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12514 Also, add stub code to each that makes building them as self-contained
12515 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12519 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12520 implementations into applications that are completely implemented in
12521 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12522 commands that can be used to configure what shared-library to load and
12523 to control aspects of the way it is handled. Also, made an update to
12524 the [README-Engine.md](README-Engine.md) file
12525 that brings its information up-to-date and
12526 provides some information and instructions on the "dynamic" ENGINE
12527 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12531 * Make it possible to unload ranges of ERR strings with a new
12532 "ERR_unload_strings" function.
12536 * Add a copy() function to EVP_MD.
12540 * Make EVP_MD routines take a context pointer instead of just the
12541 md_data void pointer.
12545 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12546 that the digest can only process a single chunk of data
12547 (typically because it is provided by a piece of
12548 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12549 is only going to provide a single chunk of data, and hence the
12550 framework needn't accumulate the data for oneshot drivers.
12554 * As with "ERR", make it possible to replace the underlying "ex_data"
12555 functions. This change also alters the storage and management of global
12556 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12557 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12558 index counters. The API functions that use this state have been changed
12559 to take a "class_index" rather than pointers to the class's local STACK
12560 and counter, and there is now an API function to dynamically create new
12561 classes. This centralisation allows us to (a) plug a lot of the
12562 thread-safety problems that existed, and (b) makes it possible to clean
12563 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12564 such data would previously have always leaked in application code and
12565 workarounds were in place to make the memory debugging turn a blind eye
12566 to it. Application code that doesn't use this new function will still
12567 leak as before, but their memory debugging output will announce it now
12568 rather than letting it slide.
12570 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12571 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12572 has a return value to indicate success or failure.
12576 * Make it possible to replace the underlying "ERR" functions such that the
12577 global state (2 LHASH tables and 2 locks) is only used by the "default"
12578 implementation. This change also adds two functions to "get" and "set"
12579 the implementation prior to it being automatically set the first time
12580 any other ERR function takes place. Ie. an application can call "get",
12581 pass the return value to a module it has just loaded, and that module
12582 can call its own "set" function using that value. This means the
12583 module's "ERR" operations will use (and modify) the error state in the
12584 application and not in its own statically linked copy of OpenSSL code.
12588 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12589 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12590 the operation, and provides a more encapsulated way for external code
12591 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12592 to use these functions rather than manually incrementing the counts.
12594 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12598 * Add EVP test program.
12602 * Add symmetric cipher support to ENGINE. Expect the API to change!
12606 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12607 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12608 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12609 These allow a CRL to be built without having to access X509_CRL fields
12610 directly. Modify 'ca' application to use new functions.
12614 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12615 bug workarounds. Rollback attack detection is a security feature.
12616 The problem will only arise on OpenSSL servers when TLSv1 is not
12617 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12618 Software authors not wanting to support TLSv1 will have special reasons
12619 for their choice and can explicitly enable this option.
12621 *Bodo Moeller, Lutz Jaenicke*
12623 * Rationalise EVP so it can be extended: don't include a union of
12624 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12625 (similar to those existing for EVP_CIPHER_CTX).
12630 EVP_MD_CTX_init(&md); /* new function call */
12631 EVP_DigestInit(&md, EVP_sha1());
12632 EVP_DigestUpdate(&md, in, len);
12633 EVP_DigestFinal(&md, out, NULL);
12634 EVP_MD_CTX_cleanup(&md); /* new function call */
12638 * Make DES key schedule conform to the usual scheme, as well as
12639 correcting its structure. This means that calls to DES functions
12640 now have to pass a pointer to a des_key_schedule instead of a
12641 plain des_key_schedule (which was actually always a pointer
12644 des_key_schedule ks;
12646 des_set_key_checked(..., &ks);
12647 des_ncbc_encrypt(..., &ks, ...);
12649 (Note that a later change renames 'des_...' into 'DES_...'.)
12653 * Initial reduction of linker bloat: the use of some functions, such as
12654 PEM causes large amounts of unused functions to be linked in due to
12655 poor organisation. For example pem_all.c contains every PEM function
12656 which has a knock on effect of linking in large amounts of (unused)
12657 ASN1 code. Grouping together similar functions and splitting unrelated
12658 functions prevents this.
12662 * Cleanup of EVP macros.
12666 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12667 correct `_ecb suffix`.
12671 * Add initial OCSP responder support to ocsp application. The
12672 revocation information is handled using the text based index
12673 use by the ca application. The responder can either handle
12674 requests generated internally, supplied in files (for example
12675 via a CGI script) or using an internal minimal server.
12679 * Add configuration choices to get zlib compression for TLS.
12683 * Changes to Kerberos SSL for RFC 2712 compliance:
12684 1. Implemented real KerberosWrapper, instead of just using
12685 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12686 2. Implemented optional authenticator field of KerberosWrapper.
12688 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12689 and authenticator structs; see crypto/krb5/.
12691 Generalized Kerberos calls to support multiple Kerberos libraries.
12692 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12693 via Richard Levitte*
12695 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12696 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12697 values for each of the key sizes rather than having just
12698 parameters (and 'speed' generating keys each time).
12702 * Speed up EVP routines.
12705 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12706 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12707 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12708 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12710 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12711 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12712 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12715 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12717 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12721 * Added the OS2-EMX target.
12723 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12725 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12726 New functions to support `NCONF` routines in extension code.
12727 New function `CONF_set_nconf()`
12728 to allow functions which take an `NCONF` to also handle the old `LHASH`
12729 structure: this means that the old `CONF` compatible routines can be
12730 retained (in particular w.rt. extensions) without having to duplicate the
12731 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12735 * Enhance the general user interface with mechanisms for inner control
12736 and with possibilities to have yes/no kind of prompts.
12740 * Change all calls to low-level digest routines in the library and
12741 applications to use EVP. Add missing calls to HMAC_cleanup() and
12742 don't assume HMAC_CTX can be copied using memcpy().
12744 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12746 * Add the possibility to control engines through control names but with
12747 arbitrary arguments instead of just a string.
12748 Change the key loaders to take a UI_METHOD instead of a callback
12749 function pointer. NOTE: this breaks binary compatibility with earlier
12750 versions of OpenSSL [engine].
12751 Adapt the nCipher code for these new conditions and add a card insertion
12756 * Enhance the general user interface with mechanisms to better support
12757 dialog box interfaces, application-defined prompts, the possibility
12758 to use defaults (for example default passwords from somewhere else)
12759 and interrupts/cancellations.
12763 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12764 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12768 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12769 tidy up some unnecessarily weird code in 'sk_new()').
12771 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12773 * Change the key loading routines for ENGINEs to use the same kind
12774 callback (pem_password_cb) as all other routines that need this
12779 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12780 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12781 than this minimum value is recommended.
12785 * New random seeder for OpenVMS, using the system process statistics
12786 that are easily reachable.
12790 * Windows apparently can't transparently handle global
12791 variables defined in DLLs. Initialisations such as:
12793 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12795 won't compile. This is used by the any applications that need to
12796 declare their own ASN1 modules. This was fixed by adding the option
12797 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12798 needed for static libraries under Win32.
12802 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12803 setting of purpose and trust fields. New X509_STORE trust and
12804 purpose functions and tidy up setting in other SSL functions.
12808 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12809 structure. These are inherited by X509_STORE_CTX when it is
12810 initialised. This allows various defaults to be set in the
12811 X509_STORE structure (such as flags for CRL checking and custom
12812 purpose or trust settings) for functions which only use X509_STORE_CTX
12813 internally such as S/MIME.
12815 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12816 trust settings if they are not set in X509_STORE. This allows X509_STORE
12817 purposes and trust (in S/MIME for example) to override any set by default.
12819 Add command line options for CRL checking to smime, s_client and s_server
12824 * Initial CRL based revocation checking. If the CRL checking flag(s)
12825 are set then the CRL is looked up in the X509_STORE structure and
12826 its validity and signature checked, then if the certificate is found
12827 in the CRL the verify fails with a revoked error.
12829 Various new CRL related callbacks added to X509_STORE_CTX structure.
12831 Command line options added to 'verify' application to support this.
12833 This needs some additional work, such as being able to handle multiple
12834 CRLs with different times, extension based lookup (rather than just
12835 by subject name) and ultimately more complete V2 CRL extension
12840 * Add a general user interface API (crypto/ui/). This is designed
12841 to replace things like des_read_password and friends (backward
12842 compatibility functions using this new API are provided).
12843 The purpose is to remove prompting functions from the DES code
12844 section as well as provide for prompting through dialog boxes in
12845 a window system and the like.
12849 * Add "ex_data" support to ENGINE so implementations can add state at a
12850 per-structure level rather than having to store it globally.
12854 * Make it possible for ENGINE structures to be copied when retrieved by
12855 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12856 This causes the "original" ENGINE structure to act like a template,
12857 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12858 operational state can be localised to each ENGINE structure, despite the
12859 fact they all share the same "methods". New ENGINE structures returned in
12860 this case have no functional references and the return value is the single
12861 structural reference. This matches the single structural reference returned
12862 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12867 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12868 needs to match any other type at all we need to manually clear the
12873 * Changes to the "openssl engine" utility to include;
12874 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12875 about an ENGINE's available control commands.
12876 - executing control commands from command line arguments using the
12877 '-pre' and '-post' switches. '-post' is only used if '-t' is
12878 specified and the ENGINE is successfully initialised. The syntax for
12879 the individual commands are colon-separated, for example;
12880 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12884 * New dynamic control command support for ENGINEs. ENGINEs can now
12885 declare their own commands (numbers), names (strings), descriptions,
12886 and input types for run-time discovery by calling applications. A
12887 subset of these commands are implicitly classed as "executable"
12888 depending on their input type, and only these can be invoked through
12889 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12890 can be based on user input, config files, etc). The distinction is
12891 that "executable" commands cannot return anything other than a boolean
12892 result and can only support numeric or string input, whereas some
12893 discoverable commands may only be for direct use through
12894 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12895 pointers, or other custom uses. The "executable" commands are to
12896 support parameterisations of ENGINE behaviour that can be
12897 unambiguously defined by ENGINEs and used consistently across any
12898 OpenSSL-based application. Commands have been added to all the
12899 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12900 control over shared-library paths without source code alterations.
12904 * Changed all ENGINE implementations to dynamically allocate their
12905 ENGINEs rather than declaring them statically. Apart from this being
12906 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12907 this also allows the implementations to compile without using the
12908 internal engine_int.h header.
12912 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12913 'const' value. Any code that should be able to modify a RAND_METHOD
12914 should already have non-const pointers to it (ie. they should only
12915 modify their own ones).
12919 * Made a variety of little tweaks to the ENGINE code.
12920 - "atalla" and "ubsec" string definitions were moved from header files
12921 to C code. "nuron" string definitions were placed in variables
12922 rather than hard-coded - allowing parameterisation of these values
12923 later on via ctrl() commands.
12924 - Removed unused "#if 0"'d code.
12925 - Fixed engine list iteration code so it uses ENGINE_free() to release
12926 structural references.
12927 - Constified the RAND_METHOD element of ENGINE structures.
12928 - Constified various get/set functions as appropriate and added
12929 missing functions (including a catch-all ENGINE_cpy that duplicates
12930 all ENGINE values onto a new ENGINE except reference counts/state).
12931 - Removed NULL parameter checks in get/set functions. Setting a method
12932 or function to NULL is a way of cancelling out a previously set
12933 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12934 and doesn't justify the extra error symbols and code.
12935 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12936 flags from engine_int.h to engine.h.
12937 - Changed prototypes for ENGINE handler functions (init(), finish(),
12938 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12942 * Implement binary inversion algorithm for BN_mod_inverse in addition
12943 to the algorithm using long division. The binary algorithm can be
12944 used only if the modulus is odd. On 32-bit systems, it is faster
12945 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12946 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12947 up to 450 bits. In 64-bit environments, the binary algorithm
12948 appears to be advantageous for much longer moduli; here we use it
12949 for moduli up to 2048 bits.
12953 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12954 could not support the combine flag in choice fields.
12958 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12959 extensions from a certificate request to the certificate.
12963 * Allow multiple 'certopt' and 'nameopt' options to be separated
12964 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12965 file: this allows the display of the certificate about to be
12966 signed to be customised, to allow certain fields to be included
12967 or excluded and extension details. The old system didn't display
12968 multicharacter strings properly, omitted fields not in the policy
12969 and couldn't display additional details such as extensions.
12973 * Function EC_POINTs_mul for multiple scalar multiplication
12974 of an arbitrary number of elliptic curve points
12975 \sum scalars[i]*points[i],
12976 optionally including the generator defined for the EC_GROUP:
12977 scalar*generator + \sum scalars[i]*points[i].
12979 EC_POINT_mul is a simple wrapper function for the typical case
12980 that the point list has just one item (besides the optional
12985 * First EC_METHODs for curves over GF(p):
12987 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12988 operations and provides various method functions that can also
12989 operate with faster implementations of modular arithmetic.
12991 EC_GFp_mont_method() reuses most functions that are part of
12992 EC_GFp_simple_method, but uses Montgomery arithmetic.
12994 *Bodo Moeller; point addition and point doubling
12995 implementation directly derived from source code provided by
12996 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12998 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12999 crypto/ec/ec_lib.c):
13001 Curves are EC_GROUP objects (with an optional group generator)
13002 based on EC_METHODs that are built into the library.
13004 Points are EC_POINT objects based on EC_GROUP objects.
13006 Most of the framework would be able to handle curves over arbitrary
13007 finite fields, but as there are no obvious types for fields other
13008 than GF(p), some functions are limited to that for now.
13012 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13013 that the file contains a complete HTTP response.
13017 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13018 change the def and num file printf format specifier from "%-40sXXX"
13019 to "%-39s XXX". The latter will always guarantee a space after the
13020 field while the former will cause them to run together if the field
13021 is 40 of more characters long.
13025 * Constify the cipher and digest 'method' functions and structures
13026 and modify related functions to take constant EVP_MD and EVP_CIPHER
13031 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13032 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13036 * Modify `EVP_Digest*()` routines so they now return values. Although the
13037 internal software routines can never fail additional hardware versions
13042 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13044 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13045 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13048 ERR_R_NESTED_ASN1_ERROR
13050 ERR_R_MISSING_ASN1_EOS
13051 were 4 .. 9, conflicting with
13052 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13054 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13055 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13057 Add new error code 'ERR_R_INTERNAL_ERROR'.
13061 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13066 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13067 sets the subject name for a new request or supersedes the
13068 subject name in a given request. Formats that can be parsed are
13069 'CN=Some Name, OU=myOU, C=IT'
13071 'CN=Some Name/OU=myOU/C=IT'.
13073 Add options '-batch' and '-verbose' to 'openssl req'.
13075 *Massimiliano Pala <madwolf@hackmasters.net>*
13077 * Introduce the possibility to access global variables through
13078 functions on platform were that's the best way to handle exporting
13079 global variables in shared libraries. To enable this functionality,
13080 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13081 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13082 is normally done by Configure or something similar).
13084 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13085 in the source file (foo.c) like this:
13087 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13088 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13090 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13091 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13093 OPENSSL_DECLARE_GLOBAL(int,foo);
13094 #define foo OPENSSL_GLOBAL_REF(foo)
13095 OPENSSL_DECLARE_GLOBAL(double,bar);
13096 #define bar OPENSSL_GLOBAL_REF(bar)
13098 The #defines are very important, and therefore so is including the
13099 header file everywhere where the defined globals are used.
13101 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13102 of ASN.1 items, but that structure is a bit different.
13104 The largest change is in util/mkdef.pl which has been enhanced with
13105 better and easier to understand logic to choose which symbols should
13106 go into the Windows .def files as well as a number of fixes and code
13107 cleanup (among others, algorithm keywords are now sorted
13108 lexicographically to avoid constant rewrites).
13112 * In BN_div() keep a copy of the sign of 'num' before writing the
13113 result to 'rm' because if rm==num the value will be overwritten
13114 and produce the wrong result if 'num' is negative: this caused
13115 problems with BN_mod() and BN_nnmod().
13119 * Function OCSP_request_verify(). This checks the signature on an
13120 OCSP request and verifies the signer certificate. The signer
13121 certificate is just checked for a generic purpose and OCSP request
13126 * Add OCSP_check_validity() function to check the validity of OCSP
13127 responses. OCSP responses are prepared in real time and may only
13128 be a few seconds old. Simply checking that the current time lies
13129 between thisUpdate and nextUpdate max reject otherwise valid responses
13130 caused by either OCSP responder or client clock inaccuracy. Instead
13131 we allow thisUpdate and nextUpdate to fall within a certain period of
13132 the current time. The age of the response can also optionally be
13133 checked. Two new options -validity_period and -status_age added to
13138 * If signature or public key algorithm is unrecognized print out its
13139 OID rather that just UNKNOWN.
13143 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13144 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13145 ID to be generated from the issuer certificate alone which can then be
13146 passed to OCSP_id_issuer_cmp().
13150 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13151 ASN1 modules to export functions returning ASN1_ITEM pointers
13152 instead of the ASN1_ITEM structures themselves. This adds several
13153 new macros which allow the underlying ASN1 function/structure to
13154 be accessed transparently. As a result code should not use ASN1_ITEM
13155 references directly (such as &X509_it) but instead use the relevant
13156 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13157 use of the new ASN1 code on platforms where exporting structures
13158 is problematical (for example in shared libraries) but exporting
13159 functions returning pointers to structures is not.
13163 * Add support for overriding the generation of SSL/TLS session IDs.
13164 These callbacks can be registered either in an SSL_CTX or per SSL.
13165 The purpose of this is to allow applications to control, if they wish,
13166 the arbitrary values chosen for use as session IDs, particularly as it
13167 can be useful for session caching in multiple-server environments. A
13168 command-line switch for testing this (and any client code that wishes
13169 to use such a feature) has been added to "s_server".
13171 *Geoff Thorpe, Lutz Jaenicke*
13173 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13174 of the form `#if defined(...) || defined(...) || ...` and
13175 `#if !defined(...) && !defined(...) && ...`. This also avoids
13176 the growing number of special cases it was previously handling.
13180 * Make all configuration macros available for application by making
13181 sure they are available in opensslconf.h, by giving them names starting
13182 with `OPENSSL_` to avoid conflicts with other packages and by making
13183 sure e_os2.h will cover all platform-specific cases together with
13185 Additionally, it is now possible to define configuration/platform-
13186 specific names (called "system identities"). In the C code, these
13187 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13188 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13189 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13194 * New option -set_serial to 'req' and 'x509' this allows the serial
13195 number to use to be specified on the command line. Previously self
13196 signed certificates were hard coded with serial number 0 and the
13197 CA options of 'x509' had to use a serial number in a file which was
13202 * New options to 'ca' utility to support V2 CRL entry extensions.
13203 Currently CRL reason, invalidity date and hold instruction are
13204 supported. Add new CRL extensions to V3 code and some new objects.
13208 * New function EVP_CIPHER_CTX_set_padding() this is used to
13209 disable standard block padding (aka PKCS#5 padding) in the EVP
13210 API, which was previously mandatory. This means that the data is
13211 not padded in any way and so the total length much be a multiple
13212 of the block size, otherwise an error occurs.
13216 * Initial (incomplete) OCSP SSL support.
13220 * New function OCSP_parse_url(). This splits up a URL into its host,
13221 port and path components: primarily to parse OCSP URLs. New -url
13222 option to ocsp utility.
13226 * New nonce behavior. The return value of OCSP_check_nonce() now
13227 reflects the various checks performed. Applications can decide
13228 whether to tolerate certain situations such as an absent nonce
13229 in a response when one was present in a request: the ocsp application
13230 just prints out a warning. New function OCSP_add1_basic_nonce()
13231 this is to allow responders to include a nonce in a response even if
13232 the request is nonce-less.
13236 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13237 skipped when using openssl x509 multiple times on a single input file,
13238 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13242 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13243 set string type: to handle setting ASN1_TIME structures. Fix ca
13244 utility to correctly initialize revocation date of CRLs.
13248 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13249 the clients preferred ciphersuites and rather use its own preferences.
13250 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13251 Internet Explorer by ensuring unchanged hash method during stepup.
13252 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13256 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13257 to aes and add a new 'exist' option to print out symbols that don't
13262 * Additional options to ocsp utility to allow flags to be set and
13263 additional certificates supplied.
13267 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13268 OCSP client a number of certificate to only verify the response
13273 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13274 handle the new API. Currently only ECB, CBC modes supported. Add new
13277 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13278 Encryption Standard (AES) Ciphersuites for Transport Layer
13279 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13280 not enabled by default and were not part of the "ALL" ciphersuite
13281 alias because they were not yet official; they could be
13282 explicitly requested by specifying the "AESdraft" ciphersuite
13283 group alias. In the final release of OpenSSL 0.9.7, the group
13284 alias is called "AES" and is part of "ALL".)
13286 *Ben Laurie, Steve Henson, Bodo Moeller*
13288 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13289 request to response.
13293 * Functions for OCSP responders. OCSP_request_onereq_count(),
13294 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13295 extract information from a certificate request. OCSP_response_create()
13296 creates a response and optionally adds a basic response structure.
13297 OCSP_basic_add1_status() adds a complete single response to a basic
13298 response and returns the OCSP_SINGLERESP structure just added (to allow
13299 extensions to be included for example). OCSP_basic_add1_cert() adds a
13300 certificate to a basic response and OCSP_basic_sign() signs a basic
13301 response with various flags. New helper functions ASN1_TIME_check()
13302 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13303 (converts ASN1_TIME to GeneralizedTime).
13307 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13308 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13309 structure from a certificate. X509_pubkey_digest() digests the public_key
13310 contents: this is used in various key identifiers.
13314 * Make sk_sort() tolerate a NULL argument.
13316 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13318 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13319 passed by the function are trusted implicitly. If any of them signed the
13320 response then it is assumed to be valid and is not verified.
13324 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13325 to data. This was previously part of the PKCS7 ASN1 code. This
13326 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13327 *Steve Henson, reported by Kenneth R. Robinette
13328 <support@securenetterm.com>*
13330 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13331 routines: without these tracing memory leaks is very painful.
13332 Fix leaks in PKCS12 and PKCS7 routines.
13336 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13337 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13338 effectively meant GeneralizedTime would never be used. Now it
13339 is initialised to -1 but X509_time_adj() now has to check the value
13340 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13341 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13342 *Steve Henson, reported by Kenneth R. Robinette
13343 <support@securenetterm.com>*
13345 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13346 result in a zero length in the ASN1_INTEGER structure which was
13347 not consistent with the structure when d2i_ASN1_INTEGER() was used
13348 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13349 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13350 where it did not print out a minus for negative ASN1_INTEGER.
13354 * Add summary printout to ocsp utility. The various functions which
13355 convert status values to strings have been renamed to:
13356 OCSP_response_status_str(), OCSP_cert_status_str() and
13357 OCSP_crl_reason_str() and are no longer static. New options
13358 to verify nonce values and to disable verification. OCSP response
13359 printout format cleaned up.
13363 * Add additional OCSP certificate checks. These are those specified
13364 in RFC2560. This consists of two separate checks: the CA of the
13365 certificate being checked must either be the OCSP signer certificate
13366 or the issuer of the OCSP signer certificate. In the latter case the
13367 OCSP signer certificate must contain the OCSP signing extended key
13368 usage. This check is performed by attempting to match the OCSP
13369 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13370 in the OCSP_CERTID structures of the response.
13374 * Initial OCSP certificate verification added to OCSP_basic_verify()
13375 and related routines. This uses the standard OpenSSL certificate
13376 verify routines to perform initial checks (just CA validity) and
13377 to obtain the certificate chain. Then additional checks will be
13378 performed on the chain. Currently the root CA is checked to see
13379 if it is explicitly trusted for OCSP signing. This is used to set
13380 a root CA as a global signing root: that is any certificate that
13381 chains to that CA is an acceptable OCSP signing certificate.
13385 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13386 extensions from a separate configuration file.
13387 As when reading extensions from the main configuration file,
13388 the '-extensions ...' option may be used for specifying the
13391 *Massimiliano Pala <madwolf@comune.modena.it>*
13393 * New OCSP utility. Allows OCSP requests to be generated or
13394 read. The request can be sent to a responder and the output
13395 parsed, outputted or printed in text form. Not complete yet:
13396 still needs to check the OCSP response validity.
13400 * New subcommands for 'openssl ca':
13401 `openssl ca -status <serial>` prints the status of the cert with
13402 the given serial number (according to the index file).
13403 `openssl ca -updatedb` updates the expiry status of certificates
13406 *Massimiliano Pala <madwolf@comune.modena.it>*
13408 * New '-newreq-nodes' command option to CA.pl. This is like
13409 '-newreq', but calls 'openssl req' with the '-nodes' option
13410 so that the resulting key is not encrypted.
13412 *Damien Miller <djm@mindrot.org>*
13414 * New configuration for the GNU Hurd.
13416 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13418 * Initial code to implement OCSP basic response verify. This
13419 is currently incomplete. Currently just finds the signer's
13420 certificate and verifies the signature on the response.
13424 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13425 value of OPENSSLDIR. This is available via the new '-d' option
13426 to 'openssl version', and is also included in 'openssl version -a'.
13430 * Allowing defining memory allocation callbacks that will be given
13431 file name and line number information in additional arguments
13432 (a `const char*` and an int). The basic functionality remains, as
13433 well as the original possibility to just replace malloc(),
13434 realloc() and free() by functions that do not know about these
13435 additional arguments. To register and find out the current
13436 settings for extended allocation functions, the following
13437 functions are provided:
13439 CRYPTO_set_mem_ex_functions
13440 CRYPTO_set_locked_mem_ex_functions
13441 CRYPTO_get_mem_ex_functions
13442 CRYPTO_get_locked_mem_ex_functions
13444 These work the same way as CRYPTO_set_mem_functions and friends.
13445 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13446 extended allocation function is enabled.
13447 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13448 a conventional allocation function is enabled.
13450 *Richard Levitte, Bodo Moeller*
13452 * Finish off removing the remaining LHASH function pointer casts.
13453 There should no longer be any prototype-casting required when using
13454 the LHASH abstraction, and any casts that remain are "bugs". See
13455 the callback types and macros at the head of lhash.h for details
13456 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13460 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13461 If /dev/[u]random devices are not available or do not return enough
13462 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13464 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13465 /etc/entropy will be queried once each in this sequence, querying stops
13466 when enough entropy was collected without querying more sockets.
13470 * Change the Unix RAND_poll() variant to be able to poll several
13471 random devices, as specified by DEVRANDOM, until a sufficient amount
13472 of data has been collected. We spend at most 10 ms on each file
13473 (select timeout) and read in non-blocking mode. DEVRANDOM now
13474 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13475 (previously it was just the string "/dev/urandom"), so on typical
13476 platforms the 10 ms delay will never occur.
13477 Also separate out the Unix variant to its own file, rand_unix.c.
13478 For VMS, there's a currently-empty rand_vms.c.
13482 * Move OCSP client related routines to ocsp_cl.c. These
13483 provide utility functions which an application needing
13484 to issue a request to an OCSP responder and analyse the
13485 response will typically need: as opposed to those which an
13486 OCSP responder itself would need which will be added later.
13488 OCSP_request_sign() signs an OCSP request with an API similar
13489 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13490 response. OCSP_response_get1_basic() extracts basic response
13491 from response. OCSP_resp_find_status(): finds and extracts status
13492 information from an OCSP_CERTID structure (which will be created
13493 when the request structure is built). These are built from lower
13494 level functions which work on OCSP_SINGLERESP structures but
13495 won't normally be used unless the application wishes to examine
13496 extensions in the OCSP response for example.
13498 Replace nonce routines with a pair of functions.
13499 OCSP_request_add1_nonce() adds a nonce value and optionally
13500 generates a random value. OCSP_check_nonce() checks the
13501 validity of the nonce in an OCSP response.
13505 * Change function OCSP_request_add() to OCSP_request_add0_id().
13506 This doesn't copy the supplied OCSP_CERTID and avoids the
13507 need to free up the newly created id. Change return type
13508 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13509 This can then be used to add extensions to the request.
13510 Deleted OCSP_request_new(), since most of its functionality
13511 is now in OCSP_REQUEST_new() (and the case insensitive name
13512 clash) apart from the ability to set the request name which
13513 will be added elsewhere.
13517 * Update OCSP API. Remove obsolete extensions argument from
13518 various functions. Extensions are now handled using the new
13519 OCSP extension code. New simple OCSP HTTP function which
13520 can be used to send requests and parse the response.
13524 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13525 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13526 uses the special reorder version of SET OF to sort the attributes
13527 and reorder them to match the encoded order. This resolves a long
13528 standing problem: a verify on a PKCS7 structure just after signing
13529 it used to fail because the attribute order did not match the
13530 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13531 it uses the received order. This is necessary to tolerate some broken
13532 software that does not order SET OF. This is handled by encoding
13533 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13534 to produce the required SET OF.
13538 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13539 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13540 files to get correct declarations of the ASN.1 item variables.
13544 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13545 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13546 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13547 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13548 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13549 ASN1_ITEM and no wrapper functions.
13553 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13554 replace the old function pointer based I/O routines. Change most of
13555 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13559 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13560 lines, recognize more "algorithms" that can be deselected, and make
13561 it complain about algorithm deselection that isn't recognised.
13565 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13566 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13567 to use new functions. Add NO_ASN1_OLD which can be set to remove
13568 some old style ASN1 functions: this can be used to determine if old
13569 code will still work when these eventually go away.
13573 * New extension functions for OCSP structures, these follow the
13574 same conventions as certificates and CRLs.
13578 * New function X509V3_add1_i2d(). This automatically encodes and
13579 adds an extension. Its behaviour can be customised with various
13580 flags to append, replace or delete. Various wrappers added for
13581 certificates and CRLs.
13585 * Fix to avoid calling the underlying ASN1 print routine when
13586 an extension cannot be parsed. Correct a typo in the
13587 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13591 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13592 entries for variables.
13596 * Add functionality to `apps/openssl.c` for detecting locking
13597 problems: As the program is single-threaded, all we have
13598 to do is register a locking callback using an array for
13599 storing which locks are currently held by the program.
13603 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13604 SSL_get_ex_data_X509_STORE_idx(), which is used in
13605 ssl_verify_cert_chain() and thus can be called at any time
13606 during TLS/SSL handshakes so that thread-safety is essential.
13607 Unfortunately, the ex_data design is not at all suited
13608 for multi-threaded use, so it probably should be abolished.
13612 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13614 *Broadcom, tweaked and integrated by Geoff Thorpe*
13616 * Move common extension printing code to new function
13617 X509V3_print_extensions(). Reorganise OCSP print routines and
13618 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13622 * New function X509_signature_print() to remove duplication in some
13627 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13628 set (this was treated exactly the same as SET OF previously). This
13629 is used to reorder the STACK representing the structure to match the
13630 encoding. This will be used to get round a problem where a PKCS7
13631 structure which was signed could not be verified because the STACK
13632 order did not reflect the encoded order.
13636 * Reimplement the OCSP ASN1 module using the new code.
13640 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13641 for its ASN1 operations. The old style function pointers still exist
13642 for now but they will eventually go away.
13646 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13647 completely replaces the old ASN1 functionality with a table driven
13648 encoder and decoder which interprets an ASN1_ITEM structure describing
13649 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13650 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13651 has also been converted to the new form.
13655 * Change BN_mod_exp_recp so that negative moduli are tolerated
13656 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13657 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13658 for negative moduli.
13662 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13663 of not touching the result's sign bit.
13667 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13672 * Changed the LHASH code to use prototypes for callbacks, and created
13673 macros to declare and implement thin (optionally static) functions
13674 that provide type-safety and avoid function pointer casting for the
13675 type-specific callbacks.
13679 * Added Kerberos Cipher Suites to be used with TLS, as written in
13681 *Veers Staats <staatsvr@asc.hpc.mil>,
13682 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13684 * Reformat the FAQ so the different questions and answers can be divided
13685 in sections depending on the subject.
13689 * Have the zlib compression code load ZLIB.DLL dynamically under
13694 * New function BN_mod_sqrt for computing square roots modulo a prime
13695 (using the probabilistic Tonelli-Shanks algorithm unless
13696 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13697 be handled deterministically).
13699 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13701 * Make BN_mod_inverse faster by explicitly handling small quotients
13702 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13703 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13707 * New function BN_kronecker.
13711 * Fix BN_gcd so that it works on negative inputs; the result is
13712 positive unless both parameters are zero.
13713 Previously something reasonably close to an infinite loop was
13714 possible because numbers could be growing instead of shrinking
13715 in the implementation of Euclid's algorithm.
13719 * Fix BN_is_word() and BN_is_one() macros to take into account the
13720 sign of the number in question.
13722 Fix BN_is_word(a,w) to work correctly for w == 0.
13724 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13725 because its test if the absolute value of 'a' equals 'w'.
13726 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13727 it exists mostly for use in the implementations of BN_is_zero(),
13728 BN_is_one(), and BN_is_word().
13732 * New function BN_swap.
13736 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13737 the exponentiation functions are more likely to produce reasonable
13738 results on negative inputs.
13742 * Change BN_mod_mul so that the result is always non-negative.
13743 Previously, it could be negative if one of the factors was negative;
13744 I don't think anyone really wanted that behaviour.
13748 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13749 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13750 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13751 and add new functions:
13760 BN_mod_lshift1_quick
13762 BN_mod_lshift_quick
13764 These functions always generate non-negative results.
13766 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13767 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13769 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13770 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13771 be reduced modulo `m`.
13773 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13776 The following entry accidentally appeared in the CHANGES file
13777 distributed with OpenSSL 0.9.7. The modifications described in
13778 it do *not* apply to OpenSSL 0.9.7.
13780 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13781 was actually never needed) and in BN_mul(). The removal in BN_mul()
13782 required a small change in bn_mul_part_recursive() and the addition
13783 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13784 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13785 bn_sub_words() and bn_add_words() except they take arrays with
13791 * In 'openssl passwd', verify passwords read from the terminal
13792 unless the '-salt' option is used (which usually means that
13793 verification would just waste user's time since the resulting
13794 hash is going to be compared with some given password hash)
13795 or the new '-noverify' option is used.
13797 This is an incompatible change, but it does not affect
13798 non-interactive use of 'openssl passwd' (passwords on the command
13799 line, '-stdin' option, '-in ...' option) and thus should not
13800 cause any problems.
13804 * Remove all references to RSAref, since there's no more need for it.
13808 * Make DSO load along a path given through an environment variable
13809 (SHLIB_PATH) with shl_load().
13813 * Constify the ENGINE code as a result of BIGNUM constification.
13814 Also constify the RSA code and most things related to it. In a
13815 few places, most notable in the depth of the ASN.1 code, ugly
13816 casts back to non-const were required (to be solved at a later
13821 * Make it so the openssl application has all engines loaded by default.
13825 * Constify the BIGNUM routines a little more.
13829 * Add the following functions:
13831 ENGINE_load_cswift()
13833 ENGINE_load_atalla()
13834 ENGINE_load_nuron()
13835 ENGINE_load_builtin_engines()
13837 That way, an application can itself choose if external engines that
13838 are built-in in OpenSSL shall ever be used or not. The benefit is
13839 that applications won't have to be linked with libdl or other dso
13840 libraries unless it's really needed.
13842 Changed 'openssl engine' to load all engines on demand.
13843 Changed the engine header files to avoid the duplication of some
13844 declarations (they differed!).
13848 * 'openssl engine' can now list capabilities.
13852 * Better error reporting in 'openssl engine'.
13856 * Never call load_dh_param(NULL) in s_server.
13860 * Add engine application. It can currently list engines by name and
13861 identity, and test if they are actually available.
13865 * Improve RPM specification file by forcing symbolic linking and making
13866 sure the installed documentation is also owned by root.root.
13868 *Damien Miller <djm@mindrot.org>*
13870 * Give the OpenSSL applications more possibilities to make use of
13871 keys (public as well as private) handled by engines.
13875 * Add OCSP code that comes from CertCo.
13879 * Add VMS support for the Rijndael code.
13883 * Added untested support for Nuron crypto accelerator.
13887 * Add support for external cryptographic devices. This code was
13888 previously distributed separately as the "engine" branch.
13890 *Geoff Thorpe, Richard Levitte*
13892 * Rework the filename-translation in the DSO code. It is now possible to
13893 have far greater control over how a "name" is turned into a filename
13894 depending on the operating environment and any oddities about the
13895 different shared library filenames on each system.
13899 * Support threads on FreeBSD-elf in Configure.
13903 * Fix for SHA1 assembly problem with MASM: it produces
13904 warnings about corrupt line number information when assembling
13905 with debugging information. This is caused by the overlapping
13908 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13911 NCONF_get_number() has no error checking at all. As a replacement,
13912 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13913 promoted strongly. The old NCONF_get_number is kept around for
13914 binary backward compatibility.
13915 Make it possible for methods to load from something other than a BIO,
13916 by providing a function pointer that is given a name instead of a BIO.
13917 For example, this could be used to load configuration data from an
13922 * Fix for non blocking accept BIOs. Added new I/O special reason
13923 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13924 with non blocking I/O was not possible because no retry code was
13925 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13930 * Added the beginnings of Rijndael support.
13934 * Fix for bug in DirectoryString mask setting. Add support for
13935 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13936 to allow certificate printing to more controllable, additional
13937 'certopt' option to 'x509' to allow new printing options to be
13942 * Clean old EAY MD5 hack from e_os.h.
13946 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13948 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13949 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13951 *Joe Orton, Steve Henson*
13953 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13955 * Fix additional bug revealed by the NISCC test suite:
13957 Stop bug triggering large recursion when presented with
13958 certain ASN.1 tags ([CVE-2003-0851])
13962 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13964 * Fix various bugs revealed by running the NISCC test suite:
13966 Stop out of bounds reads in the ASN1 code when presented with
13967 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13969 If verify callback ignores invalid public key errors don't try to check
13970 certificate signature with the NULL public key.
13974 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13975 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13980 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13981 extra data after the compression methods not only for TLS 1.0
13982 but also for SSL 3.0 (as required by the specification).
13984 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13986 * Change X509_certificate_type() to mark the key as exported/exportable
13987 when it's 512 *bits* long, not 512 bytes.
13991 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13993 * Countermeasure against the Klima-Pokorny-Rosa extension of
13994 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13995 a protocol version number mismatch like a decryption error
13996 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14000 * Turn on RSA blinding by default in the default implementation
14001 to avoid a timing attack. Applications that don't want it can call
14002 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14003 They would be ill-advised to do so in most cases.
14005 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14007 * Change RSA blinding code so that it works when the PRNG is not
14008 seeded (in this case, the secret RSA exponent is abused as
14009 an unpredictable seed -- if it is not unpredictable, there
14010 is no point in blinding anyway). Make RSA blinding thread-safe
14011 by remembering the creator's thread ID in rsa->blinding and
14012 having all other threads use local one-time blinding factors
14013 (this requires more computation than sharing rsa->blinding, but
14014 avoids excessive locking; and if an RSA object is not shared
14015 between threads, blinding will still be very fast).
14019 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14021 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14022 via timing by performing a MAC computation even if incorrect
14023 block cipher padding has been found. This is a countermeasure
14024 against active attacks where the attacker has to distinguish
14025 between bad padding and a MAC verification error. ([CVE-2003-0078])
14027 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14028 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14029 Martin Vuagnoux (EPFL, Ilion)*
14031 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14033 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14034 memory from its contents. This is done with a counter that will
14035 place alternating values in each byte. This can be used to solve
14036 two issues: 1) the removal of calls to memset() by highly optimizing
14037 compilers, and 2) cleansing with other values than 0, since those can
14038 be read through on certain media, for example a swap space on disk.
14042 * Bugfix: client side session caching did not work with external caching,
14043 because the session->cipher setting was not restored when reloading
14044 from the external cache. This problem was masked, when
14045 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14046 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14050 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14051 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14053 *Zeev Lieber <zeev-l@yahoo.com>*
14055 * Undo an undocumented change introduced in 0.9.6e which caused
14056 repeated calls to OpenSSL_add_all_ciphers() and
14057 OpenSSL_add_all_digests() to be ignored, even after calling
14062 * Change the default configuration reader to deal with last line not
14063 being properly terminated.
14067 * Change X509_NAME_cmp() so it applies the special rules on handling
14068 DN values that are of type PrintableString, as well as RDNs of type
14069 emailAddress where the value has the type ia5String.
14071 *stefank@valicert.com via Richard Levitte*
14073 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14074 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14075 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14076 the bitwise-OR of the two for use by the majority of applications
14077 wanting this behaviour, and update the docs. The documented
14078 behaviour and actual behaviour were inconsistent and had been
14079 changing anyway, so this is more a bug-fix than a behavioural
14082 *Geoff Thorpe, diagnosed by Nadav Har'El*
14084 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14085 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14089 * Fix initialization code race conditions in
14090 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14091 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14092 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14093 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14094 ssl2_get_cipher_by_char(),
14095 ssl3_get_cipher_by_char().
14097 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14099 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14100 the cached sessions are flushed, as the remove_cb() might use ex_data
14101 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14102 (see [openssl.org #212]).
14104 *Geoff Thorpe, Lutz Jaenicke*
14106 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14107 length, instead of the encoding length to d2i_ASN1_OBJECT.
14111 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14113 * [In 0.9.6g-engine release:]
14114 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14116 *Lynn Gazis <lgazis@rainbow.com>*
14118 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14120 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14121 and get fix the header length calculation.
14122 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14123 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14125 * Use proper error handling instead of 'assertions' in buffer
14126 overflow checks added in 0.9.6e. This prevents DoS (the
14127 assertions could call abort()).
14129 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14131 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14133 * Add various sanity checks to asn1_get_length() to reject
14134 the ASN1 length bytes if they exceed sizeof(long), will appear
14135 negative or the content length exceeds the length of the
14138 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14140 * Fix cipher selection routines: ciphers without encryption had no flags
14141 for the cipher strength set and where therefore not handled correctly
14142 by the selection routines (PR #130).
14146 * Fix EVP_dsa_sha macro.
14151 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14152 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14153 that was added in OpenSSL 0.9.6d.
14155 As the countermeasure turned out to be incompatible with some
14156 broken SSL implementations, the new option is part of SSL_OP_ALL.
14157 SSL_OP_ALL is usually employed when compatibility with weird SSL
14158 implementations is desired (e.g. '-bugs' option to 's_client' and
14159 's_server'), so the new option is automatically set in many
14164 * Changes in security patch:
14166 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14167 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14168 Air Force Materiel Command, USAF, under agreement number
14171 * Add various sanity checks to asn1_get_length() to reject
14172 the ASN1 length bytes if they exceed sizeof(long), will appear
14173 negative or the content length exceeds the length of the
14174 supplied buffer. ([CVE-2002-0659])
14176 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14178 * Assertions for various potential buffer overflows, not known to
14179 happen in practice.
14181 *Ben Laurie (CHATS)*
14183 * Various temporary buffers to hold ASCII versions of integers were
14184 too small for 64 bit platforms. ([CVE-2002-0655])
14185 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14187 * Remote buffer overflow in SSL3 protocol - an attacker could
14188 supply an oversized session ID to a client. ([CVE-2002-0656])
14190 *Ben Laurie (CHATS)*
14192 * Remote buffer overflow in SSL2 protocol - an attacker could
14193 supply an oversized client master key. ([CVE-2002-0656])
14195 *Ben Laurie (CHATS)*
14197 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14199 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14200 encoded as NULL) with id-dsa-with-sha1.
14202 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14204 * Check various `X509_...()` return values in `apps/req.c`.
14206 *Nils Larsch <nla@trustcenter.de>*
14208 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14209 an end-of-file condition would erroneously be flagged, when the CRLF
14210 was just at the end of a processed block. The bug was discovered when
14211 processing data through a buffering memory BIO handing the data to a
14212 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14213 <ptsekov@syntrex.com> and Nedelcho Stanev.
14217 * Implement a countermeasure against a vulnerability recently found
14218 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14219 before application data chunks to avoid the use of known IVs
14220 with data potentially chosen by the attacker.
14224 * Fix length checks in ssl3_get_client_hello().
14228 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14229 to prevent ssl3_read_internal() from incorrectly assuming that
14230 ssl3_read_bytes() found application data while handshake
14231 processing was enabled when in fact s->s3->in_read_app_data was
14232 merely automatically cleared during the initial handshake.
14234 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14236 * Fix object definitions for Private and Enterprise: they were not
14237 recognized in their shortname (=lowercase) representation. Extend
14238 obj_dat.pl to issue an error when using undefined keywords instead
14239 of silently ignoring the problem (Svenning Sorensen
14240 <sss@sss.dnsalias.net>).
14244 * Fix DH_generate_parameters() so that it works for 'non-standard'
14245 generators, i.e. generators other than 2 and 5. (Previously, the
14246 code did not properly initialise the 'add' and 'rem' values to
14247 BN_generate_prime().)
14249 In the new general case, we do not insist that 'generator' is
14250 actually a primitive root: This requirement is rather pointless;
14251 a generator of the order-q subgroup is just as good, if not
14256 * Map new X509 verification errors to alerts. Discovered and submitted by
14257 Tom Wu <tom@arcot.com>.
14261 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14262 returning non-zero before the data has been completely received
14263 when using non-blocking I/O.
14265 *Bodo Moeller; problem pointed out by John Hughes*
14267 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14269 *Ben Laurie, Lutz Jaenicke*
14271 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14272 Yoram Zahavi <YoramZ@gilian.com>).
14276 * Add information about CygWin 1.3 and on, and preserve proper
14277 configuration for the versions before that.
14279 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14281 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14282 check whether we deal with a copy of a session and do not delete from
14283 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14284 <izhar@checkpoint.com>.
14288 * Do not store session data into the internal session cache, if it
14289 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14290 flag is set). Proposed by Aslam <aslam@funk.com>.
14294 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14299 * [In 0.9.6d-engine release:]
14300 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14302 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14304 * Add the configuration target linux-s390x.
14306 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14308 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14309 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14310 variable as an indication that a ClientHello message has been
14311 received. As the flag value will be lost between multiple
14312 invocations of ssl3_accept when using non-blocking I/O, the
14313 function may not be aware that a handshake has actually taken
14314 place, thus preventing a new session from being added to the
14317 To avoid this problem, we now set s->new_session to 2 instead of
14318 using a local variable.
14320 *Lutz Jaenicke, Bodo Moeller*
14322 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14323 if the SSL_R_LENGTH_MISMATCH error is detected.
14325 *Geoff Thorpe, Bodo Moeller*
14327 * New 'shared_ldflag' column in Configure platform table.
14331 * Fix EVP_CIPHER_mode macro.
14333 *"Dan S. Camper" <dan@bti.net>*
14335 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14336 type, we must throw them away by setting rr->length to 0.
14338 *D P Chang <dpc@qualys.com>*
14340 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14342 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14343 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14344 worked incorrectly for those cases where range = `10..._2` and
14345 `3*range` is two bits longer than range.)
14349 * Only add signing time to PKCS7 structures if it is not already
14354 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14355 OBJ_ld_ce should be OBJ_id_ce.
14356 Also some ip-pda OIDs in crypto/objects/objects.txt were
14357 incorrect (cf. RFC 3039).
14359 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14361 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14362 returns early because it has nothing to do.
14364 *Andy Schneider <andy.schneider@bjss.co.uk>*
14366 * [In 0.9.6c-engine release:]
14367 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14369 *Andy Schneider <andy.schneider@bjss.co.uk>*
14371 * [In 0.9.6c-engine release:]
14372 Add support for Cryptographic Appliance's keyserver technology.
14373 (Use engine 'keyclient')
14375 *Cryptographic Appliances and Geoff Thorpe*
14377 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14378 is called via tools/c89.sh because arguments have to be
14379 rearranged (all '-L' options must appear before the first object
14382 *Richard Shapiro <rshapiro@abinitio.com>*
14384 * [In 0.9.6c-engine release:]
14385 Add support for Broadcom crypto accelerator cards, backported
14388 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14390 * [In 0.9.6c-engine release:]
14391 Add support for SureWare crypto accelerator cards from
14392 Baltimore Technologies. (Use engine 'sureware')
14394 *Baltimore Technologies and Mark Cox*
14396 * [In 0.9.6c-engine release:]
14397 Add support for crypto accelerator cards from Accelerated
14398 Encryption Processing, www.aep.ie. (Use engine 'aep')
14400 *AEP Inc. and Mark Cox*
14402 * Add a configuration entry for gcc on UnixWare.
14404 *Gary Benson <gbenson@redhat.com>*
14406 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14407 messages are stored in a single piece (fixed-length part and
14408 variable-length part combined) and fix various bugs found on the way.
14412 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14413 instead. BIO_gethostbyname() does not know what timeouts are
14414 appropriate, so entries would stay in cache even when they have
14416 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14418 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14419 faced with a pathologically small ClientHello fragment that does
14420 not contain client_version: Instead of aborting with an error,
14421 simply choose the highest available protocol version (i.e.,
14422 TLS 1.0 unless it is disabled). In practice, ClientHello
14423 messages are never sent like this, but this change gives us
14424 strictly correct behaviour at least for TLS.
14428 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14429 never resets s->method to s->ctx->method when called from within
14430 one of the SSL handshake functions.
14432 *Bodo Moeller; problem pointed out by Niko Baric*
14434 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14435 (sent using the client's version number) if client_version is
14436 smaller than the protocol version in use. Also change
14437 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14438 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14439 the client will at least see that alert.
14443 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14448 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14449 client receives HelloRequest while in a handshake.
14451 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14453 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14454 should end in 'break', not 'goto end' which circumvents various
14455 cleanups done in state SSL_ST_OK. But session related stuff
14456 must be disabled for SSL_ST_OK in the case that we just sent a
14459 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14460 before just sending a HelloRequest.
14462 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14464 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14465 reveal whether illegal block cipher padding was found or a MAC
14466 verification error occurred. (Neither SSLerr() codes nor alerts
14467 are directly visible to potential attackers, but the information
14468 may leak via logfiles.)
14470 Similar changes are not required for the SSL 2.0 implementation
14471 because the number of padding bytes is sent in clear for SSL 2.0,
14472 and the extra bytes are just ignored. However ssl/s2_pkt.c
14473 failed to verify that the purported number of padding bytes is in
14478 * Add OpenUNIX-8 support including shared libraries
14479 (Boyd Lynn Gerber <gerberb@zenez.com>).
14483 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14484 'wristwatch attack' using huge encoding parameters (cf.
14485 James H. Manger's CRYPTO 2001 paper). Note that the
14486 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14487 encoding parameters and hence was not vulnerable.
14491 * BN_sqr() bug fix.
14493 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14495 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14496 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14497 followed by modular reduction.
14499 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14501 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14502 equivalent based on BN_pseudo_rand() instead of BN_rand().
14506 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14507 This function was broken, as the check for a new client hello message
14508 to handle SGC did not allow these large messages.
14509 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14513 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14517 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14518 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14522 * Rework the configuration and shared library support for Tru64 Unix.
14523 The configuration part makes use of modern compiler features and
14524 still retains old compiler behavior for those that run older versions
14525 of the OS. The shared library support part includes a variant that
14526 uses the RPATH feature, and is available through the special
14527 configuration target "alpha-cc-rpath", which will never be selected
14530 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14532 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14533 with the same message size as in ssl3_get_certificate_request().
14534 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14535 messages might inadvertently be reject as too long.
14537 *Petr Lampa <lampa@fee.vutbr.cz>*
14539 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14543 * Modified SSL library such that the verify_callback that has been set
14544 specifically for an SSL object with SSL_set_verify() is actually being
14545 used. Before the change, a verify_callback set with this function was
14546 ignored and the verify_callback() set in the SSL_CTX at the time of
14547 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14548 to allow the necessary settings.
14552 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14553 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14554 done automatically (in contradiction to the requirements of the C
14555 standard). This made problems when used from OpenSSH.
14559 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14560 dh->length and always used
14562 BN_rand_range(priv_key, dh->p).
14564 BN_rand_range() is not necessary for Diffie-Hellman, and this
14565 specific range makes Diffie-Hellman unnecessarily inefficient if
14566 dh->length (recommended exponent length) is much smaller than the
14567 length of dh->p. We could use BN_rand_range() if the order of
14568 the subgroup was stored in the DH structure, but we only have
14573 BN_rand(priv_key, l, ...)
14575 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14582 RSA_eay_public_encrypt
14583 RSA_eay_private_decrypt
14584 RSA_eay_private_encrypt (signing)
14585 RSA_eay_public_decrypt (signature verification)
14587 (default implementations for RSA_public_encrypt,
14588 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14589 always reject numbers >= n.
14593 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14594 to synchronize access to 'locking_thread'. This is necessary on
14595 systems where access to 'locking_thread' (an 'unsigned long'
14596 variable) is not atomic.
14600 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14601 *before* setting the 'crypto_lock_rand' flag. The previous code had
14602 a race condition if 0 is a valid thread ID.
14604 *Travis Vitek <vitek@roguewave.com>*
14606 * Add support for shared libraries under Irix.
14608 *Albert Chin-A-Young <china@thewrittenword.com>*
14610 * Add configuration option to build on Linux on both big-endian and
14611 little-endian MIPS.
14613 *Ralf Baechle <ralf@uni-koblenz.de>*
14615 * Add the possibility to create shared libraries on HP-UX.
14619 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14621 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14622 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14623 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14624 PRNG state recovery was possible based on the output of
14625 one PRNG request appropriately sized to gain knowledge on
14626 'md' followed by enough consecutive 1-byte PRNG requests
14627 to traverse all of 'state'.
14629 1. When updating 'md_local' (the current thread's copy of 'md')
14630 during PRNG output generation, hash all of the previous
14631 'md_local' value, not just the half used for PRNG output.
14633 2. Make the number of bytes from 'state' included into the hash
14634 independent from the number of PRNG bytes requested.
14636 The first measure alone would be sufficient to avoid
14637 Markku-Juhani's attack. (Actually it had never occurred
14638 to me that the half of 'md_local' used for chaining was the
14639 half from which PRNG output bytes were taken -- I had always
14640 assumed that the secret half would be used.) The second
14641 measure makes sure that additional data from 'state' is never
14642 mixed into 'md_local' in small portions; this heuristically
14643 further strengthens the PRNG.
14647 * Fix crypto/bn/asm/mips3.s.
14651 * When only the key is given to "enc", the IV is undefined. Print out
14652 an error message in this case.
14656 * Handle special case when X509_NAME is empty in X509 printing routines.
14660 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14661 positive and less than q.
14665 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14666 used: it isn't thread safe and the add_lock_callback should handle
14669 *Paul Rose <Paul.Rose@bridge.com>*
14671 * Verify that incoming data obeys the block size in
14672 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14678 *Ulf Möller, Bodo Möller*
14680 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14681 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14682 when fixing the server behaviour for backwards-compatible 'client
14683 hello' messages. (Note that the attack is impractical against
14684 SSL 3.0 and TLS 1.0 anyway because length and version checking
14685 means that the probability of guessing a valid ciphertext is
14686 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14689 Before 0.9.5, the countermeasure (hide the error by generating a
14690 random 'decryption result') did not work properly because
14691 ERR_clear_error() was missing, meaning that SSL_get_error() would
14692 detect the supposedly ignored error.
14694 Both problems are now fixed.
14698 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14699 (previously it was 1024).
14703 * Fix for compatibility mode trust settings: ignore trust settings
14704 unless some valid trust or reject settings are present.
14708 * Fix for blowfish EVP: its a variable length cipher.
14712 * Fix various bugs related to DSA S/MIME verification. Handle missing
14713 parameters in DSA public key structures and return an error in the
14714 DSA routines if parameters are absent.
14718 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14719 in the current directory if neither $RANDFILE nor $HOME was set.
14720 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14721 caused some confusion to Windows users who haven't defined $HOME.
14722 Thus RAND_file_name() is changed again: e_os.h can define a
14723 DEFAULT_HOME, which will be used if $HOME is not set.
14724 For Windows, we use "C:"; on other platforms, we still require
14725 environment variables.
14727 * Move 'if (!initialized) RAND_poll()' into regions protected by
14728 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14729 having multiple threads call RAND_poll() concurrently.
14733 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14734 combination of a flag and a thread ID variable.
14735 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14736 flag), *other* threads can enter ssleay_add_bytes without obeying
14737 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14738 that they do not hold after the first thread unsets add_do_not_lock).
14742 * Change bctest again: '-x' expressions are not available in all
14743 versions of 'test'.
14747 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14749 * Fix a couple of memory leaks in PKCS7_dataDecode()
14751 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14753 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14754 the default extension for executables, if any. Also, make the perl
14755 scripts that use symlink() to test if it really exists and use "cp"
14756 if it doesn't. All this made OpenSSL compilable and installable in
14761 * Fix for asn1_GetSequence() for indefinite length constructed data.
14762 If SEQUENCE is length is indefinite just set c->slen to the total
14763 amount of data available.
14765 *Steve Henson, reported by shige@FreeBSD.org*
14767 *This change does not apply to 0.9.7.*
14769 * Change bctest to avoid here-documents inside command substitution
14770 (workaround for FreeBSD /bin/sh bug).
14771 For compatibility with Ultrix, avoid shell functions (introduced
14772 in the bctest version that searches along $PATH).
14776 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14777 with des_encrypt() defined on some operating systems, like Solaris
14782 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14783 On the Importance of Eliminating Errors in Cryptographic
14784 Computations, J. Cryptology 14 (2001) 2, 101-119,
14785 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14789 * MIPS assembler BIGNUM division bug fix.
14793 * Disabled incorrect Alpha assembler code.
14797 * Fix PKCS#7 decode routines so they correctly update the length
14798 after reading an EOC for the EXPLICIT tag.
14802 *This change does not apply to 0.9.7.*
14804 * Fix bug in PKCS#12 key generation routines. This was triggered
14805 if a 3DES key was generated with a 0 initial byte. Include
14806 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14807 (but broken) behaviour.
14811 * Enhance bctest to search for a working bc along $PATH and print
14814 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14816 * Fix memory leaks in err.c: free err_data string if necessary;
14817 don't write to the wrong index in ERR_set_error_data.
14821 * Implement ssl23_peek (analogous to ssl23_read), which previously
14826 * Replace rdtsc with `_emit` statements for VC++ version 5.
14828 *Jeremy Cooper <jeremy@baymoo.org>*
14830 * Make it possible to reuse SSLv2 sessions.
14834 * In copy_email() check for >= 0 as a return value for
14835 X509_NAME_get_index_by_NID() since 0 is a valid index.
14837 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14839 * Avoid coredump with unsupported or invalid public keys by checking if
14840 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14841 PKCS7_verify() fails with non detached data.
14845 * Don't use getenv in library functions when run as setuid/setgid.
14846 New function OPENSSL_issetugid().
14850 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14851 due to incorrect handling of multi-threading:
14853 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14855 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14857 3. Count how many times MemCheck_off() has been called so that
14858 nested use can be treated correctly. This also avoids
14859 inband-signalling in the previous code (which relied on the
14860 assumption that thread ID 0 is impossible).
14864 * Add "-rand" option also to s_client and s_server.
14868 * Fix CPU detection on Irix 6.x.
14869 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14870 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14872 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14877 *This change does not apply to 0.9.7.*
14879 * Use the cached encoding of an X509_NAME structure rather than
14880 copying it. This is apparently the reason for the libsafe "errors"
14881 but the code is actually correct.
14885 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14886 Bleichenbacher's DSA attack.
14887 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14888 to be set and top=0 forces the highest bit to be set; top=-1 is new
14889 and leaves the highest bit random.
14891 *Ulf Moeller, Bodo Moeller*
14893 * In the `NCONF_...`-based implementations for `CONF_...` queries
14894 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14895 a temporary CONF structure with the data component set to NULL
14896 (which gives segmentation faults in lh_retrieve).
14897 Instead, use NULL for the CONF pointer in CONF_get_string and
14898 CONF_get_number (which may use environment variables) and directly
14899 return NULL from CONF_get_section.
14903 * Fix potential buffer overrun for EBCDIC.
14907 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14908 keyUsage if basicConstraints absent for a CA.
14912 * Make SMIME_write_PKCS7() write mail header values with a format that
14913 is more generally accepted (no spaces before the semicolon), since
14914 some programs can't parse those values properly otherwise. Also make
14915 sure BIO's that break lines after each write do not create invalid
14920 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14921 macros previously used would not encode an empty SEQUENCE OF
14922 and break the signature.
14926 *This change does not apply to 0.9.7.*
14928 * Zero the premaster secret after deriving the master secret in
14933 * Add some EVP_add_digest_alias registrations (as found in
14934 OpenSSL_add_all_digests()) to SSL_library_init()
14935 aka OpenSSL_add_ssl_algorithms(). This provides improved
14936 compatibility with peers using X.509 certificates
14937 with unconventional AlgorithmIdentifier OIDs.
14941 * Fix for Irix with NO_ASM.
14943 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14945 * ./config script fixes.
14947 *Ulf Moeller, Richard Levitte*
14949 * Fix 'openssl passwd -1'.
14953 * Change PKCS12_key_gen_asc() so it can cope with non null
14954 terminated strings whose length is passed in the passlen
14955 parameter, for example from PEM callbacks. This was done
14956 by adding an extra length parameter to asc2uni().
14958 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14960 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14961 call failed, free the DSA structure.
14965 * Fix to uni2asc() to cope with zero length Unicode strings.
14966 These are present in some PKCS#12 files.
14970 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14971 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14972 when writing a 32767 byte record.
14974 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14976 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14977 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14979 (RSA objects have a reference count access to which is protected
14980 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14981 so they are meant to be shared between threads.)
14982 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14983 "Reddie, Steven" <Steven.Reddie@ca.com>*
14985 * Fix a deadlock in CRYPTO_mem_leaks().
14989 * Use better test patterns in bntest.
14993 * rand_win.c fix for Borland C.
14997 * BN_rshift bugfix for n == 0.
15001 * Add a 'bctest' script that checks for some known 'bc' bugs
15002 so that 'make test' does not abort just because 'bc' is broken.
15006 * Store verify_result within SSL_SESSION also for client side to
15007 avoid potential security hole. (Re-used sessions on the client side
15008 always resulted in verify_result==X509_V_OK, not using the original
15009 result of the server certificate verification.)
15013 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15014 SSL3_RT_APPLICATION_DATA, return 0.
15015 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15020 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15021 releases, have been re-implemented by renaming the previous
15022 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15023 and ssl3_read_internal, respectively, and adding 'peek' parameters
15024 to them. The new ssl[23]_{read,peek} functions are calls to
15025 ssl[23]_read_internal with the 'peek' flag set appropriately.
15026 A 'peek' parameter has also been added to ssl3_read_bytes, which
15027 does the actual work for ssl3_read_internal.
15031 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15032 the method-specific "init()" handler. Also clean up ex_data after
15033 calling the method-specific "finish()" handler. Previously, this was
15034 happening the other way round.
15038 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15039 The previous value, 12, was not always sufficient for BN_mod_exp().
15043 * Make sure that shared libraries get the internal name engine with
15044 the full version number and not just 0. This should mark the
15045 shared libraries as not backward compatible. Of course, this should
15046 be changed again when we can guarantee backward binary compatibility.
15050 * Fix typo in get_cert_by_subject() in by_dir.c
15052 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15054 * Rework the system to generate shared libraries:
15056 - Make note of the expected extension for the shared libraries and
15057 if there is a need for symbolic links from for example libcrypto.so.0
15058 to libcrypto.so.0.9.7. There is extended info in Configure for
15061 - Make as few rebuilds of the shared libraries as possible.
15063 - Still avoid linking the OpenSSL programs with the shared libraries.
15065 - When installing, install the shared libraries separately from the
15070 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15072 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15073 and not in SSL_clear because the latter is also used by the
15074 accept/connect functions; previously, the settings made by
15075 SSL_set_read_ahead would be lost during the handshake.
15077 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15079 * Correct util/mkdef.pl to be selective about disabled algorithms.
15080 Previously, it would create entries for disabled algorithms no
15085 * Added several new manual pages for SSL_* function.
15089 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15091 * In ssl23_get_client_hello, generate an error message when faced
15092 with an initial SSL 3.0/TLS record that is too small to contain the
15093 first two bytes of the ClientHello message, i.e. client_version.
15094 (Note that this is a pathologic case that probably has never happened
15095 in real life.) The previous approach was to use the version number
15096 from the record header as a substitute; but our protocol choice
15097 should not depend on that one because it is not authenticated
15098 by the Finished messages.
15102 * More robust randomness gathering functions for Windows.
15104 *Jeffrey Altman <jaltman@columbia.edu>*
15106 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15107 not set then we don't setup the error code for issuer check errors
15108 to avoid possibly overwriting other errors which the callback does
15109 handle. If an application does set the flag then we assume it knows
15110 what it is doing and can handle the new informational codes
15115 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15116 a general "ANY" type, as such it should be able to decode anything
15117 including tagged types. However it didn't check the class so it would
15118 wrongly interpret tagged types in the same way as their universal
15119 counterpart and unknown types were just rejected. Changed so that the
15120 tagged and unknown types are handled in the same way as a SEQUENCE:
15121 that is the encoding is stored intact. There is also a new type
15122 "V_ASN1_OTHER" which is used when the class is not universal, in this
15123 case we have no idea what the actual type is so we just lump them all
15128 * On VMS, stdout may very well lead to a file that is written to
15129 in a record-oriented fashion. That means that every write() will
15130 write a separate record, which will be read separately by the
15131 programs trying to read from it. This can be very confusing.
15133 The solution is to put a BIO filter in the way that will buffer
15134 text until a linefeed is reached, and then write everything a
15135 line at a time, so every record written will be an actual line,
15136 not chunks of lines and not (usually doesn't happen, but I've
15137 seen it once) several lines in one record. BIO_f_linebuffer() is
15140 Currently, it's a VMS-only method, because that's where it has
15141 been tested well enough.
15145 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15146 it can return incorrect results.
15147 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15148 but it was in 0.9.6-beta[12].)
15152 * Disable the check for content being present when verifying detached
15153 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15154 include zero length content when signing messages.
15158 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15159 BIO_ctrl (for BIO pairs).
15163 * Add DSO method for VMS.
15167 * Bug fix: Montgomery multiplication could produce results with the
15172 * Add RPM specification openssl.spec and modify it to build three
15173 packages. The default package contains applications, application
15174 documentation and run-time libraries. The devel package contains
15175 include files, static libraries and function documentation. The
15176 doc package contains the contents of the doc directory. The original
15177 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15181 * Add a large number of documentation files for many SSL routines.
15183 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15185 * Add a configuration entry for Sony News 4.
15187 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15189 * Don't set the two most significant bits to one when generating a
15190 random number < q in the DSA library.
15194 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15195 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15196 the underlying transport is blocking) if a handshake took place.
15197 (The default behaviour is needed by applications such as s_client
15198 and s_server that use select() to determine when to use SSL_read;
15199 but for applications that know in advance when to expect data, it
15200 just makes things more complicated.)
15204 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15209 * Add a few more EBCDIC conditionals that make `req` and `x509`
15210 work better on such systems.
15212 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15214 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15215 Update PKCS12_parse() so it copies the friendlyName and the
15216 keyid to the certificates aux info.
15220 * Fix bug in PKCS7_verify() which caused an infinite loop
15221 if there was more than one signature.
15223 *Sven Uszpelkat <su@celocom.de>*
15225 * Major change in util/mkdef.pl to include extra information
15226 about each symbol, as well as presenting variables as well
15227 as functions. This change means that there's n more need
15228 to rebuild the .num files when some algorithms are excluded.
15232 * Allow the verify time to be set by an application,
15233 rather than always using the current time.
15237 * Phase 2 verify code reorganisation. The certificate
15238 verify code now looks up an issuer certificate by a
15239 number of criteria: subject name, authority key id
15240 and key usage. It also verifies self signed certificates
15241 by the same criteria. The main comparison function is
15242 X509_check_issued() which performs these checks.
15244 Lot of changes were necessary in order to support this
15245 without completely rewriting the lookup code.
15247 Authority and subject key identifier are now cached.
15249 The LHASH 'certs' is X509_STORE has now been replaced
15250 by a STACK_OF(X509_OBJECT). This is mainly because an
15251 LHASH can't store or retrieve multiple objects with
15252 the same hash value.
15254 As a result various functions (which were all internal
15255 use only) have changed to handle the new X509_STORE
15256 structure. This will break anything that messed round
15257 with X509_STORE internally.
15259 The functions X509_STORE_add_cert() now checks for an
15260 exact match, rather than just subject name.
15262 The X509_STORE API doesn't directly support the retrieval
15263 of multiple certificates matching a given criteria, however
15264 this can be worked round by performing a lookup first
15265 (which will fill the cache with candidate certificates)
15266 and then examining the cache for matches. This is probably
15267 the best we can do without throwing out X509_LOOKUP
15268 entirely (maybe later...).
15270 The X509_VERIFY_CTX structure has been enhanced considerably.
15272 All certificate lookup operations now go via a get_issuer()
15273 callback. Although this currently uses an X509_STORE it
15274 can be replaced by custom lookups. This is a simple way
15275 to bypass the X509_STORE hackery necessary to make this
15276 work and makes it possible to use more efficient techniques
15277 in future. A very simple version which uses a simple
15278 STACK for its trusted certificate store is also provided
15279 using X509_STORE_CTX_trusted_stack().
15281 The verify_cb() and verify() callbacks now have equivalents
15282 in the X509_STORE_CTX structure.
15284 X509_STORE_CTX also has a 'flags' field which can be used
15285 to customise the verify behaviour.
15289 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15290 excludes S/MIME capabilities.
15294 * When a certificate request is read in keep a copy of the
15295 original encoding of the signed data and use it when outputting
15296 again. Signatures then use the original encoding rather than
15297 a decoded, encoded version which may cause problems if the
15298 request is improperly encoded.
15302 * For consistency with other BIO_puts implementations, call
15303 buffer_write(b, ...) directly in buffer_puts instead of calling
15306 In BIO_puts, increment b->num_write as in BIO_write.
15308 *Peter.Sylvester@EdelWeb.fr*
15310 * Fix BN_mul_word for the case where the word is 0. (We have to use
15311 BN_zero, we may not return a BIGNUM with an array consisting of
15312 words set to zero.)
15316 * Avoid calling abort() from within the library when problems are
15317 detected, except if preprocessor symbols have been defined
15318 (such as REF_CHECK, BN_DEBUG etc.).
15322 * New openssl application 'rsautl'. This utility can be
15323 used for low-level RSA operations. DER public key
15324 BIO/fp routines also added.
15328 * New Configure entry and patches for compiling on QNX 4.
15330 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15332 * A demo state-machine implementation was sponsored by
15333 Nuron (<http://www.nuron.com/>) and is now available in
15334 demos/state_machine.
15338 * New options added to the 'dgst' utility for signature
15339 generation and verification.
15343 * Unrecognized PKCS#7 content types are now handled via a
15344 catch all ASN1_TYPE structure. This allows unsupported
15345 types to be stored as a "blob" and an application can
15346 encode and decode it manually.
15350 * Fix various signed/unsigned issues to make a_strex.c
15351 compile under VC++.
15353 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15355 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15356 length if passed a buffer. ASN1_INTEGER_to_BN failed
15357 if passed a NULL BN and its argument was negative.
15359 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15361 * Modification to PKCS#7 encoding routines to output definite
15362 length encoding. Since currently the whole structures are in
15363 memory there's not real point in using indefinite length
15364 constructed encoding. However if OpenSSL is compiled with
15365 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15369 * Added BIO_vprintf() and BIO_vsnprintf().
15373 * Added more prefixes to parse for in the strings written
15374 through a logging bio, to cover all the levels that are available
15375 through syslog. The prefixes are now:
15377 PANIC, EMERG, EMR => LOG_EMERG
15378 ALERT, ALR => LOG_ALERT
15379 CRIT, CRI => LOG_CRIT
15380 ERROR, ERR => LOG_ERR
15381 WARNING, WARN, WAR => LOG_WARNING
15382 NOTICE, NOTE, NOT => LOG_NOTICE
15383 INFO, INF => LOG_INFO
15384 DEBUG, DBG => LOG_DEBUG
15386 and as before, if none of those prefixes are present at the
15387 beginning of the string, LOG_ERR is chosen.
15389 On Win32, the `LOG_*` levels are mapped according to this:
15391 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15392 LOG_WARNING => EVENTLOG_WARNING_TYPE
15393 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15397 * Made it possible to reconfigure with just the configuration
15398 argument "reconf" or "reconfigure". The command line arguments
15399 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15400 and are retrieved from there when reconfiguring.
15406 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15408 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15412 * The obj_dat.pl script was messing up the sorting of object
15413 names. The reason was that it compared the quoted version
15414 of strings as a result "OCSP" > "OCSP Signing" because
15415 " > SPACE. Changed script to store unquoted versions of
15416 names and add quotes on output. It was also omitting some
15417 names from the lookup table if they were given a default
15418 value (that is if SN is missing it is given the same
15419 value as LN and vice versa), these are now added on the
15420 grounds that if an object has a name we should be able to
15421 look it up. Finally added warning output when duplicate
15422 short or long names are found.
15426 * Changes needed for Tandem NSK.
15428 *Scott Uroff <scott@xypro.com>*
15430 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15431 RSA_padding_check_SSLv23(), special padding was never detected
15432 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15433 version rollback attacks was not effective.
15435 In s23_clnt.c, don't use special rollback-attack detection padding
15436 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15437 client; similarly, in s23_srvr.c, don't do the rollback check if
15438 SSL 2.0 is the only protocol enabled in the server.
15442 * Make it possible to get hexdumps of unprintable data with 'openssl
15443 asn1parse'. By implication, the functions ASN1_parse_dump() and
15444 BIO_dump_indent() are added.
15448 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15449 these print out strings and name structures based on various
15450 flags including RFC2253 support and proper handling of
15451 multibyte characters. Added options to the 'x509' utility
15452 to allow the various flags to be set.
15456 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15457 Also change the functions X509_cmp_current_time() and
15458 X509_gmtime_adj() work with an ASN1_TIME structure,
15459 this will enable certificates using GeneralizedTime in validity
15460 dates to be checked.
15464 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15465 negative public key encodings) on by default,
15466 NO_NEG_PUBKEY_BUG can be set to disable it.
15470 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15471 content octets. An i2c_ASN1_OBJECT is unnecessary because
15472 the encoding can be trivially obtained from the structure.
15476 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15477 not read locks (`CRYPTO_r_[un]lock`).
15481 * A first attempt at creating official support for shared
15482 libraries through configuration. I've kept it so the
15483 default is static libraries only, and the OpenSSL programs
15484 are always statically linked for now, but there are
15485 preparations for dynamic linking in place.
15486 This has been tested on Linux and Tru64.
15490 * Randomness polling function for Win9x, as described in:
15491 Peter Gutmann, Software Generation of Practically Strong
15496 * Fix so PRNG is seeded in req if using an already existing
15501 * New options to smime application. -inform and -outform
15502 allow alternative formats for the S/MIME message including
15503 PEM and DER. The -content option allows the content to be
15504 specified separately. This should allow things like Netscape
15505 form signing output easier to verify.
15509 * Fix the ASN1 encoding of tags using the 'long form'.
15513 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15514 STRING types. These convert content octets to and from the
15515 underlying type. The actual tag and length octets are
15516 already assumed to have been read in and checked. These
15517 are needed because all other string types have virtually
15518 identical handling apart from the tag. By having versions
15519 of the ASN1 functions that just operate on content octets
15520 IMPLICIT tagging can be handled properly. It also allows
15521 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15522 and ASN1_INTEGER are identical apart from the tag.
15526 * Change the handling of OID objects as follows:
15528 - New object identifiers are inserted in objects.txt, following
15529 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15530 - objects.pl is used to process obj_mac.num and create a new
15532 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15535 This is currently kind of a hack, and the perl code in objects.pl
15536 isn't very elegant, but it works as I intended. The simplest way
15537 to check that it worked correctly is to look in obj_dat.h and
15538 check the array nid_objs and make sure the objects haven't moved
15539 around (this is important!). Additions are OK, as well as
15540 consistent name changes.
15544 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15548 * Addition of the command line parameter '-rand file' to 'openssl req'.
15549 The given file adds to whatever has already been seeded into the
15550 random pool through the RANDFILE configuration file option or
15551 environment variable, or the default random state file.
15555 * mkstack.pl now sorts each macro group into lexical order.
15556 Previously the output order depended on the order the files
15557 appeared in the directory, resulting in needless rewriting
15562 * Patches to make OpenSSL compile under Win32 again. Mostly
15563 work arounds for the VC++ problem that it treats func() as
15564 func(void). Also stripped out the parts of mkdef.pl that
15565 added extra typesafe functions: these no longer exist.
15569 * Reorganisation of the stack code. The macros are now all
15570 collected in safestack.h . Each macro is defined in terms of
15571 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15572 DEBUG_SAFESTACK is now handled in terms of function casts,
15573 this has the advantage of retaining type safety without the
15574 use of additional functions. If DEBUG_SAFESTACK is not defined
15575 then the non typesafe macros are used instead. Also modified the
15576 mkstack.pl script to handle the new form. Needs testing to see
15577 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15578 the default if no major problems. Similar behaviour for ASN1_SET_OF
15579 and PKCS12_STACK_OF.
15583 * When some versions of IIS use the 'NET' form of private key the
15584 key derivation algorithm is different. Normally MD5(password) is
15585 used as a 128 bit RC4 key. In the modified case
15586 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15587 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15588 as the old Netscape_RSA functions except they have an additional
15589 'sgckey' parameter which uses the modified algorithm. Also added
15590 an -sgckey command line option to the rsa utility. Thanks to
15591 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15592 algorithm to openssl-dev.
15596 * The evp_local.h macros were using 'c.##kname' which resulted in
15597 invalid expansion on some systems (SCO 5.0.5 for example).
15598 Corrected to 'c.kname'.
15600 *Phillip Porch <root@theporch.com>*
15602 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15603 a STACK of email addresses from a certificate or request, these look
15604 in the subject name and the subject alternative name extensions and
15605 omit any duplicate addresses.
15609 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15610 This makes DSA verification about 2 % faster.
15614 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15615 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15616 plus overhead for 1024 bit moduli).
15617 This makes exponentiations about 0.5 % faster for 1024 bit
15618 exponents (as measured by "openssl speed rsa2048").
15622 * Rename memory handling macros to avoid conflicts with other
15624 Malloc => OPENSSL_malloc
15625 Malloc_locked => OPENSSL_malloc_locked
15626 Realloc => OPENSSL_realloc
15627 Free => OPENSSL_free
15631 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15632 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15636 * CygWin32 support.
15638 *John Jarvie <jjarvie@newsguy.com>*
15640 * The type-safe stack code has been rejigged. It is now only compiled
15641 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15642 by default all type-specific stack functions are "#define"d back to
15643 standard stack functions. This results in more streamlined output
15644 but retains the type-safety checking possibilities of the original
15649 * The STACK code has been cleaned up, and certain type declarations
15650 that didn't make a lot of sense have been brought in line. This has
15651 also involved a cleanup of sorts in safestack.h to more correctly
15652 map type-safe stack functions onto their plain stack counterparts.
15653 This work has also resulted in a variety of "const"ifications of
15654 lots of the code, especially `_cmp` operations which should normally
15655 be prototyped with "const" parameters anyway.
15659 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15660 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15661 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15662 where all of 'md' is used each time the PRNG is used, but 'state'
15663 is used only indexed by a cyclic counter. As entropy may not be
15664 well distributed from the beginning, 'md' is important as a
15665 chaining variable. However, the output function chains only half
15666 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15667 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15668 in all of 'state' being rewritten, with the new values depending
15669 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15673 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15674 the handshake is continued after ssl_verify_cert_chain();
15675 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15676 can lead to 'unexplainable' connection aborts later.
15678 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15680 * Major EVP API cipher revision.
15681 Add hooks for extra EVP features. This allows various cipher
15682 parameters to be set in the EVP interface. Support added for variable
15683 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15684 setting of RC2 and RC5 parameters.
15686 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15689 Remove lots of duplicated code from the EVP library. For example *every*
15690 cipher init() function handles the 'iv' in the same way according to the
15691 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15692 for CFB and OFB modes they zero ctx->num.
15694 New functionality allows removal of S/MIME code RC2 hack.
15696 Most of the routines have the same form and so can be declared in terms
15699 By shifting this to the top level EVP_CipherInit() it can be removed from
15700 all individual ciphers. If the cipher wants to handle IVs or keys
15701 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15704 Change lots of functions like EVP_EncryptUpdate() to now return a
15705 value: although software versions of the algorithms cannot fail
15706 any installed hardware versions can.
15710 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15711 this option is set, tolerate broken clients that send the negotiated
15712 protocol version number instead of the requested protocol version
15717 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15718 i.e. non-zero for export ciphersuites, zero otherwise.
15719 Previous versions had this flag inverted, inconsistent with
15720 rsa_tmp_cb (..._TMP_RSA_CB).
15722 *Bodo Moeller; problem reported by Amit Chopra*
15724 * Add missing DSA library text string. Work around for some IIS
15725 key files with invalid SEQUENCE encoding.
15729 * Add a document (doc/standards.txt) that list all kinds of standards
15730 and so on that are implemented in OpenSSL.
15734 * Enhance c_rehash script. Old version would mishandle certificates
15735 with the same subject name hash and wouldn't handle CRLs at all.
15736 Added -fingerprint option to crl utility, to support new c_rehash
15741 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15745 * Fix for SSL server purpose checking. Server checking was
15746 rejecting certificates which had extended key usage present
15747 but no ssl client purpose.
15749 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15751 * Make PKCS#12 code work with no password. The PKCS#12 spec
15752 is a little unclear about how a blank password is handled.
15753 Since the password in encoded as a BMPString with terminating
15754 double NULL a zero length password would end up as just the
15755 double NULL. However no password at all is different and is
15756 handled differently in the PKCS#12 key generation code. NS
15757 treats a blank password as zero length. MSIE treats it as no
15758 password on export: but it will try both on import. We now do
15759 the same: PKCS12_parse() tries zero length and no password if
15760 the password is set to "" or NULL (NULL is now a valid password:
15761 it wasn't before) as does the pkcs12 application.
15765 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15766 perror when PEM_read_bio_X509_REQ fails, the error message must
15767 be obtained from the error queue.
15771 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15772 it in ERR_remove_state if appropriate, and change ERR_get_state
15773 accordingly to avoid race conditions (this is necessary because
15774 thread_hash is no longer constant once set).
15778 * Bugfix for linux-elf makefile.one.
15782 * RSA_get_default_method() will now cause a default
15783 RSA_METHOD to be chosen if one doesn't exist already.
15784 Previously this was only set during a call to RSA_new()
15785 or RSA_new_method(NULL) meaning it was possible for
15786 RSA_get_default_method() to return NULL.
15790 * Added native name translation to the existing DSO code
15791 that will convert (if the flag to do so is set) filenames
15792 that are sufficiently small and have no path information
15793 into a canonical native form. Eg. "blah" converted to
15794 "libblah.so" or "blah.dll" etc.
15798 * New function ERR_error_string_n(e, buf, len) which is like
15799 ERR_error_string(e, buf), but writes at most 'len' bytes
15800 including the 0 terminator. For ERR_error_string_n, 'buf'
15803 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15805 * CONF library reworked to become more general. A new CONF
15806 configuration file reader "class" is implemented as well as a
15807 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15808 old `CONF_*` functions are still there, but are reimplemented to
15809 work in terms of the new functions. Also, a set of functions
15810 to handle the internal storage of the configuration data is
15811 provided to make it easier to write new configuration file
15812 reader "classes" (I can definitely see something reading a
15813 configuration file in XML format, for example), called `_CONF_*`,
15814 or "the configuration storage API"...
15816 The new configuration file reading functions are:
15818 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15819 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15821 NCONF_default, NCONF_WIN32
15823 NCONF_dump_fp, NCONF_dump_bio
15825 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15826 NCONF_new creates a new CONF object. This works in the same way
15827 as other interfaces in OpenSSL, like the BIO interface.
15828 `NCONF_dump_*` dump the internal storage of the configuration file,
15829 which is useful for debugging. All other functions take the same
15830 arguments as the old `CONF_*` functions with the exception of the
15831 first that must be a `CONF *` instead of a `LHASH *`.
15833 To make it easier to use the new classes with the old `CONF_*` functions,
15834 the function CONF_set_default_method is provided.
15838 * Add '-tls1' option to 'openssl ciphers', which was already
15839 mentioned in the documentation but had not been implemented.
15840 (This option is not yet really useful because even the additional
15841 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15845 * Initial DSO code added into libcrypto for letting OpenSSL (and
15846 OpenSSL-based applications) load shared libraries and bind to
15847 them in a portable way.
15849 *Geoff Thorpe, with contributions from Richard Levitte*
15851 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15853 * Make sure _lrotl and _lrotr are only used with MSVC.
15855 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15856 (the default implementation of RAND_status).
15858 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15859 to '-clrext' (= clear extensions), as intended and documented.
15860 *Bodo Moeller; inconsistency pointed out by Michael Attili
15861 <attili@amaxo.com>*
15863 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15864 was larger than the MD block size.
15866 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15868 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15869 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15870 using the passed key: if the passed key was a private key the result
15871 of X509_print(), for example, would be to print out all the private key
15876 * des_quad_cksum() byte order bug fix.
15877 *Ulf Möller, using the problem description in krb4-0.9.7, where
15878 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15880 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15883 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15885 * For easily testing in shell scripts whether some command
15886 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15887 returns with exit code 0 iff no command of the given name is available.
15888 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15889 the output goes to stdout and nothing is printed to stderr.
15890 Additional arguments are always ignored.
15892 Since for each cipher there is a command of the same name,
15893 the 'no-cipher' compilation switches can be tested this way.
15895 ('openssl no-XXX' is not able to detect pseudo-commands such
15896 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15900 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15904 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15905 is set; it will be thrown away anyway because each handshake creates
15907 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15908 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15909 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15910 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15914 * New s_client option -ign_eof: EOF at stdin is ignored, and
15915 'Q' and 'R' lose their special meanings (quit/renegotiate).
15916 This is part of what -quiet does; unlike -quiet, -ign_eof
15917 does not suppress any output.
15921 * Add compatibility options to the purpose and trust code. The
15922 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15923 accepts a certificate or CA, this was the previous behaviour,
15924 with all the associated security issues.
15926 X509_TRUST_COMPAT is the old trust behaviour: only and
15927 automatically trust self signed roots in certificate store. A
15928 new trust setting X509_TRUST_DEFAULT is used to specify that
15929 a purpose has no associated trust setting and it should instead
15930 use the value in the default purpose.
15934 * Fix the PKCS#8 DSA private key code so it decodes keys again
15935 and fix a memory leak.
15939 * In util/mkerr.pl (which implements 'make errors'), preserve
15940 reason strings from the previous version of the .c file, as
15941 the default to have only downcase letters (and digits) in
15942 automatically generated reasons codes is not always appropriate.
15946 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15947 using strerror. Previously, ERR_reason_error_string() returned
15948 library names as reason strings for SYSerr; but SYSerr is a special
15949 case where small numbers are errno values, not library numbers.
15953 * Add '-dsaparam' option to 'openssl dhparam' application. This
15954 converts DSA parameters into DH parameters. (When creating parameters,
15955 DSA_generate_parameters is used.)
15959 * Include 'length' (recommended exponent length) in C code generated
15960 by 'openssl dhparam -C'.
15964 * The second argument to set_label in perlasm was already being used
15965 so couldn't be used as a "file scope" flag. Moved to third argument
15970 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15971 instead of RAND_bytes for encryption IVs and salts.
15975 * Include RAND_status() into RAND_METHOD instead of implementing
15976 it only for md_rand.c Otherwise replacing the PRNG by calling
15977 RAND_set_rand_method would be impossible.
15981 * Don't let DSA_generate_key() enter an infinite loop if the random
15982 number generation fails.
15986 * New 'rand' application for creating pseudo-random output.
15990 * Added configuration support for Linux/IA64
15992 *Rolf Haberrecker <rolf@suse.de>*
15994 * Assembler module support for Mingw32.
15998 * Shared library support for HPUX (in shlib/).
16000 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16002 * Shared library support for Solaris gcc.
16004 *Lutz Behnke <behnke@trustcenter.de>*
16006 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16008 * PKCS7_encrypt() was adding text MIME headers twice because they
16009 were added manually and by SMIME_crlf_copy().
16013 * In bntest.c don't call BN_rand with zero bits argument.
16015 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16017 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16018 case was implemented. This caused BN_div_recp() to fail occasionally.
16022 * Add an optional second argument to the set_label() in the perl
16023 assembly language builder. If this argument exists and is set
16024 to 1 it signals that the assembler should use a symbol whose
16025 scope is the entire file, not just the current function. This
16026 is needed with MASM which uses the format label:: for this scope.
16028 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16030 * Change the ASN1 types so they are typedefs by default. Before
16031 almost all types were #define'd to ASN1_STRING which was causing
16032 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16037 * Change names of new functions to the new get1/get0 naming
16038 convention: After 'get1', the caller owns a reference count
16039 and has to call `..._free`; 'get0' returns a pointer to some
16040 data structure without incrementing reference counters.
16041 (Some of the existing 'get' functions increment a reference
16042 counter, some don't.)
16043 Similarly, 'set1' and 'add1' functions increase reference
16044 counters or duplicate objects.
16048 * Allow for the possibility of temp RSA key generation failure:
16049 the code used to assume it always worked and crashed on failure.
16053 * Fix potential buffer overrun problem in BIO_printf().
16054 *Ulf Möller, using public domain code by Patrick Powell; problem
16055 pointed out by David Sacerdote <das33@cornell.edu>*
16057 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16058 RAND_egd() and RAND_status(). In the command line application,
16059 the EGD socket can be specified like a seed file using RANDFILE
16064 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16065 Some CAs (e.g. Verisign) distribute certificates in this form.
16069 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16070 list to exclude them. This means that no special compilation option
16071 is needed to use anonymous DH: it just needs to be included in the
16076 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16077 EVP_MD_type. The old functionality is available in a new macro called
16078 EVP_MD_md(). Change code that uses it and update docs.
16082 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16083 where the `void *` argument is replaced by a function pointer argument.
16084 Previously `void *` was abused to point to functions, which works on
16085 many platforms, but is not correct. As these functions are usually
16086 called by macros defined in OpenSSL header files, most source code
16087 should work without changes.
16091 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16092 sections with information on -D... compiler switches used for
16093 compiling the library so that applications can see them. To enable
16094 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16095 must be defined. E.g.,
16096 #define OPENSSL_ALGORITHM_DEFINES
16097 #include <openssl/opensslconf.h>
16098 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16100 *Richard Levitte, Ulf and Bodo Möller*
16102 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16107 * Change the 'other' type in certificate aux info to a STACK_OF
16108 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16109 the required ASN1 format: arbitrary types determined by an OID.
16113 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16114 argument to 'req'. This is not because the function is newer or
16115 better than others it just uses the work 'NEW' in the certificate
16116 request header lines. Some software needs this.
16120 * Reorganise password command line arguments: now passwords can be
16121 obtained from various sources. Delete the PEM_cb function and make
16122 it the default behaviour: i.e. if the callback is NULL and the
16123 usrdata argument is not NULL interpret it as a null terminated pass
16124 phrase. If usrdata and the callback are NULL then the pass phrase
16125 is prompted for as usual.
16129 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16130 the support is automatically enabled. The resulting binaries will
16131 autodetect the card and use it if present.
16133 *Ben Laurie and Compaq Inc.*
16135 * Work around for Netscape hang bug. This sends certificate request
16136 and server done in one record. Since this is perfectly legal in the
16137 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16138 the bugs/SSLv3 entry for more info.
16142 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16146 * Add -rand argument to smime and pkcs12 applications and read/write
16151 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16155 * Add command line password options to the remaining applications.
16159 * Bug fix for BN_div_recp() for numerators with an even number of
16164 * More tests in bntest.c, and changed test_bn output.
16168 * ./config recognizes MacOS X now.
16172 * Bug fix for BN_div() when the first words of num and divisor are
16173 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16177 * Add support for various broken PKCS#8 formats, and command line
16178 options to produce them.
16182 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16183 get temporary BIGNUMs from a BN_CTX.
16187 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16192 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16193 include a #define from the old name to the new. The original intent
16194 was that statically linked binaries could for example just call
16195 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16196 link with digests. This never worked because SSLeay_add_all_digests()
16197 and SSLeay_add_all_ciphers() were in the same source file so calling
16198 one would link with the other. They are now in separate source files.
16202 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16206 * Use a less unusual form of the Miller-Rabin primality test (it used
16207 a binary algorithm for exponentiation integrated into the Miller-Rabin
16208 loop, our standard modexp algorithms are faster).
16212 * Support for the EBCDIC character set completed.
16214 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16216 * Source code cleanups: use const where appropriate, eliminate casts,
16217 use `void *` instead of `char *` in lhash.
16221 * Bugfix: ssl3_send_server_key_exchange was not restartable
16222 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16223 this the server could overwrite ephemeral keys that the client
16228 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16229 using 50 iterations of the Rabin-Miller test.
16231 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16232 iterations of the Rabin-Miller test as required by the appendix
16233 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16234 As BN_is_prime_fasttest includes trial division, DSA parameter
16235 generation becomes much faster.
16237 This implies a change for the callback functions in DSA_is_prime
16238 and DSA_generate_parameters: The callback function is called once
16239 for each positive witness in the Rabin-Miller test, not just
16240 occasionally in the inner loop; and the parameters to the
16241 callback function now provide an iteration count for the outer
16242 loop rather than for the current invocation of the inner loop.
16243 DSA_generate_parameters additionally can call the callback
16244 function with an 'iteration count' of -1, meaning that a
16245 candidate has passed the trial division test (when q is generated
16246 from an application-provided seed, trial division is skipped).
16250 * New function BN_is_prime_fasttest that optionally does trial
16251 division before starting the Rabin-Miller test and has
16252 an additional BN_CTX * argument (whereas BN_is_prime always
16253 has to allocate at least one BN_CTX).
16254 'callback(1, -1, cb_arg)' is called when a number has passed the
16255 trial division stage.
16259 * Fix for bug in CRL encoding. The validity dates weren't being handled
16264 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16268 * New function BN_pseudo_rand().
16272 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16273 bignum version of BN_from_montgomery() with the working code from
16274 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16279 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16280 made it impossible to use the same SSL_SESSION data structure in
16281 SSL2 clients in multiple threads.
16285 * The return value of RAND_load_file() no longer counts bytes obtained
16286 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16287 to seed the PRNG (previously an explicit byte count was required).
16289 *Ulf Möller, Bodo Möller*
16291 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16292 used `char *` instead of `void *` and had casts all over the place.
16296 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16300 * Retain source code compatibility for BN_prime_checks macro:
16301 BN_is_prime(..., BN_prime_checks, ...) now uses
16302 BN_prime_checks_for_size to determine the appropriate number of
16303 Rabin-Miller iterations.
16307 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16308 DH_CHECK_P_NOT_SAFE_PRIME.
16309 (Check if this is true? OpenPGP calls them "strong".)
16313 * Merge the functionality of "dh" and "gendh" programs into a new program
16314 "dhparam". The old programs are retained for now but will handle DH keys
16315 (instead of parameters) in future.
16319 * Make the ciphers, s_server and s_client programs check the return values
16320 when a new cipher list is set.
16324 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16325 ciphers. Before when the 56bit ciphers were enabled the sorting was
16328 The syntax for the cipher sorting has been extended to support sorting by
16329 cipher-strength (using the strength_bits hard coded in the tables).
16330 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16332 Fix a bug in the cipher-command parser: when supplying a cipher command
16333 string with an "undefined" symbol (neither command nor alphanumeric
16334 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16335 an error is flagged.
16337 Due to the strength-sorting extension, the code of the
16338 ssl_create_cipher_list() function was completely rearranged. I hope that
16339 the readability was also increased :-)
16341 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16343 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16344 for the first serial number and places 2 in the serial number file. This
16345 avoids problems when the root CA is created with serial number zero and
16346 the first user certificate has the same issuer name and serial number
16351 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16352 the new code. Add documentation for this stuff.
16356 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16357 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16358 structures and behave in an analogous way to the X509v3 functions:
16359 they shouldn't be called directly but wrapper functions should be used
16362 So we also now have some wrapper functions that call the X509at functions
16363 when passed certificate requests. (TO DO: similar things can be done with
16364 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16365 things. Some of these need some d2i or i2d and print functionality
16366 because they handle more complex structures.)
16370 * Add missing #ifndefs that caused missing symbols when building libssl
16371 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16372 NO_RSA in `ssl/s2*.c`.
16374 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16376 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16377 has a return value which indicates the quality of the random data
16378 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16379 error queue. New function RAND_pseudo_bytes() generates output that is
16380 guaranteed to be unique but not unpredictable. RAND_add is like
16381 RAND_seed, but takes an extra argument for an entropy estimate
16382 (RAND_seed always assumes full entropy).
16386 * Do more iterations of Rabin-Miller probable prime test (specifically,
16387 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16388 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16389 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16390 false-positive rate of at most 2^-80 for random input.
16394 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16398 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16399 in the 0.9.5 release), this returns the chain
16400 from an X509_CTX structure with a dup of the stack and all
16401 the X509 reference counts upped: so the stack will exist
16402 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16405 Also make SSL_SESSION_print() print out the verify return
16410 * Add manpage for the pkcs12 command. Also change the default
16411 behaviour so MAC iteration counts are used unless the new
16412 -nomaciter option is used. This improves file security and
16413 only older versions of MSIE (4.0 for example) need it.
16417 * Honor the no-xxx Configure options when creating .DEF files.
16421 * Add PKCS#10 attributes to field table: challengePassword,
16422 unstructuredName and unstructuredAddress. These are taken from
16423 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16424 international characters are used.
16426 More changes to X509_ATTRIBUTE code: allow the setting of types
16427 based on strings. Remove the 'loc' parameter when adding
16428 attributes because these will be a SET OF encoding which is sorted
16433 * Initial changes to the 'req' utility to allow request generation
16434 automation. This will allow an application to just generate a template
16435 file containing all the field values and have req construct the
16438 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16439 used all over the place including certificate requests and PKCS#7
16440 structures. They are currently handled manually where necessary with
16441 some primitive wrappers for PKCS#7. The new functions behave in a
16442 manner analogous to the X509 extension functions: they allow
16443 attributes to be looked up by NID and added.
16445 Later something similar to the X509V3 code would be desirable to
16446 automatically handle the encoding, decoding and printing of the
16447 more complex types. The string types like challengePassword can
16448 be handled by the string table functions.
16450 Also modified the multi byte string table handling. Now there is
16451 a 'global mask' which masks out certain types. The table itself
16452 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16453 is useful when for example there is only one permissible type
16454 (as in countryName) and using the mask might result in no valid
16459 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16460 SSL_get_peer_finished to allow applications to obtain the latest
16461 Finished messages sent to the peer or expected from the peer,
16462 respectively. (SSL_get_peer_finished is usually the Finished message
16463 actually received from the peer, otherwise the protocol will be aborted.)
16465 As the Finished message are message digests of the complete handshake
16466 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16467 be used for external authentication procedures when the authentication
16468 provided by SSL/TLS is not desired or is not enough.
16472 * Enhanced support for Alpha Linux is added. Now ./config checks if
16473 the host supports BWX extension and if Compaq C is present on the
16474 $PATH. Just exploiting of the BWX extension results in 20-30%
16475 performance kick for some algorithms, e.g. DES and RC4 to mention
16476 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16481 * Add support for MS "fast SGC". This is arguably a violation of the
16482 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16483 weak crypto and after checking the certificate is SGC a second one
16484 with strong crypto. MS SGC stops the first handshake after receiving
16485 the server certificate message and sends a second client hello. Since
16486 a server will typically do all the time consuming operations before
16487 expecting any further messages from the client (server key exchange
16488 is the most expensive) there is little difference between the two.
16490 To get OpenSSL to support MS SGC we have to permit a second client
16491 hello message after we have sent server done. In addition we have to
16492 reset the MAC if we do get this second client hello.
16496 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16497 if a DER encoded private key is RSA or DSA traditional format. Changed
16498 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16499 format DER encoded private key. Newer code should use PKCS#8 format which
16500 has the key type encoded in the ASN1 structure. Added DER private key
16501 support to pkcs8 application.
16505 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16506 ciphersuites has been selected (as required by the SSL 3/TLS 1
16507 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16508 is set, we interpret this as a request to violate the specification
16509 (the worst that can happen is a handshake failure, and 'correct'
16510 behaviour would result in a handshake failure anyway).
16514 * In SSL_CTX_add_session, take into account that there might be multiple
16515 SSL_SESSION structures with the same session ID (e.g. when two threads
16516 concurrently obtain them from an external cache).
16517 The internal cache can handle only one SSL_SESSION with a given ID,
16518 so if there's a conflict, we now throw out the old one to achieve
16523 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16524 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16525 some routines that use cipher OIDs: some ciphers do not have OIDs
16526 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16531 * Simplify the trust setting structure and code. Now we just have
16532 two sequences of OIDs for trusted and rejected settings. These will
16533 typically have values the same as the extended key usage extension
16534 and any application specific purposes.
16536 The trust checking code now has a default behaviour: it will just
16537 check for an object with the same NID as the passed id. Functions can
16538 be provided to override either the default behaviour or the behaviour
16539 for a given id. SSL client, server and email already have functions
16540 in place for compatibility: they check the NID and also return "trusted"
16541 if the certificate is self signed.
16545 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16546 traditional format into an EVP_PKEY structure.
16550 * Add a password callback function PEM_cb() which either prompts for
16551 a password if usr_data is NULL or otherwise assumes it is a null
16552 terminated password. Allow passwords to be passed on command line
16553 environment or config files in a few more utilities.
16557 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16558 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16559 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16560 Update documentation.
16564 * Support for ASN1 "NULL" type. This could be handled before by using
16565 ASN1_TYPE but there wasn't any function that would try to read a NULL
16566 and produce an error if it couldn't. For compatibility we also have
16567 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16568 don't allocate anything because they don't need to.
16572 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16575 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16577 * Rebuild of the memory allocation routines used by OpenSSL code and
16578 possibly others as well. The purpose is to make an interface that
16579 provide hooks so anyone can build a separate set of allocation and
16580 deallocation routines to be used by OpenSSL, for example memory
16581 pool implementations, or something else, which was previously hard
16582 since Malloc(), Realloc() and Free() were defined as macros having
16583 the values malloc, realloc and free, respectively (except for Win32
16584 compilations). The same is provided for memory debugging code.
16585 OpenSSL already comes with functionality to find memory leaks, but
16586 this gives people a chance to debug other memory problems.
16588 With these changes, a new set of functions and macros have appeared:
16590 CRYPTO_set_mem_debug_functions() [F]
16591 CRYPTO_get_mem_debug_functions() [F]
16592 CRYPTO_dbg_set_options() [F]
16593 CRYPTO_dbg_get_options() [F]
16594 CRYPTO_malloc_debug_init() [M]
16596 The memory debug functions are NULL by default, unless the library
16597 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16598 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16599 gives the standard debugging functions that come with OpenSSL) or
16600 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16601 provided by the library user) must be used. When the standard
16602 debugging functions are used, CRYPTO_dbg_set_options can be used to
16603 request additional information:
16604 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16605 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16607 Also, things like CRYPTO_set_mem_functions will always give the
16608 expected result (the new set of functions is used for allocation
16609 and deallocation) at all times, regardless of platform and compiler
16612 To finish it up, some functions that were never use in any other
16613 way than through macros have a new API and new semantic:
16615 CRYPTO_dbg_malloc()
16616 CRYPTO_dbg_realloc()
16619 All macros of value have retained their old syntax.
16621 *Richard Levitte and Bodo Moeller*
16623 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16624 ordering of SMIMECapabilities wasn't in "strength order" and there
16625 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16630 * Some ASN1 types with illegal zero length encoding (INTEGER,
16631 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16633 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16635 * Merge in my S/MIME library for OpenSSL. This provides a simple
16636 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16637 functionality to handle multipart/signed properly) and a utility
16638 called 'smime' to call all this stuff. This is based on code I
16639 originally wrote for Celo who have kindly allowed it to be
16640 included in OpenSSL.
16644 * Add variants des_set_key_checked and des_set_key_unchecked of
16645 des_set_key (aka des_key_sched). Global variable des_check_key
16646 decides which of these is called by des_set_key; this way
16647 des_check_key behaves as it always did, but applications and
16648 the library itself, which was buggy for des_check_key == 1,
16649 have a cleaner way to pick the version they need.
16653 * New function PKCS12_newpass() which changes the password of a
16658 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16659 dynamic mix. In both cases the ids can be used as an index into the
16660 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16661 functions so they accept a list of the field values and the
16662 application doesn't need to directly manipulate the X509_TRUST
16667 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16672 * Modify the way the V3 extension code looks up extensions. This now
16673 works in a similar way to the object code: we have some "standard"
16674 extensions in a static table which is searched with OBJ_bsearch()
16675 and the application can add dynamic ones if needed. The file
16676 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16677 updated whenever a new extension is added to the core code and kept
16678 in ext_nid order. There is a simple program 'tabtest.c' which checks
16679 this. New extensions are not added too often so this file can readily
16680 be maintained manually.
16682 There are two big advantages in doing things this way. The extensions
16683 can be looked up immediately and no longer need to be "added" using
16684 X509V3_add_standard_extensions(): this function now does nothing.
16685 Side note: I get *lots* of email saying the extension code doesn't
16686 work because people forget to call this function.
16687 Also no dynamic allocation is done unless new extensions are added:
16688 so if we don't add custom extensions there is no need to call
16689 X509V3_EXT_cleanup().
16693 * Modify enc utility's salting as follows: make salting the default. Add a
16694 magic header, so unsalted files fail gracefully instead of just decrypting
16695 to garbage. This is because not salting is a big security hole, so people
16696 should be discouraged from doing it.
16700 * Fixes and enhancements to the 'x509' utility. It allowed a message
16701 digest to be passed on the command line but it only used this
16702 parameter when signing a certificate. Modified so all relevant
16703 operations are affected by the digest parameter including the
16704 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16705 DSA key was used because it didn't fix the digest.
16709 * Initial certificate chain verify code. Currently tests the untrusted
16710 certificates for consistency with the verify purpose (which is set
16711 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16713 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16714 this is because it will reject chains with invalid extensions whereas
16715 every previous version of OpenSSL and SSLeay made no checks at all.
16717 Trust code: checks the root CA for the relevant trust settings. Trust
16718 settings have an initial value consistent with the verify purpose: e.g.
16719 if the verify purpose is for SSL client use it expects the CA to be
16720 trusted for SSL client use. However the default value can be changed to
16721 permit custom trust settings: one example of this would be to only trust
16722 certificates from a specific "secure" set of CAs.
16724 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16725 which should be used for version portability: especially since the
16726 verify structure is likely to change more often now.
16728 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16729 to set them. If not set then assume SSL clients will verify SSL servers
16732 Two new options to the verify program: -untrusted allows a set of
16733 untrusted certificates to be passed in and -purpose which sets the
16734 intended purpose of the certificate. If a purpose is set then the
16735 new chain verify code is used to check extension consistency.
16739 * Support for the authority information access extension.
16743 * Modify RSA and DSA PEM read routines to transparently handle
16744 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16745 public keys in a format compatible with certificate
16746 SubjectPublicKeyInfo structures. Unfortunately there were already
16747 functions called *_PublicKey_* which used various odd formats so
16748 these are retained for compatibility: however the DSA variants were
16749 never in a public release so they have been deleted. Changed dsa/rsa
16750 utilities to handle the new format: note no releases ever handled public
16751 keys so we should be OK.
16753 The primary motivation for this change is to avoid the same fiasco
16754 that dogs private keys: there are several incompatible private key
16755 formats some of which are standard and some OpenSSL specific and
16756 require various evil hacks to allow partial transparent handling and
16757 even then it doesn't work with DER formats. Given the option anything
16758 other than PKCS#8 should be dumped: but the other formats have to
16759 stay in the name of compatibility.
16761 With public keys and the benefit of hindsight one standard format
16762 is used which works with EVP_PKEY, RSA or DSA structures: though
16763 it clearly returns an error if you try to read the wrong kind of key.
16765 Added a -pubkey option to the 'x509' utility to output the public key.
16766 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16767 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16768 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16769 that do the same as the `EVP_PKEY_assign_*()` except they up the
16770 reference count of the added key (they don't "swallow" the
16775 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16776 CRLs would fail if the file contained no certificates or no CRLs:
16777 added a new function to read in both types and return the number
16778 read: this means that if none are read it will be an error. The
16779 DER versions of the certificate and CRL reader would always fail
16780 because it isn't possible to mix certificates and CRLs in DER format
16781 without choking one or the other routine. Changed this to just read
16782 a certificate: this is the best we can do. Also modified the code
16783 in `apps/verify.c` to take notice of return codes: it was previously
16784 attempting to read in certificates from NULL pointers and ignoring
16785 any errors: this is one reason why the cert and CRL reader seemed
16786 to work. It doesn't check return codes from the default certificate
16787 routines: these may well fail if the certificates aren't installed.
16791 * Code to support otherName option in GeneralName.
16795 * First update to verify code. Change the verify utility
16796 so it warns if it is passed a self signed certificate:
16797 for consistency with the normal behaviour. X509_verify
16798 has been modified to it will now verify a self signed
16799 certificate if *exactly* the same certificate appears
16800 in the store: it was previously impossible to trust a
16801 single self signed certificate. This means that:
16802 openssl verify ss.pem
16803 now gives a warning about a self signed certificate but
16804 openssl verify -CAfile ss.pem ss.pem
16809 * For servers, store verify_result in SSL_SESSION data structure
16810 (and add it to external session representation).
16811 This is needed when client certificate verifications fails,
16812 but an application-provided verification callback (set by
16813 SSL_CTX_set_cert_verify_callback) allows accepting the session
16814 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16815 but returns 1): When the session is reused, we have to set
16816 ssl->verify_result to the appropriate error code to avoid
16819 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16821 * Fix a bug in the new PKCS#7 code: it didn't consider the
16822 case in PKCS7_dataInit() where the signed PKCS7 structure
16823 didn't contain any existing data because it was being created.
16825 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16827 * Add a salt to the key derivation routines in enc.c. This
16828 forms the first 8 bytes of the encrypted file. Also add a
16829 -S option to allow a salt to be input on the command line.
16833 * New function X509_cmp(). Oddly enough there wasn't a function
16834 to compare two certificates. We do this by working out the SHA1
16835 hash and comparing that. X509_cmp() will be needed by the trust
16840 * SSL_get1_session() is like SSL_get_session(), but increments
16841 the reference count in the SSL_SESSION returned.
16843 *Geoff Thorpe <geoff@eu.c2.net>*
16845 * Fix for 'req': it was adding a null to request attributes.
16846 Also change the X509_LOOKUP and X509_INFO code to handle
16847 certificate auxiliary information.
16851 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16856 * Add the possibility to add extra information to the memory leak
16857 detecting output, to form tracebacks, showing from where each
16858 allocation was originated: CRYPTO_push_info("constant string") adds
16859 the string plus current file name and line number to a per-thread
16860 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16861 is like calling CYRPTO_pop_info() until the stack is empty.
16862 Also updated memory leak detection code to be multi-thread-safe.
16866 * Add options -text and -noout to pkcs7 utility and delete the
16867 encryption options which never did anything. Update docs.
16871 * Add options to some of the utilities to allow the pass phrase
16872 to be included on either the command line (not recommended on
16873 OSes like Unix) or read from the environment. Update the
16874 manpages and fix a few bugs.
16878 * Add a few manpages for some of the openssl commands.
16882 * Fix the -revoke option in ca. It was freeing up memory twice,
16883 leaking and not finding already revoked certificates.
16887 * Extensive changes to support certificate auxiliary information.
16888 This involves the use of X509_CERT_AUX structure and X509_AUX
16889 functions. An X509_AUX function such as PEM_read_X509_AUX()
16890 can still read in a certificate file in the usual way but it
16891 will also read in any additional "auxiliary information". By
16892 doing things this way a fair degree of compatibility can be
16893 retained: existing certificates can have this information added
16894 using the new 'x509' options.
16896 Current auxiliary information includes an "alias" and some trust
16897 settings. The trust settings will ultimately be used in enhanced
16898 certificate chain verification routines: currently a certificate
16899 can only be trusted if it is self signed and then it is trusted
16904 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16905 The problem was that one of the replacement routines had not been working
16906 since SSLeay releases. For now the offending routine has been replaced
16907 with non-optimised assembler. Even so, this now gives around 95%
16908 performance improvement for 1024 bit RSA signs.
16912 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16913 handling. Most clients have the effective key size in bits equal to
16914 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16915 A few however don't do this and instead use the size of the decrypted key
16916 to determine the RC2 key length and the AlgorithmIdentifier to determine
16917 the effective key length. In this case the effective key length can still
16918 be 40 bits but the key length can be 168 bits for example. This is fixed
16919 by manually forcing an RC2 key into the EVP_PKEY structure because the
16920 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16921 the key length and effective key length are equal.
16925 * Add a bunch of functions that should simplify the creation of
16926 X509_NAME structures. Now you should be able to do:
16927 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16928 and have it automatically work out the correct field type and fill in
16929 the structures. The more adventurous can try:
16930 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16931 and it will (hopefully) work out the correct multibyte encoding.
16935 * Change the 'req' utility to use the new field handling and multibyte
16936 copy routines. Before the DN field creation was handled in an ad hoc
16937 way in req, ca, and x509 which was rather broken and didn't support
16938 BMPStrings or UTF8Strings. Since some software doesn't implement
16939 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16940 using the dirstring_type option. See the new comment in the default
16941 openssl.cnf for more info.
16945 * Make crypto/rand/md_rand.c more robust:
16946 - Assure unique random numbers after fork().
16947 - Make sure that concurrent threads access the global counter and
16948 md serializably so that we never lose entropy in them
16949 or use exactly the same state in multiple threads.
16950 Access to the large state is not always serializable because
16951 the additional locking could be a performance killer, and
16952 md should be large enough anyway.
16956 * New file `apps/app_rand.c` with commonly needed functionality
16957 for handling the random seed file.
16959 Use the random seed file in some applications that previously did not:
16961 dsaparam -genkey (which also ignored its '-rand' option),
16964 x509 (when signing).
16965 Except on systems with /dev/urandom, it is crucial to have a random
16966 seed file at least for key creation, DSA signing, and for DH exchanges;
16967 for RSA signatures we could do without one.
16969 gendh and gendsa (unlike genrsa) used to read only the first byte
16970 of each file listed in the '-rand' option. The function as previously
16971 found in genrsa is now in app_rand.c and is used by all programs
16972 that support '-rand'.
16976 * In RAND_write_file, use mode 0600 for creating files;
16977 don't just chmod when it may be too late.
16981 * Report an error from X509_STORE_load_locations
16982 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16986 * New function ASN1_mbstring_copy() this copies a string in either
16987 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16988 into an ASN1_STRING type. A mask of permissible types is passed
16989 and it chooses the "minimal" type to use or an error if not type
16994 * Add function equivalents to the various macros in asn1.h. The old
16995 macros are retained with an `M_` prefix. Code inside the library can
16996 use the `M_` macros. External code (including the openssl utility)
16997 should *NOT* in order to be "shared library friendly".
17001 * Add various functions that can check a certificate's extensions
17002 to see if it usable for various purposes such as SSL client,
17003 server or S/MIME and CAs of these types. This is currently
17004 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17005 verification. Also added a -purpose flag to x509 utility to
17006 print out all the purposes.
17010 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17015 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17016 for, obtain and decode and extension and obtain its critical flag.
17017 This allows all the necessary extension code to be handled in a
17018 single function call.
17022 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17023 platforms. See crypto/rc4/rc4_enc.c for further details.
17027 * New -noout option to asn1parse. This causes no output to be produced
17028 its main use is when combined with -strparse and -out to extract data
17029 from a file (which may not be in ASN.1 format).
17033 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17034 when producing the local key id.
17036 *Richard Levitte <levitte@stacken.kth.se>*
17038 * New option -dhparam in s_server. This allows a DH parameter file to be
17039 stated explicitly. If it is not stated then it tries the first server
17040 certificate file. The previous behaviour hard coded the filename
17045 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17046 a public key to be input or output. For example:
17047 openssl rsa -in key.pem -pubout -out pubkey.pem
17048 Also added necessary DSA public key functions to handle this.
17052 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17053 in the message. This was handled by allowing
17054 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17056 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17058 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17059 to the end of the strings whereas this didn't. This would cause problems
17060 if strings read with d2i_ASN1_bytes() were later modified.
17062 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17064 * Fix for base64 decode bug. When a base64 bio reads only one line of
17065 data and it contains EOF it will end up returning an error. This is
17066 caused by input 46 bytes long. The cause is due to the way base64
17067 BIOs find the start of base64 encoded data. They do this by trying a
17068 trial decode on each line until they find one that works. When they
17069 do a flag is set and it starts again knowing it can pass all the
17070 data directly through the decoder. Unfortunately it doesn't reset
17071 the context it uses. This means that if EOF is reached an attempt
17072 is made to pass two EOFs through the context and this causes the
17073 resulting error. This can also cause other problems as well. As is
17074 usual with these problems it takes *ages* to find and the fix is
17075 trivial: move one line.
17077 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17079 * Ugly workaround to get s_client and s_server working under Windows. The
17080 old code wouldn't work because it needed to select() on sockets and the
17081 tty (for keypresses and to see if data could be written). Win32 only
17082 supports select() on sockets so we select() with a 1s timeout on the
17083 sockets and then see if any characters are waiting to be read, if none
17084 are present then we retry, we also assume we can always write data to
17085 the tty. This isn't nice because the code then blocks until we've
17086 received a complete line of data and it is effectively polling the
17087 keyboard at 1s intervals: however it's quite a bit better than not
17088 working at all :-) A dedicated Windows application might handle this
17089 with an event loop for example.
17093 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17094 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17095 will be called when RSA_sign() and RSA_verify() are used. This is useful
17096 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17097 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17098 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17099 This necessitated the support of an extra signature type NID_md5_sha1
17100 for SSL signatures and modifications to the SSL library to use it instead
17101 of calling RSA_public_decrypt() and RSA_private_encrypt().
17105 * Add new -verify -CAfile and -CApath options to the crl program, these
17106 will lookup a CRL issuers certificate and verify the signature in a
17107 similar way to the verify program. Tidy up the crl program so it
17108 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17109 less strict. It will now permit CRL extensions even if it is not
17110 a V2 CRL: this will allow it to tolerate some broken CRLs.
17114 * Initialize all non-automatic variables each time one of the openssl
17115 sub-programs is started (this is necessary as they may be started
17116 multiple times from the "OpenSSL>" prompt).
17118 *Lennart Bang, Bodo Moeller*
17120 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17121 removing all other RSA functionality (this is what NO_RSA does). This
17122 is so (for example) those in the US can disable those operations covered
17123 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17128 * Non-copying interface to BIO pairs.
17129 (still largely untested)
17133 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17134 ASCII string. This was handled independently in various places before.
17138 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17139 UTF8 strings a character at a time.
17143 * Use client_version from client hello to select the protocol
17144 (s23_srvr.c) and for RSA client key exchange verification
17145 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17149 * Add various utility functions to handle SPKACs, these were previously
17150 handled by poking round in the structure internals. Added new function
17151 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17152 print, verify and generate SPKACs. Based on an original idea from
17153 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17157 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17161 * Allow the config file extension section to be overwritten on the
17162 command line. Based on an original idea from Massimiliano Pala
17163 <madwolf@comune.modena.it>. The new option is called -extensions
17164 and can be applied to ca, req and x509. Also -reqexts to override
17165 the request extensions in req and -crlexts to override the crl extensions
17170 * Add new feature to the SPKAC handling in ca. Now you can include
17171 the same field multiple times by preceding it by "XXXX." for example:
17174 this is the same syntax as used in the req config file.
17178 * Allow certificate extensions to be added to certificate requests. These
17179 are specified in a 'req_extensions' option of the req section of the
17180 config file. They can be printed out with the -text option to req but
17181 are otherwise ignored at present.
17185 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17186 data read consists of only the final block it would not decrypted because
17187 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17188 A misplaced 'break' also meant the decrypted final block might not be
17189 copied until the next read.
17193 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17194 a few extra parameters to the DH structure: these will be useful if
17195 for example we want the value of 'q' or implement X9.42 DH.
17199 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17200 provides hooks that allow the default DSA functions or functions on a
17201 "per key" basis to be replaced. This allows hardware acceleration and
17202 hardware key storage to be handled without major modification to the
17203 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17204 associated functions.
17208 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17209 as "read only": it can't be written to and the buffer it points to will
17210 not be freed. Reading from a read only BIO is much more efficient than
17211 a normal memory BIO. This was added because there are several times when
17212 an area of memory needs to be read from a BIO. The previous method was
17213 to create a memory BIO and write the data to it, this results in two
17214 copies of the data and an O(n^2) reading algorithm. There is a new
17215 function BIO_new_mem_buf() which creates a read only memory BIO from
17216 an area of memory. Also modified the PKCS#7 routines to use read only
17221 * Bugfix: ssl23_get_client_hello did not work properly when called in
17222 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17223 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17224 but a retry condition occurred while trying to read the rest.
17228 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17229 NID_pkcs7_encrypted by default: this was wrong since this should almost
17230 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17231 the encrypted data type: this is a more sensible place to put it and it
17232 allows the PKCS#12 code to be tidied up that duplicated this
17237 * Changed obj_dat.pl script so it takes its input and output files on
17238 the command line. This should avoid shell escape redirection problems
17243 * Initial support for certificate extension requests, these are included
17244 in things like Xenroll certificate requests. Included functions to allow
17245 extensions to be obtained and added.
17249 * -crlf option to s_client and s_server for sending newlines as
17250 CRLF (as required by many protocols).
17254 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17256 * Install libRSAglue.a when OpenSSL is built with RSAref.
17258 *Ralf S. Engelschall*
17260 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17262 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17264 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17269 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17270 DH parameters/keys (q is lost during that conversion, but the resulting
17271 DH parameters contain its length).
17273 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17274 much faster than DH_generate_parameters (which creates parameters
17275 where `p = 2*q + 1`), and also the smaller q makes DH computations
17276 much more efficient (160-bit exponentiation instead of 1024-bit
17277 exponentiation); so this provides a convenient way to support DHE
17278 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17279 utter importance to use
17280 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17282 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17283 when such DH parameters are used, because otherwise small subgroup
17284 attacks may become possible!
17288 * Avoid memory leak in i2d_DHparams.
17292 * Allow the -k option to be used more than once in the enc program:
17293 this allows the same encrypted message to be read by multiple recipients.
17297 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17298 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17299 it will always use the numerical form of the OID, even if it has a short
17304 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17305 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17306 otherwise bn_mod_exp was called. In the case of hardware keys for example
17307 no private key components need be present and it might store extra data
17308 in the RSA structure, which cannot be accessed from bn_mod_exp.
17309 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17310 private key operations.
17314 * Added support for SPARC Linux.
17318 * pem_password_cb function type incompatibly changed from
17319 typedef int pem_password_cb(char *buf, int size, int rwflag);
17321 ....(char *buf, int size, int rwflag, void *userdata);
17322 so that applications can pass data to their callbacks:
17323 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17324 additional void * argument, which is just handed through whenever
17325 the password callback is called.
17327 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17329 New function SSL_CTX_set_default_passwd_cb_userdata.
17331 Compatibility note: As many C implementations push function arguments
17332 onto the stack in reverse order, the new library version is likely to
17333 interoperate with programs that have been compiled with the old
17334 pem_password_cb definition (PEM_whatever takes some data that
17335 happens to be on the stack as its last argument, and the callback
17336 just ignores this garbage); but there is no guarantee whatsoever that
17339 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17340 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17341 problems not only on Windows, but also on some Unix platforms.
17342 To avoid problematic command lines, these definitions are now in an
17343 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17344 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17348 * MIPS III/IV assembler module is reimplemented.
17352 * More DES library cleanups: remove references to srand/rand and
17353 delete an unused file.
17357 * Add support for the free Netwide assembler (NASM) under Win32,
17358 since not many people have MASM (ml) and it can be hard to obtain.
17359 This is currently experimental but it seems to work OK and pass all
17360 the tests. Check out INSTALL.W32 for info.
17364 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17365 without temporary keys kept an extra copy of the server key,
17366 and connections with temporary keys did not free everything in case
17371 * New function RSA_check_key and new openssl rsa option -check
17372 for verifying the consistency of RSA keys.
17374 *Ulf Moeller, Bodo Moeller*
17376 * Various changes to make Win32 compile work:
17377 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17378 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17379 comparison" warnings.
17380 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17384 * Add a debugging option to PKCS#5 v2 key generation function: when
17385 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17386 derived keys are printed to stderr.
17390 * Copy the flags in ASN1_STRING_dup().
17392 *Roman E. Pavlov <pre@mo.msk.ru>*
17394 * The x509 application mishandled signing requests containing DSA
17395 keys when the signing key was also DSA and the parameters didn't match.
17397 It was supposed to omit the parameters when they matched the signing key:
17398 the verifying software was then supposed to automatically use the CA's
17399 parameters if they were absent from the end user certificate.
17401 Omitting parameters is no longer recommended. The test was also
17402 the wrong way round! This was probably due to unusual behaviour in
17403 EVP_cmp_parameters() which returns 1 if the parameters match.
17404 This meant that parameters were omitted when they *didn't* match and
17405 the certificate was useless. Certificates signed with 'ca' didn't have
17408 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17410 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17411 The interface is as follows:
17412 Applications can use
17413 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17414 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17415 "off" is now the default.
17416 The library internally uses
17417 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17418 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17419 to disable memory-checking temporarily.
17421 Some inconsistent states that previously were possible (and were
17422 even the default) are now avoided.
17424 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17425 with each memory chunk allocated; this is occasionally more helpful
17426 than just having a counter.
17428 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17430 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17435 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17436 which largely parallels "options", but is for changing API behaviour,
17437 whereas "options" are about protocol behaviour.
17438 Initial "mode" flags are:
17440 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17441 a single record has been written.
17442 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17443 retries use the same buffer location.
17444 (But all of the contents must be
17449 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17452 * Fix problems with no-hmac etc.
17454 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17456 * New functions RSA_get_default_method(), RSA_set_method() and
17457 RSA_get_method(). These allows replacement of RSA_METHODs without having
17458 to mess around with the internals of an RSA structure.
17462 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17463 Also really enable memory leak checks in openssl.c and in some
17466 *Chad C. Mulligan, Bodo Moeller*
17468 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17469 up the length of negative integers. This has now been simplified to just
17470 store the length when it is first determined and use it later, rather
17471 than trying to keep track of where data is copied and updating it to
17473 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17475 * Add a new function PKCS7_signatureVerify. This allows the verification
17476 of a PKCS#7 signature but with the signing certificate passed to the
17477 function itself. This contrasts with PKCS7_dataVerify which assumes the
17478 certificate is present in the PKCS#7 structure. This isn't always the
17479 case: certificates can be omitted from a PKCS#7 structure and be
17480 distributed by "out of band" means (such as a certificate database).
17484 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17485 function prototypes in pem.h, also change util/mkdef.pl to add the
17486 necessary function names.
17490 * mk1mf.pl (used by Windows builds) did not properly read the
17491 options set by Configure in the top level Makefile, and Configure
17492 was not even able to write more than one option correctly.
17493 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17497 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17498 file to be loaded from a BIO or FILE pointer. The BIO version will
17499 for example allow memory BIOs to contain config info.
17503 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17504 Whoever hopes to achieve shared-library compatibility across versions
17505 must use this, not the compile-time macro.
17506 (Exercise 0.9.4: Which is the minimum library version required by
17508 Note: All this applies only to multi-threaded programs, others don't
17513 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17514 through a BIO pair triggered the default case, i.e.
17515 SSLerr(...,SSL_R_UNKNOWN_STATE).
17519 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17520 can use the SSL library even if none of the specific BIOs is
17525 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17526 for the encoded length.
17528 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17530 * Add initial documentation of the X509V3 functions.
17534 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17535 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17536 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17537 secure PKCS#8 private key format with a high iteration count.
17541 * Fix determination of Perl interpreter: A perl or perl5
17542 *directory* in $PATH was also accepted as the interpreter.
17544 *Ralf S. Engelschall*
17546 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17547 wrong with it but it was very old and did things like calling
17548 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17549 unusual formatting.
17553 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17554 to use the new extension code.
17558 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17559 with macros. This should make it easier to change their form, add extra
17560 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17565 * Add to configuration table a new entry that can specify an alternative
17566 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17567 according to Mark Crispin <MRC@Panda.COM>.
17571 * DES CBC did not update the IV. Weird.
17575 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17576 Changing the behaviour of the former might break existing programs --
17577 where IV updating is needed, des_ncbc_encrypt can be used.
17580 * When bntest is run from "make test" it drives bc to check its
17581 calculations, as well as internally checking them. If an internal check
17582 fails, it needs to cause bc to give a non-zero result or make test carries
17583 on without noticing the failure. Fixed.
17587 * DES library cleanups.
17591 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17592 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17593 ciphers. NOTE: although the key derivation function has been verified
17594 against some published test vectors it has not been extensively tested
17595 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17600 * Instead of "mkdir -p", which is not fully portable, use new
17601 Perl script "util/mkdir-p.pl".
17605 * Rewrite the way password based encryption (PBE) is handled. It used to
17606 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17607 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17608 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17609 the 'parameter' field of the AlgorithmIdentifier is passed to the
17610 underlying key generation function so it must do its own ASN1 parsing.
17611 This has also changed the EVP_PBE_CipherInit() function which now has a
17612 'parameter' argument instead of literal salt and iteration count values
17613 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17617 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17618 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17619 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17620 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17621 value was just used as a "magic string" and not used directly its
17622 value doesn't matter.
17626 * Introduce some semblance of const correctness to BN. Shame C doesn't
17631 * "linux-sparc64" configuration (ultrapenguin).
17633 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17634 "linux-sparc" configuration.
17636 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17638 * config now generates no-xxx options for missing ciphers.
17642 * Support the EBCDIC character set (work in progress).
17643 File ebcdic.c not yet included because it has a different license.
17645 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17647 * Support BS2000/OSD-POSIX.
17649 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17651 * Make callbacks for key generation use `void *` instead of `char *`.
17655 * Make S/MIME samples compile (not yet tested).
17659 * Additional typesafe stacks.
17663 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17667 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17669 * New configuration variant "sco5-gcc".
17671 * Updated some demos.
17673 *Sean O Riordain, Wade Scholine*
17675 * Add missing BIO_free at exit of pkcs12 application.
17679 * Fix memory leak in conf.c.
17683 * Updates for Win32 to assembler version of MD5.
17687 * Set #! path to perl in `apps/der_chop` to where we found it
17688 instead of using a fixed path.
17692 * SHA library changes for irix64-mips4-cc.
17696 * Improvements for VMS support.
17700 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17702 * Bignum library bug fix. IRIX 6 passes "make test" now!
17703 This also avoids the problems with SC4.2 and unpatched SC5.
17705 *Andy Polyakov <appro@fy.chalmers.se>*
17707 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17708 These are required because of the typesafe stack would otherwise break
17709 existing code. If old code used a structure member which used to be STACK
17710 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17711 sk_num or sk_value it would produce an error because the num, data members
17712 are not present in STACK_OF. Now it just produces a warning. sk_set
17713 replaces the old method of assigning a value to sk_value
17714 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17715 that does this will no longer work (and should use sk_set instead) but
17716 this could be regarded as a "questionable" behaviour anyway.
17720 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17721 correctly handle encrypted S/MIME data.
17725 * Change type of various DES function arguments from des_cblock
17726 (which means, in function argument declarations, pointer to char)
17727 to des_cblock * (meaning pointer to array with 8 char elements),
17728 which allows the compiler to do more typechecking; it was like
17729 that back in SSLeay, but with lots of ugly casts.
17731 Introduce new type const_des_cblock.
17735 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17736 problems: find RecipientInfo structure that matches recipient certificate
17737 and initialise the ASN1 structures properly based on passed cipher.
17741 * Belatedly make the BN tests actually check the results.
17745 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17746 to and from BNs: it was completely broken. New compilation option
17747 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17748 key elements as negative integers.
17752 * Reorganize and speed up MD5.
17754 *Andy Polyakov <appro@fy.chalmers.se>*
17758 *Richard Levitte <richard@levitte.org>*
17760 * New option -out to asn1parse to allow the parsed structure to be
17761 output to a file. This is most useful when combined with the -strparse
17762 option to examine the output of things like OCTET STRINGS.
17766 * Make SSL library a little more fool-proof by not requiring any longer
17767 that `SSL_set_{accept,connect}_state` be called before
17768 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17769 in many applications because usually everything *appeared* to work as
17770 intended anyway -- now it really works as intended).
17774 * Move openssl.cnf out of lib/.
17778 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17779 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17780 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17782 *Ralf S. Engelschall*
17784 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17785 handle PKCS#7 enveloped data properly.
17787 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17789 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17790 copying pointers. The cert_st handling is changed by this in
17791 various ways (and thus what used to be known as ctx->default_cert
17792 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17793 any longer when s->cert does not give us what we need).
17794 ssl_cert_instantiate becomes obsolete by this change.
17795 As soon as we've got the new code right (possibly it already is?),
17796 we have solved a couple of bugs of the earlier code where s->cert
17797 was used as if it could not have been shared with other SSL structures.
17799 Note that using the SSL API in certain dirty ways now will result
17800 in different behaviour than observed with earlier library versions:
17801 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17802 does not influence s as it used to.
17804 In order to clean up things more thoroughly, inside SSL_SESSION
17805 we don't use CERT any longer, but a new structure SESS_CERT
17806 that holds per-session data (if available); currently, this is
17807 the peer's certificate chain and, for clients, the server's certificate
17808 and temporary key. CERT holds only those values that can have
17809 meaningful defaults in an SSL_CTX.
17813 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17814 from the internal representation. Various PKCS#7 fixes: remove some
17815 evil casts and set the enc_dig_alg field properly based on the signing
17820 * Allow PKCS#12 password to be set from the command line or the
17821 environment. Let 'ca' get its config file name from the environment
17822 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17827 * Allow certificate policies extension to use an IA5STRING for the
17828 organization field. This is contrary to the PKIX definition but
17829 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17834 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17835 without disallowing inline assembler and the like for non-pedantic builds.
17839 * Support Borland C++ builder.
17841 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17847 * SHA-1 cleanups and performance enhancements.
17849 *Andy Polyakov <appro@fy.chalmers.se>*
17851 * Sparc v8plus assembler for the bignum library.
17853 *Andy Polyakov <appro@fy.chalmers.se>*
17855 * Accept any -xxx and +xxx compiler options in Configure.
17859 * Update HPUX configuration.
17863 * Add missing `sk_<type>_unshift()` function to safestack.h
17865 *Ralf S. Engelschall*
17867 * New function SSL_CTX_use_certificate_chain_file that sets the
17868 "extra_cert"s in addition to the certificate. (This makes sense
17869 only for "PEM" format files, as chains as a whole are not
17874 * Support verify_depth from the SSL API.
17875 x509_vfy.c had what can be considered an off-by-one-error:
17876 Its depth (which was not part of the external interface)
17877 was actually counting the number of certificates in a chain;
17878 now it really counts the depth.
17882 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17883 instead of X509err, which often resulted in confusing error
17884 messages since the error codes are not globally unique
17885 (e.g. an alleged error in ssl3_accept when a certificate
17886 didn't match the private key).
17888 * New function SSL_CTX_set_session_id_context that allows to set a default
17889 value (so that you don't need SSL_set_session_id_context for each
17890 connection using the SSL_CTX).
17894 * OAEP decoding bug fix.
17898 * Support INSTALL_PREFIX for package builders, as proposed by
17903 * New Configure options "threads" and "no-threads". For systems
17904 where the proper compiler options are known (currently Solaris
17905 and Linux), "threads" is the default.
17909 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17913 * Install various scripts to $(OPENSSLDIR)/misc, not to
17914 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17915 such as /usr/local/bin.
17919 * "make linux-shared" to build shared libraries.
17921 *Niels Poppe <niels@netbox.org>*
17923 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17927 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17928 extension adding in x509 utility.
17932 * Remove NOPROTO sections and error code comments.
17936 * Partial rewrite of the DEF file generator to now parse the ANSI
17941 * New Configure options --prefix=DIR and --openssldir=DIR.
17945 * Complete rewrite of the error code script(s). It is all now handled
17946 by one script at the top level which handles error code gathering,
17947 header rewriting and C source file generation. It should be much better
17948 than the old method: it now uses a modified version of Ulf's parser to
17949 read the ANSI prototypes in all header files (thus the old K&R definitions
17950 aren't needed for error creation any more) and do a better job of
17951 translating function codes into names. The old 'ASN1 error code embedded
17952 in a comment' is no longer necessary and it doesn't use .err files which
17953 have now been deleted. Also the error code call doesn't have to appear all
17954 on one line (which resulted in some large lines...).
17958 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17962 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17963 0 (which usually indicates a closed connection), but continue reading.
17967 * Fix some race conditions.
17971 * Add support for CRL distribution points extension. Add Certificate
17972 Policies and CRL distribution points documentation.
17976 * Move the autogenerated header file parts to crypto/opensslconf.h.
17980 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17981 8 of keying material. Merlin has also confirmed interop with this fix
17982 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17984 *Merlin Hughes <merlin@baltimore.ie>*
17986 * Fix lots of warnings.
17988 *Richard Levitte <levitte@stacken.kth.se>*
17990 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17991 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17993 *Richard Levitte <levitte@stacken.kth.se>*
17995 * Fix problems with sizeof(long) == 8.
17997 *Andy Polyakov <appro@fy.chalmers.se>*
17999 * Change functions to ANSI C.
18003 * Fix typos in error codes.
18005 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18007 * Remove defunct assembler files from Configure.
18011 * SPARC v8 assembler BIGNUM implementation.
18013 *Andy Polyakov <appro@fy.chalmers.se>*
18015 * Support for Certificate Policies extension: both print and set.
18016 Various additions to support the r2i method this uses.
18020 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18021 return a const string when you are expecting an allocated buffer.
18025 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18026 types DirectoryString and DisplayText.
18030 * Add code to allow r2i extensions to access the configuration database,
18031 add an LHASH database driver and add several ctx helper functions.
18035 * Fix an evil bug in bn_expand2() which caused various BN functions to
18036 fail when they extended the size of a BIGNUM.
18040 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18041 support typesafe stack.
18045 * Fix typo in SSL_[gs]et_options().
18047 *Nils Frostberg <nils@medcom.se>*
18049 * Delete various functions and files that belonged to the (now obsolete)
18050 old X509V3 handling code.
18054 * New Configure option "rsaref".
18058 * Don't auto-generate pem.h.
18062 * Introduce type-safe ASN.1 SETs.
18066 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18068 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18070 * Introduce type-safe STACKs. This will almost certainly break lots of code
18071 that links with OpenSSL (well at least cause lots of warnings), but fear
18072 not: the conversion is trivial, and it eliminates loads of evil casts. A
18073 few STACKed things have been converted already. Feel free to convert more.
18074 In the fullness of time, I'll do away with the STACK type altogether.
18078 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18079 specified in `<certfile>` by updating the entry in the index.txt file.
18080 This way one no longer has to edit the index.txt file manually for
18081 revoking a certificate. The -revoke option does the gory details now.
18083 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18085 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18086 `-text` option at all and this way the `-noout -text` combination was
18087 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18089 *Ralf S. Engelschall*
18091 * Make sure a corresponding plain text error message exists for the
18092 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18093 verify callback function determined that a certificate was revoked.
18095 *Ralf S. Engelschall*
18097 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18098 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18099 all available ciphers including rc5, which was forgotten until now.
18100 In order to let the testing shell script know which algorithms
18101 are available, a new (up to now undocumented) command
18102 `openssl list-cipher-commands` is used.
18106 * Bugfix: s_client occasionally would sleep in select() when
18107 it should have checked SSL_pending() first.
18111 * New functions DSA_do_sign and DSA_do_verify to provide access to
18112 the raw DSA values prior to ASN.1 encoding.
18116 * Tweaks to Configure
18118 *Niels Poppe <niels@netbox.org>*
18120 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18125 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18129 * New config option to avoid instructions that are illegal on the 80386.
18130 The default code is faster, but requires at least a 486.
18134 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18135 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18136 same as SSL2_VERSION anyway.
18140 * New "-showcerts" option for s_client.
18144 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18145 application. Various cleanups and fixes.
18149 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18150 modify error routines to work internally. Add error codes and PBE init
18151 to library startup routines.
18155 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18156 packing functions to asn1 and evp. Changed function names and error
18157 codes along the way.
18161 * PKCS12 integration: and so it begins... First of several patches to
18162 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18163 objects to objects.h
18167 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18168 and display support for Thawte strong extranet extension.
18172 * Add LinuxPPC support.
18174 *Jeff Dubrule <igor@pobox.org>*
18176 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18177 bn_div_words in alpha.s.
18179 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18181 * Make sure the RSA OAEP test is skipped under -DRSAref because
18182 OAEP isn't supported when OpenSSL is built with RSAref.
18184 *Ulf Moeller <ulf@fitug.de>*
18186 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18187 so they no longer are missing under -DNOPROTO.
18189 *Soren S. Jorvang <soren@t.dk>*
18191 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18193 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18194 doesn't work when the session is reused. Coming soon!
18198 * Fix a security hole, that allows sessions to be reused in the wrong
18199 context thus bypassing client cert protection! All software that uses
18200 client certs and session caches in multiple contexts NEEDS PATCHING to
18201 allow session reuse! A fuller solution is in the works.
18203 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18205 * Some more source tree cleanups (removed obsolete files
18206 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18207 permission on "config" script to be executable) and a fix for the INSTALL
18210 *Ulf Moeller <ulf@fitug.de>*
18212 * Remove some legacy and erroneous uses of malloc, free instead of
18215 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18217 * Make rsa_oaep_test return non-zero on error.
18219 *Ulf Moeller <ulf@fitug.de>*
18221 * Add support for native Solaris shared libraries. Configure
18222 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18223 if someone would make that last step automatic.
18225 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18227 * ctx_size was not built with the right compiler during "make links". Fixed.
18231 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18232 except NULL ciphers". This means the default cipher list will no longer
18233 enable NULL ciphers. They need to be specifically enabled e.g. with
18234 the string "DEFAULT:eNULL".
18238 * Fix to RSA private encryption routines: if p < q then it would
18239 occasionally produce an invalid result. This will only happen with
18240 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18244 * Be less restrictive and allow also `perl util/perlpath.pl
18245 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18246 because this way one can also use an interpreter named `perl5` (which is
18247 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18248 installed as `perl`).
18250 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18252 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18254 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18256 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18257 advapi32.lib to Win32 build and change the pem test comparison
18258 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18259 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18260 and crypto/des/ede_cbcm_enc.c.
18264 * DES quad checksum was broken on big-endian architectures. Fixed.
18268 * Comment out two functions in bio.h that aren't implemented. Fix up the
18269 Win32 test batch file so it (might) work again. The Win32 test batch file
18270 is horrible: I feel ill....
18274 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18275 in e_os.h. Audit of header files to check ANSI and non ANSI
18276 sections: 10 functions were absent from non ANSI section and not exported
18277 from Windows DLLs. Fixed up libeay.num for new functions.
18281 * Make `openssl version` output lines consistent.
18283 *Ralf S. Engelschall*
18285 * Fix Win32 symbol export lists for BIO functions: Added
18286 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18287 to ms/libeay{16,32}.def.
18289 *Ralf S. Engelschall*
18291 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18292 fine under Unix and passes some trivial tests I've now added. But the
18293 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18294 added to make sure no one expects that this stuff really works in the
18295 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18296 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18299 *Ralf S. Engelschall*
18301 * Fix the generation of two part addresses in perl.
18303 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18305 * Add config entry for Linux on MIPS.
18307 *John Tobey <jtobey@channel1.com>*
18309 * Make links whenever Configure is run, unless we are on Windoze.
18313 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18314 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18319 * Add a useful kludge to allow package maintainers to specify compiler and
18320 other platforms details on the command line without having to patch the
18321 Configure script every time: One now can use
18322 `perl Configure <id>:<details>`,
18323 i.e. platform ids are allowed to have details appended
18324 to them (separated by colons). This is treated as there would be a static
18325 pre-configured entry in Configure's %table under key `<id>` with value
18326 `<details>` and `perl Configure <id>` is called. So, when you want to
18327 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18328 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18329 now, which overrides the FreeBSD-elf entry on-the-fly.
18331 *Ralf S. Engelschall*
18333 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18337 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18338 on the `perl Configure ...` command line. This way one can compile
18339 OpenSSL libraries with Position Independent Code (PIC) which is needed
18340 for linking it into DSOs.
18342 *Ralf S. Engelschall*
18344 * Remarkably, export ciphers were totally broken and no-one had noticed!
18349 * Cleaned up the LICENSE document: The official contact for any license
18350 questions now is the OpenSSL core team under openssl-core@openssl.org.
18351 And add a paragraph about the dual-license situation to make sure people
18352 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18353 to the OpenSSL toolkit.
18355 *Ralf S. Engelschall*
18357 * General source tree makefile cleanups: Made `making xxx in yyy...`
18358 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18359 Additionally cleaned up the `make links` target: Remove unnecessary
18360 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18361 to speed processing and no longer clutter the display with confusing
18362 stuff. Instead only the actually done links are displayed.
18364 *Ralf S. Engelschall*
18366 * Permit null encryption ciphersuites, used for authentication only. It used
18367 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18368 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18373 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18374 signed attributes when verifying signatures (this would break them),
18375 the detached data encoding was wrong and public keys obtained using
18376 X509_get_pubkey() weren't freed.
18380 * Add text documentation for the BUFFER functions. Also added a work around
18381 to a Win95 console bug. This was triggered by the password read stuff: the
18382 last character typed gets carried over to the next fread(). If you were
18383 generating a new cert request using 'req' for example then the last
18384 character of the passphrase would be CR which would then enter the first
18389 * Added the new 'Includes OpenSSL Cryptography Software' button as
18390 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18391 button and can be used by applications based on OpenSSL to show the
18392 relationship to the OpenSSL project.
18394 *Ralf S. Engelschall*
18396 * Remove confusing variables in function signatures in files
18397 ssl/ssl_lib.c and ssl/ssl.h.
18399 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18401 * Don't install bss_file.c under PREFIX/include/
18403 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18405 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18406 functions that return function pointers and has support for NT specific
18407 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18408 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18409 unsigned to signed types: this was killing the Win32 compile.
18413 * Add new certificate file to stack functions,
18414 SSL_add_dir_cert_subjects_to_stack() and
18415 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18416 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18417 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18418 This means that Apache-SSL and similar packages don't have to mess around
18419 to add as many CAs as they want to the preferred list.
18423 * Experiment with doxygen documentation. Currently only partially applied to
18425 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18426 openssl.doxy as the configuration file.
18430 * Get rid of remaining C++-style comments which strict C compilers hate.
18432 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18434 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18435 compiled in by default: it has problems with large keys.
18439 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18440 DH private keys and/or callback functions which directly correspond to
18441 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18442 is needed for applications which have to configure certificates on a
18443 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18445 For the RSA certificate situation is makes no difference, but
18446 for the DSA certificate situation this fixes the "no shared cipher"
18447 problem where the OpenSSL cipher selection procedure failed because the
18448 temporary keys were not overtaken from the context and the API provided
18449 no way to reconfigure them.
18450 The new functions now let applications reconfigure the stuff and they
18451 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18452 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18453 non-public-API function ssl_cert_instantiate() is used as a helper
18454 function and also to reduce code redundancy inside ssl_rsa.c.
18456 *Ralf S. Engelschall*
18458 * Move s_server -dcert and -dkey options out of the undocumented feature
18459 area because they are useful for the DSA situation and should be
18460 recognized by the users.
18462 *Ralf S. Engelschall*
18464 * Fix the cipher decision scheme for export ciphers: the export bits are
18465 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18466 SSL_EXP_MASK. So, the original variable has to be used instead of the
18467 already masked variable.
18469 *Richard Levitte <levitte@stacken.kth.se>*
18471 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18473 *Richard Levitte <levitte@stacken.kth.se>*
18475 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18476 from `int` to `unsigned int` because it is a length and initialized by
18477 EVP_DigestFinal() which expects an `unsigned int *`.
18479 *Richard Levitte <levitte@stacken.kth.se>*
18481 * Don't hard-code path to Perl interpreter on shebang line of Configure
18482 script. Instead use the usual Shell->Perl transition trick.
18484 *Ralf S. Engelschall*
18486 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18487 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18488 -noout -modulus` as it's already the case for `openssl rsa -noout
18489 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18490 currently the public key is printed (a decision which was already done by
18491 `openssl dsa -modulus` in the past) which serves a similar purpose.
18492 Additionally the NO_RSA no longer completely removes the whole -modulus
18493 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18496 *Ralf S. Engelschall*
18498 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18499 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18501 *Arne Ansper <arne@ats.cyber.ee>*
18503 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18504 to be added. Now both 'req' and 'ca' can use new objects defined in the
18509 * Add cool BIO that does syslog (or event log on NT).
18511 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18513 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18514 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18515 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18516 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18520 * Add preliminary config info for new extension code.
18524 * Make RSA_NO_PADDING really use no padding.
18526 *Ulf Moeller <ulf@fitug.de>*
18528 * Generate errors when private/public key check is done.
18532 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18533 for some CRL extensions and new objects added.
18537 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18538 key usage extension and fuller support for authority key id.
18542 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18543 padding method for RSA, which is recommended for new applications in PKCS
18544 #1 v2.0 (RFC 2437, October 1998).
18545 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18546 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18547 against Bleichbacher's attack on RSA.
18548 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18551 * Updates to the new SSL compression code
18553 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18555 * Fix so that the version number in the master secret, when passed
18556 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18557 (because the server will not accept higher), that the version number
18558 is 0x03,0x01, not 0x03,0x00
18560 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18562 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18563 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18564 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18568 * Support for RAW extensions where an arbitrary extension can be
18569 created by including its DER encoding. See `apps/openssl.cnf` for
18574 * Make sure latest Perl versions don't interpret some generated C array
18575 code as Perl array code in the crypto/err/err_genc.pl script.
18577 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18579 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18580 not many people have the assembler. Various Win32 compilation fixes and
18581 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18582 build instructions.
18586 * Modify configure script 'Configure' to automatically create crypto/date.h
18587 file under Win32 and also build pem.h from pem.org. New script
18588 util/mkfiles.pl to create the MINFO file on environments that can't do a
18589 'make files': perl util/mkfiles.pl >MINFO should work.
18593 * Major rework of DES function declarations, in the pursuit of correctness
18594 and purity. As a result, many evil casts evaporated, and some weirdness,
18595 too. You may find this causes warnings in your code. Zapping your evil
18596 casts will probably fix them. Mostly.
18600 * Fix for a typo in asn1.h. Bug fix to object creation script
18601 obj_dat.pl. It considered a zero in an object definition to mean
18602 "end of object": none of the objects in objects.h have any zeros
18603 so it wasn't spotted.
18605 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18607 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18608 Masking (CBCM). In the absence of test vectors, the best I have been able
18609 to do is check that the decrypt undoes the encrypt, so far. Send me test
18610 vectors if you have them.
18614 * Correct calculation of key length for export ciphers (too much space was
18615 allocated for null ciphers). This has not been tested!
18619 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18620 message is now correct (it understands "crypto" and "ssl" on its
18621 command line). There is also now an "update" option. This will update
18622 the util/ssleay.num and util/libeay.num files with any new functions.
18624 perl util/mkdef.pl crypto ssl update
18625 it will update them.
18629 * Overhauled the Perl interface:
18630 - ported BN stuff to OpenSSL's different BN library
18631 - made the perl/ source tree CVS-aware
18632 - renamed the package from SSLeay to OpenSSL (the files still contain
18633 their history because I've copied them in the repository)
18634 - removed obsolete files (the test scripts will be replaced
18635 by better Test::Harness variants in the future)
18637 *Ralf S. Engelschall*
18639 * First cut for a very conservative source tree cleanup:
18640 1. merge various obsolete readme texts into doc/ssleay.txt
18641 where we collect the old documents and readme texts.
18642 2. remove the first part of files where I'm already sure that we no
18643 longer need them because of three reasons: either they are just temporary
18644 files which were left by Eric or they are preserved original files where
18645 I've verified that the diff is also available in the CVS via "cvs diff
18646 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18647 the crypto/md/ stuff).
18649 *Ralf S. Engelschall*
18651 * More extension code. Incomplete support for subject and issuer alt
18652 name, issuer and authority key id. Change the i2v function parameters
18653 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18654 what that's for :-) Fix to ASN1 macro which messed up
18655 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18659 * Preliminary support for ENUMERATED type. This is largely copied from the
18664 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18666 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18668 * Make sure `make rehash` target really finds the `openssl` program.
18670 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18672 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18673 like to hear about it if this slows down other processors.
18677 * Add CygWin32 platform information to Configure script.
18679 *Alan Batie <batie@aahz.jf.intel.com>*
18681 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18683 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18685 * New program nseq to manipulate netscape certificate sequences
18689 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18694 * Fixes to BN code. Previously the default was to define BN_RECURSION
18695 but the BN code had some problems that would cause failures when
18696 doing certificate verification and some other functions.
18698 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18700 * Add ASN1 and PEM code to support netscape certificate sequences.
18704 * Add ASN1 and PEM code to support netscape certificate sequences.
18708 * Add several PKIX and private extended key usage OIDs.
18712 * Modify the 'ca' program to handle the new extension code. Modify
18713 openssl.cnf for new extension format, add comments.
18717 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18718 and add a sample to openssl.cnf so req -x509 now adds appropriate
18723 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18724 error code, add initial support to X509_print() and x509 application.
18728 * Takes a deep breath and start adding X509 V3 extension support code. Add
18729 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18730 stuff is currently isolated and isn't even compiled yet.
18734 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18735 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18736 Removed the versions check from X509 routines when loading extensions:
18737 this allows certain broken certificates that don't set the version
18738 properly to be processed.
18742 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18743 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18744 can still be regenerated with "make depend".
18748 * Spelling mistake in C version of CAST-128.
18750 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18752 * Changes to the error generation code. The perl script err-code.pl
18753 now reads in the old error codes and retains the old numbers, only
18754 adding new ones if necessary. It also only changes the .err files if new
18755 codes are added. The makefiles have been modified to only insert errors
18756 when needed (to avoid needlessly modifying header files). This is done
18757 by only inserting errors if the .err file is newer than the auto generated
18758 C file. To rebuild all the error codes from scratch (the old behaviour)
18759 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18760 or delete all the .err files.
18764 * CAST-128 was incorrectly implemented for short keys. The C version has
18765 been fixed, but is untested. The assembler versions are also fixed, but
18766 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18767 to regenerate it if needed.
18768 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18769 Hagino <itojun@kame.net>*
18771 * File was opened incorrectly in randfile.c.
18773 *Ulf Möller <ulf@fitug.de>*
18775 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18776 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18777 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18778 al: it's just almost always a UTCTime. Note this patch adds new error
18779 codes so do a "make errors" if there are problems.
18783 * Correct Linux 1 recognition in config.
18785 *Ulf Möller <ulf@fitug.de>*
18787 * Remove pointless MD5 hash when using DSA keys in ca.
18789 *Anonymous <nobody@replay.com>*
18791 * Generate an error if given an empty string as a cert directory. Also
18792 generate an error if handed NULL (previously returned 0 to indicate an
18793 error, but didn't set one).
18795 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18797 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18801 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18802 parameters. This was causing a warning which killed off the Win32 compile.
18806 * Remove C++ style comments from crypto/bn/bn_local.h.
18808 *Neil Costigan <neil.costigan@celocom.com>*
18810 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18811 based on a text string, looking up short and long names and finally
18812 "dot" format. The "dot" format stuff didn't work. Added new function
18813 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18814 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18815 OID is not part of the table.
18819 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18820 X509_LOOKUP_by_alias().
18824 * Sort openssl functions by name.
18828 * Get the `gendsa` command working and add it to the `list` command. Remove
18829 encryption from sample DSA keys (in case anyone is interested the password
18834 * Make *all* `*_free` functions accept a NULL pointer.
18836 *Frans Heymans <fheymans@isaserver.be>*
18838 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18841 *Anonymous <nobody@replay.com>*
18843 * s_server should send the CAfile as acceptable CAs, not its own cert.
18845 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18847 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18849 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18851 * Temp key "for export" tests were wrong in s3_srvr.c.
18853 *Anonymous <nobody@replay.com>*
18855 * Add prototype for temp key callback functions
18856 SSL_CTX_set_tmp_{rsa,dh}_callback().
18860 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18861 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18865 * X509_name_add_entry() freed the wrong thing after an error.
18867 *Arne Ansper <arne@ats.cyber.ee>*
18869 * rsa_eay.c would attempt to free a NULL context.
18871 *Arne Ansper <arne@ats.cyber.ee>*
18873 * BIO_s_socket() had a broken should_retry() on Windoze.
18875 *Arne Ansper <arne@ats.cyber.ee>*
18877 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18879 *Arne Ansper <arne@ats.cyber.ee>*
18881 * Make sure the already existing X509_STORE->depth variable is initialized
18882 in X509_STORE_new(), but document the fact that this variable is still
18883 unused in the certificate verification process.
18885 *Ralf S. Engelschall*
18887 * Fix the various library and `apps/` files to free up pkeys obtained from
18888 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18892 * Fix reference counting in X509_PUBKEY_get(). This makes
18893 demos/maurice/example2.c work, amongst others, probably.
18895 *Steve Henson and Ben Laurie*
18897 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18898 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18899 are no longer created. This way we have a single and consistent command
18900 line interface `openssl <command>`, similar to `cvs <command>`.
18902 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18904 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18905 BIT STRING wrapper always have zero unused bits.
18909 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18913 * Make the top-level INSTALL documentation easier to understand.
18917 * Makefiles updated to exit if an error occurs in a sub-directory
18918 make (including if user presses ^C) [Paul Sutton]
18920 * Make Montgomery context stuff explicit in RSA data structure.
18924 * Fix build order of pem and err to allow for generated pem.h.
18928 * Fix renumbering bug in X509_NAME_delete_entry().
18932 * Enhanced the err-ins.pl script so it makes the error library number
18933 global and can add a library name. This is needed for external ASN1 and
18934 other error libraries.
18938 * Fixed sk_insert which never worked properly.
18942 * Fix ASN1 macros so they can handle indefinite length constructed
18943 EXPLICIT tags. Some non standard certificates use these: they can now
18948 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18949 into a single doc/ssleay.txt bundle. This way the information is still
18950 preserved but no longer messes up this directory. Now it's new room for
18951 the new set of documentation files.
18953 *Ralf S. Engelschall*
18955 * SETs were incorrectly DER encoded. This was a major pain, because they
18956 shared code with SEQUENCEs, which aren't coded the same. This means that
18957 almost everything to do with SETs or SEQUENCEs has either changed name or
18958 number of arguments.
18960 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18962 * Fix test data to work with the above.
18966 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18967 was already fixed by Eric for 0.9.1 it seems.
18969 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18971 * Autodetect FreeBSD3.
18975 * Fix various bugs in Configure. This affects the following platforms:
18979 unixware-2.0-pentium
18984 * Eliminate generated files from CVS. Reorder tests to regenerate files
18985 before they are needed.
18989 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18993 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18995 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18996 changed SSLeay to OpenSSL in version strings.
18998 *Ralf S. Engelschall*
19000 * Some fixups to the top-level documents.
19004 * Fixed the nasty bug where rsaref.h was not found under compile-time
19005 because the symlink to include/ was missing.
19007 *Ralf S. Engelschall*
19009 * Incorporated the popular no-RSA/DSA-only patches
19010 which allow to compile a RSA-free SSLeay.
19012 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19014 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19015 when "ssleay" is still not found.
19017 *Ralf S. Engelschall*
19019 * Added more platforms to Configure: Cray T3E, HPUX 11,
19021 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19023 * Updated the README file.
19025 *Ralf S. Engelschall*
19027 * Added various .cvsignore files in the CVS repository subdirs
19028 to make a "cvs update" really silent.
19030 *Ralf S. Engelschall*
19032 * Recompiled the error-definition header files and added
19033 missing symbols to the Win32 linker tables.
19035 *Ralf S. Engelschall*
19037 * Cleaned up the top-level documents;
19038 o new files: CHANGES and LICENSE
19039 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19040 o merged COPYRIGHT into LICENSE
19041 o removed obsolete TODO file
19042 o renamed MICROSOFT to INSTALL.W32
19044 *Ralf S. Engelschall*
19046 * Removed dummy files from the 0.9.1b source tree:
19047 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19048 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19049 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19050 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19051 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19053 *Ralf S. Engelschall*
19055 * Added various platform portability fixes.
19059 * The Genesis of the OpenSSL rpject:
19060 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19061 Young and Tim J. Hudson created while they were working for C2Net until
19064 *The OpenSSL Project*
19066 ### Changes between 0.9.0b and 0.9.1b [not released]
19068 * Updated a few CA certificates under certs/
19072 * Changed some BIGNUM api stuff.
19076 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19077 DGUX x86, Linux Alpha, etc.
19081 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19082 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19087 * Add -strparse option to asn1pars program which parses nested
19090 *Dr Stephen Henson <shenson@bigfoot.com>*
19092 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19096 * DSA fix for "ca" program.
19100 * Added "-genkey" option to "dsaparam" program.
19104 * Added RIPE MD160 (rmd160) message digest.
19108 * Added -a (all) option to "ssleay version" command.
19112 * Added PLATFORM define which is the id given to Configure.
19116 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19120 * Extended the ASN.1 parser routines.
19124 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19128 * Added a BN_CTX to the BN library.
19132 * Fixed the weak key values in DES library
19136 * Changed API in EVP library for cipher aliases.
19140 * Added support for RC2/64bit cipher.
19144 * Converted the lhash library to the crypto/mem.c functions.
19148 * Added more recognized ASN.1 object ids.
19152 * Added more RSA padding checks for SSL/TLS.
19156 * Added BIO proxy/filter functionality.
19160 * Added extra_certs to SSL_CTX which can be used
19161 send extra CA certificates to the client in the CA cert chain sending
19162 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19166 * Now Fortezza is denied in the authentication phase because
19167 this is key exchange mechanism is not supported by SSLeay at all.
19171 * Additional PKCS1 checks.
19175 * Support the string "TLSv1" for all TLS v1 ciphers.
19179 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19180 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19184 * Fixed a few memory leaks.
19188 * Fixed various code and comment typos.
19192 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19193 bytes sent in the client random.
19195 *Edward Bishop <ebishop@spyglass.com>*
19199 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19200 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19201 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19202 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19203 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19204 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19205 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19206 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19207 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19208 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19209 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19210 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19211 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19212 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19213 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19214 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19215 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19216 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19217 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19218 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19219 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19220 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19221 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19222 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19223 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19224 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19225 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19226 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19227 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19228 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19229 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19230 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19231 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19232 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19233 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19234 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19235 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19236 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19237 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19238 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19239 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19240 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19241 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19242 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19243 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19244 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19245 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19246 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19247 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19248 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19249 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19250 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19251 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19252 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19253 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19254 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19255 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19256 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19257 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19258 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19259 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19260 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19261 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19262 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19263 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19264 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19265 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19266 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19267 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19268 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19269 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19270 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19271 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19272 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19273 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19274 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19275 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19276 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19277 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19278 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19279 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19280 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19281 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19282 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19283 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19284 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19285 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19286 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19287 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19288 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19289 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19290 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19291 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19292 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19293 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19294 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19295 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19296 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19297 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19298 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19299 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19300 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19301 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19302 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19303 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19304 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19305 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19306 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19307 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19308 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19309 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19310 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19311 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19312 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19313 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19314 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19315 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19316 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19317 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19318 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19319 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19320 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19321 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19322 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19323 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19324 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19325 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19326 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19327 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19328 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19329 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19330 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19331 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19332 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19333 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19334 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19335 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19336 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19337 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19338 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19339 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19340 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19341 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19342 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19343 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19344 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19345 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19346 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19347 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19348 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19349 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19350 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19351 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19352 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19353 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19354 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19355 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19356 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19357 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19358 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19359 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19360 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655